Introduction

Iron deficiency chlorosis is a common abiotic stress affecting plants in many areas of the world (Victor Del Rio et al. 2008), particularly in calcareous soils. In such soils, bicarbonate has been regarded to as a major factor inducing Fe deficiency (Mengel 1994). In fact, HCO3 − anions impair both Fe uptake by roots and its translocation inside the plants to the different organs (Römheld and Marschner 1986). In general, Fe deficiency chlorosis is associated with high pH in calcareous soils and low soil Fe availability (Chen et al. 2007). Dicots and non-graminaceous monocots (Strategy I plants) respond to Fe limitation by inducing specific mechanisms to mobilize insoluble Fe compounds. The main adaptive responses to Fe deficiency are: 1) rhizosphere acidification through the activation of the plasma membrane H+-ATPase of root cells (Hinsinger et al. 2003); 2) enhanced reduction of Fe3+ to Fe2+ at the plasma membrane level by a NADPH-dependent Fe (III)-chelate reductase (FCR) (Chaney et al. 1972) and 3) the increased Fe3+ transport activity across the plasma membrane (Schikora et al. 2006). It has been found that the glycolytic metabolism is involved in the response to this nutritional disorder. The increase found in PEPC activity under Fe deficiency condition could be linked to the biosynthesis of substrates for FCR and H+-ATPase and organic acids (Zocchi 2006). In fact, the activation of Fe3+ reduction and the H+ extrusion in Fe-deficient plants need an increased rate of NAD(P)H and ATP regeneration in comparison with non Fe-deficient ones.

A common consequence of Fe deficiency is the increased production of ROS, leading to major cellular damages such as DNA alteration, oxidation of proteins and lipid peroxidation (De Vos et al 1992). Recently, a few studies have explored the role of secondary metabolic pathways in plant response to oxidative stress. A case in point is the phenylpropanoid pathway, which is responsible for the synthesis of polyphenols (Dixon and Paiva 1995). Phenolic acids are among the most widespread classes of secondary metabolites and are important in the plant-soil systems (Hu et al. 2005). They are mainly synthesized to protect plants from ROS, their antioxidant activity being due to their redox properties (Arora et al. 2000). Moreover they are involved in the response to Fe deficiency in many Strategy I species, acting as reductants and chelators of Fe(III) when released in the rhizosphere (Olsen et al. 1981; Römheld and Marschner 1983). Some of the key enzymes catalyzing the biosynthesis of polyphenols include shikimate dehydrogenase (SKDH), shikimate kinase (SK), and phenylalanine ammonia lyase (PAL).

Many studies have focused on the glycolytic metabolism responses and antioxidant defence systems of plants to Fe deficiency (Espen et al. 2000; Ranieri et al. 1999). On the contrary, less information is available so far on the possibility to use these biochemical parameters as determinant to select Fe-efficient genotypes to Fe chlorosis. With this aim in the first part of the work we investigated the possibility to use glycolytic metabolic responses as biochemical markers for the diagnosis of Fe deficiency tolerance in two lines of Medicago ciliaris: the tolerant line TN11.11 and the sensible one TN8.7. The second part of the work deals with the role of secondary metabolism in the response of plants to Fe deficiency. The two lines were selected on the basis of their different tolerance to Fe deficiency, which was assessed in previous works (M’sehli et al. 2008).

Materials and methods

Plant material and growth conditions

Two lines of Medicago ciliaris L. were used: the TN11.11 line (from Mateur) and the TN8.7 one (from Soliman). The lines were created by three generations of spontaneous selfing in the greenhouse. Seeds were obtained from the Laboratory of Interaction Legume-microorganisms, Biotechnology Center at the Technopark of Borj-Cedria (CBBC), Tunisia.

Seeds were germinated and grown for 3 days in Agriperlite moistened with 0.1 mM CaSO4. Three-day-old seedlings were transferred to a half strength aerated nutrient solution for 7 days and then similar sized seedlings were selected and cultured as groups of ten plants in 10 L of full strength aerated nutrient solution. The composition of nutrient solution was as follows: 1.25 mM Ca(NO3)2, 1.25 mM KNO3, 0.5 mM MgSO4, 0.25 mM KH2PO4, 10 μM H3BO3, 1 μM MnSO4, 0.5 μM ZnSO4, 0.05 μM (NH4)6Mo7O24 and 0.4 μM CuSO4.

Three Fe treatments were established as follows: presence of 0.03 mM Fe (+Fe), pH 6.2; absence of Fe (−Fe), pH 6.2; presence of 0.03 mM Fe plus 0.5 g L−1 CaCO3 plus 10 mM NaHCO3, which brought the pH to 8.3 (+Bic.). Fe was supplied in the form of Fe(III)-EDTA. NaHCO3 and CaCO3 were included in the nutrient solution to mimic the effects of calcareous soils. The solution was renewed every 5 days. During this period the pH of the nutrient solution never rose over 6.3 for the +Fe treatment and slightly decreased for the −Fe treatment. Plants were maintained in a growth chamber with the day/night regime of 16/8 h, 24°C/18°C and a relative humidity of 70%.

Soluble protein extraction and cytosolic enzyme assays

Roots of 10-days-old plants grown under the different treatments were harvested, rinsed and homogenized in one volume of a buffer containing 50 mM Tris-HCl (pH 7.5), 10 mM MgCl2, 10% (v/v) glycerol, 1 mM ethylene diamine tetraacetic acid (EDTA) and 14 mM ß-mercaptoethanol; 1 mM phenylmethylsulphonyl fluoride (PMSF) and 10 μg mL−1 leupeptin were added to avoid or minimize proteolysis (De Nisi and Zocchi 2000). The homogenate was filtered through four layers of gauze and centrifuged at 13,000 g for 15 min and the supernatant was again centrifuged at 100,000 g for 30 min. The soluble extracted proteins were dialysed against the same homogenization buffer and used for the activity assays directly or after storing in liquid N2.

Phosphoenolpyruvate carboxylase (PEPC) (EC 4.1.1.31) was determined as reported by De Nisi and Zocchi (2000). Glyceraldehyde-3-phosphate dehydrogenase (G3PDH) (EC1.2.1.12), phosphofructokinase 1 (PFK) (EC 2.7.1.11) and pyruvate kinase (PK) (EC 2.7.1.40) activities were determined as reported by Espen et al. (2000). Glucose-6-phosphate dehydrogenase (G6P-DH) (EC 1.1.1.49) was assayed according to Rabotti et al. (1995). The reactions were started by adding aliquots of the protein extracts. All enzymatic assays were performed at 25°C in 1 mL final volume. Oxidation of NADH or reduction of NADP+ was followed spectrophotometrically at 340 nm.

PEPC immunocharacterization

Soluble protein extracted from roots of 10-day-old plants grown under the different treatments was loaded on a discontinuous SDS-polyacrylamide gel (3.75% [w/v] acrylamide stacking gel and 9% [w/v] acrylamide separating gel) (De Nisi and Zocchi 2000). After SDS-PAGE, proteins were electrophoretically transferred to PVDF membrane filters (Sigma) using a semi-dry blotting system with a buffer containing 10 mM 3-cyclohexylamino-1-propane sulphonic acid (pH 11) and 10% (v/v) methanol for 1.5 h at room temperature at a current intensity of 0.8 mA cm−2. Polyclonal antibodies raised against a PEPC isoform of sorghum were used (a kind gift from Dr J. Vidal, Univrsité de Paris-Sud). Antiserum was diluted 1:1,000 in TBS-T buffer [20 mM Tris-HCl (pH 7.5), 200 mM NaCl, 0.05% (w/v) Tween 20] and incubation was carried out overnight at 4°C. After rinsing with TBS-T, PVDF membranes were incubated at room temperature for 2 h with a 1:25,000 diluted secondary antibody (alkaline phosphatase-conjugated anti-rabbit IgG, Sigma).

Lipid peroxidation

Samples of fresh material (roots and shoots) of 0.2 g were homogenized in 4 mL of 1% (w/v) trichloroacetic acid (TCA) solution. The homogenate was filtered through four layers of gauze and then centrifuged at 12,000 g for 15 min. One mL aliquots of the supernatants were added to 3 mL of 0.5% (w/v) thiobarbituric acid (TBA) in 20% (w/v) TCA and the tubes were incubated in a shaking water bath at 95°C for 2 h. The reaction was stopped by placing the reaction tubes into an ice bath. The tubes were subsequently centrifuged at 9,000 g for 10 min, the absorbance of the supernatant was measured at 532 and 600 nm (Cakmak and Horst 1991) and the concentration of the malonyldialdehyde (MDA)-TBA complex produced was calculated using the molar extinction coefficient of 155 mM−1 cm−1.

Determination of phenols

Roots and leaves (1 g) were collected from 10-day-old plants and homogenized in a mortar containing methanol and water (1:1) as extraction medium. The homogenate was centrifuged at 10,000 g for 10 min and the supernatant treated according to Swain and Hillis (1959). Phenol concentration was determined by spectrophotometry at 750 nm referring to a gallic acid calibration curve.

Shikimate pathway enzymes

Fresh tissue was homogenized with quartz sand and liquid nitrogen in a grinding medium containing 0.1 M K-phosphate buffer (pH 7.4), 0.5 mM dithiotreitol (DTT), 2 mM l-cysteine, 2 mM EDTA, 8 mM β-mercaptoethanol and 0.5 g polyvinyl polypyrrolidone (PVPP). The homogenates were filtered individually through four layers of gauze and centrifuged at 19,000 g for 20 min at 0–4 °C (Diáz et al. 1997).

The shikimate kinase (SK) (EC 2.7.1.71) was assayed at 25°C by coupling the release of ADP to the oxidation of NADH using pyruvate kinase (EC 2.7.1.40) and lactate dehydrogenase (EC 1.1.1.27) as coupling enzymes according to Krell et al. (2001). Shikimate-dependent oxidation of NADH was monitored at 340 nm. The assay mixture contained 50 mM triethanolamine hydrochloride/KOH buffer at pH 7.0, 50 mM KCl, 5 mM MgCl2, 1.6 mM shikimic acid, 5 mM ATP, 1 mM phosphoenolpyruvate, 0.1 mM NADH, 3 units/mL pyruvate kinase, and 2.5 units/mL lactate dehydrogenase. Kinetic parameters were estimated using nonlinear regression to the Michaelis–Menten equation (Microcal Software); the errors in the parameters were less than 5%.

The shikimate dehydrogenase (SDH) (EC 1.1.1.25) was assayed at 25°C by monitoring the reduction of NADP+ at 340 nm according to Chaudhuri and Coggins (1985). The assay mixture (total volume 1 mL) contained 100 mM Na2CO3 (pH 10.6), 4 mM shikimic acid and 2 mM NADP+.

The extract enzyme preparation for phenylalanine ammonia-lyase (PAL) (EC 4.3.1.5) was obtained by homogenizing 0.20 g fresh tissue in 15 mL of an extraction medium containing 20 mM ß-mercaptoethanol, 0.1 M sodium borate buffer (pH 8.8) and 5% (m/v) polyvinyl polypyrrolidone (PVPP). After filtration through four layers of gauze, the homogenate was centrifuged at 12,000 g for 20 min. The enzyme activity was determined by adding 1 mL of the enzyme extract to a reaction medium containing 1 mL of 0.1 M sodium borate buffer (pH 8.8) and 1 mL of 0.1 M L-phenylalanine. After incubation for 1 h at 30°C, the reaction was stopped by adding 0.1 mL of 6 N HCl and the absorbance was determined at 290 nm (Cahill and Mc Comb 1992).

Protein determination

Protein concentration was determined according to Bradford (1976) by using the BioRad reagent and BSA as a standard.

Statistical analysis

A two-way analysis of variance (ANOVA), with lines and treatments as factors, was performed for the whole data using the STATI-CF statistical program. Means were compared using the Tukey-Kramer test at p < 0.05 when significant differences were found. Data shown are means of five replicates for each treatment.

Results

Activity and immunodetection analysis of PEPC

The activity of PEPC was much higher in Fe-deficient root extracts than in the control for both lines, this increase being higher in the bicarbonate induced deficiency compared to the direct Fe deficiency (Fig. 1). The highest increase in PEPC activity was found in TN11.11 plants grown in the presence of bicarbonate.

Fig. 1
figure 1

(a) PEPC activity (nmol NADH min−1mg−1protein) in roots of two lines of Medicago ciliaris plants grown in the presence of Fe (+Fe), in the absence of Fe (−Fe), or in the presence of Fe plus bicarbonate (+Bic) during the treatment period (10 days). Values are the means of 5 replicates ± SD. In the case of significant interaction between Fe treatments and lines, means followed by different letters are significantly different at P < 0.05 according to Tukey-Kramer test. SDS-PAGE (b) and immunoblotting (c) analysis of PEPC of two lines of Medicago ciliaris plants grown in the presence of Fe (+Fe), in the absence of Fe (−Fe), or in the presence of Fe plus bicarbonate (+Bic) during the treatment period (10 days)

Full size image

Western blot analysis of soluble proteins obtained from roots of Medicago plants grown under the different treatments are shown in Fig. 1c. In all samples the antibody raised against a PEPC isoenzyme of sorghum reacts against two polypeptides with apparent molecular masses of 103 and 108 kDa. In the lanes with the proteins extracted from plants grown in the presence of bicarbonate both polypeptides are enhanced respect to the control, the increase being much stronger for the polypeptide of 103 kDa (Fig. 1c). In the −Fe treatment, only a slight enhancement of the 103 kDa polypeptide was detectable in TN11.11 (Fig. 1c). The SDS-PAGE of the soluble proteins extracted from TN11.11 and TN8.7 roots is shown in Fig. 1b.

Cytosolic enzyme assay (G3PDH, PFK, PK and G6PDH)

Four cytosolic enzymes (G3PDH, PK, PFK and G6PDH) were assayed in extracts of Fe-sufficient and Fe-deficient roots of Medicago. Enzymatic activities increased markedly with Fe deficiency (in both −Fe and +Bic. treatments) for both lines (Fig. 2a, b, c and d). The highest activities were found in TN11.11 roots grown in the presence of bicarbonate.

Fig. 2
figure 2

Pyruvate kinase (PK) (a), Phosphofructo kinase (PFK) (b), Glucose 6-phosphate dehydrogenase (G6PDH) (C) and Glyceraldehyde 3-phosphphate dehydrogenase (G3PDH) (D) activities of two lines of Medicago ciliaris plants grown in the presence of Fe (+Fe), in the absence of Fe (−Fe), or in the presence of Fe plus bicarbonate (+Bic) during the treatment period (10 days). Values are the means of 5 replicates ± SD. In the case of significant interaction between Fe treatment and line, means followed by different letters are significantly different at P < 0.05 according to Tukey-Kramer test

Full size image

Lipid peroxidation

Under Fe deficiency conditions lipid peroxidation, measured by the MDA-TBA complex concentration in plant tissues (roots and leaves) was increased in both parts of the plants of the two lines (Fig. 3a and b). Both Fe deficiency conditions induced higher increases of lipid peroxidation in TN8.7 plants than in TN11.11 ones, more pronounced in the absence of Fe then in the presence of bicarbonate. In the absence of Fe (−Fe treatment) a considerable increase of lipid peroxidation in leaves, more pronounced in TN8.7 (+104% in leaf system) than in TN11.11 (+70%) (Fig. 3a) was observed.

Fig. 3
figure 3

Iron deficiency effects on MDA concentration in leaves (a) and in roots (b) (μg g− 1 FW) of 10-day-old Medicago ciliaris plants grown in the presence of Fe (+Fe), in the absence of Fe (−Fe), or in the presence of Fe plus bicarbonate (+Bic) during the treatment period (10 days). Values are the means of 5 replicates ± SD. In the case of significant interaction between Fe treatment and line, means followed by different letters are significantly different at P < 0.05 according to Tukey-Kramer test

Full size image

Phenol content and shikimate pathway

A common response to Fe deficiency in Strategy I plants is the increased synthesis and release of phenolics (Olsen et al. 1981; Marschner and Römheld 1995). The concentration of phenolics in Fe-deficient tissues of Medicago plants (Fig. 7a and b) underwent a significant increase, especially when Fe deficiency was induced by bicarbonate. In both lines a higher content was found in leaves than in roots, more likely because M. ciliaris lines are able to exude these compounds from the roots into the rhizosphere (M’sehli et al. 2008), so accumulation of phenolics in the roots does not occur. Notably, the increase in concentration of phenolics in all plant parts was higher for the tolerant line (TN11.11) than for the sensible one (TN8.7).

As shikimate is a precursor of the biosynthesis of aromatic compounds, and consequently of phenols, some enzymatic activities with a key role in the shikimate pathway were assayed in Medicago tissues grown in Fe deficiency.

As shown in Figs. 4a, b, 5a, b, 6a and b, SK, SDH and PAL activities were higher in leaves than in roots regardless the treatments and the lines. In comparison to the control, Fe deficiency, especially when induced by bicarbonate, led to a significant increase in the activity of these enzymes for both lines. The highest activities of SK, SDH and PAL were found in the tolerant line (TN11.11), characterized by a higher phenol content than the sensitive one (TN8.7) (Fig. 7).

Fig. 4
figure 4

Shikimate kinase activity (SK) in leaves (a) and in roots (b) (nmol min−1mg−1 protein) of 10-day-old Medicago ciliaris plants grown in the presence of Fe (+Fe), in the absence of Fe (−Fe), or in the presence of Fe plus bicarbonate (+Bic) during the treatment period (10 days). Values are the means of 5 replicates ± SD. In the case of significant interaction between Fe treatment and line, means followed by different letters are significantly different at P < 0.05 according to Tukey-Kramer test

Full size image
Fig. 5
figure 5

Shikimate dehydrogenase activity (SDH) in leaves (a) and in roots (b) (nmol min−1mg−1 protein) of 10-day-old Medicago ciliaris plants grown in the presence of Fe (+Fe), in the absence of Fe (−Fe), or in the presence of Fe plus bicarbonate (+Bic) during the treatment period (10 days). Values are the means of 5 replicates ± SD. In the case of significant interaction between Fe treatment and line, means followed by different letters are significantly different at P < 0.05 according to Tukey-Kramer test

Full size image
Fig. 6
figure 6

Phenylalanine ammonia-lyase activity (PAL) in leaves (a) and in roots (b) (nmol min−1g−1 FW) of 10-day-old Medicago ciliaris plants grown in the presence of Fe (+Fe), in the absence of Fe (−Fe), or in the presence of Fe plus bicarbonate (+Bic) during the treatment period (10 days). Values are the means of 5 replicates ± SD. In the case of significant interaction between Fe treatment and line, means followed by different letters are significantly different at P < 0.05 according to Tukey-Kramer test

Full size image
Fig. 7
figure 7

Amounts of phenols in leaves (a) and in roots (b) (μg.g − 1 FW) of 10-day-old Medicago ciliaris plants grown in the presence of Fe (+Fe), in the absence of Fe (−Fe), or in the presence of Fe plus bicarbonate (+Bic) during the treatment period (10 days). Values are the means of 5 replicates ± SD. In the case of significant interaction between Fe treatment and line, means followed by different letters are significantly different at P < 0.05 according to Tukey-Kramer test

Full size image

Discussion

The metabolic changes occurring in two lines of medics differently affected by Fe deficiency conditions were investigated in this work. TN11.11, coming from a calcareous soil, with high lime content and high pH value, is characterized by a good tolerance to Fe deficiency, attributable to a high aptitude to induce the Strategy I responses; TN8.7, selected from a soil poor in lime content and with a low pH value, was shown to be more sensitive and less efficient in responding to such constraint (M’sehli et al. 2008).

The activation of reduction processes and the increased extrusion of protons under Fe deficiency require an increased rate of NAD(P)H and ATP regeneration in comparison with non Fe stressed conditions. Recharging of these substrates implies the acceleration of the carbohydrate catabolism, i.e. glycolysis and oxidative pentose phosphate pathway, as already found in different species (Sijmons and Bienfait 1983; Rabotti et al. 1995; Espen et al. 2000; López-Millán et al. 2000). Moreover, the protogenic reactions catalyzed by hexokinase (HK), PFK and G3PDH will generate the H+ required by H+-ATPase, contributing to re-equilibrate the cytosolic pH (Sakano 1998; Zocchi 2006). In the tolerant line TN11.11 the increased H+-ATPase activity in Fe limiting conditions (M’sehli et al. 2008) is accompanied by an up-regulation of glycolysis and pentose phosphate pathway, shown by the activation of PK, PFK, G3P-DH and G6P-DH. The same does not occur in the sensitive line TN8.7, in which the weak activation of the two main responses to Fe deficiency is reflected by a lower activity of these enzymes.

The increased proton extrusion by Fe-deficient roots has been shown to be linked to an increased synthesis and accumulation of organic acids, in particular citric and malic acids (Landsberg 1981; De Vos et al. 1986). In fact, according to the pH-stat theory (Davies 1973), proton efflux by H+-ATPase activity causes cytoplasm alkalinization, which in turn activates the dark fixation of CO2 catalyzed by PEPC, and consequently enhances organic acid synthesis (Rabotti et al. 1995). Previous results (M’sehli et al. 2008) showed that the total pools of root organic anions (citrate and malate) were higher in Fe-deficient Medicago roots than in the controls both in TN11.11 and in TN8.7. The highest values of organic acid concentration were found in the tolerant line TN11.11, characterized by a high proton extrusion activity as reported before for grapevine (Ollat et al. 2003), woody species (Sun et al. 1987; Rombolà et al. 2002) and herbaceous plants (Abadia et al. 2002). Since citrate accumulation is more pronounced than malate in both lines of Medicago ciliaris and under both Fe deficient treatments, citrate could be taken into consideration as a biochemical marker of Fe chlorosis tolerance, as suggested by Ollat et al. (2003) and Rombolà et al. (2002) for other species. Lambers et al. (2002) and Veneklaas et al. (2003) found that the proportion of malate to citrate in the rhizosphere varied as a function of soil pH. The increased organic acid pools in Fe-deficient roots could be linked to a coordinated increase in the activities of PEPC and malate dehydrogenase (MDH): PEPC catalyzes the carboxylation of PEP to oxaloacetate, which could be then reduced to malate via cytosolic MDH; malate could be transported into the mitochondria via the dicarboxylic acid shuttle and converted to citrate by citrate synthase, as suggested by López-Millán et al. 2000. In agreement with this hypothesis, Fe-deficient Medicago roots (both −Fe and +Bic. treatments) have the ability to enhance the PEPC activity (Fig. 1), other than the MDH activity (M’sehli et al. 2008). In addition, as shown by immunoblotting (Fig. 1c), the enhancement of PEPC activity is attributable to an increase in protein expression, as previously found in cucumber (De Nisi and Zocchi 2000) and in sugar beet (Andaluz et al. 2002). The tolerant line TN11.11 showed the highest capacity to enhance the PEPC activity, associated with greater increase in the level of protein detected by immunoblotting when compared to the sensitive TN8.7 line (Fig. 1c). Plants grown in the presence of bicarbonate witnessed the highest PEPC and MDH activities differentiating themselves from the other treatments. The greatest increase in the PEPC activity occurred upon addition of bicarbonate and lime to the growing medium. This observation could be explained by the fact that PEPC catalyses the incorporation of bicarbonate into a C3 organic acid, PEP, producing oxaloacetate. Since bicarbonate has long been considered the major causal factor of Fe deficiency chlorosis in calcareous soils, the HCO3 − fixing enzyme, PEPC, has been proposed as a component of adaptive strategies of plants to cope with poor external Fe availability (Rombolà et al. 2005) in calcareous soils.

The possibility that PEPC may fix C in Fe-deficient roots would depend on the existence of a source of PEP in such roots. PEP could come via glycolysis from compounds previously synthesized and/or stored within the plant. As observed in the present study, the activity of some enzymes (G3PDH, PFK and PK) involved in the glycolytic pathway was increased in Fe-deficient roots for both lines, as reported for other plant species (Espen et al. 2000; Abadía et al. 2002). Maas et al. (1988) observed that both phloem and root sugar concentrations have been reported to increase with Fe deficiency. Overall, the most important consequences of up regulation of glycolysis under Fe deficiency conditions are (Zocchi 2006): i) ATP synthesis to sustain the increased H+-ATPase activity; ii) formation of reducing equivalents for the FCR; iii) formation of phosphoenolpyruvate (PEP); iv) a contribution to the regulation of cytosolic pH, due to the protogenic reactions catalyzed by HK, PFK and G3PDH (Sakano 1998). As the greatest metabolic changes have been detected in the Fe efficient line TN11.11, such aptitude to sustain the Strategy I responses by shifting metabolism from the anabolic to the catabolic pathways should be taken into consideration in screening Fe efficient Medic lines. Moreover, the bicarbonate treatment has shown to be more effective in inducing such metabolic changes, confirming that imposition of bicarbonate-induced Fe deficiency is a suitable method to screen Medic lines tolerant to calcareous soils.

Involvement of secondary metabolism in the response of Medicago ciliaris to Fe deficiency

Iron deficiency can induce oxidative stress mainly by impairing the electron transport chain functionality both at the mitochondrion and at the chloroplast level (Pascal and Douce 1993; Terry and Abadía 1986) and by decreasing the antioxidant activity of Fe-dependent enzymes such as catalases, peroxidases and Fe-superoxide dismutases (Iturbe-Ormaetxe et al. 1995; Ranieri et al. 1999). Hernandez et al. (2001) proposed MDA content in plant tissues, which is a measure of lipid peroxidation, as a reliable indicator of oxidative stress. As observed in the present study, Fe deficiency elicited an increase in MDA content for both Medicago genotypes. However, in the susceptible line (TN8.7) Fe deficiency induced a higher degree of lipid peroxidation in comparison with the tolerant one (TN11.11), suggesting a less efficient defence system against oxidative stress for TN8.7 than for TN11.11. Similar results were observed in grapevine genotypes (Ksouri et al. 2006) where MDA contents were significantly higher in the young leaves of the sensitive variety.

Both lines grown in the presence of bicarbonate showed a lower degree of lipid peroxidation in comparison with those grown in the −Fe treatment, suggesting that the former condition induces a less intense oxidative stress: in the presence of bicarbonate the Strategy I responses can actually lead to Fe uptake, as the ion is indeed present in the nutrient solution, resulting in a less dramatic Fe deficiency than in the complete absence of Fe.

An increase in the synthesis of secondary metabolites such as phenolic compounds is a common response to environmental stress in plants (Dixon and Paiva 1995). In our work Fe deficiency induced the synthesis of phenols in both lines especially in leaves, showing genotypic differences between sensitive and tolerant line. In agreement with our findings Ksouri et al. (2006) reported that the absence of Fe led to a significant increase of the phenol index in tolerant grapevine genotypes, contrasting with constant values in sensitive ones. Among their biological activities, phenolics are involved in the response to oxidative stress as potential scavengers of free radicals. Anti-oxidative properties of polyphenols arise from their high reactivity as hydrogen or electron donors and from the ability of the polyphenol-derived radical to stabilize and delocalize the unpaired electron (chain-breaking function) (Blokhina et al. 2003). Another mechanism underlying the anti-oxidative properties of phenols is the ability of flavonoids to alter peroxidation kinetics by modification of the lipid packing order and to decrease fluidity of the membranes which could hinder diffusion of free radicals and restrict peroxidative reactions (Arora et al. 2000). Our results show that the bicarbonate treatment induces a higher phenolic synthesis, other than a lower oxidative stress respect to the −Fe one. The effectiveness of the Strategy I responses in the bicarbonate treatment could result in a lower oxidative stress both directly and by allowing a more efficient anti-oxidant defence system.

Under Fe deficiency conditions phenols play at the same time two roles, as antioxidant metabolites and also as chelators for Fe when they are exuded into the rhizosphere (M’sehli et al. 2008). The observed stimulatory effect of Fe deficiency on the total phenol content led us to investigate on the activity of three enzymes belonging to the shikimate pathway, involved in phenolic synthesis: SK, SDH and PAL. In both genotypes, Fe deficiency caused an increase in the activity of these enzymes, higher in the tolerant line (TN11.11) when compared to the sensible one (TN8.7), higher in leaves than in roots. The shikimate pathway converts simple carbohydrates to aromatic amino acids, such as phenylalanine, which is the first substrate for the phenylpropanoid pathway. Phenylalanine is required for the synthesis of various phenolic compounds (Herrmann 1995). PAL (a key enzyme in phenylpropanoid pathway) is widely distributed in plants and is believed to be the main enzyme responsible for phenolic synthesis, conferring increased tolerance to a wide array of stresses (Grace and Logan 2000). The activation of phenylpropanoid pathway by biotic stress is well established (Hahlbrock and Scheel 1989; Dixon and Paiva 1995). However, the ability of a wide range of abiotic stresses to stimulate phenylpropanoid pathway is less widely appreciated, these abiotic stresses include high light (Beggs and Wellman 1994), low temperature (Solecka et al. 1999), phosphate deficiency (Trull et al. 1997) and aluminium toxicity (Yamamoto et al. 1998).

Conclusion

The present work demonstrates that the different Fe deficiency tolerance found in Medicago ciliaris lines (TN11.11 and TN8.7) can be ascribed to the different aptitude to a) modulate the carbohydrate metabolism, sustaining the primary Strategy I responses; b) increase secondary metabolism pathways, leading to more efficient Fe mobilization and reducing oxidative damage. The genotypic variability in such aptitude should then be considered in screening programs for the selection of new Fe-efficient lines. Moreover, the presence of bicarbonate and lime in the nutrient solution, which better simulates environmental conditions occurring in calcareous soils, could be adopted in such programs.