Wound-induced Oxidative Responses in Mountain Birch Leaves

Abstract

• Aims The aim of the study was to examine oxidative responses in subarctic mountain birch, Betula pubescens subsp. czerepanovii, induced by herbivory and manual wounding.

• Methods Herbivory-induced changes in polyphenoloxidase, peroxidase and catalase activities in birch leaves were determined. A cytochemical dye, 3,3-diaminobenzidine, was used for the in situ and in vivo detection of H₂O₂ accumulation as a response to herbivory and wounding. To localize peroxidase activity in leaves, 10 mm H₂O₂ was applied to the dye reagent.

• Key Results Feeding by autumnal moth, Epirrita autumnata, larvae caused an induction in polyphenoloxidase and peroxidase activities within 24 h, and a concomitant decrease in the activity of antioxidative catalases in wounded leaves. Wounding also induced H₂O₂ accumulation, which may have both direct and indirect defensive properties against herbivores. Wound sites and guard cells showed a high level of peroxidase activity, which may efficiently restrict invasion by micro-organisms.

• Conclusion Birch oxidases together with their substrates may form an important front line in defence against herbivores and pathogens.

INTRODUCTION

Plants use multiple strategies to defend against plant-eating animals. These strategies include direct defensive traits, such as physical obstacles or chemical compounds (Scriber and Slansky, 1981; Lucas et al., 2000), and indirect strategies, such as attracting a pest's natural enemies with volatile cues from the plant (Alborn et al., 1997). Chemical compounds may act as feeding deterrents or they may have toxic or anti-nutritive properties (e.g. Scriber and Slansky, 1981). The emphasis in ecological studies of plant–herbivory interaction has long been on the secondary compounds of host plants. Secondary compounds can be divided into three groups: terpenes, phenolics and nitrogen-containing compounds. Phenolics are probably the most common and most widely studied group of compounds presumed to provide defence against herbivores in northern deciduous trees.

The defence of mountain birch against its main defoliator, autumnal moth (Epirrita autumnata) larvae, has received considerable attention over several decades. During high-density years, these moth larvae may cause severe damage to mountain birch forest in North European Lapland (Tenow, 1972), impacting the entire ecosystem of the damage area. Folivory is known to trigger both rapid induced resistance within the same season (e.g. Haukioja and Niemelä, 1979; Hanhimäki, 1989) and delayed induced resistance in the following season or seasons (e.g. Haukioja and Neuvonen, 1985; Ruohomäki et al., 1996; Kaitaniemi and Ruohomäki, 2001) in mountain birches, but the mechanisms behind this resistance are not fully understood. During the most severe E. autumnata outbreaks, this resistance is not enough to prevent death of the birch trees. This does not mean, however, that the inducible defences in question are inconsequential for E. autumnata or other herbivores in a non-outbreak situation.

The chemical composition of birch leaves has been characterized in detail (e.g. Salminen et al., 1999, 2002; Ossipov et al., 2001). However, attempts to link the autumnal moth's performance to certain phenolic compounds have thus far failed; results have been inconsistent and variable between studies (see Haukioja, 2003). One obvious reason is that the chemical composition of mountain birch leaves changes very rapidly early in the season (Kause et al., 1999; Riipi et al., 2002) during larval growth, and different chemicals may be noxious at different stages of leaf and larval development.

Another explanation is that phenolics require enzymes to activate their defensive properties (Felton et al., 1989). For example, peroxidases (PODs) and polyphenol oxidases (PPOs) are induced by herbivore attacks (e.g. Felton et al., 1989; Tscharntke et al., 2001). These enzymes catalyse the conversion of plant diphenols, such as chlorogenic acid, a main individual phenolic compound in mountain birch leaves (Nurmi et al., 1996), to highly reactive quinones. These quinones bind to electron-rich moieties of amino acids and proteins, decreasing their assimilation in the herbivore digestive tract and thus leading to malnutrition in developing insects (Felton et al., 1989).

The classic plant quiacol PODs (class III peroxidases; EC 1.11.17) belong to a large multigene family and take part in a wide range of physiological processes, including lignification, suberization, cross-linking of cell-wall proteins, auxin metabolism, defence against pathogen attack and oxidative stress (Hiraga et al., 2001). The PPO family also contains several soluble and membrane-bound isomers, which may oxidize a variety of phenolic compounds. Catechol oxidases or diphenolases (EC 1.10.2.2) use o-diphenols as substrates, converting them to o-quinones. Tyrosinases or monophenoloxidases (EC 1.14.18.1) in turn oxidize monophenols to o-diphenols (reviewed in Martinez and Whitaker, 1995). Plant PODs use hydrogen peroxide, H₂O₂, as a co-substrate, whereas PPOs use molecular oxygen, O₂, and may be inhibited by endogenous H₂O₂.

Pathogen infection may cause a hypersensitive response (HR) in plants, characterized by a rapid oxidative burst of reactive oxygen species (ROS): hydrogen peroxide, the hydroxyl radical and the superoxide radical. ROS play multiple roles in plant defence against pathogens as well as in other stress reactions (Levine et al., 1994; Laloi et al., 2004; Mittler et al., 2004). ROS such as H₂O₂ have direct defensive effects, inducing the peroxidation of lipids and the oxidation of proteins and DNA (Wu et al., 1995); H₂O₂ also acts as a local signal of the HR and as a diffusible signal activating genes in the adjacent cells (Alvarez et al., 1998). Likewise, wounding triggers the production of ROS in plant tissues (Bi and Felton, 1995; Orozco-Cárdenas et al., 2001; Gould et al., 2002; Pellinen et al., 2002; Musetti et al., 2005).

This study examined the oxidative defence of subarctic mountain birch in response to herbivory by larvae of Epirrita autumnata. Activities of PPOs, PODs and catalase (CAT, EC 1.11.1.6) were determined for wounded and unwounded leaves in a natural stand of mountain birch trees growing near the Kevo Subarctic Research Station. A cytochemical dye, 3,3-diaminobenzidine (DAB) was used for in situ and in vivo detection of the accumulation of H₂O₂ as a response to herbivory and manual wounding. DAB polymerizes instantly to a reddish-brown polymer at sites of POD activity in the presence of H₂O₂. DAB staining has been used in the in situ detection of short-lived H₂O₂ in several studies (e.g. Thordal-Christensen et al., 1997; Orozco-Cárdenas et al., 2001; Pellinen et al., 2002; Hernández et al., 2004). In addition, by applying exogenous H₂O₂ to the dye reagent it was possible to localize in vivo and in situ POD activity in birch leaves.

MATERIALS AND METHODS

All experiments were conducted at the Kevo Subarctic Research Station of the University of Turku in the summer of 2003. The research station is located in northernmost Finland (69°45′N, 27°00′E), and its subarctic summers are characterized by 24 h of daylight from mid May to late July.

Experimental organisms

Mountain birches, Betula pubescens subsp. czerepanovii (Orlova) Hämet-Ahti, are phenotypically highly variable trees as a result of the hybrid nature of the species: mountain birch is an introgressive hybrid of Betula pubescens and dwarf birch, Betula nana (Vaarama and Valanne, 1973). Mountain birch varies phenotypically from small shrubs to large monocormic trees, but in the study area the most common growth form is a rather low, polycormic tree with 5–10 trunks.

The larvae of Epirrita autumnata (Borkhausen) (Lepidoptera: Geometridae) hatch in early season at the bud break of birches, and larval development is synchronized with the development of mountain birch leaves (Kaitaniemi et al., 1997). Larvae collected in Norwegian Lapland (in the phenoloxidase experiment), as well as laboratory strains were used; the latter were a local Kevo strain and a strain of Norwegian origin (used in the DAB staining studies). Phratora polaris (Schneider) (Coleoptera: Chrysomelidae) is a polyphagous leaf beetle, the larvae and adults of which feed on the leaves of mountain birches. Adult leaf beetles were collected from a nearby mountain birch forest.

Experimental designs and sampling

For the first experiment, 24 mountain birches growing close to the research station were selected. Twelve pairs of birches were selected; in each pair, one was randomly allocated to serve as a wounded tree and the other as a control tree (external control). In the wounding treatment, 20 E. autumnata larvae (4th instar) per tree were released on two branches of the same trunk. A third branch on the same tree but on another trunk was selected as an internal control, without introduced larvae. The larvae were allowed to feed freely on the birch foliage for 24 h, after which leaf samples were collected. After this 24-h experiment, the internal control branches did not contain any larvae. Three types of sample were collected from the wounded trees: (1) leaves damaged by larva, (2) adjacent unwounded leaves and (3) internal control leaves from the unwounded trunk. In the case of the external control trees, leaves were collected from two branches of the same trunk. All leaf samples were frozen immediately in liquid nitrogen and stored at −80 °C prior to analysis.

Three naturally growing mountain birches were selected for the indoor experiment. Small branches from each tree were cut and taken to the laboratory. Six moth larvae (4th instar) and six leaf beetles were placed on their own branches of each tree. Shortly (1–5 min) after the insects started feeding, two wounded leaves from each branch were taken for DAB staining. At the same time, manual wounding was conducted with a cork bore and the wounded leaves and unwounded control leaves were stained with DAB in the presence and absence of H₂O₂.

In the outdoor experiment, three mountain birch trees were again used (other than in the indoor experiment); one trunk per tree was selected to act as the treatment trunk and another as a control trunk. Two short shoot leaves were collected from both trunks as 0-h controls. Immediately after the first sampling, 25 moth larvae (4th instar) were released into the foliage of the treatment trunks. Samples (wounded and unwounded control leaves) were then collected 6, 12 and 24 h after release of the larvae into the foliage and stained with DAB in the presence and absence of H₂O₂.

Protein extraction procedure

Frozen birch leaves (1·5 g f. wt) were immersed in 20 mL of cold extraction buffer containing 1 mm PMSF (phenylmethanesulfonyl fluoride; Sigma-Aldrich), 1 mm DTT (dl-dithiothreitol; Sigma) and 1 % Triton X-100 (Acros) in 0·2 m potassium buffer, pH 7·5. To remove phenolics from the leaf extract, 4·7 % (w/v) PVPP (polyvinylpyrrolidone; Sigma) and 2·5 % (w/v) of anion exchange resin (AG^® 1-X8 resin, 200–400 mesh chloride form, BioRad) were added to the extraction buffer. The leaves were cut with scissors into small pieces and homogenized in an Ultra Turrax homogenizer twice for 30 s in a cold room (8 °C). The homogenates were incubated via shaking on an ice bath for 1 h. The samples were centrifuged at 17 000g for 20 min at 4 °C. The supernatant was filtered through cheesecloth and further purified in Econo-Pac^® 10 DG columns (Bio-Gel^® P-6DG polyacrylamide gel, BioRad). Proteins were eluted from the columns with 4 mL of buffer B (0·05 m potassium buffer, pH 7·5). The samples were stored at −80 °C prior to activity measurements. The absence of phenolic compounds in extracts was checked with a high-performance liquid chromatograph (HPLC) using the method of Salminen et al. (1999). A phenolic-free enzyme extract is essential, as endogenous phenolics may distort the enzyme activity measurements.

Enzyme activity measurements

All enzymatic measurements were made in triplicate. The enzymatic methods used were based on several pilot experiments, during which the appropriate conditions for enzyme activities were tested. Enzyme activities were measured with a spectrophotometer that was combined with an external water bath to control the reaction temperature. All measurements were performed at 25 °C.

Acidic PPO activity with 80 mm catechol (Sigma) as a substrate and the rate of change in A₄₂₀ (Coseteng and Lee, 1987) were measured. The reaction mixture contained 375 µL buffer C (0·250 m K-phosphate buffer, pH 5·8), 177 µL buffer B and 188 µL catechol in dH₂O. The reaction mixtures were placed in microcentrifuge tubes and incubated on a warm block at 25 °C prior to the assay. Reactions were started by the addition of 10 µL of enzyme extract (kept on ice). The increase in absorbance was followed for 3 min (initial reaction velocity).

Both PPOs and PODs use diphenols such as catechol as substrate. To measure total quinone production combined PPO and POD activity were measured by adding 6·5 nm H₂O₂ (30 % J.T. Baker) to the reaction solution. Activity was otherwise measured as presented above.

We measured POD activity with 60 mm guaiacol (90 %, Sigma) as a substrate and 20 mm H₂O₂ as a co-substrate, and monitored the rate of change in A₄₇₀ with the method described in Sakharov and Ardilla (1999) with slight modifications. The reaction mixture contained 375 µL buffer C, 183 µL buffer B, 167 µL dH₂O, 15·5 µL H₂O₂, 5·5 µL guaiacol and 10 µL of enzyme extract. The increase in absorbance was followed for 3·5 min.

CAT activity was measured with 80 mm H₂O₂ as a substrate and the decrease in A₂₄₀ was followed for 3 min with the method described in Matsumura et al. (2002). The reaction mixture contained 375 µL buffer C, 183 µL buffer B, 125 µL dH₂O, 62 µL H₂O₂ and 5 µL of enzyme.

Alkaline PPO activity was measured with 10 mm l-DOPA (Sigma) in 0·100 m phosphate buffer (pH 7·5) as a substrate, using the method presented in Rantala et al. (2002). The increase in A₄₉₀ over 30 min was read on a Multilabel counter (1420 Victor, Wallac Oy, Turku, Finland).

The amount of protein in the samples was measured in duplicate with the multilabel counter, using the BioRad protein assay method based on the Bradford method (Bradford, 1976). All enzyme activities were expressed as units per milligram of protein; a single unit was the amount of enzyme required to increase the absorbance by 0·001 min⁻¹.

Histochemical detection of H₂O₂ and peroxidase in birch leaves

The DAB staining system presented in Pellinen et al. (2002), with some modifications, was used. To localize H₂O₂ accumulation, birch leaves were vacuum-infiltrated with 0·1 % (w/v) DAB (Sigma) in 10 mm MES (Sigma), pH 5·8, for 40 min. Air was extracted twice and released so that the leaves were completely saturated and immersed in the reagent, and the reaction was allowed to continue in a vacuum for 40 min. To test DAB-specificity, three control treatments were conducted: (1) manually wounded leaves incubated in the presence of 0·1 % DAB in 10 mm MES in the absence of catalase, (2) manually wounded leaves incubated in the presence of 0·1 % DAB in 10 mm MES in the presence of catalase (Sigma: 2000 U mL⁻¹) and (3) manually wounded leaves incubated in 10 mm MES but in the absence of 0·1 % DAB. To localize POD activity in birch leaves, 10 mm H₂O₂ was added to half of the reactions. The leaves were cleared by incubation in phenol (Sigma)/lactic acid (Acros)/dH₂O (1 : 1 : 1 v/v/v) for 2 d in the dark at room temperature. The samples were stored in 96 % ethanol before photography.

Statistical analyses

Differences in enzymatic activities between wounded leaves, adjacent unwounded leaves and unwounded internal control leaves were tested with one-way ANOVA followed by Tukey HSD for multiple comparisons. Differences between these individual treatments and the external controls were tested with independent-sample t-tests. The normal distribution of observations was tested with the Kolmogorov–Smirnov test and the equality of variances with Levene's test. All tests were performed with SPSS (2002) for Windows 11.5.1 (SPSS, Chicago, IL, USA).

RESULTS

Wound-induced changes in enzyme activities

In the pilot experiment, the PPOs of mountain birch were found to have different pH optima for different substrates. The highest catechol oxidase activity was found at pH 5·8, whereas the highest conversion rate of l-DOPA to quinones was found at pH 7·4. Pure PPO activity was as low as 30 % of the combined PPO/POD activity, indicating that PODs play an important role in the formation of quinones in mountain birch leaves (Fig. 1).

Fig. 1.

Effect of Epirrita autumnata feeding on foliar oxidative enzyme activities in experiments conducted in natural mountain birch stands. Twelve pairs of birches were selected; one served as a wounded tree and the other as a control tree (outer control). Twenty E. autumnata larvae were released on to two branches of the same trunk (damaged and adjacent undamaged leaves). A third branch from another trunk served as an internal control. Guaiacol PODs and CATs (A), acidic PPO and combined PPO/POD (B) and alkaline PPOs (C) activities were measured 24 h after release of larvae into the foliage. Significant differences between treatments within a tree are marked with different lower-case letters; significant differences between external control and other treatments are marked with asterisks (P ≤ 0·05). Bars indicate s.e. Abbreviations: POD, peroxidase; PPO, polyphenoloxidase; CAT, catalase.

Guaiacol POD activity (ANOVA F = 3·651, P = 0·037) and alkaline PPO activity (ANOVA F = 3·692, P = 0·036) were increased by herbivory compared with values found in the internal control leaves, i.e. leaves collected from a separate, unwounded trunk of the same tree (Fig. 1A and C). Activities were increased significantly in damaged leaves (Tukey HSD: POD, P = 0·028; PPO, P = 0·037). In adjacent unwounded leaves (collected from the same short shoot as the wounded leaves), no statistically significant increase in enzyme activities was detected. Neither acidic PPO nor combined PPO/POD activities (measured with catechol as substrate, pH 5·8) were significantly increased in either wounded or adjacent leaves compared with the internal control (Fig. 1B).

When enzyme activities of the internal control leaves were compared with those of external control leaves, i.e. leaves collected in separate unwounded trees, no significant differences were detected except for the activity of acidic PPO (t-test F = 25·34, P = 0·034, Fig. 1B). Similarly, enzyme activities in adjacent undamaged leaves did not differ from those of external control leaves, except for the activity of acidic PPO (t-test F = 47·833, P = 0·012). By contrast, combined PPO/POD (t-test F = 1·85, P = 0·047) and both acidic (t-test F = 37·82, P = 0·001) and alkaline PPO (t-test F = 2·21, P = 0·024) activities were significantly higher in damaged leaves than in external control leaves (Fig. 1B and C). In addition, POD activity in damaged leaves tended to be higher than in external control leaves (Fig. 1A). Interestingly, the activity of CATs, enzymes that reduce oxidative stress in plants by removing H₂O₂ from the tissues, was significantly decreased in wounded leaves in comparison with external control leaves (t-test F = 1·61, P = 0·003, Fig. 1A).

In situ localization of H₂O₂ accumulation

The specificity of DAB staining for H₂O₂ was tested by incubating wounded leaves in 0·1 % DAB in the absence (Fig. 2A) and presence (Fig. 2B) of catalase. Catalase inhibited staining at the edges of the wound but not staining of surrounding small veins probably because catalase was not able to penetrate efficiently enough into the small veins to prevent the staining reaction totally (Fig. 2B). In the absence of DAB, no colour reaction was detected (Fig. 2C).

Fig. 2.

Staining controls. In a leaf stained with 0·1 % DAB in 10 mm MES in the absence of catalase, H₂O₂ accumulation is shown as a brownish coloration at the edges of the wound (arrow) and in veins (v) surrounding the injury (A) In a leaf stained with 0·1 % DAB in 10 mm MES in the presence of catalase (2000 U), there is no brownish coloration at the edges of the wound but this is found in minor veins (v) surrounding the injury (B) In a DAB control (10 mm MES), no H₂O₂ accumulation is detected (C) Scale bars = 1 mm.

In the outdoor experiment, accumulation of H₂O₂ in natural conditions over 24 h was investigated. In unwounded control leaves, no indications of H₂O₂ accumulation were detected; all samples were similar to those shown in Fig. 3A and B. In wounded leaves, H₂O₂ accumulated in tissues and minor veins around the wounds (Fig. 3C and D). In some leaves, a clear ‘track’ of H₂O₂ from the wound site to the major vein was detected (Fig. 3C and D), suggesting systemic H₂O₂ accumulation along vascular bundles. No visible differences in H₂O₂ accumulation in wounded leaves were detected between the 6-, 12- and 24-h time points (data not shown). The explanation for this lack of difference may be that accumulation of H₂O₂ occurs very rapidly; in contrast to the laboratory test the actual time of the injury was unknown. Unwounded leaves from the same short shoot as the wounded leaves were also collected and stained, but no systemic accumulation of H₂O₂ in these unwounded leaves was detected (data not shown).

Fig. 3.

Hydrogen peroxide accumulation in response to Epirrita autumnata feeding in the outdoor experiment. In control, uninjured leaves, no H₂O₂ accumulation is detected (A and B). In leaves wounded by E. autumnata (C and D), H₂O₂ accumulation is shown as a brownish coloration at the edges of wounds, minor veins (v) and a major vein (mv). No visible differences in the staining pattern of wounded leaves were detected after 6, 12 and 24 h (6 h: C and D). Scale bars = 1 mm (A and C), 100 µm (B and D).

In the indoor experiment, we studied whether the type of wound may cause different patterns in H₂O₂ accumulation in birch leaves. All the treatments—manual wounding (Fig. 4A and B) and feeding by a moth larva (Fig. 4C and D) and a leaf beetle (Fig. 4E and F)—caused the release of H₂O₂ in tissues and minor veins around the injury, but no visible differences were detected between wound-types in the amount of H₂O₂ accumulation. This reaction was rapid; samples were collected and stained immediately after wounding. The leaf beetle, P. polaris, avoided feeding on leaf veins; this is indicated by the reticular feeding marks in Fig. 4E. This might be a strategy by the beetle for avoiding the harmful H₂O₂ that accumulated strongly in vascular bundles. E. autumnata larvae, by contrast, did not show any selective feeding pattern (Fig. 4C). Systemic accumulation of H₂O₂ in major veins similar to that found in the outdoor experiment was not detected in the indoor experiment with excised branches.

Fig. 4.

Hydrogen peroxide accumulation in response to manual wounding and herbivory in the indoor experiment. Leaves were wounded manually with a cork bore (A and B), or by feeding by Epirrita autumnata larvae (C and D) or Phratora polaris beetles (F and F). H₂O₂ accumulation is seen as a brownish coloration at the edges of wounds and major veins (mv) and in minor veins (v). Glands (gl) in major veins (mv) were also stained. Leaf beetles avoided feeding on veins, as indicated by the reticular feeding marks (E and F). Scale bars = 1 mm (A, C and E), 100 µm (B, D and F).

In situ localization of POD activity

DAB staining in the presence of 10 mm H₂O₂ revealed a strong constitutive POD activity: in the outdoor experiment, control and wounded leaves were stained similarly throughout the leaf lamina and surrounded by strongly stained leaf edges (Fig. 5A and B). No visible differences were detected after 6, 12 and 24 h (data not shown). A more detailed examination revealed that POD activity was localized in the lower epidermis (Fig. 5C and D), in the guard cells and tissues around the stomata (Fig. 5D), and in the mesophyll cells, forming cavity-like structures beneath the epidermis (Fig. 5D; arrow). The glands on major veins also showed POD activity (Fig. 5C and D).

Fig. 5.

In situ localization of peroxidase (POD) activity in mountain birch leaves by DAB staining in the presence of 10 mm H₂O₂ in the outdoor experiment. Control leaves (A) and leaves wounded by Epirrita autumnata (B) were similarly coloured by brownish spots throughout the leaf lamina. Leaf edges were especially intensively stained. No visible differences were detected in the staining pattern of wounded leaves after 0, 6, 12 and 24 h (6 h: B). A more detailed examination of control leaves (C and D) shows that guard cells (gc) in the lower epidermis contained high POD activity. Beneath the epidermis are brownish cavity-like structures (arrow in D), indicating POD activity in spongy mesophyll. Glands (gl) on major veins also contained POD activity. Scale bars = 100 µm.

In the indoor experiment, the type of injury had no visible effect on POD activity (Fig. 6). All types of wounds were surrounded with thick, dark brown multi-cell layers, apparently produced by the cross-linking of polymers catalysed by PODs in the presence of surplus H₂O₂. POD activity was localized mainly on the lower epidermis. The edges of veins near the injury also showed POD activity (Fig. 6B, D and F).

Fig. 6.

In situ localization of peroxidase (POD) activity in mountain birch leaves by DAB staining in the presence of 10 mm H₂O₂ in the indoor experiment. Leaves were wounded manually with cork bore (A and B), or by feeding by Epirrita autumnata larvae (C and D) or Phratora polaris beetles (E and F). Injuries are surrounded by thick, dark brown multi-cell layers, apparently produced by cross-linking of polymers catalysed by PODs in the presence of surplus H₂O₂. Leaf beetles avoided feeding on veins, as is apparent from the reticular feeding marks (E and F). Scale bars = 1 mm (A, C and E), 100 µm (B, D and F).

DISCUSSION

Herbivory induces PPO and POD activity but depresses CAT activity in wounded leaves

Feeding by Epirrita autumnata larvae induced both POD and PPO activities in wounded mountain birch leaves and led to the accumulation of H₂O₂ around the wound site. Concomitantly, the activity of CATs, antioxidant enzymes that remove H₂O₂ from plant tissues, was decreased. Similar results were obtained by Bi and Felton (1995), in which herbivory induced the activities of oxidative enzymes, including PODs, and concomitantly decreased the levels of antioxidants, including CAT activity. We did not measure the levels of antioxidants other than CAT, but it is likely that the concomitant increase in H₂O₂ and decrease in CAT activity increased oxidative stress in wounded birch leaves. Transgenic tobaccos with reduced CAT activity were unable to detoxify the accumulation of H₂O₂ under photorespiratory conditions, demonstrating the central role of CAT in the detoxifying of H₂O₂ (Dat et al., 2003). H₂O₂ may have direct defensive properties against herbivores and pathogens, and H₂O₂ and other ROS are suggested to act as feeding deterrents against herbivores (Lamb and Dixon, 1997) and to cause oxidative damage in the herbivore midgut (Bi and Felton, 1995). In addition, H₂O₂ acts as a substrate of PODs, and an increase in substrate availability may consequently increase POD activity (see Thordal-Christensen et al., 1997).

The induction of PPOs by insect damage is a well-documented phenomenon (e.g. Felton et al., 1989; Felton and Duffey, 1991; Stout et al., 1999; Haruta et al., 2001; Tscharntke et al., 2001), and POD activity has also been shown to increase as a response to herbivory or wounding (e.g. Bi and Felton, 1995; Tscharntke et al., 2001; Kranthi et al., 2003). Felton et al. (1989) found a strong negative correlation between PPO activity and the growth of Heliothis zea. In addition, chlorogenic acid, a substrate of PPOs, correlated negatively with the growth of H. zea. Chlorogenic acid is the most abundant phenolic compound found in mountain birch leaves, reaching concentrations of up to 4 % (d. wt) (Nurmi et al., 1996); it is thus the most obvious substrate of birch oxidases.

Accumulation of H₂O₂ probably mediates induction in birch oxidases

Acidic PPO activity was found to be significantly higher in the internal control leaves than in the external controls. Twenty-four hours may have been too short a time to activate systemic induction in the other enzymes. In hybrid poplar, maximal mRNA levels of PPO were detected between 12 and 24 h in wounded leaves; in unwounded leaves induction was observed 36 h after wounding (Haruta et al., 2001). It is possible that different enzymes or isoenzymes are induced with a different delay or even via distinct signalling systems.

H₂O₂ may act as a diffusible signal in mountain birch leaves, mediating the activation of defensive genes and the observed increase in oxidase activities. Plants have parallel, interactive and/or competitive signalling pathways induced by different or even the same types of stress (Fidantsef et al., 1999; Stout et al., 1999; León et al., 2001). The best known are probably the jasmonate (JA) pathway induced by wounding and the salicylic acid (SA) pathway that mediates signalling in defence against pathogens. In both pathways, H₂O₂ plays an important role as a signal molecule mediating the induction of defence genes (Lamb and Dixon, 1997; Fidantsef et al., 1999; Kawano, 2003). In silver birch (Betula pendula), wounding, pathogen inoculation and ozone exposure all induced H₂O₂ production, but the induced gene expression differed between stresses (Pellinen et al., 2002). Generation of H₂O₂ can be produced by different sources; the primary source, however, is suggested to be O₂^−• produced by plasma membrane NADPH-oxidase (e.g. Levine et al., 1994; Orozco-Cárdenas et al., 2001; Laloi et al., 2004; Mittler et al., 2004) but extracellular PODs are also suggested to mediate the production of H₂O₂ (Kawano, 2003).

The accumulation of H₂O₂ was localized in the cells and minor veins around the wound site. This accumulation evidently occurred rapidly, as was found also by Gould et al. (2002), given that sampling in the laboratory experiment was conducted immediately after wounding of the leaves. No systemic H₂O₂ accumulation was detected in adjacent unwounded leaves, similar to that found by Orozco-Cárdenas et al. (2001) in the tomato, but there was a comparable systematic accumulation of H₂O₂ in the major veins of wounded leaves in the outdoor experiment with longer exposure time but not in the indoor experiment when sampling was conducted immediately after the injury. According to the model of Orozco-Cárdenas et al. (2001), H₂O₂ is produced in vascular bundles and diffuses out to mesophyll cells, where it activates the expression of PPO and proteinase inhibitors.

Peroxidase activity together with accumulation of H₂O₂ may represent an efficient strategy to suspend pathogen invasion

The type of wound did not affect the release of H₂O₂ in tissues around the wound site; all wound types caused H₂O₂ accumulation. However, the two types of herbivore fed on leaves with different strategies. The leaf beetle, P. polaris, avoided leaf veins, as shown by the reticular feeding marks. This may be a strategy to avoid the harmful H₂O₂ that accumulates strongly in vascular bundles. Moth larvae, E. autumnata, by contrast, did not show any selective feeding pattern. This is consistent with personal observations by ourselves and other researchers in the outbreak area: the moth larvae fed on all leaves, not only from birches but also from other plants (except for conifer needles), indicating that species is able to detoxify the defensive compounds of different plant species.

Wounding exposes plants to invasion by pathogens, and thus a rapid and an efficient wound-healing system is essential. In the enzyme activity measurements, guaiacol PODs showed a high constitutive activity. This high constitutive level was also apparent in the histochemical in situ DAB staining in both control and wounded leaves in the presence of surplus H₂O₂. POD activity was localized in lower epidermis, guard cells and mesophyll. Similarly, in the leaves of Prunus armeniaca, POD activity was localized in mesophyll cells but was also found in xylem and phloem tissue (Musetti et al., 2005). PODs catalyse the cross-linking of polymers, such as proteins, in the cell wall (Lamb and Dixon, 1997); together with their substrates they efficiently seal wound sites in mountain birch leaves, as seen in the multi-cell brown layers around the wound. Wounds were also surrounded by a brownish layer in the absence of extra H₂O₂, indicating the endogenous production of H₂O₂, but in the presence of 10 mm H₂O₂ wound sealing was much more efficient. According to Thordal-Christensen et al. (1997), the restricting component in DAB staining in tomato is H₂O₂, not POD activity, and this seems to be true in birch as well. It is possible that in the wild, too, the amount of available H₂O₂ is the restricting component in POD-mediated reactions. Thus, an increase in the amount of H₂O₂ would up-regulate POD-mediated reactions such as quinone production and wound sealing. The high POD activity in guard cells and mesophyll cells next to intercellular spaces may represent an efficient strategy to suspend pathogen invasion via stomata and intercellular spaces (see Orozco-Cárdenas et al., 2001). H₂O₂ produced by NADPH oxidase (ABA-mediated pathway) or extracellular PODs (SA pathway) is also involved in the regulation of stomatal closure (Kawano, 2003; Laloi et al., 2004).

CONCLUSIONS

Oxidative changes—an increase in birch oxidases together with an increase in H₂O₂—may form an important front line in mountain birch defence against herbivores and pathogens. Our latest results show that birch phenoloxidases play an important role in defence against multiple pest species (T. Ruuhola et al., unpubl. data). However, herbivores have co-evolved with their host plants and have apparently adapted to feed on plants that produce defensive metabolites (Karban and Agrawal, 2002); in other words, herbivores have different mechanisms to suppress oxidative stress. These include several antioxidative enzymes, such as catalase, and other antioxidants (Felton and Duffey, 1991; Felton and Summers, 1993; Barbehenn et al., 2001). Autumnal moth larvae, like many other Lepidopteran larvae, have an alkaline gut that conducts the oxidation of phenolics and formation of oxygen radicals (Felton et al., 1992; Appel, 1993). Oxidative or other defences of mountain birch seem to be relatively ineffective against E. autumnata, as shown by its high growth rate and fecundity on mountain birch. However, the performance of moth larvae is negatively associated with foliar PPO activities (T. Ruuhola et al., unpublished data). This success at the individual level may in turn play a role in the cyclical population dynamics of E. autumnata, characterized by phases of rapid population growth. The next question is: how do moth larvae cope with birch defences? An efficient ROS detoxifying mechanism could offer some explanation for the seeming ability of E. autumnata to break down the defence of mountain birch so effectively.