Plant responses to abiotic stresses: heavy metal‐induced oxidative stress and protection by mycorrhization

Abstract

The aim of this review is to assess the mode of action and role of antioxidants as protection from heavy metal stress in roots, mycorrhizal fungi and mycorrhizae. Based on their chemical and physical properties three different molecular mechanisms of heavy metal toxicity can be distinguished: (a) production of reactive oxygen species by autoxidation and Fenton reaction; this reaction is typical for transition metals such as iron or copper, (b) blocking of essential functional groups in biomolecules, this reaction has mainly been reported for non‐redox‐reactive heavy metals such as cadmium and mercury, (c) displacement of essential metal ions from biomolecules; the latter reaction occurs with different kinds of heavy metals. Transition metals cause oxidative injury in plant tissue, but a literature survey did not provide evidence that this stress could be alleviated by increased levels of antioxidative systems. The reason may be that transition metals initiate hydroxyl radical production, which can not be controlled by antioxidants. Exposure of plants to non‐redox reactive metals also resulted in oxidative stress as indicated by lipid peroxidation, H ₂ O ₂ accumulation, and an oxidative burst. Cadmium and some other metals caused a transient depletion of GSH and an inhibition of antioxidative enzymes, especially of glutathione reductase. Assessment of antioxidative capacities by metabolic modelling suggested that the reported diminution of antioxidants was sufficient to cause H ₂ O ₂ accumulation. The depletion of GSH is apparently a critical step in cadmium sensitivity since plants with improved capacities for GSH synthesis displayed higher Cd tolerance. Available data suggest that cadmium, when not detoxified rapidly enough, may trigger, via the disturbance of the redox control of the cell, a sequence of reactions leading to growth inhibition, stimulation of secondary metabolism, lignification, and finally cell death. This view is in contrast to the idea that cadmium results in unspecific necrosis. Plants in certain mycorrhizal associations are less sensitive to cadmium stress than non‐mycorrhizal plants. Data about antioxidative systems in mycorrhizal fungi in pure culture and in symbiosis are scarce. The present results indicate that mycorrhization stimulated the phenolic defence system in the Paxillus–Pinus mycorrhizal symbiosis. Cadmium‐induced changes in mycorrhizal roots were absent or smaller than those in non‐mycorrhizal roots. These observations suggest that although changes in rhizospheric conditions were perceived by the root part of the symbiosis, the typical Cd‐induced stress responses of phenolics were buffered. It is not known whether mycorrhization protected roots from Cd‐induced injury by preventing access of cadmium to sensitive extra‐ or intracellular sites, or by excreted or intrinsic metal‐chelators, or by other defence systems. It is possible that mycorrhizal fungi provide protection via GSH since higher concentrations of this thiol were found in pure cultures of the fungi than in bare roots. The development of stress‐tolerant plant‐mycorrhizal associations may be a promising new strategy for phytoremediation and soil amelioration measures.

Antioxidant systems, heavy metals, mycorrhiza, oxidative stress.

Introduction

To date an unprecedented, rapid change in environmental conditions is observed, which is likely to override the adaptive potential of plants, especially that of tree species with their long reproductive cycles. These environmental changes mainly originate from anthropogenic activities, which have caused air and soil pollution, acid precipitation, soil degradation, salinity, increasing UV‐B radiation, climate change, etc. In addition, plants are exposed to natural climatic or edaphic stresses, for example, high irradiation, heat, chilling, late frost, drought, flooding, and nutrient imbalances. Some of these stress factors may fluctuate significantly in intensity and duration on time scales of hours, days, seasons, or years; others may change slowly and gradually affect plant growth conditions. Since plants are sessile organisms and have only limited mechanisms for stress avoidance, they need flexible means for acclimation to changing environmental conditions. In order to improve a plant's protection, it is important to understand the mechanisms contributing to stress tolerance.

A common consequence of most abiotic and biotic stresses is that they result, at some stage of stress exposure, in an increased production of reactive oxygen species ( Polle and Rennenberg, 1993 ). The successive reduction of molecular oxygen to H ₂ O yields the intermediates O ₂^•− , HO ^• and H ₂ O ₂ , which are potentially toxic, because they are relatively reactive compared with O ₂ . Reactive oxygen species may lead to the unspecific oxidation of proteins and membrane lipids or may cause DNA injury. As a consequence, tissues injured by oxidative stress generally contain increased concentrations of carbonylated proteins and malondialdehyde and show an increased production of ethylene ( Dean et al ., 1993 ; Ames et al ., 1993 ).

For a long time reactive oxygen species have been considered mainly as dangerous molecules, whose levels need to be kept as low as possible. Now this opinion is changing rapidly. It has been realized that reactive oxygen species play important roles in the plant's defence system against pathogens (‘oxidative burst’, Alvarez and Lamb, 1997 ; Doke, 1997 ; Bolwell et al ., 2002 ), mark certain developmental stages such as tracheary element formation, lignification and other cross‐linking processes in the cell wall (‘programmed cell death’, Jacobson, 1996 ; Teichmann, 2001 ; Fath et al ., 2002 ) and act as intermediate signalling molecules to regulate the expression of genes ( May et al ., 1998 ; Karpinski et al ., 1999 ; Neill et al ., 2002 ; Vranova et al ., 2002 ). Because of these multiple functions of activated oxygen, it is necessary for cells to control the level of reactive oxygen molecules tightly, but not to eliminate them completely.

The control of oxidant levels is achieved by antioxidative systems. These defence systems are composed of metabolites such as ascorbate, glutathione, tocopherol, etc., and enzymatic scavengers of activated oxygen such as SODs, peroxidases and catalases ( Noctor and Foyer, 1998 ; Asada, 1999 ). The maintenance of ascorbate in its reduced form is achieved by monodehydroascorbate radical reductase (MDAR) and NAD(P)H or ferredoxin as reductant or by the operation of the ascorbate–glutathione pathway ( Foyer and Halliwell, 1976 ; Borraccino et al ., 1986 ; Miyake and Asada, 1994 ). In the latter pathway the reduction of dehydroascorbate is coupled to the oxidation of glutathione (GSH), which, in turn, is reduced by glutathione reductase by oxidation of NADPH ( Foyer and Halliwell, 1976 ). Antioxidant systems and their significance for the acclimation of plants to air pollution and climatic stresses have been reviewed frequently with emphasis on the responses of leaves ( Smirnoff and Pallanca, 1996 ; Polle, 1996 , 1997 , 1998 ; Smirnoff, 1996 ; Noctor and Foyer, 1998 ; Asada, 1999 ). Less attention has been paid to soil‐borne stresses and their effects in roots.

In soils influenced by human activities a range of different problems such as overexploitation, salinity, acidification, and contamination by various pollutants have been reported. Increasing emissions of heavy metals are dangerous because they may get into the food chain with risks for human health ( Lantsy and Mackenzie, 1979 ; Galloway et al ., 1982 ; Angelone and Bini, 1992 ). For the recultivation of degraded soils and the reclamation of industrial sites, stress‐tolerant plants are required. Biotechnological efforts are underway to improve plant stress tolerance and the ability to extract pollutants from the soil with the aim of using plants for soil clean‐up ( Salt et al ., 1995 ). In order to devise new strategies for phytoremediation and improved tolerance, it is important to understand the basic principles as to how the pollutants are taken up and act at the cellular and tissue level. In the present study the occurrence and mode of action of metal pollutants will be briefly reviewed, and the role of antioxidants as defence systems will be discussed. By applying metabolic modelling, oxidant fluxes will be calculated as an estimate of oxidative stress levels and for the prediction of efficient compensation mechanisms in roots. A further question that will be addressed is whether there is evidence that mycorrhizal symbionts improve plant performance under heavy metal stress through increased antioxidative systems.

Occurrence, chemical and physical properties of heavy metals and their mode of action

Heavy metals are defined as metals with a density higher than 5 g cm ⁻³ . 53 of the 90 naturally occurring elements are heavy metals ( Weast, 1984 ), but not all of them are of biological importance. Based on their solubility under physiological conditions, 17 heavy metals may be available for living cells and of importance for organism and ecosystems ( Weast, 1984 ). Among these metals, Fe, Mo and Mn are important as micronutrients. Zn, Ni, Cu, V, Co, W, and Cr are toxic elements with high or low importance as trace elements. As, Hg, Ag, Sb, Cd, Pb, and U have no known function as nutrients and seem to be more or less toxic to plants and micro‐organisms ( Godbold and Hüttermann, 1985 ; Breckle, 1991 ; Nies, 1999 ).

In most terrestrial ecosystems, there are two main sources of heavy metals: the underlying parent material and the atmosphere. The concentrations of heavy metals in soils depend on the weathering of the bedrock and on atmospheric inputs of metals. Natural sources are volcanoes and continental dusts. Anthropogenic activities like mining, combustion of fossil fuels, metal‐working industries, phosphate fertilizers, etc., lead to the emission of heavy metals and the accumulation of these compounds in ecosystems ( Lantsy and Mackensie, 1979 ; Galloway et al ., 1982 ; Angelone and Bini, 1992 ). It has been estimated that, for example, the anthropogenic emissions of Cd are in the range of 30 000 t per year ( di Toppi et al ., 1999 ). In unpolluted soil Cd is present at concentrations of 0.1–0.5 mg kg ⁻¹ , but in Great Britain, in heavily polluted soils of sewage sludge, concentrations of up to 150 mg kg ⁻¹ have been found ( Jackson and Alloway, 1991 ). In the soil, mobile and immobilized fractions have to be distinguished since heavy metals bind to inorganic and organic soil compounds and to the humus. The solubility and mobility of metals is affected by adsorption, desorption, and complexation processes, which in turn are dependent on the soil type.

The availability of heavy metals to plants and, thus, their toxicity depends on complex rhizospheric reactions involving not only exchange processes between soil and plants but also microbial activities. In this respect, mycorrhizal fungi appear to play a central modulating role (see below). Access of heavy metals to bare roots is confined to the first few millimetres of the root tip. Within the cortex the metals are transported in the apoplastic space according to their concentration gradient and also accumulate in the cell walls ( Arduini et al ., 1996 ). Toxic effects are exerted at the plasma membrane and within the cell. Two different uptake routes have been reported: (a) passive uptake, only driven by the concentration gradient across the membrane and (b) inducible substrate‐specific and energy‐dependent uptake ( Nies, 1999 ; Williams et al ., 2000 ). A common transmembrane transporter was found for Cd, Cu, and Ni ( Clarkson and Lüttge, 1989 ). The uptake of these compounds was competitively inhibited by K, Ca, and Mg ( Clarkson and Lüttge, 1989 ). Active and passive transport systems have also been reported for Cd and Ni in roots of spruce and soybean ( Cataldo et al ., 1978 , 1981 ; Godbold, 1991 ). Measurements in the authors’ laboratory indicated that the phase of net accumulation of Cd in the root tip was only short in pine (24 h) suggesting that a steady‐state flux between import and export rates was acquired relatively quickly ( Schützendübel et al ., 2001 ).

To understand the mode of action leading to heavy metal toxicity in living cells, their chemical properties have to be considered. Most of the heavy metals are transition metals with an incompletely filled δ‐orbital present as cations under physiological conditions. The physiological redox range of aerobic cells stretches from −420 mV to +800 mV. Therefore, heavy metals of biological significance can be divided into two groups of redox active and inactive metals. Metals with lower redox potentials than those of biological molecules can not participate in biological redox reactions (Table 1 ). Autoxidation of redox active metals such as Fe ²⁺ or Cu ⁺ results in O ₂^•− formation and subsequently in H ₂ O ₂ and OH ^• production via Fenton‐type reactions. Cellular injury by this type of mechanism is well‐documented for iron ( Halliwell and Gutteridge, 1986 ; Imlay et al ., 1988 ), copper ( Li and Trush, 1993a, b ) as well as other metals ( Jones et al ., 1991 ; Lund et al ., 1991 , 1993 ; Shi and Dalal, 1993 ; Shi et al ., 1993 ).

Another important mechanism of heavy metal toxicity is their ability to bind strongly to oxygen, nitrogen and sulphur atoms ( Nieboer and Richardson, 1980 ). This binding affinity is related to free enthalpy of the formation of the product of metal and ligand. Table 2 shows a range of heavy metal cations with increasing affinity for sulphides and the low solubility of these products. Because of these features, heavy metals can inactivate enzymes by binding to cysteine residues. Direct effects of cadmium on the sulphydryl homeostasis of cells and inhibition of enzymes have been reported for mammalian and animal cells ( Canesi et al ., 1998 ; Chrestensen et al ., 2000 ).

Many enzymes contain metals in positions important for their activity. The displacement of one metal by another will normally also lead to inhibition or loss of enzyme activities. Divalent cations such as Co ²⁺ , Ni ²⁺ , and Zn ²⁺ were found to displace Mg ²⁺ in ribulose‐1,5‐bisphosphate‐carboxylase/oxygenase and resulted in loss of activity ( Wildner and Henkel, 1979 ; van Assche and Clijsters, 1986 ). Displacement of Ca ²⁺ by Cd ²⁺ in the protein calmodulin, important in cellular signalling, led to an inhibition in the calmodulin‐dependent phosphodiesterase activity in radish ( Rivetta et al ., 1997 ).

These examples show that, according to their chemical and physical properties, three different molecular mechanisms of metal toxicity can be distinguished: (a) production of reactive oxygen species by autoxidation and Fenton reaction, (b) blocking of essential functional groups in biomolecules, and (c) displacement of essential metal ions from biomolecules.

Table 1. 

Electrochemical potentials (mV) of heavy metals in aqueous media (pH 7, 25 °C, after Weast, 1984 )

Metal cation	Redox potential (mV)
Zn 2+	−1.18
Cd 2+	−0.82
Ni 2+	−0.65
Pb 2+	−0.55
Cu 2+	−0.26
Fe 2+	+0.35
Hg 2+	+0.43
Ag 2+	+1.57

Table 2. 

Free energy of formation of metal‐sulphides (Δ Ff° ) from free metals in Joules at 25 °C and their solubility (K _sp ) (after Weast, 1984 )

Compound	Ksp	Δ Ff °
Ag 2 S	6.7×10 −50	−9.3
HgS	1.6×10 −52	−11.6
CuS	6.3×10 −36	−11.7
CoS	2.0×10 −25	−19.8
CdS	8.0×10 −27	−33.6
ZnS	1.6×10 −23	−47.4
MnS	2.5×10 −13	−49.9
PbS	8.0×10 −28	−62.2
NiS	1.0×10 −24	−184.9

Heavy metals and antioxidative defences

There is ample evidence that exposure of plants to excess concentrations of redox active heavy metals such as Fe and Cu results in oxidative injury ( De Vos et al ., 1992 ; Gallego et al ., 1996 ; Weckx and Clijsters, 1996 ; Mazhoudi et al ., 1997 ; Yamamoto et al ., 1997 ). The ability of plants to increase antioxidative protection to combat negative consequences of heavy metal stress appears to be limited since many studies showed that exposure to elevated concentrations of redox reactive metals resulted in decreased and not in increased activities of antioxidative enzymes (Table 3 ). Growth with excess Fe resulted in increased O ₂^•− and HO ^• ‐production ( Caro and Puntarulo, 1996 ). Autoxidation and Fenton reaction may cause the oxidative loss of defence enzymes. For example, catalase activity is directly inhibited by O ₂^•− ( Kono and Fridovich, 1982 ). Cu‐Zn‐superoxide dismutase is fragmented by HO ^• ‐radicals ( Casano et al ., 1997 ). If uptake of excess Fe ²⁺ or Cu ⁺ preferentially drives the formation of HO ^• ‐radicals, protection mediated by antioxidative enzymes is unlikely ( Polle, 1997 ). The authors of the present paper have not been able to find literature data providing evidence that elevated levels of antioxidant enzymes protect from excess copper or iron, whereas there are reports that overexpression of iron‐ or copper‐chelators, for example, of metallothioneins and ferritin, protect against metal‐induced oxidative injury ( Fabisiak et al ., 1999 ).

Interestingly, the occurrence of activated oxygen and symptoms of oxidative injury have also been observed in plants exposed to heavy metals, which do not belong to the group of transition metals (Cd: Gallego et al ., 1996 ; Lozano‐Rodriguez et al ., 1997 ; Chaoui et al ., 1997 ; Cho and Park, 1999 ; Piqueras et al ., 1999 ; Romero‐Puertas et al ., 1999 ; Schützendübel et al ., 2001 ; Zn: Weckx and Clijsters, 1997 ; Prasad et al ., 1999 ; Rao and Sresty, 2000 ; Ni: Baccouch et al ., 1998 ; Rao and Sresty, 2000 ). Since these metals do not interfere directly with cellular oxygen metabolism, the question arises as to the reasons of the observed oxidative stress. Exposure to heavy metals also provoked pronounced responses of antioxidative systems, but the direction of the response was dependent on the plant species and tissue analysed, the metal used for the treatment and the intensity of the stress (Table 3 ). However, some common reaction patterns can be found. In most cases, exposure to heavy metals initially resulted in a severe depletion of GSH (Cd: Rauvolfia serpentina : Grill et al ., 1987 ; pine: Schützendübel et al ., 2001 ; carrot: di Toppi et al ., 1999 ; tobacco: Vögeli‐Lange and Wagner, 1996 ; Cu: Silene cucubalus : de Vos et al ., 1992 ; Cu or Cd: Arabidopsis : Xiang and Oliver, 1998 ; Ni and Zn: pigeonpea: Rao and Sresty, 2000 ; Fe, Cu or Cd: sunflower leaves: Gallego et al ., 1996 ). This is a common response to Cd caused by an increased consumption of glutathione for phytochelatin production ( Zenk, 1996 ; Mehra and Tripathi, 1999 ). The significance of phytochelatins for protection from heavy metals has frequently been reviewed ( Rauser, 1995 ; Zenk, 1996 ; Mehra and Tripathi, 1999 ) and, therefore, will be summarized here only briefly. Phytochelatins sequester heavy metals. For Cd, the formation of Cd–thiolate (Cd–S) complexes in phytochelatins has been shown ( Strasdeit et al ., 1991 ). The chelated metals are transported to the tonoplast, taken up by active transport systems, and deposited in the vacuole ( Tommasini et al ., 1998 ; Rea, 1999 ). This mechanism contributes to the protection from heavy metal toxicity in several plant species and in some fungi as well ( Ishikawa et al ., 1997 ). In pine, which is a relatively Cd‐sensitive species, the vacuolar Cd‐concentrations in cells of root tips were as high as 20 mM, even though the exposure medium contained only 50 μM Cd (Fritz, unpublished data). In general, the glutathione pool recovered after prolonged Cd‐exposure, frequently to levels above those of controls ( Vögeli‐Lange and Wagner, 1996 ; Xiang and Oliver, 1998 ; Arisi et al ., 2000 ; Schützendübel et al ., 2001 ). The ability to synthesize glutathione appears to be crucial for protection from cadmium, as shown by the increased tolerance of plants with elevated levels of GSH as well as a decreased tolerance in plants with diminished levels of GSH ( Howden et al ., 1995 ; Zhu et al ., 1999 a, b ). However, the threshold required to enhance protection seems to be plant‐specific since the amelioration of growth under cadmium stress by elevated GSH has not been observed in all cases ( Arisi et al ., 2000 ).

Since glutathione is also an important component for the redox balance of the cell, as it is involved in the regulation of the cell cycle, the detoxification of oxidants, and acts as a transport form of reduced sulphur ( Bergmann and Rennenberg, 1993 ; May et al ., 1998 ; Noctor and Foyer, 1998 ; Vernoux et al ., 2000 ), it may be suspected that a short‐term lack of GSH may favour the accumulation of reactive oxygen and disturb developmental processes. The idea, that Cd and perhaps also other toxic metals, act in cells through a depletion of antioxidative defences is further supported by the observation that glutathione reductase, ascorbate peroxidase and catalase activities were inhibited at time scales similar to those found for the depletion of GSH (Fig. 1 ). Heavy metal‐induced loss in glutathione reductase has frequently been observed: in pea by Zn, Cu and Fe ( Bielawski and Joy, 1986 ), in sunflower by Fe, Cu and Cd ( Gallego et al ., 1996 ), in Lemna minor by Cu ( Teisseire and Guy, 2000 ). Glutathione reductase contains a highly conserved disulphide bridge between Cys ⁷⁶ and Cys ⁸¹ ( Creissen et al ., 1992 ; Lee et al ., 1998 ), which may undergo cleavage by heavy metals. The sensitivity of glutathione reductase to direct inhibition by Cd was shown in in vitro assays (Fig. 2 ). If EDTA, a chelator of divalent cations, was added, glutathione reductase activity was recovered (data not shown). Currently, it is unknown whether roots exposed to Cd accumulate sufficiently high free concentrations of this compound for direct interaction with glutathione reductase in situ . However, it is tempting to speculate that the initial decrease in glutathione reductase activity may have been caused by Cd‐binding when the concentrations of GSH were severely diminished (Fig. 1B ). The activities of defence enzymes recovered after prolonged Cd‐exposure (Fig. 1C ). The observed increase of thiol concentrations above those of controls (Fig. 1C ) may be necessary to protect sensitive enzymes.

‘Unspecific’ peroxidases, i.e. enzymes oxidizing phenolic substrates such as guaiacol, were also affected by exposure to cadmium (Table 3 ). In pine roots, ‘unspecific’ peroxidases were not inhibited by Cd, but increased slowly with a time pattern clearly distinct from that observed for the constituents of the SOD–ascorbate–glutathione pathway ( Schützendübel et al ., 2001 ). ‘Unspecific’ peroxidase activities were elevated in root tips, which showed increased concentrations of phenolics and lignification in response to Cd ( Schützendübel et al ., 2001 ). The final result of Cd in roots resembles that of plant tissues exposed to pathogens. During pathogenic attack, plant cells display an increased production of reactive oxygen species (oxidative burst) followed by secondary defence reactions ( Alvarez and Lamb, 1997 ). These responses lead to mechanical strengthening of cell walls including lignification to prevent intrusion of the pathogen ( Alvarez and Lamb, 1997 ). The scheme that Cd induces defence pathways resulting in cell wall rigidification is also consistent with the observation that root growth stops or is significantly inhibited after Cd‐exposure ( Punz and Sieghardt, 1993 ; Kahle, 1993 ). It may be suspected that, as a result of Cd‐induced defence reactions, lignified root tips have also lost their capacity for nutrient uptake, and, thus, their ability to sustain plant growth. This would lead to growth retardation at the whole‐plant level (Fig. 3 ).

Fig. 1. 

Antioxidative enzymes and antioxidants in pine roots ( Pinus sylvestris ). (A) Mean volume activities of enzyme and concentrations of antioxidants were calculated on the basis of the water content of the roots (93%) of control plants. Cat, catalase; SOD*, superoxide dismutase (indicated as concentration); APX, ascorbate peroxidase; MDAR, monodehydroascorbate radical reductase; DAR, dehydroascorbate reductase; GR, glutathione reductase; ASC, ascorbate (black)+dehydroascorbate (grey); GSH, GSH (black)+GSSG (grey). Data are means of five replicates. Nd, not detected. (B) Changes in enzyme activities and antioxidant concentrations relative to controls after 6 h (B) and 96 h (C) exposure to 50 μM cadmium.

Fig. 2. 

Inhibition of glutathione reductase by cadmium. Glutathione reductase (EC 1.6.4.2) from bakers yeast was incubated with different concentrations of Cd for 30 min and then tested. Data are means of three replicates (±SD). Different letters indicate significant differences at P ≤0.05 as determined by ANOVA followed by a multiple range test (LSD).

Fig. 3. 

Growth inhibition in pine seedlings ( Pinus sylvestris L.) exposed to different Cd‐concentrations. Pine seedlings were grown in sand‐culture under a light regime of 17 h and 200 μE at 20 °C. Nine‐week‐old plants were exposed for 21 d to different concentrations of CdSO ₄ in the nutrient solution. Data are means of six replicates (±SD). Different letters indicate significant differences at P ≤0.05 as determined by ANOVA followed by a multiple range test (LSD).

Table 3. 

Relative activities of antioxidative enzymes in different plant species exposed to heavy metals

Enzyme activities of controls were set as 100%. When it was not possible to determine relative enzyme activity, the trend is given (+=increase).

Enzyme	Compound	Concentration (μM)	Species/organ	Relative enzyme activity (%)	Exposure time (h)	Reference
Superoxide	Hg		Lycopersicon esculentum			Cho and Park, 2000
dismutase		10	Roots	120	240
		10	Leaves	150	240
	Cd		Hordeum vulgare			Patra and Panda, 1998
		10	Leaves	+	48
	Cd		Helianthus annuus			Gallego et al ., 1999
		50	Leaves	100	96
	Cd		Helianthus annuus			Gallego et al ., 1996
		500	Leaves	77	12
	Cd		Nicotiana tabacum			Piqueras et al ., 1999
		5000	BY2 cell culture	180	0.08
	Fe		Phaseolus vulgaris			Shainberg et al ., 2000
		900	Leaves	+	24
	Cu		Helianthus annuus			Chaoui et al ., 1997
		500	Leaves	118	12
	Cu		Phaseolus vulgaris			Weckx and Clijsters, 1996
		630	Leaves	83	48
	Zn		Cajanus cajan			Rao and Sresty, 2000
		5000	Leaves	230	144
	Zn		Brassica juncea			Prasad et al ., 1999
		5000	Shoots	220	240
Ascorbate	Cd		Phaseolus vulgaris			Chaoui et al ., 1997
peroxidase		5	Roots	100	96
	Cd		Phaseolus vulgaris			Chaoui et al ., 1997
		5	Leaves	210	96
	Cd		Helianthus annuus			Gallego et al ., 1999
		50	Leaves	117	96
	Cd		Helianthus annuus			Gallego et al ., 1996
		500	Leaves	75	12
	Cd		Nicotiana tabacum			Piqueras et al ., 1999
		5000	BY2 cell culture	170	0.16
	Fe		Phaseolus vulgaris			Shainberg et al ., 2000
		900	Leaves	480	24
	Cu		Phaseolus vulgaris			Gupta et al ., 1999
		15	Roots	130	48
	Cu		Lemna minor			Teisseire and Guy, 2000
		10		71	24
	Cu		Lycopersicon esculentum			Mazhoudi et al ., 1997
		50	Roots	100	168
			Leaves	38
	Cu		Helianthus annuus			Gallego et al ., 1996
		500	Leaves	50	12
	Zn		Phaseolus vulgaris			Gallego et al ., 1999
		612	Roots	+	48
	Zn		Phaseolus vulgaris			Chaoui et al ., 1997
		100	Leaves	144	96
	Zn		Brassica juncea			Prasad et al ., 1999
		5000	Shoots	260	240
Catalase	Hg		Phaseolus aureus			Shaw, 1995
		5	Leaves	113	48
			Roots	0	48
	Hg		Lycopersicon esculentum			Cho and Park, 2000
		10	Leaves	100	240
			Roots	140	240
	Cd		Phaseolus vulgaris			Chaoui et al ., 1997
		5	Roots	75	96
	Cd		Phaseolus vulgaris			Chaoui et al ., 1997
		5	Leaves	75	96
	Cd		Hordeum vulgare			Gallego et al ., 1999
		50	Leaves	120	96
	Cd		Helianthus annuus			Patra and Panda, 1998
		100	Leaves	+	48
	Cd		Helianthus annuus			Gallego et al ., 1996
		500	Leaves	70	12
	Cd		Nicotiana tabacum			Piqueras et al ., 1999
Catalase		5000	BY2 cell culture	75	0.08
	Cu		Lemna minor			Teisseire and Guy, 2000
		10		347	24
	Cu		Lycopersicon esculentum			Mazhoudi et al ., 1997
		50	Roots	76	168
			Leaves	100
	Cu		Helianthus annuus			Gallego et al ., 1996
		500	Leaves	33	12
	Cu		Oryza sativa			Chen and Kao, 1999
		10000	Leaves	46	24
	Zn		Cajanus cajan			Rao and Sresty, 2000
		5000	Leaves	75	144
	Zn		Brassica juncea			Prasad et al ., 1999
		5000	Shoots	1200	240
Glutathione	Cd		Phaseolus vulgaris			Chaoui et al ., 1997
reductase		5	Leaves	80	96
	Cd		Helianthus annuus			Gallego et al ., 1999
		50	Leaves	120	96
	Cd		Helianthus annuus			Patra and Panda, 1998
		500	Leaves	80	12
	Fe		Phaseolus vulgaris			Shainberg et al ., 2000
		900	Leaves	155	24
	Cu		Lemna minor			Teisseire and Guy, 2000
		10		68	24
	Cu		Phaseolus vulgaris			Gupta et al ., 1999
		15	Roots	119	24
	Cu		Helianthus annuus			Gallego et al ., 1996
		500	Leaves	54	12
	Cu		Oryza sativa			Chen and Kao, 1999
		10000	Leaves	33	24
MDHAR	Zn		Phaseolus vulgaris			Chaoui et al ., 1997
		100	Leaves	150	96
	Zn		Brassica juncea			Prasad et al ., 1999
		5000	Shoots	258	240
DHAR	Cd		Phaseolus vulgaris			Gupta et al ., 1999
		15	Roots	120	96
	Cd		Brassica juncea			Prasad et al ., 1999
		5000	Shoots	280	240
	Cd		Helianthus annuus			Gallego et al ., 1999
		50	Leaves	120	96
	Cd		Helianthus annuus			Gallego et al ., 1996
		500	Leaves	75	12
	Cu		Helianthus annuus			Gallego et al ., 1996
		500	Leaves	50	12
	Zn		Brassica juncea			Prasad et al ., 1999
		5000	Shoots	570	240
POD	Hg		Phaseolus aureus			Shaw, 1995
		5	Leaves	210	24
	Hg		Lycopersicon esculentum			Cho and Park, 2000
		10	Roots	2.5	240
			Leaves	100	240
	Cd		Helianthus annuus			Gallego et al ., 1999
		50	Leaves	116	96
	Cd		Nicotiana tabacum			Piqueras et al ., 1999
		5000	BY2 cell culture	100	0.16
	Cu		Lycopersicon esculentum			Mazhoudi et al ., 1997
		50	Roots	130	168
			Leaves	100
	Cu		Lemna minor			Teisseire and Guy, 2000
		10		166	24
	Zn		Phaseolus vulgaris			Chaoui et al ., 1997
		100	Leaves	110	96
	Zn		Brassica juncea			Prasad et al ., 1999
		5000	Shoots	400	240
	Zn		Cajanus cajan			Rao and Sresty, 2000
		5000	Leaves	156	144

Metabolic modelling as a means to predict changes in oxidant levels from measured Cd‐induced changes in ‘antioxidative capacities’

An intriguing question is whether Cd induces an ‘oxidative burst’ similar to that reported for pathogens or whether the concentrations of reactive oxygen increase because of the initial depletion of GSH and inhibition of protective enzymes. To find out whether the observed decreases in antioxidative defences would be sufficient to explain H ₂ O ₂ ‐accumulation, quantitative estimates of the ‘antioxidative capacity’ are necessary. As a first step towards an assessment of the oxidant scavenging efficiency, mean concentrations of antioxidants and volume‐related activities of defence enzymes in healthy root tips were calculated (Fig. 1A ). By contrast to needles, pine roots contained relatively low concentrations of soluble antioxidants. Even more striking was that the redox status of the antioxidant pool was also very low, which means that the concentrations of reduced antioxidants (ascorbate and GSH) were low relative to their oxidized counterparts (dehydroascorbate and GSSG).

A metabolic model, which can be used to calculate oxidant scavenging activity ( Polle, 2001 ), suggested that the ratio of GSH/GSSG was especially sensitive to changes in oxidative stress and, thus, reflects the steady‐state flux of oxidants and reductants under normal conditions. Given this presumption, the measured concentrations of antioxidants and activities of protective enzymes and their known biochemical properties ( Polle, 2001 ) can be used to provide a semi‐quantitative estimate of the ‘antioxidative capacity’ and intrinsic oxidative stress under ‘normal’ conditions. To find the stress rate, which would result in the measured mean GSSH/GSH ratio, an increasing production rate of O ₂^•− radicals was simulated (as in Polle, 2001 ). This condition, indicating mean intrinsic stress in ‘normal’ root tips, was fulfilled at a O ₂^•− production rate of 93 μM s ⁻¹ (Fig. 4 ). Currently, this model is coarse because it describes only an average situation not taking into account the subcellular distribution of antioxidant systems and possible differences in intrinsic stress exposure. For comparison, O ₂^•− production rates of 120–250 μM s ⁻¹ were estimated in chloroplasts under ‘normal’ conditions and up to 720 μM s ⁻¹ under stress ( Asada, 1999 ). These considerations show that the above model provides an estimate in the appropriate range.

In a second modelling step, a situation was envisaged where the ‘antioxidative capacity’ was severely diminished after Cd‐exposure (as indicated for 6 h of Cd‐treatment in Fig. 1B ). It was assumed that the production rate of O ₂^•− remained unchanged (93 μM s ⁻¹ ), i.e. no ‘oxidative burst’. When the model was run under these conditions, a significant accumulation of H ₂ O ₂ was predicted (Fig. 5C ), while the steady‐state concentrations of O ₂^•− and monodehydroascorbate radicals were somewhat decreased (Fig. 5A , B ) because of increased activities of superoxide dismutase and monodehydroascorbate radical reductase (Fig. 1B ). In fact, accumulation of H ₂ O ₂ has been observed in Cd‐exposed roots ( Schützendübel et al ., 2001 ) and in Cd‐exposed tobacco suspension cultures ( Piqueras et al ., 1999 ). It was suggested that Cd triggered an ‘oxidative burst’ as in pathogenesis because they detected H ₂ O ₂ in the culture medium ( Piqueras et al ., 1999 ). However, they also found a significant inhibition of catalase and since H ₂ O ₂ is membrane permeable, the site of H ₂ O ₂ generation in response to Cd may not be totally clear.

In conclusion, the above estimates suggest that a significant intracellular H ₂ O ₂ accumulation can be expected after Cd exposure, simply because of the Cd‐induced depletion of GSH and inhibition of antioxidative enzymes. In pine roots, H ₂ O ₂ was detected at an early stage (6 h after Cd addition), when the roots still appeared visibly fully viable and no lipid peroxidation was found ( Schützendübel et al ., 2001 ). H ₂ O ₂ disappeared within some hours but, thereafter, the differentiation of protoxylem elements became apparent in unusual places of the previous elongation zone of root tips ( Schützendübel et al ., 2001 ). Taking all these observations together, the following hypothetical framework may be suggested (Fig. 6 ): Cd induces a transient loss in ‘antioxidative capacity’, perhaps accompanied by a stimulation of oxidant producing enzymes, which results in intrinsic H ₂ O ₂ accumulation. H ₂ O ₂ , then, would act as a signalling molecule triggering secondary defences. These, in turn, would cause an untimely cell wall rigidification and lignification, thereby, decreasing cellular viability and finally resulting in cell death (Fig. 6 ). This view would be clearly distinct from the alternative idea that Cd results in unspecific necrosis and is also supported by the observation that Cd‐exposed cells show a distinct pattern of DNA‐fragmentation typical for programmed cell death ( Fojtova and Kovarik, 2000 ).

Fig. 4. 

Simulated changes in the steady‐state concentrations of GSSG by increasing O ₂^•− production rates. The black line indicates a steady‐state O ₂^•− production rate, which results in the initial GSSG concentration. The calculations are based on the model SHAG‐ENZ ( Polle, 2001 ), which was enlarged by a component for catalase. The activities of antioxidant enzymes and concentrations of antioxidants shown in Fig. 1A were used for the model calculations. It was assumed that the measured enzyme activities reflected V_max . The concentration of SOD was calculated by conversion of the measured units to a concentration on the basis of the relationship: 500 units nmol ⁻¹ of enzyme protein. It was assumed that the supply of NAD(P)H was not limiting.

Fig. 5. 

Simulated oxidant concentrations in the absence and presence of 50 μM cadmium. (A) Superoxide radials, (B) monodehydroascorbate radical, (C) hydrogen peroxide. The calculations were performed as described in Fig. 4 using activities of enzymes and concentrations of antioxidants in controls or those obtained after 6 h of cadmium exposure.

Fig. 6. 

Hypothetical view of cadmium action on the cellular redox control; for further explanations, see text.

Heavy metals and stress responses in mycorrhizal symbiosis

Under natural conditions, roots of many plant species, especially those of trees are associated with mycorrhizal symbionts. This modifies the response of plants to heavy metals significantly. Several studies have dealt with a possible alleviation of metal toxicity by mycorrhization, but only a few presented direct evidence for such effects ( Hartley et al ., 1997 ; Leyval et al ., 1997 ; Jentschke and Godbold, 2000 ). In the present study non‐mycorrhizal pine seedlings in the presence of Cd‐concentrations ≥15 μM showed 35% diminished biomass as compared to controls (Fig. 3 ). By contrast, a remarkable protection of plant performance against the negative effects of Cd, i.e. no biomass reduction was observed, when the seedlings were 80% associated with a strain of the mycorrhizal fungus Paxillus involutus isolated from a heavy metal‐polluted site (Fig. 7 ). It is still unclear whether the observed alleviation is a consequence of better nutrition, a fungal influence on the physiological stress reaction of the plant or simply hindered access of heavy metals to the root surface caused by the fungal sheath around the root surface ( Jentschke and Godbold, 2000 ). In Hebeloma crustuliniforme the latter suggestion has some support ( Frey et al ., 2000 ), whereas Paxillus involutus also shows significant Cd accumulation in the vacuole ( Blaudez et al ., 2000 ).

Little is known about the heavy metal‐induced stress responses of mycorrhiza‐building basidiomycetes in pure cultures and in association with their hosts. In most fungi, metallothioneins are induced to detoxify the metals in a reaction similar to that found in animal cells ( Mehra and Winge, 1991 ). In yeast, heavy metals also caused oxidative stress ( Mannazu et al ., 2000 ). Some scarce data also suggest that heavy metals affect antioxidative systems in mycorrhizal fungi. For example, an inhibition of Mn‐SOD was found in pure cultures of Rhizopogon roseolus treated with 300 μM Cd ( Miszalski et al ., 1996 ). By contrast, an increase in Mn‐SOD activity was found in Cd‐treated Paxillus involutus cultures ( Jacob et al ., 2001 ). The antioxidative systems of mycorrhizal fungi revealed important differences in comparison with plant tissues. For example, in pure cultures of Laccaria laccata , Suillus bovinus , and Paxillus involutus typical ‘unspecific’ peroxidase activities were not detected ( Münzenberger et al ., 1997 ; Schützendübel et al ., 2001 ) neither was ascorbate peroxidase activity or ascorbate as a potential substrate (T Ott and A Schützendübel, unpublished data). Much higher concentrations of glutathione were found in pure cultures of Suillus bovinus and Paxillus involutus than in pine roots (2–10 μmol g ⁻¹ fresh weight compared with 0.2–1 μmol g ⁻¹ fresh weight plant tissues, Schützendübel et al ., 2001 ).

The question arises as to whether the whole‐mycorrhizal association and each individual partner ( Paxillus–Pinus) exposed to Cd at concentrations, which did not result in a loss of the degree of mycorrhization and only in small growth reduction (Fig. 7 ), show stress reactions similar to those found in bare roots (Fig. 1B ). Initially ‘total’ SOD activities in mycorrhizal roots were similar to those of non‐mycorrhizal roots and remained unaffected by Cd. In the mycorrhizal roots SOD activities increased with time, whereas in the non‐mycorrhizal roots and in Cd‐treated roots (±mycorrhiza) SOD activities remained low after prolonged Cd‐exposure (Fig. 8B ). The activities of ‘unspecific’ peroxidases, which can be used as a marker for the root‐specific response in the mycorrhizal association, initially was neither affected by the fungal symbiont nor by Cd‐exposure (Fig. 8A ). However, after 14 d, POD activities were increased in non‐mycorrhizal Cd‐exposed roots (as observed previously in hydroponically grown bare roots) but not in mycorrhizal Cd‐exposed roots (Fig. 8A ). This observation suggests that the stress reaction is diminished or perhaps the stress not perceived in mycorrhizal roots.

Analysis of the effects of mycorrhization and Cd‐exposure on soluble and cell wall‐bound phenolics in the Paxillus–Pinus symbiosis supports this idea (Fig. 9 ). Cd‐treatment resulted in increased concentrations of ‘total’ soluble phenolics only in non‐mycorrhizal roots but not in mycorrhizal roots (Fig. 9 ). However, mycorrhizal roots generally contained elevated concentrations of soluble phenolics as compared with non‐mycorrhizal roots (Fig. 9 ). A pattern similar to that of ‘total’ soluble phenolics was found for kaempherol‐3‐glucoside (Fig. 10A ), whereas the major plant phenolic acid, catechin, remained unaffected by both Cd‐exposure and the presence or absence of the mycorrhizal symbiont (Fig. 10A ). Increases in free p ‐coumaric acid were found in response to mycorrhization as well as in response to Cd (Fig. 10A ). Interestingly, in non‐mycorrhizal roots free ferulic acid was not detected either in the presence or absence of Cd (Fig. 10A ). By contrast, mycorrhizal roots contained low, but consistently detectable concentrations of free ferulic acid, which were slightly increased after exposure to Cd (Fig. 10A ). Since none of the phenolics were found in pure cultures of Paxillus , these compounds mark root‐specific differential responses to mycorrhization and Cd. The observed increases in ferulic acid and p ‐coumaric acid are particularly interesting in this respect because they indicate that the root part in the mycorrhizal association displayed a specific new response to Cd, which was not found in non‐mycorrhizal Cd‐exposed roots. Both ferulic acid and p ‐coumaric acid play roles in cell wall rigidification because they may become cross‐linked with lignins, proteins or carbohydrate residues decreasing the extensibility of walls ( Fry, 1986 ). Major wall‐bound phenolics were 3,4‐dihydroxybenzoic acid, p ‐coumaric acid and ferulic acid (Fig. 10B ). 3,4‐dihydroxybenzoic acid appeared at elevated concentrations in mycorrhizal roots (Fig. 10B ). The concentration of wall‐bound ferulic acid was neither affected by mycorrhization nor by Cd (Fig. 10B ), whereas bound p ‐coumaric acid increased specifically in mycorrhizal Cd‐exposed roots (Fig. 10B ).

Taken together, these results show that Paxillus–Pinus mycorrhizal associations contained higher concentrations of secondary metabolites than non‐mycorrhizal roots. This suggests that mycorrhization stimulated the defence systems. However, the Cd‐induced changes in mycorrhizal roots were absent or smaller than those in non‐mycorrhizal roots. These observations suggest that although changes in rhizospheric conditions are perceived by the root part of the symbiosis, the typical Cd‐induced stress response was significantly buffered. The mechanism by which mycorrhization protects from Cd is unclear. Metallothioneins may be involved, but this has not been investigated in mycorrhiza‐building basidiomycetes. Other chelators of heavy metals such as excreted organic acids etc. as well as binding Cd in the Hartig net, may also be involved as a protective mechanism. Since much higher concentrations of glutathione were found in the mycorrhizal fungi Suillus bovinus and Paxillus involutus than in bare pine roots, it is also possible that roots in mycorrhizal associations are ‘armed’ with a powerful physiological defence against Cd. The use of stress‐tolerant mycorrhizal fungi may be a promising strategy to develop tools for soil reclamation and amelioration.

Fig. 7. 

Mycorrhization frequency of root tips (A) and fresh mass (B) of 14‐week‐old P. sylvestris – Paxillus involutus associations treated for 2 weeks with different concentrations of CdSO ₄ in the nutrient solution. For growth conditions, see Fig. 3 . Mycorrhizal root tips and non‐mycorrhizal root tips of each plant were counted under a binocular microscope and the degree of mycorrhization was estimated as number of mycorrhizal root tips per total number of root tips ×100. Data are means of six replicates (±SD). The asterix indicates a significant difference at P ≤0.05 as determined by ANOVA followed by a multiple range test (LSD).

Fig. 8. 

POD (EC 1.11.1.7) activity (A) and SOD (EC 1.15.1.1) activity (B) in mycorrhizal root tips of 14‐week‐old P. sylvestris–Paxillus involutus mycorrhiza (myc) and non‐mycorrhizal pine roots (nm) of control seedlings or presence of 50 μM of CdSO ₄ in the nutrient solution for 1 d and 14 d (myccd, mycorrhizal seedlings+cadmium; nmcd, non‐mycorrhizal seedlings+cadmium). For growth conditions see Fig. 3 . Data are means of six replicates (±SD). Different letters indicate significant differences at P ≤0.05 as determined by ANOVA followed by a multiple range test (LSD).

Fig. 9. 

Soluble phenolics in mycorrhizal root tips of 14‐week‐old P. sylvestris–Paxillus involutus mycorrhizal (myc) and non‐mycorrhizal pine roots (nm) of control seedlings or presence of 50 μM of CdSO ₄ in the nutrient solution for 1 d and 14 d (myccd, mycorrhizal seedlings+ cadmium; nmcd, non‐mycorrhizal seedlings+cadmium). For growth conditions see Fig. 3 . Data are means of six replicates (±SD). Different letters indicate significant differences at P ≤0.05 as determined by ANOVA followed by a multiple range test (LSD).

Fig. 10. 

Soluble (A) and cell wall‐bound (B) phenolic compounds in mycorrhizal root tips of 14‐week‐old P. sylvestris–Paxillus involutus mycorrhizal (myc) and non‐mycorrhizal pine roots (nm) of control seedlings or presence of 50 μM of CdSO ₄ in the nutrient solution for 1 d and 14 d (myccd, mycorrhizal seedlings+cadmium, nmcd non‐mycorrhizal seedlings+cadmium). Abbreviations: Cat, catechin; Kä, kaempferol‐3‐glucoside; Ca, p ‐coumaric acid; Fa, ferulic acid; DHB, 3,4‐dihydroxybenzoic acid. For growth conditions see Fig. 3 . Data are means of six replicates (±SD); nd; not detected. Different letters indicate significant differences at P ≤0.05 as determined by ANOVA followed by a multiple range test (LSD).

We are grateful to Dr E Fritz and T Ott (Forest Botanical Institute, University of Göttingen) for communicating unpublished data, to D Godbold (School of Biological Sciences, University of Bangor) for helpful discussions and to C Kettner and C Rudolf for technical assistance. We acknowledge financial support by the European Community.