Silicon alleviates iron deficiency in cucumber by promoting mobilization of iron in the root apoplast



This article is corrected by:

  1. Errata: Corrigendum Volume 199, Issue 3, 866, Article first published online: 12 June 2013

Author for correspondence:

Miroslav Nikolic

Tel: +381 11 3058956



  • Root responses to lack of iron (Fe) have mainly been studied in nutrient solution experiments devoid of silicon (Si). Here we investigated how Si ameliorates Fe deficiency in cucumber (Cucumis sativus) with focus on the storage and utilization of Fe in the root apoplast.
  • A combined approach was performed including analyses of apoplastic Fe, reduction-based Fe acquisition and Fe-mobilizing compounds in roots along with the expression of related genes.
  • Si-treated plants accumulated higher concentrations of root apoplastic Fe, which rapidly decreased when Fe was withheld from the nutrient solution. Under Fe-deficient conditions, Si also increased the accumulation of Fe-mobilizing compounds in roots. Si supply stimulated root activity of Fe acquisition at the early stage of Fe deficiency stress through regulation of gene expression levels of proteins involved in Fe acquisition. However, when the period of Fe deprivation was extended, these reactions further decreased as a consequence of Si-induced enhancement of the Fe status of the plants.
  • This work provides new evidence for the beneficial role of Si in plant nutrition and clearly indicates that Si-mediated alleviation of Fe deficiency includes an increase of the apoplastic Fe pool in roots and an enhancement of Fe acquisition.


Iron (Fe) is a component of a number of proteins and enzymes with functions in key metabolic processes and Fe is accordingly an essential microelement for plants and all other living organisms. Despite being the fourth most abundant element in the earth's crust, Fe deficiency is one of the major limiting factors for crop production in calcareous soils all over the world (Vose, 1982). Strategy 1 plant species (all dicots and monocots with the exception of grasses, which belong to Strategy 2) respond to lack of Fe by undergoing both morphological and physiological changes in the roots which mobilize Fe in the rhizosphere thereby enabling more efficient acquisition. Morphological changes include an increase of the root surface characterized by the formation of lateral roots, root hairs in the apical zone as well as the development of transfer cells (Schmidt, 1999). Several components of the Fe root uptake machinery in Strategy 1 plants are upregulated and quickly respond to Fe deficit, that is plasma membrane (PM)-bound FeIII chelate reductases belonging to the Ferric Reductase Oxidase (FRO) family (Jeong & Connolly, 2009); Iron Regulated Transporter (IRT) FeII transporter (Vert et al., 2002); and PM proton pump of the HA family (P-type H+-ATPase), which mediates acidification of the root apoplast/rhizosphere (Santi & Schmidt, 2009). Under Fe starvation, many Strategy 1 species also enhance root accumulation and thus exudation of organic compounds (e.g. carboxylates, phenolics and flavins), which can increase Fe availability due to chelation and potential reduction of FeIII (e.g. Ohwaki & Sugahara, 1997; Welkie, 2000; Jin et al., 2007).

Although silicon (Si) is the second most abundant mineral element in the earth's crust, it is still not fully accepted as an essential element for higher plants. However, Si is the only known element that effectively alleviates multiple environmental stresses in many plant species (e.g. metal excess, drought, salt, waterlogging, pathogens and pests; for review, see for example Ma, 2004; Liang et al., 2007; Cooke & Leishman, 2011). Over the past few years rapid progress has been made in elucidating the mechanisms through which Si mediates metal toxicity alleviation in higher plants (e.g. Song et al., 2009, 2011; Frantz et al., 2011; Prabagar et al., 2011; Dragišić Maksimović et al., 2012). Yet, information on the relevance of Si nutrition under deficiency of Fe and other micronutrients in plants is still lacking. This is partly due to the fact that root responses to Fe deficiency have so far mainly been studied and characterized in nutrient solution experiments in which Si (e.g. bioavailable monosilicic acid) was omitted. On the other hand, Si availability in calcareous soils is very low (Liang et al., 1994) and further decreases with soil drying (Snyder et al., 2007), conditions conducive to a high risk of Fe deficiency chlorosis. Bityutskii et al. (2010) were the first to demonstrate that addition of Si to the nutrient solution could mitigate Fe deficiency chlorosis in Strategy 1 plants – cucumber (Cucumis sativus) and pumpkin (Cucurbita pepo), but not in Strategy 2 plants – corn (Zea mays) and barley (Hordeum vulgare). Very recently, Gottardi et al. (2012) reported that the acquisition of Fe by corn salad (Valerianella locusta) roots increased when Si was included in the hydroponic solution adequately supplied with Fe.

The aim of the present work was to investigate how Si ameliorates Fe deficiency in cucumber. This species was selected because it is a Strategy 1 model plant for Fe research and is also characterized as a Si-accumulating dicot (Nikolic et al., 2007). Our hypothesis is that Si has a beneficial influence in mobilizing extracellular Fe. To test this hypothesis we investigated the effect of Si on storage and utilization of Fe in the root apoplast. To gain further insight into the role of Si in root biosynthesis and thus exudation of Fe mobilizing compounds under Fe deficiency conditions, we also investigated the accumulation of carboxylates, phenolics and flavins in root tissues along with the expression of genes encoding key enzymes in the biosynthetic pathways of these compounds. Our results provide new evidence for the beneficial role of Si in plant nutrition and facilitate development of new strategies for sustainable control of Fe deficiency chlorosis in crops.

Materials and Methods

Plant materials and growth conditions

After soaking in 1 mM CaSO4 overnight, seeds of cucumber (Cucumis sativus L. cv Chinese long; kindly provided by Superior d.o.o., Velika Plana, Serbia) were germinated between two sheets of filter paper moistened with saturated CaSO4. The 4-d-old seedlings were then transferred to a complete nutrient solution (four plants per 2.5-l plastic pot) containing (in mM): 0.7 K2SO4, 0.1 KCl, 2.0 Ca(NO3)2, 0.5 MgSO4, 0.1 KH2PO4, and (in μM): 0.5 MnSO4, 0.5 ZnSO4, 0.2 CuSO4, 0.01 (NH4)6Mo7O24, 10 H3BO3. Fe was supplied as NaFeIIIEDTA at different concentrations, unless mentioned separately in the text (for detail information, see the legend of Fig. 10). In the first set of experiments plants were pre-cultured for 7 d in nutrient solution supplied with 1 μM Fe and then grown for another 7 d in either +Fe (50 μM) or in −Fe (Fe-free) nutrient solution, without (−Si) or with (+Si) supply of Si. In the second set of experiments plants were grown with 10 μM Fe in either −Si or +Si nutrient solution, and after 7 d of pre-culture Fe was excluded from both −Si and +Si nutrient solutions. In the additional experiment plants were pre-cultured in Fe/Si-free nutrient solution for 7 d and than transferred to −Si or +Si nutrient solution supplied with 1 μM Fe either as FeIIIEDTA (control) or Fe(OH)3. If applied Si was in the form of Si(OH)4 at 1.5 mM. This was freshly prepared by passing Na2SiO3 through a column filled with cation-exchange resin (Amberlite IR-120, H+ form; Fluka, Buchs SG, Switzerland). Amberlite IR-120 is also capable of removing ionic Fe forms (Schmid & Gerloff, 1961) from Na2SiO3, thus preventing any possible contamination of the nutrient solution with Fe. The concentration of soluble Si in the nutrient solution (checked daily) was in the range 1.3–1.5 mM.

The pH of nutrient solutions was adjusted to 6.0 and checked daily. The nutrient solutions were renewed completely every 2 d and continuously aerated.

Plants were grown under controlled environmental conditions in a growth chamber with a photoperiod of 16 h : 8 h and temperature regime of 24°C : 20°C (light : dark), photon flux density of 250 μmol m−2 s−1 at plant height, and relative humidity of c. 70%.

After harvest plants were divided in two parts, root and shoot, oven dried at 70°C for 48 h, weighed and hand ground.

Chlorophyll determination

Chlorophyll content in the youngest fully expanded leaves was approximated nondestructively using a portable Chlorophyll Meter SPAD-502 device (Minolta Camera Co., Osaka, Japan).

Collection of xylem sap

Xylem sap was obtained by exudation after plants were decapitated at the stem c. 2 cm above the root base. Soft rubber tubes were fixed over decapitated stem and xylem sap was collected by a micropipette for 1 h, after discarding of the exudates obtained during the first few minutes.

Determination of Fe in root apoplast, plant tissues and xylem sap

After washing in the solution containing 0.5 mM CaSO4 and 5 mM MES for 10 min, intact roots of each plant were transferred to 30 ml incubation solution containing 5 mM MES (pH 5.5), 0.5 mM CaSO4 and 1.5 mM 2,2′-bipyridyl and then incubated for 10 min under reductive conditions by adding 0.5 g solid Na-dithionite under continuous N2 bubbling through the solution. Apoplastic Fe was removed as a red FeII(bipyridyl)3 complex and determined by spectrophotometric measurement of the absorbance at 520 nm using an extinction coefficient of 8.65 mM−1 (Bienfait et al., 1985).

Dry plant material (0.2 g) was digested in 3 ml concentrated HNO3 + 2 ml H2O2 for 1 h in a microwave oven (Speedwave MWS-3+; Berghof Products + Instruments GmbH, Eningen, Germany). Samples were then transferred into 25 ml-plastic flasks and flask volume was adjusted to 25 ml with deionized H2O. Fe was determined by inductively coupled plasma optical emission spectroscopy (ICP-OES, SpectroGenesis EOP II; Spectro Analytical Instruments GmbH, Kleve, Germany).

The Fe concentration in freshly collected xylem sap was determined by ICP-OES immediately after the samples were diluted 10 times (v/v) with 1 M HNO3.

Determination of Si

Dry leaf material (0.2 g) was microwave digested with 3 ml concentrated HNO3 + 2 ml H2O2 for 1 h. Samples were diluted with c. 15 ml deionized H2O, transferred into 25 ml-plastic flasks, 1 ml concentrated HF was added, and left overnight. After addition of 2.5 ml 2% (w/v) H3BO3, the flask volume was adjusted to 25 ml with deionized H2O, and Si was determined by ICP-OES, after a final dilution of the samples of 1 : 100 (v/v) with deionized H2O.

Soluble Si in the nutrient solution was determined by ICP-OES immediately after 1-ml aliquot was diluted 100 times (v/v) with deionized H2O.

Determination of organic acids, phenolic compounds and riboflavin

Organic acids, phenolic compounds and riboflavin were extracted from the fresh root material. The samples (root tips of c. 0.5– 1 cm; 0.1 g FW) were immediately frozen in liquid N2, ground thoroughly and extracted in 1 ml of methanol:deionized H2O (3 : 1, v/v) mixture. Before analyses, all samples were filtered through 0.22 μm pore size nylon syringe filters (Phenomenex, Torrrance, CA, USA).

Separation of organic acids from different samples, that is root-tissue extract and xylem sap mixed with methanol (3 : 1, v/v), was performed using a HPLC system (Waters, Milford, MA, USA) consisting of 1525 binary pumps, thermostat and 717+ autosampler connected to the Waters 2996 diode array detector (DAD; Waters) adjusted at 210 nm. An anion exchange column, Aminex HPX-87H (Bio-Rad Laboratories, Hercules, CA, USA) 300 × 7.8 mm was used with 5 mM H2SO4 as a mobile phase. Isocratic elution was used with a flow rate of 0.6 ml min−1 at 40°C.

Phenolic compounds and riboflavin from the root-tissue extracts were quantified by reversed phase HPLC/DAD analysis. Samples were injected in the Waters HPLC system specified above. Separation of phenolics and riboflavin was performed on a Symmetry C-18 RP column 125 × 4 mm size with particle diameter of 5 μm (Waters) connected to appropriate guard column. Two mobile phases, A (0.1% phosphoric acid) and B (acetonitrile) were used at a flow of 1 ml min−1 with the following gradient profile: the first 20 min from 10% to 22% B; next 20 min of linear rise up to 40% B, followed by 5 min reverse to 10% B and additional 5 min of equilibration time. Chromatograms were recorded in 3D mode with consecutive extraction of channels at specific wavelengths for exclusive phenolic compounds (220 and 309 nm) and at 445 nm for riboflavin.

The riboflavin peak was further confirmed by reversed phase HPLC-MS analysis using the Waters HPLC system connected to EMD 1000 single quadrupole detector with ESI probe (Waters). Separation of riboflavin was performed under the same conditions as described for phenolic compounds, however using 0.1% formic instead of phosphoric acid. A post column flow splitter (ASI, Richmond, CA, USA) with 5 : 1 split ratio was used to obtain optimal mobile phase inflow of 200 μl min−1 for ESI probe. For LC-MS analysis, signals for each compound were detected in positive ESI scan (from 100 to 600 m/z) and SIR mode (377 m/z) with the following parameters: capillary voltage 3.0 kV, cone voltage 40 V, extractor and RF lens voltages were 3.0 and 0.2 V, respectively. Source and desolvation temperatures were 120°C and 400°C, respectively, with N2 flow of 600 l h−1. Supporting Information Fig. S1 clearly shows the exact match of UV and MS spectra of the peak recorded at 10.93 min (ion mass of 377 m/z, characteristic for protonated riboflavin) in both standard solution and root extract samples.

Detected metabolites were located and identified in corresponding chromatograms through comparison of separated standards (Sigma-Aldrich, St Louis, MO, USA) with authentic samples by combining their retention times and UV spectra obtained by DAD. The calibration curves were made by injection of different volumes from the stock solutions over the range of concentration observed for each of the compounds, using linear regression for the relationship of peak area sum vs concentration, under the same conditions as for the sample analysis. Each component was analysed quantitatively by the external standard method using pure compounds as references for concentration, retention time and characteristic UV spectra, respectively. Amounts of injected samples were adjusted to fit within the linear calibration range gathered by injection of selected commercial standards. Data acquisition and spectral evaluation of the peaks was processed by the Empower 2 software (Waters).

Determination of FeIII chelate reducing capacity by roots

Intact roots of cucumber plants were washed in 30 ml 0.5 mM CaSO4 solution with continuous aeration. Subsequently, roots were incubated in 30 ml of a solution containing 0.5 mM CaSO4 and 5 mM MES, 0.3 mM BPDS (bathophenathroline dissulfonic acid) and 0.25 mM FeIIIEDTA in darkness for 1 h at 25°C. Reduction rates were determined as formation of a red FeII(BPDS)3 complex by measuring the absorbance at 535 nm spectrophotometrically against blanks (without roots) and using an extinction coefficient of 22.14 mM−1 cm−1 for calculation.

Uptake of 57Fe

Enriched 57Fe2O3 (96.64% isotopic enrichment; Isoflex, San Francisco, CA, USA) was converted into 57FeCl3 by dissolving in a minimum of sub-boiled concentrated HCl (35% suprapur; Merck, Darmstadt, Germany) and repeatedly concentrated in ultra-pure H2O (18 M; Millipore, Bedford, MA, USA) under reflux. EDTA (Sigma-Aldrich) was dissolved in a sufficient amount of NaOH (1 : 3 molar ratio). 57FeCl3 was slowly added into the ligand solution (57Fe:EDTA molar ratio of 1 : 1.1) and left to equilibrate overnight by stirring with filter paper pieces to absorb possible nonchelated precipitates. Thereafter, the solution was filtered through a membrane filter (0.45 μm pore size) and adjusted to a final volume. Exposure to light was prevented during chelation to avoid any potential photodecomposition of the complex.

For the 57Fe uptake experiment, roots of intact cucumber plants were washed with Fe-free nutrient solution for 30 min and then transferred to the nontransparent plastic boxes filled with 250 ml of a freshly prepared Fe-free nutrient solution. An enriched 57FeEDTA stock solution was added to give a final Fe concentration of 10 μM. The uptake solution was buffered at pH 5.5 with 10 mM MES/KOH and continuously aerated over the uptake period of 8 h.

After the uptake period, the plants were transferred to a freshly prepared Fe-free nutrient solution for 15 min. Root apoplastic 57Fe pool was removed during reductive extraction according to a method of Bienfait et al. (1985) slightly modified as described above, and then plants were separated into roots and shoots. Samples were oven-dried at 70°C, weighed and microwave digested in 10 ml HNO3 for 57Fe determination.

57Fe was determined by quadrupole inductively coupled plasma mass spectrometry (ICP-MS, Agilent 7500ce; Agilent Technologies, Manchester, UK). The instrument was equipped with a PFA microflow nebulizer and used in hydrogen mode to eliminate spectral interference(s). Instrument settings were as described previously (Laursen et al., 2009); however, before analysis the H2 flow of the octopole ion guide was optimized to give maximum interference free 56Fe and 57Fe signals, lowest possible backgrounds and accurate 56Fe/57Fe values according to the natural abundance ratio of 43.3 (IUPAC values). This was evaluated by hydrogen ramping on nonenriched samples and standards, which yielded accuracies of > 95% of the true 56Fe/57Fe ratio at the optimal H2 flow rate (7 ml min−1). All samples were diluted to 3.5% (v/v) HNO3 before analysis and external calibration was conducted to obtain total 56Fe and 57Fe concentrations using a commercially available standard solution (P/N 4400-132565; CPI International, Amsterdam, the Netherlands). The analytical accuracy of total concentrations was evaluated using certified reference material (Spinach NCS ZC73013; China National Analysis Center for Iron and Steel, Beijing, China). Concentration data was accepted if the accuracy exceeded 90% of the reference value. The accuracy of 56Fe/57Fe isotope ratios was evaluated by analysis of four nonenriched samples resulting in an average isotope ratio of 44.3 ± 2.3. Data was processed by the Masshunter Workstation software, version B.02.01 (Agilent Technologies).

RNA extraction and semiquantitative RT-PCR

Root tissue samples (0.5–1 g FW) were frozen in liquid N2 and ground thoroughly in a mortar. RNA was isolated using the RNeasy® Mini Kit (Qiagen) as described in the RNeasy® Mini Handbook.

Before cDNA synthesis DNA was removed from RNA samples using Ambion DNA-free DNase Treatment and Removal Reagents. First strand cDNA was synthesized from 5 μg of RNA with M-MuLV reverse transcriptase (Fermentas, Vilnius, Lithuania) and random hexamer primers (Applied Biosystems, Foster City, CA, USA) according to the manufacturer's instructions. The cDNA were diluted 1 : 5 (v/v) with nuclease-free H2O. Aliquots of the same cDNA sample were used for standard and Real-time PCR with primers designed for cucumber, sequences of the primers and accession numbers are given in Table S1. For semiquantitative RT-PCR, reactions were cycled 25 times for 30 s at 95°C, 30 s at 60°C and 30 s at 72°C, employing 100 ng of cDNA as the template, 450 nM of primers and 0.5 U of Fermentas Taq Polymerase. PCR products were electrophoresed on 1.5% (w/v) agarose gels.

Real-time quantitative PCR

Real-time PCR reactions were performed in 25 μl volume containing 500 nM of each primer and 1X SYBER Green PCR master mix (Applied Biosystems). Real-time PCR was performed on the ABI Prism 7500 Sequence Detection System (Applied Biosystem) using parameters recommended by the manufacturer (i.e. 2 min at 50oC, 10 min at 95°C and 40 cycles of 95°C for 15 s and 60°C for 1 min). Each PCR reaction was carried out in triplicate and no-template controls were included. Accumulation of PCR products was detected in real time and the results were analyzed with 7500 System software (Applied Biosystems) and presented as inline image (dCt, the difference between Ct values of target gene and the reference gene). In order to determine the amplification efficiency of real-time PCRs using newly designed primer pars, cDNA diluted 5-, 10-, 20- and 40-times were used as template in PCR. Mean Ct values were plotted against Log10 of sample dilutions and slopes of amplification plots were calculated. The efficiency of the PCR was determined using formula: E = 10−1/slope−1. For each of primer pairs efficiency of reaction was between 90 and 100% (−3.6 > slope > −3.1).

Sequence data from this article can be found in the GenBank/EMBL data libraries and EST database ( under accession numbers listed in Table S1.

Statistical analysis

Data were subjected to analysis of variance using the statistical software Statistica 6 (StatSoft Inc., Tulsa, OK, USA) and means were compared by Tukey's test at 5% significance level (P < 0.05).


Interactions between Fe and Si nutrition

Silicon supply successfully suppressed the symptoms of Fe deficiency in cucumber plants pre-cultured in Si-free nutrient solution with 1 μM FeIIIEDTA for 7 d, and further grown in Fe-free nutrient solution for another 7 d (Fig. 1). By contrast, plants grown without Si became severely chlorotic (Fig. 1a), and their roots showed specific morphological changes such as formation of short lateral roots (Fig. 1b). To further document the role of Si in alleviating Fe deficiency stress, the relative chlorophyll content measured as Spectral Plant Analysis Diagnostic (SPAD)-index, root and shoot dry biomass and leaf Fe concentration were measured (Table 1). In contrast to Fe adequate (+Fe) plants, where Si did not affect any of the examined parameters, a strong enhancing effect of Si supply was observed in Fe-deficient (−Fe) plants. For instance, the SPAD readings in young expanding leaves of Si-treated (+Si) plants were three-fold higher than those grown in the absence of Si (−Si). Accordingly, the leaf Fe concentration of +Si plants was > 50% higher than in −Si plants. Si nutrition also stimulated growth of the −Fe plants, recorded as significant increases in both root and shoot dry biomass. However, chlorophyll, leaf Fe and shoot biomass measurements of the −Fe/+Si plants were all significantly lower than those of the plants adequately supplied with Fe (+Fe), irrespectively of the addition of Si. Fe nutrition did not affect the leaf Si concentration, which depended only on Si addition to the nutrient solution (Table 1).

Table 1. Interactions between Fe and Si nutrition in cucumber (Cucumis sativus)
TreatmentsSi supplyChlorophyll (SPAD-units)Dry biomass (mg per plant)Leaf Fe concentration (μg g−1 DW)Leaf Si concentration (% DW)
  1. Plants were grown in nutrient solutions +Fe (50 μM) or −Fe, in treatments with or without supply of 1.5 mM Si(OH)4 for 7 d. Data are means ± SD (n = 4). Significant differences between treatments (< 0.05) are indicated by different letters.

+Fe37.6 ± 2.1c41.7 ± 17.6c296.3 ± 94.3c88.2 ± 19.1c0.36 ± 0.05a
+36.5 ± 1.5c41.6 ± 11.5c362.7 ± 82.6c97.7 ± 28.9c1.40 ± 0.11b
−Fe8.7 ± 1.3a12.3 ± 5.4a119.8 ± 46.9a19.4 ± 3.7a0.37 ± 0.08a
+24.9 ± 4.9b29.1 ± 10.9b216.2 ± 45.2b48.1 ± 5.0b1.57 ± 0.23b
Figure 1.

Effect of Si nutrition on visual symptoms of Fe deficiency in cucumber (Cucumis sativus). Two-wk-old plants (a); root morphological changes (b). Plants were pre-cultured for 7 d in nutrient solution supplied with 1 μM FeIIIEDTA and then grown for another 7 d in Fe-free nutrient solution, without (−Si) or with (+Si) supply of 1.5 mM Si(OH)4.

Expression analyses of Strategy 1 genes involved in Fe acquisition (FRO2, ITR1, HA1) showed that Fe-deficient plants treated with Si displayed a similar expression pattern as that observed in Fe-adequate plants, whereas in the absence of Si, all of these three genes were highly upregulated in response to Fe deficiency (Fig. 2a). These changes at expression levels of FRO2 and IRT1 resulted in enhanced root FeIII reducing capacity and 57Fe uptake in the plants grown in Fe-free nutrient solution without Si supply (Fig. 2b,c). In addition, the 57Fe root-to-shoot translocation rate was about two-fold higher as compared to Si-supplied plants (Fig. 2d). However, in Fe-adequate plants, the root FeIII reducing capacity and the 57Fe uptake and translocation remained very low regardless of the Si nutritional status.

Figure 2.

Effect of Fe and Si nutrition on the gene expression of proteins involved in Fe acquisition and the reduction-based uptake and translocation of Fe by cucumber (Cucumis sativus). Expression of HA1, FRO2 and IRT1 (a); FeIII chelate reducing capacity (b); uptake (c) and root-to-shoot translocation (d) of 57Fe by cucumber plants. Plants were grown in nutrient solution +Fe (50 μM) or −Fe, in the treatments without or with supply of 1.5 mM Si(OH)4 for 7 d. Black bars, −Si; grey bars, +Si. Amplification of Actin is shown as a control for equal template loading. Uptake solution contained 10 μM 57FeIIIEDTA (96.64% enriched). Data are means ± SD (n = 3). Significant differences between treatments (< 0.05) are indicated by different letters.

Dynamics of Si-mediated alleviation of Fe deficiency chlorosis

Time-course experiments were conducted to study the dynamics of relevant physiological markers and to elucidate further the mechanism of Si-mediated amelioration of Fe deficiency stress. Plants were pre-cultured for 7 d with 10 μM Fe supply, with (+Si) or without (−Si) addition of monosilicic acid, and then transferred to an Fe-free nutrient solution, again with or without Si supply. The +Si plants showed significant increase in the total dry biomass at the end of the experiment (7 d of −Fe treatment; Fig. 3a) and maintained approximately the same high concentration of chlorophyll of c. 36 SPAD-units during the whole period of Fe deprivation (Fig. 3b). During the first 4 d of Fe withholding from the nutrient solution, no significant differences between Si treatments were observed. However, on day 5 the young leaves of the −Si plants started to develop Fe-chlorosis symptoms and after 7 d without Fe supply these plants became clearly chlorotic, whereas the +Si plants remained green (Fig. 3d), as also reflected by distinctly higher SPAD readings relative to the −Si plants (Fig. 3a). Accordingly, the leaf Fe concentration of +Si plants was significantly higher than in −Si plants throughout the entire experiment (Fig. 3c).

Figure 3.

Effect of duration of Fe deficiency and Si supply on the plant growth and leaf Fe nutritional status in cucumber (Cucumis sativus). Dry plant (root + shoot) biomass (a); chlorophyll content (b); leaf Fe concentration (c). Cucumber plants 7 d after transfer to −Fe nutrient solution (d). Black circle, −Si; grey circle, +Si. Plants were pre-cultured in complete nutrient solution (10 μM Fe) −Si or +Si (1.5 mM) for 7 d and then transferred to Fe-free nutrient solution −Si or +Si. Data are means ± SD (n = 4).

Si modulates the activities and gene expression of proteins involved in Fe acquisition

During the whole period of −Fe treatment, FeIII chelate reductase activity varied greatly depending on the Si treatments (Fig. 4a). Already on the first day following withdrawal of Fe from the nutrient solution, FeIII chelate reductase activity increased significantly, and during the following 3 d became significantly higher in Si-fed plants than those in the −Si treatment. However, in the period between 3 and 7 d after withholding Fe, root FeIII chelate reductase activity gradually decreased and at the end of the experiment it was almost three-fold lower in the plants receiving Si as compared to those grown without Si. The expression profile of the FRO2 gene in response to Fe starvation and Si addition was in close agreement with the results of FeIII chelate reductase activity measurements (Fig. 4a,b). Also the expression of two other Strategy 1-related genes, namely IRT1 and HA1, increased in Si-treated plants during the first 3 d of −Fe treatment (Fig. 4c,d). By contrast, between 3 and 7 d of Fe deprivation, the expression of all the three examined Strategy 1 genes was downregulated in Si-supplied plants.

Figure 4.

Effect of Si supply on Fe deficiency responses of cucumber (Cucumis sativus) roots during 7 d of Fe deprivation. FeIII chelate reductase activity (a); FRO2 relative expression (b); IRT1 relative expression (c); HA1 relative expression (d). −Si (black bars); +Si (grey bars). Plants were pre-cultured in complete nutrient solution (10 μM Fe) −Si or +Si (1.5 mM) for 7 d and then transferred to Fe-free nutrient solution −Si or +Si. Data are means ± SD (n = 3). Significant differences between treatments (< 0.05) are indicated by different letters.

Si increases utilization of apoplastic Fe and its movement via xylem

The root apoplastic Fe concentration gradually decreased during −Fe treatment in both the −Si and +Si plants (Fig. 5a). Interestingly, Si supply to the nutrient solution during Fe pre-culture resulted in increased concentrations of root apoplastic Fe (day 0, Fig. 5a). After the first day of Fe deficiency, continuous Si supply significantly lowered the root apoplastic Fe concentration (Fig. 5a), while the Fe concentration measured in the xylem sap markedly increased in the +Si treatment compared to the −Si treatment (Fig. 5b).

Figure 5.

Effect of duration of Fe deficiency and Si supply on Fe concentrations in the root apoplast (a) and xylem sap (b) of cucumber (Cucumis sativus). −Si (black bars); +Si (grey bars). Plants were pre-cultured in complete nutrient solution (10 μM Fe) −Si or +Si (1.5 mM) for 7 d and then transferred to Fe-free nutrient solutions −Si or +Si. Data are means ± SD (n = 4); n.d., not detected. Significant differences between treatments (< 0.05) are indicated by different letters.

The concentration of citrate in the xylem sap significantly increased after 1 d of Fe deprivation, and it was highest in the Si-treated plants as the experiment continued until 5 d after withholding Fe (Fig. 6a). At the end of experiment (7 d of Fe deprivation), the xylem concentration of citrate significantly decreased in the +Si plants, while high values were maintained in the xylem sap of −Si plants. The concentration of malate remained relatively low during the first 5 d of −Fe treatment, but increased dramatically at the end of the experiment, most pronouncedly in the −Si plants (Fig. 6b). Addition of Si also increased concentrations of citrate and malate in the root tips in response to Fe deficiency (Fig. 7a,b). The concentrations of both organic acids were significantly higher in the +Si plants during the 5 d when Fe was withheld from the nutrient solution, while at the end of experiment they were below the concentrations recorded in the −Si plants.

Figure 6.

Effect of duration of Fe deficiency and Si supply on the concentrations of citrate (a) and malate (b) in xylem sap of cucumber (Cucumis sativus). −Si (black bars); +Si (grey bars). Plants were pre-cultured in complete nutrient solution (10 μM Fe) −Si or +Si (1.5 mM) for 7 d and then transferred to Fe-free nutrient solution −Si or +Si. Data are means ± SD (n = 4). Significant differences between treatments (< 0.05) are indicated by different letters.

Figure 7.

Effect of duration of Fe deficiency and Si supply on the concentrations of citrate (a) and malate (b) in the apical root tissue of cucumber (Cucumis sativus). −Si (black bars); +Si (grey bars). Plants were pre-cultured in complete nutrient solution (10 μM Fe) −Si or +Si (1.5 mM) for 7 d and then transferred to Fe-free nutrient solution −Si or +Si. Data are means ± SD (n = 3). Significant differences between treatments (< 0.05) are indicated by different letters.

Si enhances the accumulation of Fe mobilizing compounds in roots and stimulates expression of genes related to their biosynthesis

The concentration of catechins (present as a sum of catechin and epicatechin) and gallic acid gradually increased in root tissue of −Si plants during the experiment (Fig. 8a,b). However, from days 3 to 5 of Fe deprivation, significantly higher concentrations of both phenolic compounds were detected in the root tissue of plants supplied with Si. At the end of the experiment (day 7 of Fe deprivation), the root concentrations of catechins and gallic acid in +Si plants were below the values measured in plants without Si supply. The concentration of riboflavin in the root tissues of Si-fed plants significantly increased during the first day of −Fe treatment, and thereafter steadily decreased (Fig. 8c). This time-course contrasted with that of −Si plants in which the root concentration of riboflavin gradually increased and finally became > 1.5-fold higher than at the start of the Fe deprivation period.

Figure 8.

Dynamics of accumulation of Fe mobilizing/reducing compounds in the apical root tissue of cucumber (Cucumis sativus) as affected by duration of Fe deficiency and Si supply. Catechin + epicatechin (a); gallic acid (b); riboflavin (c). −Si (black bars); +Si (grey bars). Plants were pre-cultured in complete nutrient solution (10 μM Fe) −Si or +Si (1.5 mM) for 7 d and then transferred to Fe-free nutrient solution −Si or +Si. Data are means ± SD (n = 3). Significant differences between treatments (< 0.05) are indicated by different letters.

The effect of duration of Fe deficiency stress and Si supply was examined through the changes in relative expression of: two TCA cycle-related genes, ICD and MDH (encoding isocitrate dehydrogenase and malate dehydrogenase, respectively); PEPC (encoding phosphoenolpyruvate carboxylase); and two genes involved in biosynthesis of phenolics, SK (encoding shikimate kinase; shikimate pathway) and PAL (encoding phenylalanine ammonia-lyase; phenylpropanoid pathway) in cucumber roots (Fig. 9). Si supplementation resulted in significantly higher expression levels of MDH, PEPC and SK, being upregulated at the beginning of withholding Fe from the nutrient solution (Fig. 9b–d). At the end of the experimental period of 7 d the transcript abundance of all the examined genes was decreased in the +Si plants, and they were significantly lower than in the −Si plants. By contrast, a significant increase in transcript abundance of ICD, MDH, PEPC and PAL was observed in the −Si plants at this stage (Fig. 9a–d), whereas SK showed a decreased expression (Fig. 9e).

Figure 9.

Effect of duration of Fe deficiency and Si supply on the relative expression of some genes involved in the biosynthetic pathways of carboxylates and phenolics in cucumber (Cucumis sativus) roots. ICD (a); MDH (b); PEPC (c); SK (d); PAL (e). −Si (black bars); +Si (grey bars). Plants were pre-cultured in complete nutrient solution (10 μM Fe) −Si or +Si (1.5 mM) for 7 d and then transferred to Fe-free nutrient solution −Si or +Si. Data are means ± SD (n = 3). Significant differences between treatments (< 0.05) are indicated by different letters.

In addition, the effect of Si-induced accumulation of Fe-mobilizing compounds in the root apoplast/rhizosphere was further examined by the experiment with sparingly soluble FeIII hydroxide (Fig. 10). The −Si plants grown in the nutrient solution supplied with FeIII hydroxide were chlorotic, with accordingly low amounts of Fe taken up over the period of 10-d Fe treatment (Fig. 10a,d). Supply of Si significantly increased total DW and leaf chlorophyll content and total Fe uptake of the cucumber plants subjected to FeIII hydroxide (Fig. 10b–d).

Figure 10.

Effect of Si nutrition on the utilization of Fe from sparingly soluble FeIII hydroxide by cucumber (Cucumis sativus). Plants were pre-cultured in Fe/Si-free nutrient solution for 7 d and than transferred to −Si or +Si nutrient solution supplied with 1 μM Fe either as FeIIIEDTA (control; soluble Fe source) or Fe(OH)3. FeCl3 was converted to Fe(OH)3 with 1 mM KOH, and left to equilibrate over 30 min with continuous stirring. To diminish the utilization of Fe by acidification of the root surface, the pH of the nutrient solution was adjusted to 6.5 and kept constant by adding CaCO3 (10 mg l−1). Cucumber plants at the end of experiment (10 d of Fe treatment) (a); dry plant (root + shoot) biomass (b); chlorophyll content (c); Fe uptake (d). Fe uptake was calculated as a sum of Fe amounts in the root symplast (after reductive remobilization of Fe from the root apoplast) and Fe amounts in the shoots. Data are means ± SD (n = 3). Significant differences between treatments (< 0.05) are indicated by different letters.


From the first set of experiments it was shown that Si supply to the nutrient solution prevented or at least delayed development of Fe-deficiency induced chlorosis even though no Fe had been added to the growth medium (Table 1; Fig. 1a). Root responses to Fe deficiency (e.g. root morphological changes, Fig. 1b; reductive 57Fe uptake, Fig. 2b–d) were hence indirectly influenced by Si as a consequence of the improved Fe status in the whole plant, rather than directly caused by Si-modulated expression of Strategy 1-responsive genes (FRO2, IRT1 and HA1; Fig. 2a). However, the detailed mechanisms involved in amelioration of Fe deficiency by Si nutrition could not be clarified in this first set of experiments where only the final effects of long-term Si treatment were measured (see 'Materials and Methods'). Detailed responses to Fe deficiency seemed to be obscured by the obviously improved Fe nutrition in Si-treated cucumber plants.

In order to elucidate these mechanisms we conducted a time-course experiment focusing on the dynamics of utilization of root apoplastic Fe and Strategy 1 response reactions. In plants grown without Si supply, the first visual symptoms of Fe deficiency (chlorotic young leaves) were visible from the fifth day after transfer to Fe-free nutrient solution and progressively increased, whereas Si-treated plants remained green throughout the entire experiment (Fig. 3b).

The root apoplast is of particular importance for nutrient acquisition, because the epidermal cell walls contain negatively charged sites (e.g. pectic polysaccharides abundant in carboxylic groups), which can serve as storage for most cationic nutrients (Sattelmacher, 2001). For instance, c. 80% of the total Fe in roots was located in the apoplast, and the root apoplastic Fe was suggested to be an important Fe storage pool for higher plants (Bienfait et al., 1985). In the present study, supply of Si led to significantly higher Fe accumulation in the apoplast of cucumber roots in comparison to the −Si treatment (Fig. 5a). Although Si association with cell-wall components including polysaccharides, lignins and proteins is not fully understood (Currie & Perry, 2007), Si can be strongly bound to the pectins rich in galacturonate residues in vitro (Schwarz, 1973) and may thus contribute to cross-linking of the cell wall structures. On the other hand, polysilicate can fix metal ions as chelate-like complexes (Iler, 1979), which may also occur in the plant cell wall during polymerization of orthosilicic acid. This Si-induced extension of FeIII binding sites in the root apoplast can explain the increased concentration of apoplastic Fe found in the roots of Si-treated plants during pre-culture with Fe (day 0, Fig. 5a). Shortly after the Si-treated plants were transferred to the Fe-free solution (day 1, Fig. 5a), the rapid decrease of root apoplastic Fe was accompanied by significantly higher Fe concentrations in the xylem sap of Si-treated plants (Fig. 5b), thereby enabling uninterrupted Fe movement to the upper plant parts during the first 3 d of Fe deprivation.

It has been demonstrated that a decrease in the amount of root apoplastic Fe resulted in recovery of corresponding amounts of Fe in shoots when plants were transferred from an Fe-sufficient to an Fe-deficient medium (Zhang et al., 1991). Recently, Jin et al. (2007) suggested that root exudates such as phenolics played a crucial role in facilitating the reutilization of the apoplastic Fe in roots of white clover (Trifolium repens). Although phenolic compounds are frequently reported to be the main components of root exudates in response to Fe deficiency in Strategy 1 species (e.g. Römheld & Marschner, 1983; Jin et al., 2007; for a recent review see also Cesco et al., 2010), the mechanism of reutilization of apoplastic Fe by phenolic compounds still remains unclear. A probable underlying mechanism might involve the chelation of Fe bound in root apoplast by secreted phenolics, thus making it available for the reduction-based uptake by root cells and translocation to the upper plant parts. It has been proposed that soluble Si acts as a modulator of plant resistance to pathogens, including increased production of phenolics (e.g. lignin and phytoalexins) in response to fungal infection (Fauteux et al., 2006). To date, Si-modulated changes in phenol metabolism have been reported in Al-exposed corn roots (Kidd et al., 2001) as well as in the roots of cucumber subjected to high Mn concentrations (Dragišić Maksimović et al., 2007). The shikimate pathway is typically described as a link coupling carbohydrate metabolism with the biosynthesis of aromatic compounds via the phenylpropanoid pathway (Herrmann & Weaver, 1999). An increased synthesis of phenylpropanoids and terminal flavonoids in the symplast may then facilitate their secretion into the apoplast. In the present work, Si application led to an increase in the expression of transcripts related to both shikimate and phenylpropanoid pathways (SK and PAL, respectively) in the root tips 1–3 d after Fe deprivation (Fig. 9d,e). Consequently, Si-induced accumulation of flavonoid-type of phenolics (e.g. catechins) in the apical root tissue was observed during the very early stage of Fe deficiency (Fig. 8a). Additionally, riboflavin secretion from roots induced by Fe deficiency has also been reported in some plant species (e.g. Venkat Raju et al., 1972; Susín et al., 1996; Welkie, 2000). The clear role of riboflavin and other flavin-like compounds in Fe-deficient plants is still unknown, although various suggestions have been made as to their function including accumulation in root cells to act as a cofactor or a component of redox reactions, or export to the rhizosphere to enhance Fe availability (Rodríguez-Celma et al., 2011). Gallic acid appears not only to be an Fe chelating agent (Fazary et al., 2009), but also shows a strong potential to reduce FeIII in plant tissues (Mehrotra & Gupta, 1990). Therefore, Si-induced root accumulation of gallic acid and riboflavin at the beginning of withholding Fe from the nutrient solution (Fig. 8b,c) and thus their excretion in the apoplast, might also enhance both the reutilization of apoplastic FeIII and its reduction preceding Fe2+ transport across a PM via IRT. The enhanced root Fe-mobilizing potential of Si-treated plants was further confirmed by the increased Fe uptake and the concomitant regreening of cucumber plants subjected to sparingly soluble FeIII hydroxide (Fig. 10a,d).

FeIII chelate reducing capacity was significantly higher in the roots of Si-supplied cucumber plants up to 3 d following Fe exclusion from the nutrient solution; afterwards it continuously decreased to the end of experiment (Fig. 4a). The expression of Fe deficiency-responsive genes such as FRO2, IRT1 and HA1 followed the same pattern, which might indicate a Si-induced expression of these genes at the beginning of Fe deficiency stress (Fig. 4b–d). Later during the experiment, as the Si-treated plants maintained a critical Fe status for growth due to Si-mediated Fe reutilization from the root apoplast (see Fig. 5a,b), the expression of FRO2, IRT1 and HA1 were all downregulated, most probably by an Fe-dependent signal. It is for this reason that the Si effect was masked in the first set of experiments where all Fe-responsive root reactions of cucumber plants supplied with Si were already downregulated (see Fig. 2).

Among carboxylates present in the xylem sap, citrate has been considered as the most likely major candidate for Fe xylem transport (Rellán-Álvarez et al., 2010). It is also possible that under specific conditions (e.g. different plant species or varying Fe status) some other chelators (e.g. other carboxylates abundant in the xylem sap and/or nicotianamine) might form stable FeIII complexes in the xylem sap. In the present study, the concentration of citrate in the xylem sap increased as Fe deficiency symptoms developed (Figs 3b, 6a) and was significantly higher in Si-fed plants, particularly at the first day of Fe deprivation. The ratio of Fe:citrate in the xylem sap was 1 : 700 in +Si plants, while in −Si plants it exceeded 1 : 3000 (calculated from Figs 5b, 6a). An increase in the Fe:citrate molar ratio has been shown to inhibit FeIII-citrate reduction in leaves (González-Vallejo et al., 1999; Nikolic & Römheld, 1999); thus, Fe uptake in the young chlorotic leaves of −Si plants seems to be further impaired. However, increased citrate concentration in the xylem sap of +Si plants, not exceeding the Fe:citrate molar ratio of 1 : 1000, appears to be of great importance for both facilitating long-distance transport and utilization of Fe by leaf cells.

Increased carboxylate concentrations (citrate and malate; Fig. 7a,b) as well as the enhanced expression of TCA cycle/PEPC-related transcripts in the root tissues of Si-treated cucumber plants (Fig. 9a–c) are in agreement with the findings of increased citrate concentration in xylem sap (see Fig. 6). Until recently, there has been little information in the literature on the contrasting effect of Si on organic acid metabolism. For instance, while Wang et al. (2004) reported that Si did not affect Al-induced exudation of organic acid anions from the corn root apices, Si has been found to enhance the root concentration of succinate and both root and shoot concentrations of malate in Al-exposed corn plants (Barceló et al., 1993). Overall, Si has been shown to be responsible for changes in metabolite pools under certain stress conditions (e.g. pathogen attack, metal toxicity; Fauteux et al., 2006; Führs et al., 2012), even though it remained unclear from these studies whether this was a primary or a secondary effect of Si.

Our results clearly show for the first time that, under Fe-deficient conditions, Si can influence the carboxylate, shikimate and phenylpropanoid metabolism as well as the accumulation of riboflavin in cucumber roots. It is obvious that Si, either in solid form of plant opal or by its mere presence as a soluble orthosilicic acid, cannot directly influence the metabolome of unstressed plants (Epstein, 2009), but Si can act as an enhancer or positive modulator of cell responses to Fe deficiency. Experimental data related to impact of Si on gene transcription is scarce. Recent transcriptome microarray analysis with an Arabidopsis-powdery mildew system has demonstrated that Si supply does not impact expression of defence-related genes in the absence of pathogen stress and thus any potential advantage of Si supply might not be apparent unless plants are stressed (Fauteux et al., 2006). Possibly, Si has the potential to enable an individual plant to react adaptively as biotic or abiotic conditions change. Because Si does not have an effect on gene expression per se, in the unstressed (+Fe) plants (see Fig. 2a), one may only speculate as to how Si might indirectly affect transcription factors leading to transcriptional activation of Fe deficiency associated genes. In fact, our results also show the ability of Si to modulate root activity of Fe acquisition at the early stage of Fe deficiency stress through regulation of gene expression levels of the proteins (i.e. FRO2, IRT1 and HA1) involved in this process. However, if the plants are subjected to Fe deficiency for a longer period of time, these reactions (enhanced at the beginning of Fe deprivation) decrease as a consequence of Si-based increase of plant Fe status.

In conclusion, the presented results indicate that the role of Si in alleviation of Fe deficiency chlorosis includes an increase of the apoplastic Fe pool in the roots, and an enhancement of Fe mobilization in the roots due to Si-mediated biosynthesis of Fe chelating compounds. In perspective, these findings may be of practical importance in the development of new sustainable measures for controlling Fe chlorosis in calcareous soils, which in general are low in available Si.


This research was supported by the Serbian Ministry of Education, Science and Technological Development (ON-173028 and ON-173005) and in part by the Danish Ministry of Science, Technology and Development (2104-08-0039 Bio4Bio/DSF). We thank Petar Ilic for technical assistance during the experiments, Mrs Nina Nikolic and Dr Ernest A. Kirkby (University of Leeds, UK) for correction of the English, and Dr Nikolai Petrovitch Bityutskii (St. Petersburg State University, Russia) for providing the copy of his article.