Studies of Iron (Fe) uptake mechanisms by plant roots have focussed on Fe(III)-siderophores or Fe(II) transport systems. Iron deficency also enhances root secretion of flavins and phenolics. However, the nature of these compounds, their transport outside the roots and their role in Fe nutrition are largely unknown.
We used HPLC/ESI-MS (TOF) and HPLC/ESI-MS/MS (ion trap) to characterize fluorescent phenolic-type compounds accumulated in roots or exported to the culture medium of Arabidopsis plants in response to Fe deficiency. Wild-type and mutant plants altered either in phenylpropanoid biosynthesis or in the ABCG37 (PDR9) ABC transporter were grown under standard or Fe-deficient nutrition conditions and compared.
Fe deficiency upregulates the expression of genes encoding enzymes of the phenylpropanoid pathway and leads to the synthesis and secretion of phenolic compounds belonging to the coumarin family. The ABCG37 gene is also upregulated in response to Fe deficiency and coumarin export is impaired in pdr9 mutant plants.
Therefore it can be concluded that: Fe deficiency induces the secretion of coumarin compounds by Arabidopsis roots; the ABCG37 ABC transporter is required for this secretion to take place; and these compounds improved plant Fe nutrition.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Iron (Fe) availability for plants depends on the physico-chemical properties of the soil (Lindsay & Schwab, 1982). Calcareous soils cover 30% of the earth surface (Vose, 1982) and favour the formation of scarcely soluble Fe3+ oxy-hydroxides. Consequently, plants grown on these soils are often Fe deficient and develop interveinal chlorosis. According to the plant family considered, that is, Graminaceae vs other species, plants have evolved two different mechanisms for mining Fe from the soil. One involves chelation of Fe3+ by small organic molecules belonging to the mugineic acids (MAs) family, and the other involves Fe3+ reduction to Fe2+ before transport across the plasmalemma of root epidermal cells (Morrissey & Guerinot, 2009). MAs are synthesized from S-adenosyl methionine, and are secreted into the rhizosphere by the TOM transporter (Nozoye et al., 2011). The resulting Fe3+–MAs complexes are taken up by the roots via YS1 transporters (Curie et al., 2001). Nongrass plants respond to Fe deficiency with both morphological and physiological changes (Schmidt, 1999). The AHA2 H+-ATPase acidifies the rhizosphere of Arabidopsis iron (Fe) deficient plants, facilitating Fe solubilization (Santi & Schmidt, 2009). An enhanced Fe3+ reduction capacity of the roots (Yi & Guerinot, 1996) involves the FRO2 root Fe3+ chelate reductase (Robinson et al., 1999). Then Fe2+ is transported across the root plasma membrane via the IRT1 divalent metal transporter (Eide et al., 1996; Vert et al., 2002).
In addition, it has been known for many years that plants also respond to Fe deficiency by enhancing root secretion of organic compounds, including flavins and phenolics (Römheld & Marschner, 1983; Susín et al., 1993; Rodríguez-Celma et al., 2011; Donnini et al., 2012). The role of these molecules is still unclear. Flavins could participate by reducing or complexing extracellular Fe (González-Vallejo et al., 1998; Cesco et al., 2010) or bridging the electron flow to the root Fe3+ chelate reductase (López-Millán et al., 2000; Higa et al., 2010). Back in the seventies, phenolic compounds were already considered as putative external Fe reductants and/or chelators (Brown & Ambler, 1973). More recently, it was reported that phenolics secreted by red clover contribute to solubilization and utilization of apoplasmic Fe (Jin et al., 2007), and that phenolics such as protocatechuic acid can solubilize and chelate Fe3+ and reduce it, both in vitro (Yoshino & Murakami, 1998) and within the plant (Bashir et al., 2011; Ishimaru et al., 2011).
Therefore, the chemical identification of these secreted compounds, the understanding of their biosynthetic pathways and the characterization of the root plasmalemma transporters responsible for export are of major interest to understand their possible roles in plant response to Fe deficiency. Transcriptome and proteome analyses of roots from Arabidopsis Fe-deficient plants have provided evidence that some enzymes of the phenylpropanoid pathway were upregulated, likely leading to an increase in scopoletin synthesis (Yang et al., 2010; Lan et al., 2011). Also, rice phenolics efflux transporters were characterized and shown to be essential for solubilizing apoplasmic Fe precipitated in the stele (Bashir et al., 2011; Ishimaru et al., 2011). In this paper, we characterize the fluorescent phenolic compounds accumulating in roots of Fe-starved Arabidopsis plants, as well as those secreted to the medium, and we show that their secretion is dependent upon the functionality of the ABCG37 gene (previously named PLEIOTROPIC DRUG RESISTANCE 9: PDR9) belonging to the ATP-Binding Cassette family of transporters (Ito & Gray, 2006; Strader et al., 2008; Strader & Bartel, 2009; Ruzicka et al., 2010).
Materials and Methods
Plant materials and growth conditions
Arabidopsis thaliana (L.) Heynh (ecotype Col0) seeds were surface sterilized and sown in 0.2-ml tubes containing 0.8% agar prepared in pH 5.5 Hoagland-based solution (in mM: 1 KH2PO4, 1 KNO3, 1 MgSO4, 5 CaNO3; and in μM: 50 H3BO3, 0.05 CoCl2, 0.05 CuSO4, 15 ZnSO4, 2.5 KI, 50 MnSO4, 3 Na2MoO4). Iron was added as 50 μM Fe3+-EDTA. After 3–4 d in the growth chamber, the tube bottoms were cut off and tubes placed in opaque plastic pipette tip racks (Starlab, Hamburg, Germany). Twelve plants were grown in boxes with 300-ml half-Hoagland solution (Terry, 1980) with 50 μM Fe3+-EDTA. Growth conditions were 23°C, 70% relative humidity, 8 h : 16 h, light : dark photoperiod and 220 μE m−2 s−1 PPFD (PAR). Plants were grown for 4 wk before being subjected to Fe deficiency (zero Fe in the solution) for 7 d (final pH was c. 7.0). For in vitro experiments, sterilized seeds were sown on 0.8% (w/v) agarose-solidified Hoagland media. Mutant lines pdr9-2 (SALK 050885) and pdr9-3 (SALK 078574) were obtained from the Salk center, and seeds from the comt and ccoamt-1 mutants were a generous gift of Dr Lise Jouanin (IJPB, Versailles, France; Do et al., 2007). The irt1 Arabidopsis mutant was obtained previously in our laboratory (Vert et al., 2002).
Brassica napus L. var. napus (cv Fantasio) seeds were a gift from Semences de France (In vivo Group, La Chapelle d'Armentière, France). Seeds were placed onto cheesecloth held with a plastic ring inside a Magenta vessel. Culture medium and growth conditions were identical to those of the Arabidopsis culture.
RNA extraction and quantitative RT-PCR analysis
Total RNA was extracted from Arabidopsis roots using the RNeasy Plant Mini Kit (Quiagen). One microgram RNA was treated with RQ1 DNase (Promega) before use for reverse transcription (Goscript reverse transcriptase; Promega) with oligo(dT)18 and 0.4 mM dNTPs (Promega). The cDNAs were diluted twice with water, and 1 μl of each cDNA sample was assayed by qRT-PCR in a LightCycler 480 (Roche Applied Science) using Lightcycler 480 SYBR Green master I (Roche Applied Science). Expression levels were calculated relative to the housekeeping gene PP2 (At1g13320) using the ∆∆CT method to determine the relative transcript level. The primers used in this study are presented in Supporting Information Table S1.
Sample preparations for analytical measurements
Reagents, materials and standard solutions
Reagents (eluents and phenolic standards) are shown in Table S2. For dilutions, we used analytical grade type I water (Ultramatic, Wasserlab, Pamplona, Spain). All standards were dissolved in methanol to concentrations in the range 1–5 mM and stored in darkness at −80°C until analysed. Working standard solutions were prepared before use by diluting the stock solutions with a solution 85 : 15 (v/v) of 0.1% (v/v) formic acid in methanol and 0.1% (v/v) formic acid in water.
Extraction of phenolic compounds from roots and nutrient solutions
Roots frozen in liquid N2 and stored at −80°C were extracted following the procedure described by Kai et al. (2006), with several modifications. Frozen roots (c. 100 mg) were ground in liquid N2 using a Retsch M301 ball mill (Restch, Düsseldorf, Germany) for 3 min and then phenolic compounds were extracted from the tissue with 1 ml of 100% methanol by homogenization in the same mill for 5 min. The supernatant was recovered by centrifugation (12 000 g at 4°C and 5 min), stored at −20°C, and the pellet was resuspended in 1 ml of 100% methanol, homogenised again for 5 min and the supernatant recovered. The two supernatant fractions were pooled, vacuum concentrated and diluted in methanol to a final volume of 250 μl. Extracts were filtered through polyvinylidene fluoride (PVDF) 0.45 μm ultrafree-MC centrifugal filter devices (Millipore) and stored at −80°C until analysis.
In order to examine Fe-deficiency induced changes on the phenolic compounds secreted by A. thaliana roots, nutrient solutions from hydroponically grown Arabidopsis plants (12 plants grown in 300 ml-containers) were collected 7 d after the onset of Fe treatments and filtered using PVDF 0.45 μm membrane filters. Then, phenolic compounds were retained in a SepPack C18 cartridge (Waters, Mildford, MA, USA), eluted from the cartridge with 2 ml of 100% methanol and stored at −80°C until analysis.
Phenolic compound analysis by HPLC coupled to fluorescence detection
Separations were performed with a binary HPLC pump (Waters 1525) using an analytical HPLC column (Symmetry® C18, 15 cm × 2.1 mm i.d., 5 μm spherical particle size, Waters) protected by a guard column (Symmetry® C18, 10 mm × 2.1 mm i.d., 3.5 μm spherical particle size; Waters), with a gradient mobile phase built with 0.1% (v/v) formic acid in water (A) and 0.1% (v/v) formic acid in methanol (B), and a flow rate of 0.2 ml min−1. The gradient program started at 15% B for 2 min, then increased linearly to 55% B for 13 min. This proportion was maintained for 10 min and then returned linearly to initial conditions in 5 min. The column was then allowed to stabilise for 15 min at the initial conditions before the next injection. The injection volume was 20 μl. Phenolic compounds were detected with a scanning fluorescence detector (Waters 474) using λexc 365 and λem 460 nm. Empower software (build 2154; Waters) was used to control the HPLC system and process data.
Phenolic compound analyses by HPLC/ESI-MS(TOF) and by HPLC/ESI-MS/MS(ion trap)
HPLC/ESI-MS(TOF) analysis was carried out using an Alliance 2795 HPLC system (Waters) coupled to a time-of-flight mass spectrometer (MicrOTOF, Bruker Daltonics, Bremen, Germany) equipped with an electrospray (ESI) source. Autosampler and column temperatures were 6 and 30°C, respectively, and the HPLC conditions were the same as those described above. The ESI-MS(TOF) operating conditions were optimized by direct injection of 50 μM solutions of phenolics standards at a flow rate of 250 μl h−1. Mass spectra were acquired in positive ion mode (see settings in Table S3) in the range of 50–1000 mass-to-charge ratio (m/z) units. The mass axis was calibrated externally and internally using Li-formate adducts (10 mM LiOH, 0.2% (v/v) formic acid and 50% (v/v) 2-propanol). The internal mass axis calibration was carried out by introducing the calibration solution with a divert valve at 1–3 and 40–43 min in each HPLC run. Bruker Daltonik (Bruker Biosciences Espanola, S.A. Madrid, Spain) software packages micrOTOF Control v2.2, HyStar v3.2 and Data Analysis v4.0 were used to control the ESI-MS(TOF) apparatus, interface the HPLC with the MS system and process data, respectively. Molecular formulae were assigned based on exact molecular mass with errors < 5 ppm (Bristow, 2006).
HPLC/ESI-MS/MS(ion trap) analysis was carried out with an 1100 HPLC system (Agilent Technologies, Waldbronn, Germany) coupled to an ion-trap mass spectrometer (Esquire 3000+; Bruker Daltonics) equipped with an ESI source. The HPLC conditions were those described above. ESI-ion trap-MS analysis was carried out in positive ion mode, the MS spectra were acquired in the standard/normal mass range mode and the mass axis was externally calibrated with a tuning mix (from Agilent Technologies) (see settings in Table S3). The protonated ions of interest [M+H]+ were subjected to collision induced dissociation (CID; using the He background gas present in the trap for 40 ms) to produce a first set of fragment ions, MS/MS or MS2. Subsequently, one of the fragment ions was isolated and fragmented to give the next set of fragment ions, MS3 and so on. For each precursor ion, fragmentation steps were optimized by visualizing the intensity changes of the fragmented ions (see settings in Table S3).
Metal analyses by ICP-MS
Arabidopsis thaliana roots and leaves were dried at 65°C for 2 d. Approximately 0.01–0.04 g of ground dry tissues were digested with 5.6 ml HNO3 (4.6%) and 1.4 ml H2O2 (30%) in a vessel microwave oven (Milestone Model MLS 1200, Sorisole, Italy). Samples were diluted to 10 ml with Milli-Q water and total Fe, Mn, Cu and Zn was determined by ICP-MS (Agilent 7500ce equipped with an octapole collision cell to remove polyatomic interferences; Agilent Technologies, Tokyo, Japan).
Iron deficiency induces the secretion of phenolic compounds by Brassica napus and Arabidopsis thaliana roots
The culture medium of Brassica napus grown without Fe for 7 d appeared fluorescent when exposed to 365 nm UV light (Fig. 1a). The culture medium fluorescence was detected after 3 d of Fe deficiency and increased up to 8 d (Fig. 1b). Roots and growth media (agar or nutrient solution) fluorescence was also observed when Arabidopsis plants were grown under low Fe concentrations (Fig. 1c,d), although Arabidopsis root density was lower than the one of B. napus. Furthermore, the increase in Arabidopsis root fluorescence was inversely proportional to the Fe concentration in the agar medium (Fig. 1c). The impact of Fe deficiency was evidenced by the red leaf chlorophyll fluorescence, which was much less pronounced at 0 Fe than at 50 μM Fe (Fig. 1c). Therefore, Fe deficiency induces the synthesis, root accumulation and secretion to the growth media of UV fluorescent compounds in B. napus and A. thaliana roots. Their maximum fluorescence emission, when excited at 365 nm, was observed at 420–460 nm indicating that they likely belong to the phenolic compounds family.
In order to know whether the compounds secreted to the growth medium impact plant Fe nutrition, we used the phenolic removal system described by Jin et al. (2007), which consisted of circulating the nutrient solution through a Sep-Pak C18 resin column. This system almost completely removed fluorescent compounds from the nutrient solution (Fig. S1). Removing these molecules from the hydroponic medium led to a marked chlorosis of the Arabidopsis leaves compared to the control in which the nutrient solution was re-circulated without placing the C18 resin in the circuit (Fig. 2). This simple experiment clearly demonstrates that the Fe deficiency-induced secretion of phenolics by Arabidopsis roots enhances Fe nutrition efficiency and limits leaf chlorosis.
In order to establish that the leaf chlorosis was the result of iron deficiency, and did not have any other cause, we measured Fe, Zn, Mn and Cu concentrations in young leaves and roots from Arabidopsis thaliana plants grown hydroponically with the nutrient solution circulated or not through the Sep-Pak C18 resin column. Young leaves from plants grown on the circulated medium had 60% more Fe than those of plants grown on medium passed through a Sep-Pak C18 column, supporting that the observed chlorosis was the result of Fe deficiency (Fig. 2 b). A lower (only 15%) decrease in root Fe concentration was found in plants where the culture medium was circulated through the column. The changes in the concentrations of other metals when using the column are shown in Table S4.
Involvement of the phenylpropanoid pathway and of the ABC transporter ABCG37 in the Arabidopsis Fe deficiency responses
Our above observations are consistent with the output of recent transcriptomes (Yang et al., 2010; Rodríguez-Celma et al., 2013) and proteome (Lan et al., 2011) studies in roots of Fe-deficient Arabidopsis plants, which indicated that Fe shortage induced the phenylpropanoid pathway in roots.
Also, Fe deficiency led to an increase in abundance of the Arabidopsis ABCG37 transporter transcript (Yang et al., 2010), known to be expressed predominantly in the lateral root cap and in epidermal cells of the root tip (Ito & Gray, 2006). ABCG37 has been described as a plasma membrane exporter of the auxin storage precursor indole-3-butyric acid (Strader et al., 2008; Strader & Bartel, 2009; Ruzicka et al., 2010). As stated by Yang et al. (2010) the role of ABCG37 in Fe homeostasis remains elusive, but its ability to transport the synthetic auxin analogue 2,4-dichlorophenoxyacetic acid suggests an alternative role in the export of phenolic compounds such as caffeic acid or chlorogenic acid.
As already reported by Yang et al. (2010), and more recently by Rodríguez-Celma et al. (2013) we observed an increased expression of ABCG37 and F6′H1 (encoding an oxidoreductase involved in scopoletin synthesis; Kai et al., 2008) genes in Arabidopsis roots in response to Fe deficiency (Fig. 3), both in Col0 and irt1 mutant genetic backgrounds (Fig. 3). These results indicate that the functionality of the major Fe2+ high affinity root transporter is not required for the regulation of ABCG37 and F6′H1 in response to Fe shortage. IRT1 and FRO2 are among the best documented genes induced in response to Fe deficiency (Robinson et al., 1999; Vert et al., 2002; Morrissey & Guerinot, 2009). IRT1 and FRO2 gene expression was slightly lower in pdr9 lines than in Col independently of the Fe conditions, but the induction factor in response to Fe deficiency was roughly the same, (Fig. 3). CCoAMT1 (encoding trans-caffeoyl-CoA 3-O-methyltransferase) gene expression was Fe-deficiency responsive in all the genotype tested whereas COMT (encoding caffeic acid/5-hydroxyferulic acid O-methyltransferase) gene expression was Fe-deficiency independent. The upregulation of F6′H1 in response to Fe deficiency was lost in the comt and ccoamt1 mutant plants, and was strongly decreased in one of the pdr9 allele. Whether or not the regulation of ABCG37 expression was indicative of its role in Fe acquisition was tested by comparing the growth of Arabidopsis Col0 plants with that of the pdr9-2 T-DNA KO mutant, in the presence or absence of Fe at two different pH values (Fig. 4a). In the presence of Fe at pH 5.5, no differences were observed between the two genotypes. By contrast, at pH 6.5 and in the absence of Fe the pdr9-2 seedlings were smaller and more chlorotic when compared to Col0 plants under the same conditions. At pH 6.5 Fe is less available than at pH 5.5. Consequently Col0 plants in the absence of Fe were more chlorotic at this pH than at pH 5.5, and it was even more pronounced for the pdr9-2 plants (Fig. 4a). Interestingly, when using 365 nm UV light the medium, and to a lesser extent the roots, of the pdr9-2 plantlets grown on agarose without Fe appeared much less fluorescent than those of Col0 plants (Fig. 4b). These observations establish a link, therefore, between a defect in the ABCG37 transporter, a decrease in fluorescent compounds in the medium, and the development of leaf chlorosis, in particular at acidic pH. The above observations prompted us to investigate in more detail the nature of the fluorescent compounds accumulated in roots and secreted to the medium of Col0 and pdr9 Arabidopsis plants in response to Fe deficiency. As a control we also included the Arabidopsis mutant comt. The rationale for using this mutant in our analyses was based on the fact that the enzyme COMT is involved in the methylation of several substrates in the phenylpropanoid biosynthesis pathway. The production of ferulate from cafeate is altered in comt, compromising the synthesis of feruloyl-CoA, which is o-hydroxylated by the pivotal enzyme F6′H1 for the biosynthesis of scopoletin.
ABCG37 exports scopoletin and derivatives
Accumulation and secretion of a range of fluorescent compounds by Arabidopsis roots
Methanolic extracts were obtained from both roots and compounds occurring in the nutrient solutions, and then analysed by reverse phase C18 HPLC coupled to fluorescence (with λexc 365 and λem 460 nm). Each of the peaks in the chromatograms (Fig. 5) can contain one or more fluorescent compounds (see sections below for identification).
In Col0, the chromatograms of all root extracts showed two peaks, at c. 10 and 15 min, whereas those from growth media showed an additional peak at 19 min (Fig. 5). Under Fe deficiency, the area of the 15 min peak increased in the root extracts, whereas in the growth media large increases were observed for the 15 min peak, and to a lower extent for the 19 min peak.
The chromatograms from pdr9-2 and pdr9-3 root extracts also showed two peaks at 10 and 15 min, along with three additional peaks at 8, 11.5 and 13 min (Fig. 5). Under Fe deficiency, the area of all these peaks increased, with a larger increase for the peak at 15 min in pdr9-3. These observations indicate that the pdr9 mutation caused an increase in abundance of several fluorescent compounds in roots, some of them with higher polarity than those present in Col0, and that this increase in abundance was further enhanced by Fe deficiency. In the growth media extracts, only a minor peak at 15 min was observed in pdr9-2 and pdr9-3 that was much smaller than that found at the same retention time (RT) in Col0. Therefore, the pdr9 mutation markedly impairs the secretion of the fluorescent compounds eluting at 15 min that occurs in Col0 in response to Fe deficiency.
Chromatograms of comt root extracts showed two peaks at 10 and 15 min. Although these RT were similar to those of the peaks present in Col0, peak areas were much larger in comt. Furthermore, the area of both peaks was only slightly affected by Fe deficiency. These results indicate that the comt mutation results in the accumulation in the roots of fluorescent compounds, and that Fe deficiency did not result in major changes in their abundance. No fluorescent peaks were observed in the comt growth media, indicating that the step catalysed by COMT could be involved in the synthesis of either the precursors or the actual fluorescent compounds excreted by the Col0 roots in response to Fe deficiency. Nevertheless, so far we do not understand why the comt mutant failed to secrete these fluorescent compounds, including scopoletin.
Identification of the Arabidopsis fluorescent compounds as scopoletin and derivatives
The identity of the fluorescent compounds found in the Arabidopsis root extracts and secreted to the medium was studied in detail by HPLC coupled to two MS techniques, ESI-MS(TOF) and ESI-MS/MS (ion trap), with the same HPLC conditions used in the fluorescence analysis. The HPLC/ESI-MS(TOF) analysis provides highly accurate m/z measurements (mass accuracy < 5 ppm) that allows for accurate elemental formula assignments (Bristow, 2006). The identification of the nine unknown compounds behind the fluorescence peaks observed by HPLC analysis (Fig. 5) was tackled first by using TOF data (RT, exact m/z values and elemental formulae) (Table 1). Three of the nine compounds – 6, 7 and 8 in Fig. 5 – matched the RT and m/z values of the known standards fraxetin, scopoletin and isofraxidin, respectively (three fluorescent coumarins) (Tables 1, 2; Fig. S2). We confirmed these identifications using the MS2 spectra, which matched with those obtained for the standards, both in terms of m/z values and relative intensities of the major fragment ions (Tables 1 and 2). Compound 9 had the same m/z value and elemental formula as isofraxidin (8-metoxy scopoletin or methylated fraxetin; see structures in Table 2), but eluted 0.8 min later, supporting the hypothesis that it could be a less polar isofraxidin isomer. It was identified as 5-metoxy scopoletin (see notes S1 for more details).
Table 1. Retention times (RT) and [M+H]+ exact mass-to-charge ratios (m/z), molecular formulae and error of the compounds secreted and accumulated by roots of Col0, pdr9-2, pdr9-2 and comt Arabidopsis roots in response to iron (Fe) deficiency
Error m/z (ppm)
ESI-MSnm/z (Relative intensity%)
The m/z ratios of parent and fragment ions were determined from the data in the HPLC/ESI-MS(TOF) and HPLC/ESI-MS/MS(ion trap) chromatograms, respectively. In the case of compound 3, the MS(TOF) m/z included in the table is the one measured for the Na adduct ([M+Na]+) because it was more intense than the [M+H]+; the [M+H]+ (at 357 m/z) ion was subsequently used as precursor in the MS/MS(ion trap) analyses.
Table 2. Chemical structures, retention times (RT) and [M+H]+ exact mass-to-charge ratios (m/z), molecular formula and error of a selected group of phenolic compound standards; esculin, esculetin, and fraxetin
The MS(TOF) spectra of compounds 1–5 showed additional ions of high intensity at m/z 225.0393, 193.0498, 195.0646, 239.0547 and 225.0386, respectively, all of them consistent with the loss of a glucosyl moiety (−162.0523 Da, with an error below 5 ppm; see notes S1 for more details). The identity of glucosides 1–5 was further deciphered by MS3 using as a model the fragmentation of esculin (see Fig. S3). Compounds 2, 1 and 5, 4 and 3 were identified as glucoside of scopoletin (scopolin), as two different glucosides of dihydroxyscopoletin, as a glucoside of hydroxymethoxyscopoletin, and as a glucoside of ferulic acid, respectively (see Notes S1 for more details).
Changes in scopoletin and derivatives in roots and growth media in response to Fe deficiency
TOF ion chromatograms extracted at the corresponding exact m/z values of each individual phenolic compound for the different root extract and nutrient solution samples (sampled 7 d after imposing the treatments) are shown in Fig. 6 (note some panels have different y-axes). Roots of +Fe Col0 contained minor amounts of scopoletin (7; Fig. 6f), and only traces of other derivatives, including isofraxidin and metoxyscopoletin (8, 9; Fig. 6g), glucoside of scopoletin (2; Fig. 6b), and glucoside of ferulic acid (3; Fig. 6c). The growth media of +Fe Col0 plants contained almost exclusively minor amounts of glucoside of scopoletin (2; Fig. 6b), with only traces of scopoletin being present (7; Fig. 6f). When grown in the absence of Fe, the peak areas of all the phenolics already present in +Fe Col0 roots increased several-fold (up to 20-fold), whereas those of the phenolics present in the growth medium increased markedly: large amounts of scopoletin derivatives, including isofraxidin and methoxyscopoletin (8–9; Fig. 6g) and fraxetin (6; Fig. 6e) were found. On the other hand, the amount of scopoletin also increased (7; Fig. 6f), and no changes were found for the glucoside of scopoletin (2; Fig. 6b).
Roots of +Fe pdr9-2 and pdr9-3 contained glucosides of dihidroxyscopoletin (1 and 5; Fig. 6a) and some glucoside of scopoletin (in pdr9-3; 7; Fig. 6f), as well as minor amounts of other derivatives, including the glucosides of hydroxymethoxyscopoletin (4; Fig. 6d) and scopoletin (2; Fig. 6b), and traces of glucoside of ferulic acid (3); pdr9-2 also contained traces of isofraxidin (8; Fig. 6g). The growth media of Fe-sufficient pdr9-2 and pdr9-3 plants contained only minor traces of phenolic compounds. When grown in the absence of Fe, the peak areas of all the phenolics already present in +Fe pdr9-2 and pdr9-3 increased several-fold, and fraxetin was in large amounts in pdr9-2 (6; Fig. 6e). In the growth medium, the phenolics peak areas were low in comparison with Col0, specially in pdr9-2 (in pdr9-3 the small increase was mainly due to isofraxidin and metoxyscopoletin; Fig. 6g).
Roots of +Fe comt contained large amounts of scopoletin (7; Fig. 6f) and minor amounts of glucosides of scopoletin (2; Fig. 6b) and glucoside of ferulic acid (3; Fig. 6c) and isofraxidin (8; Fig. 6g). The comt mutation fully abolished phenolics export to the medium. When grown in the absence of Fe, the concentrations of root phenolics did not change, except for the glucoside of ferulic acid (3; Fig. 6c), which increased, whereas the secretion of phenolics to the medium was also abolished.
Therefore, phenolics secreted to the growth media in Col0 consisted mainly of coumarins structurally analogous to scopoletin (7), including fraxetin (6) and the isomers isofraxidin and methoxyscopoletin (8 and 9), and export of these compounds to the medium in the pdr9 plants was very low. The glucosides of coumarins were never found in the growth media, with the sole exception of minor amounts of glucoside of scopoletin in Col0. Phenolics in roots of Col0 and comt consisted of scopoletin (7), isofraxidin and methoxyscopoletin (8 and 9), whereas in the case of the pdr9 mutants the amounts of coumarin glucosides were also large.
In general, the changes in the peak areas of the coumarins found in Fig. 6 were in line with the changes observed in the HPLC-fluorescence chromatogram in Fig. 5. Exceptions were the large amounts of fraxetin (6) seen in roots of -Fe pdr9-2 and nutrient solutions of -Fe Col0, which did not result in major fluorescence changes in the chromatograms at the corresponding RT. This is likely due to the low fluorescence yield caused by the hydroxylation (Crosby & Berthold, 1962).
Studies on reduction and chelation-based mechanisms for Fe uptake by plant roots in response to Fe deficiency have mainly focussed on the FRO/IRT (Vert et al., 2002) and YS1/YSL (Curie et al., 2001) systems. However, Fe-efficient plants also respond to Fe deficiency by enhancing the root secretion of organic compounds, such as flavins or phenolics (Römheld & Marschner, 1983; Susín et al., 1993). The role of secreted red clover phenolics in increasing Fe efficiency was recently documented (Jin et al., 2006, 2007). Consistent with this report we observed that both Brassica napus and Arabidopsis thaliana roots produced fluorescent compounds, the amount of which increased in response to Fe deficiency. Part of these compounds were secreted to the nutrient growth medium, and this secretion was enhanced by Fe shortage (Fig. 1). Furthermore, passing the culture medium of Arabidopsis plantlets through a Sep-Pak C18 column led to enhanced leaf chlorosis (Fig. 2a), and a 60% decrease in leaf Fe concentration (Fig. 2b). Scopoletin and derivatives found in the culture medium of Fe-deficient plants almost disappeared when the medium was re-circulated through the Sep-Pak C18 column (Fig. S1), and they were recovered from the column by elution with methanol (Fig. S4). These data strongly suggest that the secretion of phenolics is part of the plant response to Fe shortage, and that they participate in protection against Fe-deficiency induced chlorosis.
Recent transcriptome (Yang et al., 2010; Rodríguez-Celma et al., 2013) and proteome (Lan et al., 2011) studies of Fe-deficient Arabidopsis roots revealed the upregulation of a set of genes encoding enzymes of the phenylpropanoid pathway. Phenylalanine ammonia-lyase, one of the most upstream enzymes of this pathway, and the coumarate:CoA ligases 4CL1 and 4CL2, mediating its last step, were induced by Fe deficiency, as well as caffeoyl-CoA O-methyltransferase and F6′H1, which mediate the conversion of caffeoyl-CoA to feruloyl-CoA and the subsequent production of 6′hydroxy-feruloyl-CoA, respectively. The expression of the ABCG37 transporter gene was also upregulated under such conditions, and Yang et al. (2010) suggested that it could export phenolics compounds such as caffeic or chlorogenic acids. ABCG37 is a member of the G-subgroup of ABC transporters (Verrier et al., 2008) able to transport auxinic compounds (Ito & Gray, 2006; Ruzicka et al., 2010). The promoter sequence of the ABCG37 gene contains several putative WRKY transcription factor binding sites (Ito & Gray, 2006), suggesting a role of ABCG37 in biotic stress defence and in transport of metabolites, including ferulic or coumaric acids (Walker et al., 2003).
Involvement of specific transporters for phenolics efflux in response to Fe deficiency was recently documented. In rice the PEZ1 transporter controls the concentration of protocatechuic acid in the xylem sap, and is essential for the utilization of precipitated Fe in the stele apoplast (Ishimaru et al., 2011). PEZ2 is closely homologous to PEZ1 and seems to perform a very similar function in rice plants (Bashir et al., 2011).
More recently, comparative transcriptional profiling by RNA sequencing of roots of Medicago and Arabidopsis plants grown under Fe deficiency conditions revealed a massive upregulation of genes encoding enzymes involved in riboflavin biosynthesis in Medicago and in phenylpropanoid synthesis in Arabidopsis (Rodríguez-Celma et al., 2013). Coexpression and promoter analysis provided evidence of a tight link between these biosynthetic pathways and a co-regulation with other genes encoding proteins involved in Fe uptake. This paper also reported that mutations in the Fe(II)- and 2-oxoglutarate-dependent dioxygenase family gene F6′H, and defects in the expression of ABCG37, abolished the presence of phenolic compounds in the culture medium and compromised Fe uptake from a low bioavailability Fe source. Our data presented in Figs 1, 2 and 4, although obtained under slightly different Fe nutrition conditions (no Fe added in our study vs Fe added as a low available source in Rodriguez-Celma and co-workers study), fully confirms these findings. However, the chemical nature of the phenolic compounds accumulated in Arabidopsis roots and/or secreted in response to Fe deficiency was not investigated in that study. We also confirmed that Fe deficiency led to increase the transcript abundances of F6′H1 and ABCG37 (Fig. 3; Yang et al., 2010). This is also consistent with the report that the expression of the closest homologue of AtABCG37 in tobacco (NtPDR3) is also induced in response to Fe deficiency (Ducos et al., 2005). In our case, the role of ABCG37 in phenolic secretion by Arabidopsis roots was suggested by the fact that the culture medium of plantlets, grown without Fe added, appeared much more fluorescent than the one of the pdr9-2 mutant allele (Fig. 4b), as previously reported by Rodríguez-Celma et al. (2013). This was confirmed by HPLC analysis of these compounds (Fig. 5). The pdr9 mutation led to a Fe-deficiency dependent increase in abundance of several fluorescent compounds in roots, some of them of higher polarity than those present in Col0. In the growth media extracts of the pdr9-2 and pdr9-3 mutant plants only a minor peak eluting at 15 min was observed, much smaller than the one at the same retention time in growth media extracts of Col0 wild-type plants (Fig. 5). Therefore, the pdr9 mutation markedly impaired the root excretion of the fluorescent compounds eluting at 15 min that occurs in Col0 in response to Fe deficiency. As shown in Fig. 2, these compounds participate in Fe nutrition. Because the ABCG37 transporter is required for their secretion, it was therefore consistent to observe that pdr9-2 plantlets were more sensitive to Fe deficiency than wild-type Col0 plants (Fig. 4a), in agreement with the recent findings of Rodríguez-Celma et al. (2013). Therefore, it can be concluded that: Fe deficiency induces the secretion of fluorescent phenolic compounds by Arabidopsis roots; the ABCG37 transporter is required for this secretion to take place; and these compounds improved plant Fe nutrition.
The chemical characterization of these compounds is a prerequisite to the understanding of their function in Fe nutrition, and no such characterization has been reported so far (Rodríguez-Celma et al., 2013). Their in-depth analysis by mass spectroscopy revealed that the compounds belong to the coumarin family, and more precisely are consistent with scopoletin and several derivatives (Table 2). Are these phenolics the elusive ‘chelate’ bringing Fe(III) to the root plasmalemma ferric chelate reductase, or do they have a role independent of Fe reduction and IRT1 uptake? Their chemical characterization reported in this study should help to adress these questions. Coumarins can interact in vitro with Fe either by chelating or reducing it. For instance some o-dihydroxy-coumarins chelate Fe3+ with the same affinity than deferoxamine at neutral pH, and may reduce Fe3+ at acidic pH (Mladenka et al., 2010). Also, coumarins could play an allelopathic role (Jin et al., 2006) similar to riboflavin derivatives (Susín et al., 1993; Rodríguez-Celma et al., 2011). They are toxic for bacteria and fungi (reviewed in Gnonlonfin et al., 2012), and their secretion could decrease competition for Fe in the rhizosphere. It was known that mutations in the genes encoding F6′H1 and caffeoyl-CoA O-methyltransferase reduced the concentrations of scopoletin and its ß-glucoside scopolin in the roots (Kai et al., 2008). Indeed, analysis of Fe-deficient Arabidopsis root extracts by ultra-performance liquid chromatography-MS/MS revealed that scopoletin, but not scopolin, accumulated under these conditions (Lan et al., 2011). However, in that study these two compounds were not found in the root exudates of Fe-sufficient or Fe-deficient plants. This can be explained by the fact that the plant medium was analysed after 3 d of Fe-deficiency, a time point at which fluorescent compounds are not fully accumulated (Fig. 1). We also observed that scopoletin accumulated in the roots of Col0 in response to Fe deficiency, but in contrast to Lan et al. (2011) we also found it in the growth media, with its presence being dependent on the functionality of the ABCG37 transporter. Not only scopoletin, but also four glucosylated/hydroxylated coumarins and a glucosylated coumarin precursor accumulated in Fe-deficient roots of pdr9 plants. Coumarins synthesis in the roots in response to Fe deficiency is likely coupled to their secretion to the external medium through ABCG37 activity. When this activity is impaired, they accumulate inside the roots. The glucosylated coumarin forms are adequate substrates to be detoxified within the vacuoles, suggesting the existence of an unidentified vacuolar transporter likely upregulated in Fe-deficient Arabidopsis roots of pdr9 plants. However, at this stage it cannot be deduced whether ABCG37 excretes glucosylated and unglucosylated phenolics or whether the glucosylated phenolics are deglucosylated once excreted. Challenging these hypotheses will be the aim of future experiments.
Work supported in the framework of the European Transnational Cooperation within the PLANT-KBBE Initiative funded by the Spanish Ministry of Economy and Competitivity (MINECO EUI2008-03618 to J.A.) and the Agence Nationale de la Recherche (ANR-08-KBBE-009-01 to J-F.B.), the Spanish MINECO Projects AGL2010-16515 and AGL2012-31988 (cofinanced with FEDER) and the Aragon Government (Group A03).