Dissecting iron deficiency-induced proton extrusion in Arabidopsis roots


Author for correspondence:
Wolfgang Schmidt
Tel:+886 2 2789 2997
Email: wosh@gate.academia.sinica.edu.tw


  • • Here, we have analysed the H+-ATPase-mediated extrusion of protons across the plasma membrane (PM) of rhizodermic cells, a process that is inducible by iron (Fe) deficiency and thought to serve in the mobilization of sparingly soluble Fe sources.
  • • The induction and function of Fe-responsive PM H+-ATPases in Arabidopsis roots was investigated by gene expression analysis and by using mutants defective in the expression or function of one of the isogenes. In addition, the expression of the most responsive isogenes was investigated in natural Arabidopsis accessions that have been selected for their in vivo proton extrusion activity.
  • • Our data suggest that the rhizosphere acidification in response to Fe deficiency is chiefly mediated by AHA2, while AHA1 functions as a housekeeping isoform. The aha7 knock-out mutant plants showed a reduced frequency of root hairs, suggesting an involvement of AHA7 in the differentiation of rhizodermic cells. Acidification capacity varied among Arabidopsis accessions and was associated with a high induction of AHA2 and IRT1, a high relative growth rate and a shoot–root ratio that was unaffected by the external Fe supply.
  • • An effective regulation of the Fe-responsive genes and a stable shoot–root ratio may represent important characteristics for the Fe uptake efficiency.


Iron (Fe) is an integral catalytic component in many vital enzymatic reactions and is required for essential processes such as DNA synthesis, respiration and photosynthesis. Although Fe is highly abundant in most soils, the poor solubility of ferric hydrous oxides in aerobic environments decreases the activity of Fe well below the plant's demand. Iron deficiency-induced chlorosis causes significant yield losses of crops, about one-third of the world's arable soils are potentially Fe deficient. Since plants are the major dietary source of Fe for a large part of the world's population, low Fe levels in plants also compromises human health. According to the World Health Organization (WHO), approximately two billion people suffer worldwide from Fe anemia, which is a leading cause of early termination in pregnancy, birth defects, retarded motor development and a general reduction in fitness. Understanding the molecular mechanisms of Fe uptake and homeostasis is therefore crucial for the correction of nutritional disorders in both plants and humans.

Plants have developed sophisticated mechanisms to acquire Fe even from sources that are not readily available for uptake by plants (Curie & Briat, 2003; Schmidt, 2003; Kobayashi et al., 2006). The classical distinction made by Römheld & Marschner (1986) separates the Graminaceae (strategy II plants) from all other species. Plants attributed to the so-called strategy II rely on the excretion of iron-avid phytosiderophores and subsequent uptake of the Fe-loaded complex. In strategy I plants, Fe is acquired in a three-step process, comprising mobilization of sparingly soluble plants by plasma membrane (PM) H+-ATPase-mediated proton extrusion, reduction of ferric chelates by a PM-bound oxidoreductase, and subsequent uptake of the ferrous ion by a member of the ZIP (zinc (Zn)–Fe-regulated transporter) family of metal transporters. Genes coding for the ferric reductase and the ferrous transporter were first identified in Arabidopsis (FRO2 and IRT1; Eide et al., 1996; Robinson et al., 1999) and then later in species such as pea, tomato and cucumber (Eckhardt et al., 2001; Waters et al., 2002; Cohen et al., 2004; Li et al., 2004; Waters et al., 2007). FRO2 belongs to an eight-member gene family and contains binding sites for FAD and NADPH, consistent with a transmembrane electron transport (Robinson et al., 1999; Feng et al., 2006; Mukherjee et al., 2006). IRT1 is the major route of Fe entry into the cell, demonstrated by the chlorotic phenotype of the irt1 knockout mutant (Henriques et al., 2002; Varotto et al., 2002; Vert et al., 2002). Both FRO2 and IRT1 are Fe regulated and subject to complex transcriptional and post-translational regulation (Connolly et al., 2002, 2003; Kerkeb et al., 2008). Iron uptake in Arabidopsis is regulated by FIT (FRU, bHLH29), a bHLH-type transcription factor that controls a subset of Fe-regulated genes including FRO2 and IRT1 (Colangelo & Guerinot, 2004; Jakoby et al., 2004; Yuan et al., 2005). The activity if FRO2 and IRT1 is regulated by both local and long-range signals, integrating the demand of distant plant parts (Schmidt, 2003; Vert et al., 2003). Rice, a strategy II species, possesses a IRT1 homologue (Bughio et al., 2002; Gross et al., 2003; Ishimaru et al., 2006), which indicates that ferrous iron can be taken up in addition to phytosiderophores when available (e.g. under submerged conditions).

In strategy I plants, Fe deficiency is generally associated with an increased net flux of protons mediated by PM H+-ATPases (Schmidt, 2006). Acidification of the rhizosphere is achieved by increased abundance of H+-ATPase protein in the PM and is, in some strategy I species, associated with epidermal cells being differentiated as transfer cells (Dell’Orto et al., 2002; Schmidt et al., 2003). Since the solubility of Fe decreases up to 1000-fold for each unit increase in pH, acidification of the rhizosphere can have an enormous impact on the Fe activity in the vicinity of the roots.

In contrast to other key reactions to Fe deficiency, the molecular mechanisms underlying Fe stress-induced proton extrusion are not well characterized. In addition to its role in Fe mobilization and in preserving a microenvironment for the function of the Fe(III) chelate reductase by preventing repulsion of negatively charged Fe chelates from the cell wall, net proton extrusion may play a role in the regulation of FRO2 and IRT1 expression (Zhao & Ling, 2007). Plants possess many isoforms of PM H+-ATPases with putatively redundant functions. It was recently shown that in cucumber roots transcript levels of the H+-ATPase gene CsHA1 accumulate in response to Fe deficiency, while a close homolog, CsHA2, remained unaffected by the Fe regime (Santi et al., 2005). In Arabidopsis, the PM H+-ATPases family consists of 12 members, some of which are Fe regulated (Colangelo & Guerinot, 2004; Li et al., 2007). The function of these isoforms in the Fe stress response and their potential impact in Fe uptake efficiency has not yet been defined. This is in part because of the fact that Arabidopsis roots show a relatively weak acidification response when compared with other strategy I species such as cucumber or tomato. We show here that two PM H+-ATPases, AHA2 and AHA7, participate in the Fe-deficiency response of Arabidopsis roots. Mutant analysis revealed that AHA2 mediates the Fe-deficiency induced acidification of the rhizosphere, while the function of AHA7 appears to be associated with the development of root hairs, the formation of which is increased in response to Fe deficiency. Light and low pH were found to modulate the Fe-deficiency response by increasing the mRNA levels of AHA2 and AHA7, as well as those of IRT1 and FRO2. Assuming differences in Fe availability in their natural habitat, we have screened natural occurring Arabidopsis accessions for their ability to acidify the rhizosphere after transferring to an Fe-deficient medium. We demonstrate that the capacity for net proton extrusion is associated with increased upregulation of both the Fe transporter IRT1 and two ATPase isoforms, AHA7 and AHA2. Accessions that show a higher proton extrusion capacity are characterized by a stable shoot–root ratio upon Fe deficiency, revealing a potential link between regulation of strategy I core genes, rhizosphere acidification, shoot–root ratio and Fe efficiency.

Materials and Methods

Plant material and growth conditions

Plants were grown in a growth chamber on an agar medium as described by Estelle & Somerville (1987). All stocks of Arabidopsis (Arabidopsis thaliana L. Heynh) seeds were obtained from the Arabidopsis Biological Resource Center (ABRC, Ohio State University, Columbus, OH, USA). Plants were ecotype Col-0 if not otherwise indicated. Seeds were surface-sterilized by immersing them in 5% (v : v) NaOCl for 5 min and 96% ethanol for 7 min, followed by four rinses in sterile water. Seeds were placed onto Petri dishes and kept for 1 d at 4°C in the dark, before the plates were transferred to a growth chamber and grown at 21°C under continuous illumination (50 µmol m−2 s−1, Phillips TL lamps). The medium was composed of: 5 mm KNO3, 2 mm MgSO4, 2 mm Ca(NO3)2, 2.5 mm KH2PO4, 70 µm H3BO3, 14 µm MnCl2, 1  µm ZnSO4, 0.5  µm CuSO4, 10 µm NaCl, 0.2 µm Na2MoO4 and 50 µm Fe-EDTA, solidified with 0.3% Phytagel (Sigma-Aldrich). Sucrose (43 mm) and 4.7 mm MES were included and the pH was adjusted to 5.5. After a variable time of precultivation, plants were transferred to fresh agar medium containing agar medium either with 50 µm Fe-EDTA (+Fe plants) or without Fe and with 100  µ m 3-(2-pyridyl)-5,6-diphenyl-1,2,4-triazine sulfonate (−Fe plants). For the dark treatment, plants were grown under continuous dark for 3 d. For gene expression analysis, roots were harvested 1–3 d after replanting to the different growth conditions, and stored in RNAlater solution (Ambion) at 4°C until RNA extraction. For Fe efficiency analysis plants were sown and grown in medium with either 0.5 µm Fe or 50 µm Fe.

For chlorophyll measurement, shoots of two plants were homogenized with a pestle and mortar in 80% acetone and total chlorophyll was determined as described by Gitelson et al. (2003).

RNA analysis and real-time reverse transcriptase polymerase chain reaction (RT-PCR)

Total RNA was isolated from roots of 12–15 plants using the RNeasy Plant Mini Kit (Qiagen). About 1 µg of total DNase-treated RNA (Turbo Dnase; Ambion) was reverse-transcribed using an oligo-dT primer and the Superscript III Reverse Transcriptase (Invitrogen) in a total volume of 20 µl. The cDNA synthesis reaction mixture was diluted and c. 25 ng of the initial RNA were used for PCR. Real-time RT-PCR was performed using the Platinum SYBR Green qPCR SuperMix-UDG (Invitrogen) in a 20 µl total volume. A DNA Engine Opticon 2 (MJ Research Inc., Waltham, MA, USA) system was used. Primers were designed by using the primer3 program to give an amplicon length of between 150 bp and 250 bp. Gene specificity of primers was ensured through sequence alignment by the blastn algorithm (Altschul et al., 1997) and experimentally by analysis of the melting curves of products. To enable detection of contaminating genomic DNA, PCR was performed with RNA as template. Alpha tubulin (At5g19770, TUA) was used as a constitutive reference gene and target gene mean expression was normalized to TUA expression by considering a 100% efficiency of the amplification reaction (E = 2). Relative expression (n-fold) of the normalized target gene in the treatment was calculated by comparison with a control (+Fe) experiment. At least two biologically distinct experiments and three technical replicates were performed for each experiment. For primer sequences see Supporting Information, Table S1.

Determination of Fe3+-EDTA reduction activity

The Fe(III) chelate reductase activity was determined as previously described (Schmidt, 1994). For each measurement five seedling roots were pooled. Three independent measurements in the presence or absence of Fe(III)-EDTA were performed, values were calculated as µmol Fe2+ (g−1 FW h−1).

Acidification capacity

Net proton flux was measured by incubating 7-d-old plants that have been grown for 2 d in Fe-free media in standard nutrient solution without phytagel, sucrose and MES buffer, containing the pH indicator bromocresol purple (0.005%) either with or without 50 µm Fe-EDTA for 2 d in 96-well plates. Proton extrusion capacity was accessed by reading the absorption at 590 nm (A590) with an automated microplate reader.

Western analysis

Total protein was prepared from roots of plants grown either in Fe-sufficient or Fe-deficient medium. Extracts were obtained by grinding tissue on ice in extraction buffer as described by Connolly et al. (2002). Extracts were mixed in a 1 : 1 ratio with 8% lithium dodecyl sulfate (LDS) sample buffer (Invitrogen) containing 300 mm dithiothreitol (DTT), heated at 70°C for 10 min, then centrifuged for 10 min at 15 000 g. Total protein (c. 20 µg) was separated by denaturing 7% sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) (NuPAGE Novex Tris-acetate gels; Invitrogen), then stained with Coomassie Brilliant Blue for checking the amount of total protein, or electrotransferred to PVDF membrane filters (Immobilon-P; Millipore Co., Bedford, MA, USA) overnight at 4°C for Western analysis. Filters were blocked for 1 h in TBS-buffer (20 mm Tris-HCl, pH 7.5, 150 mm NaCl) containing 5% (w : v) skimmed milk powder and incubated for 3 h with the antiserum. An antiserum raised against the synthetic 18-mer peptide AKRRAEIARLRELHTLKG, as described in Santi et al. (2005), was used to detect PM H+-ATPases. This peptide shares a 100% identity with the corresponding C-terminal amino acid sequence of AHA1, AHA2, AHA3, AHA4 and AHA11. The peptide is 88% identical (100% similar) to AHA7. After a 1 h-incubation with the secondary antibody (horseradish peroxidase-conjugated antirabbit IgG; Sigma-Aldrich), the band corresponding to PM H+-ATPase was detected by chemiluminescence (SuperSignal West Substrate; Pierce, Rockford, IL, USA) on autoradiography film (Kodak X-Omat AR; Eastman Kodak).

Confocal microscopy

Seedlings were placed in 10 mg ml−1 propidium iodide solution (PI) for 1 min. The seedling was gently rinsed with water for 2 min. The root was removed and mounted in fresh water. The roots where then observed using a Confocal Laser Scanning Microscope (Zeiss LSM510 Meta). The peak excitation wavelength (λ) and emission λ for PI are 536 nm and 620 nm, respectively.


AHA2 and AHA7 are the major Fe-responsive H+-ATPase isoforms in roots

To identify the Fe-responsive PM H+-ATPase isogenes, real-time RT-PCR analysis of H+-ATPase transcripts in roots of Fe-sufficient and Fe-deficient plants was performed with gene-specific primers. Five out of the 12 P-type ATPase isogenes (AHA1 to AHA12) were found to be affected by Fe deficiency, with AHA2 and AHA7 representing the most responsive genes (Fig. 1a). After imposing Fe deficiency for 3 d, AHA7 and AHA2 gene expression was increased by 4.0-fold and 3.3-fold, respectively. The AHA3 and AHA4 transcripts were found to be slightly increased (c. twofold), whereas the AHA11 gene was slightly downregulated (c. twofold). Genes encoding proteins that mediate the two other major components of the strategy I response, FRO2 and IRT1, displayed a much higher induction level at the end of the experimental period. FRO2 transcripts increased c. 80 times when compared with control plants while IRT1 message was found to be c. 50-fold higher in roots of Fe-deficient plants (Fig. 1a). The transcriptional regulator gene FIT, which controls a subset of the Fe-responsive genes in Arabidopsis (Colangelo & Guerinot, 2004) was induced by 6.6-fold. AHA1, one of the two major H+-ATPase isoforms in the PM of Arabidopsis was not affected by Fe deficiency. Two other AHA isogenes, AHA8 and AHA10, whose transcripts had been previously detected in roots (Ueno et al., 2005) were not transcriptionally regulated under Fe-deficient conditions (data not shown).

Figure 1.

Changes in the abundance of AHA transcripts and protein in response to iron (Fe) starvation. (a) Relative expression of various AHA genes and genes involved in the regulation and uptake of Fe in roots of Arabidopsis plants that have been grown in media deprived of Fe for 3 d. Mean values from two independent experiments with three technical repeats each plus standard errors of technical repeats are shown. Values are first normalized to TUA (see text) and then to control (+Fe) plants. (b) Immunoblot analysis. The AHA protein was detected by hybridizing the blot with antibodies raised against the synthetic 18-mer peptide AKRRAEIARLRELHTLKG. Protein samples were prepared from roots deprived of Fe for 3 d. About 20 µg of total proteins were loaded for the immunoblot and in parallel for Coomassie Brilliant Blue staining (not shown). One out of two biologically independent experiments is shown. (c) Time-course of the relative expression of AHA genes and FIT, IRT1, and FRO2 in roots after transfer to Fe-free medium. Expression values of IRT1 and FRO2 are shown on the right vertical axis. Data are calculated as described earlier. A typical experiment is shown.

Antibodies raised against the synthetic 18-mer peptide AKRRAEIARLRELHTLKG were used to detect PM H+-ATPase protein in roots of Fe-sufficient and Fe-deficient plants. This antibody most probably binds to all the H+-ATPase isoforms under investigation. Thus, changes in protein abundance upon Fe deficiency represent the total of all Fe-responsive isoforms. The H+-ATPase protein level was found to be markedly increased upon Fe deficiency, indicating that a substantial percentage of the transcripts was translated into protein (Fig. 1b).

The effect of Fe deficiency on the relative expression of strategy I core genes was examined in time-course experiments (Fig. 1c). IRT1 and FRO2 transcripts significantly increased after 24 h. After the same period, the expression of FIT appeared to be twofold higher than in +Fe roots. By contrast, a significant increase in the message levels of AHA genes was not apparent before 72 h of Fe depletion. AHA4 and AHA11 were found to be expressed at a very low level and were not further investigated.

Based on the transcript abundance, the most represented isoforms in roots were AHA1, AHA2 and AHA7 (Fig. 2b, experiment at pH 5.5). Under Fe-deficient conditions, however, the message level of AHA2 clearly exceeded that of the housekeeping isoform AHA1, pointing to a function of AHA2 in the Fe-deficiency response. The transcript abundances of AHA3 (Fig. 2b), AHA4 and AHA11 (not shown) were one order of magnitude lower than that of AHA7 and almost two orders lower than that of AHA2, thus scarcely contributing to the overall phenotype. Interestingly, IRT1 transcripts were found to be ninefold higher than those of FRO2, although both genes exhibited a comparable level of induction (Fig. 2b).

Figure 2.

Effect of light (a) and pH (b) on mRNA levels of AHA genes and FIT, IRT1 and FRO2 in roots of iron (Fe)-sufficient and Fe-deficient Arabidopsis plants. Plants were analysed 3 d after transfer to fresh media either with or without Fe. Light experiments were conducted at pH 5.5, pH experiments at a photon flux of 50 µmol m−2 s−1. For dark treatment, plants were grown under continuous dark for 3 d with or without Fe. Mean expression values were normalized to TUA mRNA level (= 100 arbitrary units; see text). Standard errors of technical repeats are shown.

Induction of the Fe-responsive ATPase isoforms is affected by light and pH

Since the expression of FRO2 and IRT1 was shown to be diurnally regulated (Vert et al., 2003), we investigated whether this is also true for the transcription of the Fe-responsive H+-ATPase genes. Similar to what has been reported previously, a higher message level of IRT1 and FRO2 was noted when both Fe-sufficient and Fe-deficient plants were grown in light (100 µmol m−2 s−1) relative to dark-grown plants (Fig. 2a). The expression of AHA7 followed a similar pattern, indicating a common regulation of the three genes. Interestingly, the expression of FIT clearly deviated from this pattern. Iron deficiency induced FIT expression under both light regimes but the amount of expression decreased when plants were grown in light compared with dark-grown plants, in particular in the presence of Fe (fivefold lower). A slight reduction of message levels in the light relative to plants grown in darkness was also observed for Fe-sufficient plants with respect to the expression of AHA2. In both the FIT and AHA2 genes, induction by Fe deficiency was significantly higher in light (Fig. 2a). FIT and AHA2 were found to be upregulated 5.8- and 3.4-fold in light-grown plants, while the increase was markedly reduced (1.9- and 1.5-fold) when plants were grown in the dark. AHA3, although expressed at a very low level compared with AHA2, was regulated similarly to AHA2. The expression of AHA1 was neither affected by light nor by Fe depletion.

Unlike the phytosiderophore-based Fe uptake system of grasses, the strategy I-type of Fe acquisition is severely affected by the external pH, being restricted with decreasing proton activity (Schmidt, 1999). This fact underlines the significance of net proton export to generate an environment in which the Fe acquisition mechanism works most efficiently. Analysis of the expression pattern of genes involved in the strategy I response revealed a strong repression at pH 7.5 in Fe-deficient plants when compared with lower (5.5) pH (Fig. 2b). FRO2 and IRT1 were sevenfold and threefold more induced at pH 5.5, respectively. Interestingly, FIT and the AHA genes AHA2, AHA3 and AHA7 showed a similar pattern: they were negatively affected by an increase in pH. Under Fe-sufficient conditions, an increase in message level at pH 7.5 relative to pH 5.5 was observed for some of the genes under investigation (IRT1 and FRO2). The expression of AHA1 was neither affected by pH nor light, underlining its role as a housekeeping H+-ATPase isoform (Fig. 2b).

Rhizosphere acidification is mediated by AHA2 activity

In order to assess the acidification capacity in a quantitative manner, Fe-deficient plants were cultivated in 96-well plates containing Fe-free, half-strength nutrient solution supplemented with the pH indicator bromocresol purple. Net proton flux was measured by reading the optical density at 590 nm, which is the wavelength of absorption of the unprotonated form of the dye. To evaluate the importance of the various H+-ATPase isoforms for rhizosphere acidification, the capacity for net proton extrusion in homozygous mutants harboring defects in AHA1, AHA2 and AHA7 was compared with the Col-0 wild-type (Fig. 3a). In addition, mutants defective in the protein kinase PKS5 were analysed. Phosphorylation of AHA2 at Ser-931 by PKS5 inhibits the H+-ATPase activity by preventing the interaction with an activating 14-3-3 protein; in homozygous pks5 mutants PM H+-ATPase activity was shown to be derepressed (Fuglsang et al., 2007). An overview on the mutants used in this study is provided in the Supplementary Information, Fig. S1. In wild-type plants, net proton flux was markedly increased when plants were subjected to Fe deficiency (Fig. 3a). Neither under control (+Fe) nor Fe-deficient conditions, did homozygous aha1 and aha7 mutants show a decreased potential for rhizosphere acidification. With regard to AHA1, both mutants showed a substantial reduction in transcript level (3.1-fold and 7.1-fold for SALK_065288 and SALK_118350, respectively, compared with the wild-type under –Fe conditions). Among the available aha7 mutants only SALK_042485 showed a marked reduction in transcript abundance (14-fold reduction with regard to the wild type). aha2 mutants revealed a clearly lower proton flux under both conditions. Among the three aha2 mutant lines investigated, proton extrusion activity was affected to a similar extent when grown on control media, but was most reduced in the knockout line SALK_082786, followed by SALK_0220010 (17.5-fold reduced message level) and SALK_062371 (12.5-fold). In all cases, however, Fe deficiency evoked an increase in net proton efflux, suggesting a partly functional redundancy of the AHA genes and/or other mechanisms, such as an altered cation/anion uptake ratio (e.g. caused by an altered nitrogen (N) metabolism that came into play). A putatively activated AHA2 enzyme in pks5 mutants led to a higher proton pumping activity under +Fe conditions, whereas no difference with respect to the wild-type was observed for Fe-deficient pks5 plants. Together, the results clearly indicate that AHA2 accounts chiefly if not entirely for Fe-deficiency induced proton flux.

Figure 3.

In vivo acidification (a) and Fe(III) chelate reductase activity (b) of wild-type Arabidopsis plants and mutant harboring defects in the expression AHA genes or regulation of AHA activity. For the determination of the acidification capacity, 7-d-old plants were used for the experiments. Plants were incubated in the test solution for 48 h. +Fe, closed bars; −Fe, tinted bars. Error bars show standard deviation of 16 plants analysed in two independent experiments. The optical density on the unprotonated form of bromocresol is reported on the y-axis. Values marked with a different letter are significantly different from each other at P < 0.05. For Fe chelate reductase measurements, Fe-deficient plants were grown on Fe-free media for 3 d before the experiment. Error bars represent standard deviation of five plants.

No specific H+-ATPase isoform is required for ferric reductase activity

To evaluate a putative involvement of one of the H+-ATPase isoforms in the regulation of Fe acquisition, we measured Fe(III) chelate reductase activity in the Col-0 wild-type and in aha1, aha2, pks5 and aha7 mutants (Fig. 3b). With the exception of pks5 plants and the aha2 knockout line SALK_082786, which revealed a slightly upregulated reductase activity under control conditions, no significant deviations from the wild type were observed. This indicates that, although a high proton pumping activity accelerates the transmembrane flux of electrons and the reduction of ferric iron, the reaction does not depend on the activity of a particular isoform. This is supported by the time-course of expression, showing an early induction of the FRO2 gene (24 h), whereas the AHA genes were upregulated at least 48 h after transferring the plants to Fe-free media (Fig. 1c). Iron-deficient pks5 mutant plants did not show a phenotype (Fig. 3b), probably owing to the fact that the flux of electrons is restricted by other processes, such as the availability of intracellular NADPH, or by regulation of the activity according to the plant's demand.

AHA7 activity is crucial for the formation of Fe-deficiency induced root hairs

Growing plants in media deprived of Fe caused a significant increase in the number of root hairs (Figs 4, 5). Differences in the phenotype became obvious 3–4 d after transfer of the plants to Fe-free media; the phenotype was fully established after c. 6 d. In aha7 mutants, the frequency of root hairs was found to be decreased both under control and Fe-deficient conditions, the effect being more pronounced under the latter condition (Figs 4, 5b). In aha2 mutant plants, only a slight reduction in root hair density compared with the wild type was observed (Fig. 5). In contrast to aha7 plants, a clear increase in root hair frequency upon Fe deficiency was observed in roots of aha2 plants, suggesting that AHA7 function is critical for the induction of the root hair phenotype typical of Fe-deficient plants. One of the aha2 mutants under investigation, SALK_022010, revealed a marked (7.3-fold) increase in AHA7 transcript abundance. Although the reasons for this increase cannot be deduced from the site or type of insertion, the root hair density of the mutant was found to be markedly increased, in particular under control conditions (Fig. 5d), supporting a role of AHA7 in root hair formation.

Figure 4.

Root hair phenotype of iron (Fe)-sufficient and Fe-deficient wild-type and aha7 mutant Arabidopsis plants. Pictures were taken 6 d after transfer to Fe-free media. Images are of 16-d-old plants.

Figure 5.

Root hair density of wild-type (a), aha7 (b) and aha2 (c,d) mutant Arabidopsis plants. SALK_022010 revealed an eightfold increase in AHA7 transcript abundance both under control and iron (Fe)-deficient conditions. Root hairs were counted 6 d after transfer to Fe-free media. +Fe, closed bars; −Fe, tinted bars. Error bars show standard errors of 10 roots analysed in two independent experiments.

Natural accessions differ in their capacity for rhizosphere acidification

The comparably low acidification capacity of the Col-0 accession led us to investigate whether this trait varies among Arabidopsis lines collected from different habitats. Variation of the parameters investigated among natural Arabidopsis accessions was assessed by screening a collection of 96 lines obtained from the ABRC stock center (CS22660). A first round of screening was conducted by transferring 10-d-old plants to an Fe-free medium containing the pH indicator bromocresol purple (0.005%) at pH 7.0 for 24 h. Twenty-one accessions were found to have a significant acidification response at the end of the experimental period. A second screening among these 21 lines was carried out with a similar protocol but using a medium containing 0.5 mm MES buffer adjusted to pH 6.3. Under these conditions, Uod-1, Br-0, Ts-1, Lz-0, Ws-2, and Fab-4 plants were confirmed as ‘strong’ acidifiers in this screen. Six lines without a clearly pronounced acidification around the roots were randomly chosen from the remaining lines and referred to as ‘weak’ acidifiers.

The transcript abundance of AHA7 and AHA2 was analysed in Fe-sufficient and Fe-deficient roots of plants from the 12 selected accessions (Table 1). The expression of the Fe transporter gene IRT1 was taken as an indicator for the induction of the Fe-deficiency responses. Although the selection of the lines was associated with a relatively high variation in the investigated parameter, verification of the acidification capacity by quantitative measurements revealed a general trend for a higher proton extrusion activity in the strong acidifiers when the average values of both groups were considered (Table 1). Two of the ‘weak’ lines, Lov-5 and Uod-7, did show almost no net proton extrusion activity. The highest activity among the strong acidifiers was observed for Uod-1. Strong acidification was generally associated with a higher induction level of all the investigated genes under Fe-deficient conditions. AHA7 and AHA2 were on average threefold and twofold induced in the strong acidifier group, whereas the expression was 1.6-fold and 1.4-fold, respectively, in the weak acidifier group. IRT1 expression was increased by Fe depletion in both groups, but the induction was much more pronounced in the strong acidifier group (60-fold, on average, compared with 25-fold in the weak acidifiers). No such differences were evident when the transcript level was considered (Table 1). Since the net proton extrusion activity is strongly affected by the nitrate uptake pattern, the expression of the high-capacity low-affinity nitrate transporter NRT1.1 was also analysed. Low acidification was found to be associated with a slightly higher transcript abundance of NRT1.1 under both Fe regimes, while Fe deficiency did not affect its induction (Table 1). Similarly, the expression of the high-affinity nitrate transporter NRT2.1 remained unaffected by the iron regime under our experimental conditions (data not shown).

Table 1.  Induction level of iron (Fe)-responsive genes, acidification capacity and transcript abundance of genes that potentially affect net proton flux across the plasma membrane of roots from Fe-deficient Arabidopsis accessions
AccessionPrimary screening phenotypeAHA7AHA2NRT1.1IRT1In vivo acidification (0.7 – ΔA590)
Fab-4Strong301.84652.2571.3 58657.30.114
Ts-1Strong382.63851.7140.9 62659.20.164
Br-0Strong311.73651.1480.9 56340.60.217
Lz-0Strong312.03311.5261.5 47021.60.193
Lov-5Weak351.63661.3571.0 34333.50.025
Uod-7Weak231.73922.2392.0 82930.40.015
Mt-0Weak331.64911.4552.0 85428.40.120
Sq-8Weak422.53661.6241.0 62829.10.101
Sq-1Weak361.43221.1440.5 71621.50.166
Bil-5Weak241.03780.8700.8 3915.80.138
AverageStrong38 ± 43.2 ± 0.9427 ± 312.0 ± 0.334 ± 71.3 ± 0.2 825 ± 16959.6 ± 11.00.215 ± 0.039
AverageWeak32 ± 31.6 ± 0.2386 ± 231.4 ± 0.248 ± 61.2 ± 0.3 627 ± 8924.7 ± 4.10.094 ± 0.025

Acidification capacity may be linked to the performance at low Fe availability of the accessions

To investigate whether the acidification capacity of the different accessions is related to their Fe efficiency, plants were grown on media supplemented either with 50 µm (control plants) or 0.5 µm Fe-EDTA (low Fe plants). The two groups (strong and weak acidifier) did not deviate from each other with respect to the average reduction of shoot biomass production and decrease in chlorophyll content caused by Fe depletion over an experimental period of 14 d, indicating that the acidification ability is not directly linked to a tolerance to low Fe availability (Table 2). However, it was evident that the weak acidifiers have a generally lower growth rate under both +Fe and –Fe conditions and a markedly higher chlorophyll concentration. Thus, a high acidification capacity appears to be associated with a higher demand for Fe caused by a higher growth rate. Interestingly, the strong acidification trait was associated with a lack of reduction of the shoot–root ratio when grown on low-Fe media, whereas a considerable reduction (–24%) was apparent in the weak acidifiers group (Table 2).

Table 2.  Chlorophyll concentration and fresh weight in natural accessions of Arabidopsis under different iron (Fe) regimes
AccessionIn vivo acidificationShoot fresh weight (mg)Chlorophyll (µg g−1 FW)Shoot–root ratio
  1. Plants were grown for 14 d in medium with 50 µm Fe-EDTA (+Fe plants) or 0.5 µm Fe-EDTA (−Fe plants).

Uod-1Strong42.625.8 (−39%)2.10.8 (−62%)5.04.8
Ws-2Strong26.517.6 (−33%)2.61.3 (−51%)4.84.6
Fab-4Strong39.331.8 (−19%)2.31.0 (−58%)5.05.0
Ts-1Strong20.814.3 (−31%)2.91.3 (−43%)5.04.9
Br-0Strong36.214.7 (−59%)1.71.0 (−43%)5.05.0
Lz-0Strong41.029.9 (−27%)2.70.8 (−69%)5.05.0
Lov-5Weak32.612.2 (−63%)2.81.9 (−31%)5.05.0
Uod-7Weak40.018.6 (−53%)2.61.3 (−51%)6.04.1
Mt-0Weak16.411.1 (−32%)6.82.5 (−64%)4.32.9
Sq-8Weak14.19.9 (−30%)6.11.4 (−77%)5.33.0
Sq-1Weak19.914.8 (−26%)3.41.5 (−55%)5.04.8
Bil-5Weak21.115.0 (−29%)4.02.0 (−50%)6.64.6
AverageStrong41.522.4 (−46%)2.41.0 (−57%)5.04.9 (−2%)
AverageWeak30.917.3 (−43%)4.31.8 (−59%)5.44.1 (−24%)


The ATP-dependent proton pumps are canonical PM proteins that energize essential transport processes in fungal and plant cells by establishing the membrane potential and pH gradient across the PM (Palmgren, 2001; Duby & Boutry, 2009). In Arabidopsis, the most abundant PM H+-ATPase isoforms are AHA1 and AHA2, as shown by phosphoproteomics studies, expression profiling, and RT-PCR analysis (Birnbaum et al., 2003; Niittyläet al., 2007; this study). All of the PM H+-ATPase isoforms investigated in the present study were found to be enriched in PM fractions (Nelson et al., 2006), the subcellular localization of AHA10 is unknown (Baxter et al., 2005). In roots, AHA2 appears to be the prominent P-type ATPase isoform (Harper et al., 1990; this study) and has been shown to be expressed in epidermal cells including root hairs, in the cortex, and in phloem and xylem parenchyma cells (Fuglsang et al., 2007).

ATPase-driven proton extrusion is a core component of the strategy I-type response to Fe deficiency and may represent an important trait for the Fe uptake efficiency of a species or genotype. The function of rhizosphere acidification in the acquisition of Fe is associated with weakening of the Fe–O bond of iron oxides by protonation and subsequent metal detachment (Schwertmann, 1991), which increases the pool of bioavailable Fe. The functioning of the FRO2 is dependent on an acidic pH and a flux of protons across the PM (Schmidt, 1994; Zhao & Ling, 2007). Repulsion of negatively charged Fe chelates in the apoplasm can be compensated for by a net efflux of protons, thereby increasing the velocity of the reduction (Toulon et al., 1992). Finally, net proton flux energizes the membrane for the uptake of the reduced Fe and large fluctuations in the expression/activity of transporters are supposed to be associated with large fluctuations in proton pumping. This renders the strategy I type response more sensitive to the external pH than the siderophore-based strategy II, in which only the last step, the uptake of the loaded siderophore, depends on an electrochemical gradient across the PM. In line with the assumption that low proton activity decreases the efficiency of the reduction-based Fe acquisition, under Fe-deficient conditions the RNA levels of all strategy I-related genes were markedly decreased at pH 7.5. Notwithstanding in Fe-sufficient plants, increasing the pH from 5.5 to 7.5 leads to an increase in the expression of IRT1, FRO2, AHA7 and FIT (Fig. 2b), probably as a compensation for a decreased Fe uptake owing to decreased availability of Fe caused by decreased stability of the ferric chelate. This increase in message levels is in accordance with results reported for tomato roots (Zhao & Ling, 2007). Interestingly, such an increase in transcript levels was not observed for the AHA2 gene, which may be attributed to the complex post-translational regulation of PM H+-ATPases (see later) and/or to a lower responsiveness at the transcriptional level. Under Fe-deficient conditions, the transcript levels of all Strategy I-related genes were markedly decreased at pH 7.5, underlining the importance of an acidic environment for reduction-based iron acquisition.

Induction of the Strategy I core genes IRT1 and FRO2 by light has been described earlier and was attributed to photosynthetic products or other metabolites that systemically regulate the high-affinity Fe uptake system according to the shoot's demand for iron (Vert et al., 2003). In our experiments, light modulated the expression of all genes under examination except for AHA1, but did not control their Fe starvation-specific expression. This is indicative for a dual regulation by local and systemic signals for FIT, AHA2 and AHA7 also. Interestingly, the expression of FIT showed a decrease under high light intensity both under Fe-sufficient and Fe-deficient conditions (Fig. 2a). This suggests that the expression of IRT, FRO2 and AHA7 is not exclusively controlled by FIT. FIT protein is required for the accumulation of FRO2 mRNA and IRT1 protein (Colangelo & Guerinot, 2004). However, no direct binding of FIT to the promoters of FRO2 and IRT1 has been reported. The involvement of two other bHLH-type transcription factors, AtbHLH38 and AtbHLH39, in the regulation of the expression of genes involved in Fe uptake has recently been demonstrated (Yuan et al., 2008). Both AtbHLH38 and AtbHLH39 can form heterodimers with FIT, which have been shown to regulate IRT1 and FRO2 expression. It is noteworthy that similar to what has been observed when the pH was varied, the response of the AHA2 gene (and AHA3) differed from that of AHA7, IRT1 and FRO2 under Fe-sufficient conditions. Under all growth regimes applied in this study, AHA7 was found to be strictly coregulated with IRT1 and FRO2.

The PM H+-ATPases are regulated at different levels. In addition to transcriptional control, regulation of pump activity occurs at the post-translational level. Phosphorylation of the penultimate residue at the C-terminus of PM H+-ATPases allows for binding of 14-3-3 proteins, displacing autoinhibition at the C-terminus and increasing pump activity (Ferl, 2004). The proton pump interactor 1 gene (Ppi1) encodes a protein that binds to the C-terminus of PM H+-ATPases to a site different from the 14-3-3 binding site, thereby modulating pump activity (Morandini et al., 2005). Thus, changes in transcription upon Fe deficiency do not necessarily mirror changes in pump activity in a quantitative manner. However, based on transcriptional changes, an Fe-responsive PM H+-ATPase gene in the roots of cucumber (CsHA1) has been distinguished from a closely related housekeeping homolog (CsHA2; Santi et al., 2005). Accumulation of CsHA1 transcripts occurred in rhizodermal cells and was associated with Fe-deficiency induced root hair formation, whereas CsHA2 was found not to be expressed in root hairs (Santi & Schmidt, 2008). This suggests that in cucumber roots proton extrusion upon Fe starvation is predominantly mediated by CsHA1 and that transcriptional regulation is at least in part responsible for the different activity of H+-ATPase isoforms under different environmental conditions.

Based on the relative induction by Fe deficiency and the abundance of their transcripts, we propose that two PM H+-ATPase genes are involved in the Fe-deficiency response of Arabidopsis: AHA2 and AHA7 (Figs 1, 2). AHA2 activity is negatively regulated by the Salt Overly Sensitive (SOS)-like protein kinase PKS5 (Fuglsang et al., 2007). Under Fe-deficient conditions, pks5 mutants show only a slightly higher acidification response when compared with the wild type, but show a markedly more acidic pattern when grown in the presence of Fe (Fig. 3a). The lack of stimulation of proton extrusion in pks5 mutants under Fe-deficient conditions may reflect a negative feedback mechanism to avoid excessive consumption of ATP. The difference in proton extrusion capacity between pks5 plants and the wild type was reported to be most pronounced in the root apex, a zone where Fe-deficiency induced acidification is typically observed (Fuglsang et al., 2007; Santi & Schmidt, 2008). Mutants defective in AHA2 expression showed lower net proton flux under both Fe-sufficient and Fe-deficient conditions. Together, these data suggests that Fe-deficiency induced rhizosphere acidification can mainly be attributed to AHA2 activity.

In experiments conducted to assay ferric reductase (FRO2) activity in aha1, aha2, aha7 and pks5 mutants, we observed that only Fe-sufficient pks5 mutant plants and one of the aha2 mutants tested (SALK_082786) showed a deviation from the wild-type, revealing an increased activity in roots of the mutants versus the wild type (Fig. 3b). While FRO2 activity can be almost completely abolished by H+-ATPase inhibitors (Alcántara et al., 1991; Schmidt, 1994), it appears that other H+-ATPase isogenes can compensate for a defect in one of the homologues, while the derepression of AHA2 in pks5 mutants leads to an increase in FRO2 activity. SALK_082786 appeared to show increased AHA1 and AHA7 message levels (data not shown), in particular under Fe-sufficient conditions, suggesting that one or both of these isoforms cause the observed increase in reduction activity.

The physiological processes involved in Fe acquisition are associated with developmental changes (such as the formation of transfer cells and extra root hairs) leading to increased absorptive surface of the roots. While it is tempting to speculate that proton extrusion is the first step in Fe acquisition, several lines of evidence instead argue in favor of an ‘emergency’ reaction if the normal acquisition machinery becomes insufficient to meet the demand of the plant. AHA2 gene expression and proton extrusion capacity is relatively weak in Arabidopsis and is induced relatively late after the onset of Fe-deficient conditions, while the other two responses are robust and fast. Moreover, Fe(III) chelates are a much better substrate for FRO2 than Fe(III) salts or oxides (Schmidt, 1999); thus, proton extrusion will provide mainly substrates that cannot easily be acquired and will not directly contribute to the FRO2-IRT1 route of Fe uptake. In addition, proton extrusion appears to be linked to morphological alterations such as extra root hairs or transfer cells. Transfer cell formation occurs later than the induction of Fe reduction activity (Schikora & Schmidt, 2001). Interestingly, Fe deficiency causes an accumulation of H+-ATPase protein in rhizodermal transfer cells of Fe-deficient tomato roots, but neither IRT1 protein nor Fe reductase activity was found to be specifically increased in this cell type (Schmidt & Bartels, 1996; Schmidt et al., 2003; Schikora et al., 2006), suggesting close coupling between transfer cell induction and proton extrusion. Finally, AHA2 expression appears not to be controlled by FIT, suggesting an induction pathway of AHA2 that is parallel to that of IRT1 and FRO2. We therefore propose that proton extrusion is induced at later stages of Fe deficiency and is upregulated together with alterations in developmental programs for the purpose of exploring additional Fe pools in soils.

AHA7 was shown to be regulated by the transcription factor FIT, suggesting a function of AHA7 in the Fe stress response of Arabidopsis (Colangelo & Guerinot, 2004). In addition, AHA7 was reported to be part of the root hair transcriptome, being 3.6-fold more abundant in the wild type compared with the root hairless mutant rhd2 (Jones et al., 2006). It is thus tempting to suggest that AHA7 is involved in the Fe-deficiency induced formation of root hairs. Homozygous aha7 insertion mutants (salk_042485) showed a lower frequency of root hairs when compared with the wild-type, supporting this assumption. In contrast to wild-type plants, aha7 mutants did not show a clear increase in root hair density upon Fe deficiency, suggesting that the root hair phenotype of Fe-deficient plants is dependent on AHA7 function. Root hair elongation has been attributed to pH oscillations (Monshausen et al., 2007; Wu et al., 2007), suggesting that tip growth of root hairs is dependent on PM H+-ATPases. In Fe-deficient cucumber roots, Fe-deficiency induced root hair formation is closely linked to proton extrusion mediated by H+-ATPase (Santi & Schmidt, 2008). In support of a role of ATPase-mediated proton extrusion in root hair formation, in the aha2 mutant SALK_022010 a higher level of AHA7 message level was associated with an increased number of root hairs, independent on the growth conditions. This does not, of course, rule out other functions of AHA7 in the Fe-deficiency response. The onset of the high-affinity Fe uptake system is robust and requires a steep pH gradient across the PM to support the uptake of Fe. It is tempting to speculate that AHA7, closely coexpressed with the Fe uptake machinery, contributes to the energization of the PM in particular under Fe-deficient conditions.

Natural Arabidopsis accessions differ in their capacity to acidify the rhizosphere. This difference is reflected by a higher induction of AHA2, AHA7 and IRT1 gene expression in accessions that have been selected for high proton extrusion capacity (Table 1). No such correlation was found when the transcript abundance of these genes was considered. Thus, regulation of the gene activity appears to be crucial for the phenotype. This differed from the abundance of transcripts of the nitrate transporter NRT1.1 in Fe-sufficient and Fe-deficient plants. The cation/anion uptake pattern, in particular the uptake of nitrate, can contribute significantly to the proton activity in the vicinity of the roots. Higher message levels of NRT1.1 in the group of weak acidifiers may reflect a higher uptake of nitrate and may affect the in vivo acidification pattern.

The efficiency of Fe acquisition could be linked either to the ability to mobilize additional sources of Fe or to a higher rate of induction of FRO2 and IRT1. The maximal velocity of FRO2 activity was found to be positively related to the relative growth rate of the species and was not related to the maintenance of the growth at low Fe concentrations (Schmidt & Fühner, 1998), making the capacity of rhizosphere acidification a likely candidate for a determinant of Fe uptake efficiency. While no clear differences among the weak and strong acidifiers were observable regarding the reduction of chlorophyll concentration or decrease in biomass production, strong acidifiers were generally characterized by a stable shoot–root ratio. A constant shoot–root ratio may confer a competitive advantage and may represent a trait reflecting Fe efficiency. A constant shoot–root ratio is of ecological advantage and has been considered as an important parameter in phosphate and zinc efficiency (Streeter et al., 2001; Hacisalihoglu & Kochian, 2003). Similar to Fe, the bioavailability of soil zinc pools is increased by root-mediated processes that lower the pH in the rhizosphere (Hacisalihoglu & Kochian, 2003). It should be noted that the experimental setup in the present study was not designed to monitor differences in tolerance to sparingly soluble Fe pools as may occur in soils in natural habitats or agricultural systems. Thus, a potential difference in the ability to use these pools among the accessions might not be observable by the experimental approach.

In summary, we have demonstrated that Fe-deficiency induced rhizosphere acidification is mediated by AHA2, while the other major isoform in roots, AHA1, represents a housekeeping isogene that is not affected by the Fe status of the plant. AHA7 is closely coregulated with IRT1 and FRO2 and may have a function in the induction of Fe-deficiency induced formation of extra root hairs. Additional functions of AHA7 in the Strategy I response should not be excluded. A high proton extrusion capacity appears to be associated with a stable shoot–root ratio, which may confer a competitive advantage in acquiring limited resources.


We thank Drs Wenfeng Li and Thomas J. W. Yang for advice in determining proton extrusion activity, stimulating discussions and comments on the manuscript, and Chung-Wen Lin for skilful experimental help. This work was supported by an AS Pilot grant to W.S. S.S. was supported by the Area Science Park (Trieste, Italy) with a F.V.G.-D4 Project grant.