Iron (Fe) has an essential role in the biosynthesis of chlorophylls and redox cofactors, and thus chloroplast iron uptake is a process of special importance. The chloroplast ferric chelate oxidoreductase (cFRO) has a crucial role in this process but it is poorly characterized.
To study the localization and mechanism of action of cFRO, sugar beet (Beta vulgaris cv Orbis) chloroplast envelope fractions were isolated by gradient ultracentrifugation, and their purity was tested by western blotting against different marker proteins. The ferric chelate reductase (FCR) activity of envelope fractions was studied in the presence of NAD(P)H (reductants) and FAD coenzymes. Reduction of Fe(III)-ethylenediaminetetraacetic acid was monitored spectrophotometrically by the Fe(II)-bathophenanthroline disulfonate complex formation.
FCR activity, that is production of free Fe(II) for Fe uptake, showed biphasic saturation kinetics, and was clearly associated only to chloroplast inner envelope (cIE) vesicles. The reaction rate was > 2.5 times higher with NADPH than with NADH, which indicates the natural coenzyme preference of cFRO activity and its dependence on photosynthesis.
FCR activity of cIE vesicles isolated from Fe-deficient plants also showed clear biphasic kinetics, where the KM of the low affinity component was elevated, and thus this component was down-regulated.
In plants, as in all organisms, iron (Fe) plays a crucial role. Owing to the outstanding redox characteristics, Fe is often used in electron transport chains and redox reactions. The relatively high concentration in tissues and essential role in metabolic processes make Fe the most important nutrient among transition metals. A large part of shoot Fe is localized in the chloroplasts (Terry & Abadia, 1986; Morrissey & Guerinot, 2009). In the absence of Fe, the biosynthesis of essential cofactors such as cytochromes and Fe-S clusters is strongly disturbed, which inhibits the biogenesis of thylakoid complexes and induces dysfunction of electron transport and enzyme reactions (Briat et al., 2007). Iron is also essential for Chl biosynthesis, and thus Fe deficiency leads to a series of symptoms called Fe chlorosis, causing a decreased biomass production and, finally, the death of plants (Spiller et al., 1982; Duy et al., 2007; El-Jendoubi et al., 2011). Therefore, Fe uptake into the chloroplasts from the cytoplasm of the mesophyll cells is a process of prime importance. Nevertheless, the whole mechanism of chloroplast Fe uptake is not yet well known (Abadía et al., 2011; Duy et al., 2011).
Chloroplasts are semiautonomous organelles of the plant cell covered by a double envelope membrane system. Thus, their Fe acquisition may differ from the Fe uptake of eukaryotic cells as Fe should cross both membranes. The first protein found to be involved in the chloroplast Fe acquisition is the AtPIC1 (Duy et al., 2007), which is localized in the inner envelope of chloroplasts (cIE) and is often thought to be a member of the cIE protein translocon machinery (TIC21). Based on data obtained on AtPic1 overexpressing lines, PIC1 also seems to be a regulator of chloroplast Fe metabolism (Duy et al., 2011).
It has long been postulated that a reduction-based system should also be involved in Fe import into the chloroplast (Bughio et al., 1997). In Arabidopsis chloroplasts, a ferric chelate oxidoreductase (FRO) family protein, AtFRO7, was shown to have a crucial role in Fe metabolism, particularly under Fe-limiting conditions (Jeong et al., 2008). In fro7 mutants, chloroplasts have reduced ability to collect Fe from the cytoplasm, and plants were only able to survive under high Fe (Jeong et al., 2008). FROs belong to a eukaryotic protein family, a branch of the superfamily of flavocytochromes. Possibly every plant membrane system contains its own FRO enzyme (Wu et al., 2005; Jeong & Connolly, 2009). FROs are thought to pass electrons across membranes using two cooperating intramembrane heme groups (Robinson et al., 1999). The apoproteins consist of a ferric reductase-like transmembrane domain, two cytochrome b-245 domains, a ferredoxin reductase-type FAD-binding domain and an NAD(P) binding domain (Schagerlöf et al., 2006; TAIR database, http://www.arabidopsis.org). The most studied root plasma membrane FRO (FRO2 in Arabidopsis) was shown to catalyze electron transport from NADH to ferric complexes (Holden et al., 1991), and NADPH-dependent ferric chelate reductase (FCR) activity was low compared with NADH-dependent activity (Bérczi & Møller, 2000).
In spite of the importance of the chloroplast FRO (cFRO) in chloroplast Fe uptake, the localization of the enzyme is putative, and its mechanism of action is also still poorly resolved (Jeong & Connolly, 2009). Intact chloroplasts were shown to have FCR activity in vitro (Mikami et al., 2011). Jeong et al. (2008) postulated that the key enzyme, AtFRO7, is localized in the chloroplast envelope membranes. Although recent papers hypothesized that cFRO should be a component of the cIE membrane (Nouet et al., 2011; Solti et al., 2012), this has not been proven experimentally as yet. Under Fe deficiency, the chloroplast FCR (cFCR) activity was shown not to be markedly increased (Mikami et al., 2011), which may indicate a different regulation of cFRO from that of the well-known root plasma membrane FRO.
In the present study, we compared the enzymatic characteristics of cFRO in chloroplast inner and outer envelope membranes (cIE and cOE, respectively) isolated from intact chloroplasts of Fe-sufficient and Fe-deficient sugar beet plants to reveal its localization, coenzyme preference, and response to Fe deficiency.
Materials and Methods
Seeds of sugar beet (Beta vulgaris L. cv Orbis) were germinated at moderate light, and planted directly into nutrient solution in 12 l plastic containers. Plants were grown in a climate chamber (14 : 10 h light (120 μmol photons m−2 s−1) : dark periods, 24 : 22°C and 70 : 75% relative humidity). Fe-sufficient plants were cultivated in modified quarter-strength Hoagland solution (1.25 mM Ca(NO3)2, 1.25 mM KNO3, 0.5 mM MgSO4, 0.25 mM KH2PO4, 0.25 mM NaCl, 11.56 μM H3BO3, 4.6 μM MnCl2, 0.19 μM ZnSO4, 0.12 μM Na2MoO4, 0.08 μM CuSO4) with 10 μM Fe(III)-ethylenediaminetetraacetic acid (Fe(III)-EDTA) as the Fe source. In order to obtain well-developed Fe chlorosis, another group of seedlings were precultivated up to four-leaf stage in the same nutrient solution and then transferred to iron-free nutrient solution also containing 0.5% (w/v) CaCO3. Leaves that emerged during the next 3 wk were used for chloroplast isolation from both Fe-sufficient and Fe-deficient plants.
Chlorophyll fluorescence induction
Fluorescence induction measurements of leaf samples were performed using a PAM 101-102-103 Chlorophyll Fluorometer (Walz, Effeltrich, Germany). Leaves were dark-adapted for 30 min. After eliminating reduced electron carriers by a 3 s far-red light impulse (Belkhodja et al., 1998), the F0 level of fluorescence was determined by switching on the measuring light (modulation frequency of 1.6 kHz and photosynthetic photon flux density (PPFD) < 1 μmol m−2 s−1). The maximum fluorescence yields, Fm in the dark-adapted state and in the light-adapted state, were measured by applying a 0.7 s pulse of white light (PPFD of 3500 μmol m−2 s−1; light source, KL 1500 electronic; Schott, Mainz, Germany) which saturated photosystem II (PSII) electron transport thus closing all PSII traps. Maximal efficiency of PSII centres were determined as Fv/Fm = (Fm−F0)/Fm.
Isolation of chloroplasts
All procedures were carried out at 4°C. About 8 kg of leaves were used to isolate envelope membranes. On every single occasion, 50 g of leaves were homogenized in 400 ml isolating HEPES buffer (50 mM HEPES-KOH, pH 7.0, 330 mM sorbitol, 2 mM EDTA, 2 mM MgCl2, 0.1% (w/v) BSA, 0.1% (w/v) Na-ascorbate) for 2 × 3 s in a 1 l Waring blender container. The homogenate was filtered through four layers of gauze and two layers of Miracloth™ (Calbiochem-Novabiochem, San Diego, CA, USA) to remove cell wall fragments and nuclei. The filtered homogenate was centrifuged at 1500 g for 5 min in a swing-out rotor to obtain chloroplasts. To remove mitochondrial contamination, the chloroplast pellet was washed twice in 50 mM HEPES-KOH (pH 7.0), 330 mM sorbitol, 2 mM EDTA, 2 mM MgCl2, and pelleted by a centrifugation at 2400 g for 5 min in a swing-out rotor. Washed chloroplasts were resuspended in a Tricine-EDTA (TE) storage buffer (10 mM Tricine-KOH, pH 7.8, 2 mM EDTA) with 0.6 M sucrose.
The Chl content of chloroplasts was determined in 80% (v/v) acetone extracts using a UV–VIS spectrophotometer (Shimadzu, Kyoto, Japan) using the absorption coefficients of Porra et al. (1989).
Isolation of chloroplast envelope membranes
Chloroplast envelope membrane vesicles were isolated from washed chloroplasts according to the modified method of Keegstra & Yousif (1986) and Froehlich et al. (2003). Chloroplasts, suspended in TE buffer with 0.6 M sucrose, were broken in three freeze–thaw (−20/0°C) cycles. After the last thaw phase, the suspension was three times diluted with TE buffer to achieve a sucrose concentration of 0.2 M, and incubated on ice for 1 h. The 1 h incubation time proved to be crucial to reach the highest purity of cIE membranes (i.e. to reduce thylakoid cross-contamination). All other steps were carried out at 4°C. Thylakoid membranes were removed from the broken chloroplast suspension by a centrifugation at 4500 g for 15 min in a swing-out rotor. The supernatant was collected, and membranes were pelleted at 25 000 g for 65 min using a swing-out rotor (Sw40Ti) in a Beckman L7 ultracentrifuge (Beckman Coulter Inc., Brea, CA, USA). The pellet was resuspended in TE buffer containing 0.2 M sucrose, and layered on a discontinuous sucrose gradient. The gradient was composed of 3 ml 1.0 M sucrose (ρ = 1.13 g ml−1), 2 ml 0.8 M sucrose (ρ = 1.10 g ml−1), 3 ml 0.46 M sucrose (ρ = 1.06 g ml−1) and a 3 ml sample in 0.2 M sucrose (ρ = 1.03 g ml−1) in TE buffer. Envelope and thylakoid membranes were separated by ultracentrifugation at 140 000 g for 135 min in a swing-out rotor. Envelope membrane vesicle fractions were collected from the 1.0/0.8 M (quasi-cIE) and the 0.8/0.46 M (mixed cIE + cOE) gradient interface. The band at the 0.46/0.2 M interface usually contains chloroplast stroma material, and therefore it was not used in further experiments. Isolated envelope vesicles were diluted with TE buffer containing 0.2 M sucrose, and pelleted by ultracentrifugation at 40 000 g for 75 min in a swing-out rotor. Vesicle pellets were resuspended in TE buffer containing 0.2 M sucrose (protein concentration of 4–8 μg μl−1) and stored in liquid nitrogen until use.
After pelleting the chloroplasts from leaf homogenate, the supernatant was used to isolate microsomal fraction (M). Broken thylakoids were removed by a centrifugation at 4500 g for 15 min in a swing-out rotor. M fraction was pelleted by a centrifugation at 20 000 g for 30 min and also suspended in TE buffer containing 0.2 M sucrose. M fraction was also subjected to gradient separation as described before. A fraction enriched in cOE vesicles was collected from the 0.8/0.46 M gradient interface (M2).
Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and Western blotting
Chloroplasts and membrane samples were solubilized in 62.5 mM Tris-HCl, pH 6.8, 2% SDS, 2% dithiothreitol, 10% glycerol, and 0.001% bromophenol blue at room temperature for 30 min. Fifty-microgram proteins were separated according to Laemmli (1970) but in 10–18% gradient polyacrylamide gels in a MiniProtean apparatus (Bio-Rad Laboratories Inc., Hercules, CA, USA) using a constant current of 20 mA per gel at 6°C. The protein concentration of samples was determined by comparing the area density with that of a standard mixture using Phoretix 4.01 software (Phoretix International, Newcastle upon Tyne, UK).
To identify fractions and detect cross-contamination of vesicles, protein blots were carried out against a chloroplast triose-phosphate translocator (cTPT, a cIE marker; Lundmark et al., 2006), light harvesting complex apoproteins (apoLHC, a thylakoid marker), a protein component of the cOE translocon complex (TOC75 – a membrane channel protein of the TOC protein import machinery, a cOE marker, Eckart et al., 2002), and mitochondrial alternative oxidase (AOX 1/2, a mitochondrial inner envelope marker, Lang et al., 2011). Membrane proteins separated by SDS-PAGE were transferred to Hybound™-C Extra (Amersham-Pharmacia, Piscataway, NY, USA) nitrocellulose membranes in 25 mM Tris, pH 8.3, 192 mM glicine, 20% (v/v) methanol and 0.02% (m/v) SDS at 4°C using 90 V constant voltage (< 0.4 A) for 3 h. Membranes were decorated with rabbit polyclonal antibodies against apoLHCII (a gift from Dr Udo Johanningmeier, Bohum Universität, Germany), cTPT, which has a known cross-reaction to the ribulose-1,5-bisphosphate carboxylase oxygenase large subunit, RbcL, AOX 1/2, and TOC75 (antibodies were obtained from Agrisera AG, Vännäs, Sweden). Antibodies were dissolved in 20 mM Tris-HCl (pH 7.5), 0.15 M NaCl, 1% gelatine following the manufacturer's instructions. Horseradish peroxidase (HRP)-conjugated goat-anti-rabbit IgG (Bio-Rad) was used to detect bands following the manufacturer's instructions.
FCR activity measurements
The following procedure ensured that, from the envelope vesicles population, only right-side-out chloroplast envelope membrane vesicles could participate in FCR reactions. Chloroplast envelope vesicles equivalent to 50 μg protein suspended in 10 μl of TE buffer containing 0.2 M sucrose were mixed with 10 (−40) μl 10 mM NAD(P)H (depending on the concentration of Fe(III)-EDTA, see below) and 10 μl 10 mM FAD, and vesicles were loaded with these substances by using one freeze–thaw cycle (−20/0°C). Based on our previous experiments, the FCR reaction is not limited by NAD(P)H at the given concentrations.
In order to avoid any changes in the protein concentration, NAD(P)H and FAD-enclosing envelope vesicle suspensions (30–60 μl) that were not washed (with NAD(P)H and FAD inside and outside the vesicles) were diluted by HEPES buffer containing BPDS up to 1 ml. After dilution, the final reaction mixture contained 50 mM HEPES-KOH (pH 7.0), 330 mM sorbitol, 2 mM EDTA, 2 mM MgCl2, 300 μM BPDS, 100 μM FAD and 100 (−400) μM NAD(P)H (depending on the concentration of Fe(III)-EDTA), and the envelope vesicles described earlier (containing inside 3.3 (−6.6) mM NAD(P)H and 3.3 mM FAD). As a result of the procedure described earlier, the NAD(P)H and FAD concentrations in the envelope vesicles were much higher than outside the vesicles, and in line with previously reported chloroplast concentrations of NAD(P)H (Ogren & Krogmann, 1965; Harvey & Brown, 1969; Heineke et al., 1991) and FAD (Giancaspero et al., 2009) (see Supporting Information, Tables S1,S2). Above 100 μM Fe(III)-EDTA in the external solution, the concentration of NAD(P)H was raised in parallel to Fe(III)-EDTA to avoid coenzyme limitation of the blank reaction. Although FAD and NAD(P)H were present both in the vesicles and in the reaction mixture, Fe(III)-EDTA and BPDS were absent in the vesicles; thus the FCR reaction could only be mediated by right-side-out vesicles (see Notes S1, Fig. S1). To test the pH dependence of the FCR reaction, the pH of HEPES-KOH was set to 5.5, 6.0, 6.5, 7.0, 7.5, 8.0 and 8.5.
The reaction was initiated by adding Fe(III)-EDTA to the mixture. A blank reaction mixture (without envelope vesicles), which contained 50 mM HEPES-KOH (pH 7.0), 330 mM sorbitol, 2 mM MgCl2, 300 μM BPDS, 100 μM FAD with Fe(III)-EDTA and NAD(P)H in the concentrations described earlier, was used to subtract the background reaction of NAD(P)H to Fe(III)-EDTA. The absorbance change at 535 nm was followed using a UV–VIS spectrophotometer (Shimadzu) for 30 min, using an absorption coefficient of 22.14 mM−1 cm−1 (Smith et al., 1952). The reaction rate was calculated from the linear phase of the reaction.
Error bars show SD values. To compare treatments, Student's t-test was performed with InStat v. 3.00 (GraphPad Software, Inc., La Jolla, CA, USA). Origin v. 6.01 (Origin Lab, Co., Northampton, MA, USA) was used to fit sigmoid functions on data points. Boltzmann's function was used for the KM calculation.
Physiological status of plant material
Before isolating chloroplasts, the Chl content and photosynthetic activity of Fe-sufficient and Fe-deficient leaves were measured, as indicators of the iron nutrition status. The Chl content was 39.0 ± 3.8 and 18.2 ± 4.3 μg Chla + Chlb cm−2 in Fe-sufficient and Fe-deficient leaves, respectively. The Chla/b ratios were 3.21 ± 0.16 and 4.18 ± 0.10 in Fe-sufficient and Fe-deficient leaves, respectively. The maximal quantum efficiencies of PSII reaction centers (Fv/Fm) were 0.838 ± 0.009 in Fe-sufficient leaves and 0.680 ± 0.038 in Fe-deficient leaves. All the differences were significant (P <0.05).
Purity of the cIE vesicles
A fraction of yellowish color, collected from the 1.0/0.8 M sucrose gradient interface, gave a strong band in the cTPT assay. In this fraction, no AOX 1/2, TOC75 or apoLHCII, and only a weak band for RbcL (markers of mitochondria, cOE and thylakoid and chloroplast stroma contamination, respectively) appeared in the western blot assay. Thus, the 1.0/0.8 M gradient interface fraction was assigned to a pure cIE fraction (only a slight RbcL contamination appeared). In the fraction collected from the 0.8/0.46 M gradient interface, TOC75 was seen together with cTPT. Nevertheless, this fraction was free of AOX 1/2 and apoLHC bands. Thus, this fraction was assigned to a mixed envelope (cIE + cOE) fraction. In the M2, a clear band of TOC75 appeared with only very faint traces of RbcL, and apoLHC (Fig. 1a,b), so it was assigned to a cOE fraction. A weak band of AOX 1/2 was observed in the gradient pellet (mainly thylakoids) and in the unwashed chloroplast fraction (Fig. 1c). Thylakoid membranes contained a strong band of apoLHCII. In FCR assays, the 1.0/0.8 M gradient interface and M2 fractions are hereafter referred to as cIE and cOE vesicle fractions, respectively.
FCR activity of Fe-sufficient chloroplast envelope membranes
Both cIE and cOE vesicles enclosing NAD(P)H and FAD were subjected to FCR assay. The cOE fraction showed no FCR activity, whereas cIE membrane vesicles did reduce Fe(III)-EDTA at high rates (Fig. 2). Fe reduction was not measured in the absence of the reduced coenzyme, and the weak measuring light of the spectrophotometer (< 0.01 μmol m−2 s−1) had no effect on FCR activity (data not shown).
The activity and kinetics of the FCR reaction differed significantly in the presence of NADPH vs NADH (Fig. 3). Using NADPH, clear biphasic saturation kinetics were measured, with the first saturation found at KM = 4.6 ± 0.6 μM and the second saturation found at KM = 55.7 ± 2.1 μM Fe(III)-EDTA; the maximal rates (vsat) of high- and low-affinity reactions were 4.4 ± 0.1 and 17.1 ± 0.8 pmol Fe μg−1 protein min−1 Fe(III)-EDTA, respectively. In the presence of NADH, the saturation kinetics was monophasic; the sole saturation reaction rate was significantly lower (vsat = 6.6 ± 0.2 pmol Fe μg−1 protein min−1), and the KM value of the reaction was significantly higher (83.4 ± 0.9 μM Fe(III)-EDTA), than those of the second saturation measured with NADPH. FCR activity was also tested in the pH range 5.5–8.5 in the presence of NADPH coenzyme and 200 μM Fe(III)-EDTA (Fig. 4). The reaction rate was found to be maximal between pH 6.5 and 7.0. Higher pH values strongly decreased the FCR activity, whereas acidic pH affected FCR activity slightly.
FCR activity of Fe-deficient chloroplast envelope membranes
Comparing the FCR activity of Fe-sufficient and Fe-deficient cIE vesicles with NADPH as a coenzyme, some differences were found in the enzyme kinetics (Fig. 5). The FCR activity of cIE vesicles obtained from Fe-deficient plants also showed a biphasic saturation curve, with the first saturation rate being significantly lower (2.6 ± 0.2 pmol Fe μg−1 protein min−1) than that of the vesicles isolated from Fe-sufficient plants. Nevertheless, the affinity of the enzyme did not change (KM = 4.2 ± 0.1 μM Fe(III)-EDTA). The second, low-affinity saturation point (KM = 155.3 ± 8.2 μM Fe(III)-EDTA) was found to be significantly higher than that in Fe-sufficient plants, with a reaction rate that was also somewhat higher (vsat = 21.0 ± 0.7 pmol Fe μg−1 protein min−1) than in the cIE membranes of Fe-sufficient plants. In the FCR reaction kinetics measured on Fe-deficient cIE vesicles, the two saturation curves could be clearly distinguished as a result of the significant difference in the KM values.
Iron uptake of chloroplasts from the mesophyll cells is very important from the point of view of chloroplast development (Chl and cofactor synthesis, complex stabilization), which also deeply influences plant productivity (Terry & Abadia, 1986). Although it is a hot topic in studies on Fe homeostasis in plants, the only information available has to do with the process (Abadía et al., 2011). Some characteristics of cFRO, one of the key components of chloroplast iron uptake, have been revealed here on the basis of activity measurements on isolated chloroplast envelope fractions. We used Fe(III)-EDTA in the FCR assay, because in contrast to the possible natural complexes such as ferric-citrate, ferric-malate and ferric-nicotianamine, the reaction rate of NAD(P)H with Fe(III)-EDTA was low (as a result of the strong chelation of Fe(III) by EDTA, log K1 = 25.1). This allowed the cIE FCR enzyme reaction to be measured against a blank without membrane vesicles but containing NAD(P)H, FAD and Fe(III)-EDTA. With regard to the measurements of pH dependence of the FCR reaction, Fe(III)-EDTA is often thought to lose its stability above pH 7.0. However, according to previous Mössbauer spectroscopy measurements, this does not mean that Fe(III) would be released from the chelate and form insoluble FeOH3; instead, the monomeric Fe(III)-EDTA is transformed into the dimeric Fe(III)2-EDTA2 form (in a 0.01 M Fe(III)-EDTA 1 : 1.1 solution, a 100% dimeric complex was found; Homonnay et al., 2008). A similar effect was also shown at pH 7.0 (Stein & Marinsky, 1975).
Chloroplast FRO is a cIE enzyme with NADPH coenzyme preference
The cIE fraction was free of detectable cross-contaminations with cOE, thylakoid and mitochondrial membranes, while the M2 fraction was enriched in cOE membranes and showed no sign of cIE, mitochondrial and thylakoid markers. The gradient separation also excludes the possibility of plasma membrane and endoplasmic reticulum membrane contamination. As only cIE vesicles were able to reduce Fe(III)-EDTA, and the M2 fraction, containing a significant amount of cOE, gave no reaction, the cFCR activity of chloroplasts can be assumed to be associated with cIE membranes. The applied freeze–thaw technique of isolating envelope membrane vesicles in the presence of Mg(II) was shown to result in a mixed (both right-side-out and inside-out vesicles) population of envelope vesicles (Waegemann et al., 1992; Shingles & McCarty, 1995). Nevertheless, the presence of inside-out cIE vesicles practically did not affect the determination of enzyme affinity parameters. According to structural information (Jeong & Connolly, 2009; ARAMEMNON database, http://aramemnon.uni-koeln.de/), the N terminus containing the NAD(P) and FAD-binding domain of FRO enzymes is located in the cytoplasm, that is in the stroma compartment in the chloroplasts, whereas the C terminus containing the ferric-chelate-binding domain face toward the ‘external’ spaces, which is the inter-envelope space in the case of cFRO. In our experiments, Fe(III)-EDTA was only present in the external solution, and therefore inside-out vesicles, where the ferric-chelate-binding domain of cFRO must be located on the inside of the vesicles, did not contribute to the FCR reaction (see Notes S3, Fig. S1). The internal NAD(P)H concentration of membrane vesicles was comparable to NAD(P)H concentrations in chloroplast stroma (see Notes S1), and therefore the experimental setup more or less represented the in vivo conditions. In the presence of the NADPH coenzyme, the FCR reaction of Fe-sufficient cIE membranes showed biphasic saturation kinetics, whereas in the presence of NADH, the high-affinity component was not found. The saturation rate of the low-affinity FCR reaction was significantly lower, while the KM was significantly higher, with NADH than with NADPH, indicating a much higher activity of the enzyme to reduce Fe(III)-EDTA in the presence of NADPH. The maximal rate of the FCR reaction in the presence of NADPH in cIE (taking into account the efficiency of envelope membrane isolation from intact chloroplasts; see Cline et al., 1981) was comparable to that measured by Mikami et al. (2011) in intact chloroplasts (Notes S2, Table S3). Therefore, the measured in vitro cFRO reaction rate appears to be reliable. Chloroplast Fe uptake was shown to be a light- and photosynthesis-dependent process (Bughio et al., 1997; Solti et al., 2012). In contrast to the cytoplasm, the NADPH : NADH ratio is very high in chloroplasts (Ogren & Krogmann, 1965; Harvey & Brown, 1969; Heineke et al., 1991), and it increases strongly under light conditions (Heineke et al., 1991). NADP+ is reduced mainly by ferredoxin : NADP+ oxidoreductase (FNR), that is, by electrons originating from the photosynthetic electron transport. Overall this indicates that there must be a strong connection between the cFCR reaction and the activity of photosynthetic electron transport, in line with the findings of Mikami et al. (2011). As the cFCR reaction is essential to chloroplast Fe acquisition (Jeong et al., 2008), the NADPH-dependence of the FCR reaction could be one of the main reasons for the light dependence and 3-(3,4-dichlorophenyl)-1,1-dimethylurea (DCMU) sensitivity of the chloroplast Fe uptake process (Bughio et al., 1997; Solti et al., 2012). Therefore, NADPH is likely to be the electron donor for cFRO in vivo. The pH optimum found in the pH 6.5–7.0 range, as well as the slight decrease in acidic pH and the strong decrease in alkaline pH range underline that the ferric chelate reducing side of the enzyme should face the inter-envelope space. cIE membranes contain H+/ATPases establishing an inwardly directing H+ gradient across the IE and also acidifying the inter-envelope space (Neuhaus & Wagner, 2000). The dependence of the cFRO enzyme on NADPH produced in the light reaction, together with its pH preference again exclude the possibility that cFRO would face chloroplast stroma with the ferric chelate reducing side.
Mechanism of action of chloroplast FRO
In previous studies, isolated cIE membranes were only found to be permeable to free Fe(II) (Shingles et al., 2002; Duy et al., 2007; Solti et al., 2012). The only and essential cIE Fe transporter identified so far is the PIC1 Fe(II) permease, which is of cyanobacterial origin (Duy et al., 2007). According to our previous Mössbauer spectroscopy measurements, Fe(II) released by cFCR does not accumulate in high spin complexes (e.g. [Fe(H2O)6]2+) between the two envelope membranes (Solti et al., 2012), as it does in roots after reduction by FRO2 (Kovács et al., 2010). The absence of the accumulation of a ferrous-hexahydrate pool is indicative of a possibly immediate transport of the reduced Fe(II) through the cIE membranes. Therefore, cFRO may directly feed PIC1 with Fe(II), and the ferric chelate reduction site should be located in the cIE in a position facing the inter-envelope space. Consequently, the in vivo role of cFRO is the free Fe(II) production in the inter-envelope space for subsequent chloroplast Fe uptake by PIC1. The biphasic saturation of the cFRO enzyme indicates that a high- and a low-affinity isoform exist in the cIE membrane. The high-affinity isoform has a KM one order of magnitude lower than the low-affinity isoform. The different affinity of cFRO to Fe(III)-EDTA may have originated from either isoform changes or allosteric modifications (e.g. phosphorylation, redox regulation). In Arabidopsis, Jeong et al. (2008) identified a single cFRO enzyme, FRO7. Comparing the sequence homology of FROs in different dicots (Arabidopsis thaliana, Populus trichocarpa, Medicago truncatula, Vitis vinifera, ARAMEMNON database), AtFRO7 has a single predicted ortholog in these taxa, which makes it more probable that Beta vulgaris also has a single cFRO. Certainly, allosteric regulation processes, such as phosphorylation, may take part in the formation of high- and low-affinity cFRO isoforms. We still have no information on the possible allosteric modifications of cFRO (Jeong & Connolly, 2009), but this may be a major regulator of cFCR activity if only one cFRO exists.
Effect of Fe-deficiency treatment on cFCR activity
Iron-deficient treatment affected the iron nutrition status of sugar beet plants. The decreased PSII maximal quantum efficiency, together with the lowered total Chl content, indicated that the Fe-dependent biosynthesis of chlorophylls, as well as the functioning of the photosynthetic apparatus, was retarded. The elevated Chla/b ratio in Fe-deficient leaves indicates that Fe limitation in chloroplasts increases the relative amount of Chla containing reaction centres in contrast to Chla and b containing antennae. The behavior of Fe-deficient plants is in accordance with previous studies on Fe-deficient sugar beet plants (Larbi et al., 2004). cFCR activity increased in isolated barley chloroplasts but did not change in sorghum under Fe deficiency (Mikami et al., 2011). Fe deficiency induces disturbances in thylakoid development, so that their photosynthetic performance strongly decreases (Andaluz et al., 2006; Timperio et al., 2007). Thus, lower amounts of NADPH must be available for the cFCR reaction. Even considering the c. 50% decrease in the amount of NADPH in Fe-deficient chloroplasts compared with the controls, the resulting NADPH concentrations will be higher than the limiting ones.
Iron-deficient treatment also affected the cFCR activity of isolated cIE vesicles: a biphasic Fe uptake consisting of a low- and a high-affinity component was also detected in Fe-deficient cIE vesicles. The low-affinity component showed a significantly lower affinity (higher KM) to Fe(III)-EDTA, which may indicate that under low Fe concentration in the cytoplasm (Fe deficiency), the low-affinity component loses its affinity to ferric-chelates and is thus down-regulated. The KM of the high-affinity component was similar to that in the Fe-sufficient cIE vesicles, indicating that this component of cFRO has an importance under iron starvation. However, the reaction rate of this component decreased compared with Fe-sufficient samples. These changes may be connected to an altered iron allocation among organelles. A similar effect was recently found in Clamydomonas reinhardtii (Höhner et al., 2013), where iron deprivation induced a transition from photoautotrophic to partially heterotrophic metabolism, which enhanced the iron allocation to mitochondria. In Aphanocapsa (Cyanobacteria), below a critical concentration of cytoplasm iron content, the photosynthetic electron transfer was specifically depressed, whereas the influence of iron limitation on respiration was negligible (Sandmann, 1985). Thus, iron allocation to the respiratory pathway may be an important mechanism under iron starvation which also affects chloroplast iron homeostasis.
Chloroplast FRO is a cIE membrane enzyme that reduces Fe in the inter-envelope space. NADPH coenzyme preference of cFRO is in line with the light/photosynthesis dependence of chloroplast Fe uptake. Under Fe-deficiency conditions, the low-affinity isoform becomes less effective at collecting Fe from the cytoplasm.
Sugar beet (Beta vulgaris L. cv Orbis) seeds were a kind gift of Dr Javier Abadía (EEAD Aula Dei, Zaragoza, Spain). We would like to thank to Zsuzsanna Ostorics for technical assistance. This work was supported by the following grants: ERA Chemistry – OTKA NN-84307 and Chinese-Hungarian Bilateral Cooperation TÉT_12_CN-1-2012-0025.