Competition or complementation: the iron‐chelating abilities of nicotianamine and phytosiderophores

Efficient acquisition of iron (Fe) in graminaceous plant species relies on the synthesis and release of phytosiderophores, which are low-molecular weight chelators with functional carboxy-, amino- and hydroxy-groups for hexadentate metal coordination (Sugiura & Nomoto 1984). Metal-phytosiderophores formed in the rhizosphere are subsequently taken up by YS1-type membrane proteins, that energize root uptake by the cotransport of metal-phytosiderophores with protons (Schaaf et al., 2004). As Fe(III)-phytosiderophores dominate Fe(III) chelation over hydroxide formation in the pH range between 3.5 and 8.5 (von Wirén et al., 1999) and as they are mainly found in root exudates or the xylem sap (Kawai et al., 2001), phytosiderophores may adopt a major role in extracytosolic Fe chelation. Nicotianamine (NA), which occurs in all higher plants, is structurally similar to phytosiderophores and serves as a precursor in their biosynthetic pathway. Fe trafficking within the plant has been shown to have a strong dependency on NA; this has been demonstrated by severe symptoms of Fe deficiency in mutant or transgenic plant lines with low NA levels (Ling et al., 1999; Takahashi et al., 2003). NA has been shown to chelate not only Fe(III) but also Fe(II) (Benešet al., 1983), in particular at higher pH values, which may allow its function as an intracellular Fe(II) scavenger thereby protecting the cell from Fe(II)-mediated oxidative damage (von Wirén et al., 1999). The occurrence of phytosiderophores and NA in graminaceous plant species thus raises the question whether Fe(III) may undergo a chelate exchange from phytosiderophores to NA, as proposed by von Wirén et al. (1999). These authors concluded on the basis of titration and capillary electrophoresis studies that the phytosiderophore deoxymugineic acid (DMA) chelates Fe(III) preferentially at more acidic pH values (pH 3–6), while NA is a more competitive chelator at pH values > 6. Reichman & Parker (2002) reanalysed the previously published data and claimed to have identified information which partially rejects the above hypothesis and has important implications for Fe physiology and transport in plants. The two aspects which were the subject of criticism were (i) the competition between NA and DMA for Fe(III) at pH 7.0, and (ii) the relative affinity of NA for Fe(II) and Fe(III). We address these two aspects in sequence. By considering the formation of a single negatively charged complex [Fe(III)-DMA(H−1)]−, along with the neutral complex [Fe(III)-DMA]0, Reichman & Parker (2002) calculated a speciation plot, in which Fe(III)-DMA formation dominated over the pH range 2.5–9, thus prohibiting the occurrence of an Fe(III)-NA complex, when NA was present at equimolar concentrations (Reichman & Parker, 2002, Fig. 2B). It was then concluded that DMA outcompetes NA over the entire physiologically relevant pH range, in contrast to that proposed by von Wirén et al. (1999). Since the analysis of chelate formation reported by von Wirén et al. (1999) was based on experimentally determined pKa values and iron(III) affinity constants, we first reassessed the pKa values for both NA and DMA and also determined the affinity constants for Fe(III) by potentiometric titration. Our previously determined values of the Fe(III) affinity constants were based on spectrophotometric analysis of competition studies with maltol. Employing chemically synthesised compounds the same sequence of pKa values for both NA and DMA was obtained as previously reported (von Wirén et al., 1999) (Table 1). DMA has four measurable pKa values whereas NA has five. This difference is due to the terminal amino group of NA, that protonates at a pKa of 7.73, while the terminal hydroxyl group of DMA does not dissociate even at very high pH values (Table 1, Fig. 1a). There are different protonation states for both NA-Fe(III) and DMA-Fe(III) complexes and these possess different net charges, namely for DMA [Fe(III) DMA]0, [Fe(III) DMA(H−1)]− and [Fe(III) DMA·(H−1)OH]2–. We have now measured the equilibrium constants relating to each of these species for both NA and DMA by potentiometric titration as previously reported (von Wirén et al., 1999) (Table 1). With DMA we found evidence for the protonated species [Fe(III) DMA]0 over the pH range 2–7 which agrees with the earlier data produced by von Wirén et al. (2000). We found no evidence for the hydroxylated species [Fe(III) DMA·(H−1)OH]2–. We also investigated the possibility of dimer formation of the type reported for HEDTA (Gustafson & Martell, 1963) and EDTA (Schuger et al., 1969; 1972) and could find no evidence for the existence of such species at µm concentrations and below. This analysis is in agreement with the electrophoretic determination of Fe(III)-DMA species that showed a gradual shift from the negatively charged to the neutral species when lowering the pH from 7 to 5 (von Wirén et al., 2000, Fig. 3). Unfortunately, this latter study was apparently not considered by Reichman & Parker (2002). Thus the speciation plot for Fe(III)-DMA is relatively simple with the two Fe complexes [Fe(III) DMA·(H−1)]1– and [Fe(III) DMA]0 dominating over the pH range 2–8 (Fig. 1b). The pKa value for this deprotonation process is 6.3. (a) Structures of deoxymugineic acid (DMA) and nicotianamine (NA). Superscript numbers indicate the assignment of the pKa values given in Table 1. (b)–(d) Computer simulations of the pH dependence of the Fe(III) complexes of NA and DMA: (b) 1 µm Fe +10 µm DMA (= L); (c) 1 µm Fe +10 µm NA (= L′); (d) 1 µm Fe +10 µm DMA (= L) +10 µm NA (= L′). With NA there are three major Fe(III) species. As derived from the binding constants (Table 1) and the associated speciation plot, [Fe(III)-NA·H]1+ dominates at pH 2–6, while [Fe(III)-NA]0 dominates over the pH range 6–9 (Fig. 1c). This is consistent with with the neutral charge of the Fe(III)-NA complex at pH 7.0 when determined by paper electrophoresis (von Wirén et al., 1999). The hydroxylated species [Fe(III)-NA·(H−1)]1– is only present at more alkaline pH values (above 8.0). As with DMA we could find no evidence for the existence of dimer complexes at concentrations below 1 µm. The affinity constants of both NA and DMA for Fe(III) have been determined by two independent methods, namely potentiometric titration and spectrophotometric analysis of the competition of each ligand with maltol. As there are several Fe(III) complex species for NA and mugineic acid (MA), the competition analyses give an apparent formation constant (KFe3+). With these Fe(III) affinity constants it is possible to calculate the pFe3+ values independently. Both pairs of values are in good agreement (Table 2). On the basis of this new conductiometric data (Table 1) we have recalculated the speciation of Fe(III) in the simultaneous presence of NA and DMA. Consideration of the negatively charged [Fe(III)-DMA(H−1)]1–, that was omitted in von Wirén et al. (1999), will effectively decrease the relative competitiveness of Fe(III)-DMA over Fe(III)-NA. Over the 5.5–8.5 pH range it is clear that NA is predicted to dominate the co-ordination of Fe(III) (Fig. 1d), which supports the earlier analysis reported by von Wirén et al. (1999), where both DMA species have not been considered separately. A major point of Reichman & Parker's (2002) criticism pertained to the experimental approach made by von Wirén et al. (1999), in which capillary electrophoresis was employed to verify the relative ability of NA to form Fe(II) or Fe(III) complexes. Reichman & Parker (2002) raised the issue that Fe(III)-NA complexes were not detected by capillary electrophoresis, by virtue of the rapid formation of Fe(III)-hydroxides. This difference, however, is clearly indicated in Fig. 5 of von Wirén et al. (1999), where it is demonstrated that the pFe3+ value for NA at pH 7.4 is only marginally greater than the Fe(III)-hydroxide solubility curve, whereas the corresponding pFe2+ value is very well resolved from the Fe(II)-hydroxide solubility curve. Thus, although the affinity constant for Fe(III) is larger than that for Fe(II) (1018.4 vs 1012.8, respectively), competition with the hydroxide anion at pH 7.4 is much more effective for Fe(III) than for Fe(II). Since Fe(III)-hydroxide formation can certainly occur in the cytoplasm, our experimental conditions properly reflect the situation in planta and correctly emphasize the fact that Fe(II) possesses a relatively higher stability when complexed by NA. Reichman & Parker (2002) developed another argument based on theoretically derived redox potentials of NA-complexed Fe(II) and Fe(III) which led to the conclusion that thermodynamic considerations can satisfactorily explain the high stability of Fe(II), while kinetic aspects as in von Wirén et al. (1999) are not needed. This statement turns out to be purely speculative. First, Reichman & Parker's (2002) computation of redox potentials considered [Fe(III)-DMA]1– and [Fe(III)-DMA·OH]2– as the most important complexes, which is in contrast with the predicted and experimentally determined species at neutral pH (von Wirén et al., 2000). Second, Reichman & Parker's (2002) theoretically derived reduction potentials do not agree with experimentally derived values from Sugiura & Nomoto (1984), as they indicate in their correspondence ‘the reasons for this discrepancy are not known’. Third, kinetic considerations are of outmost importance when the formation of DMA complexes with different metal ions are observed. For example, the efficient formation of Cr(III)-DMA complexes is severely restricted by its slow kinetics. Thus, metals that undergo faster complex formation will dominate even if their formation constants are weaker (Hider, 1984). Until such time as there is better consistency between theory and experiment, we believe that arguments based on redox potentials associated with a complicated series of equilibria cannot provide definitive conclusions. On the basis of the concepts presented herein, the transfer of Fe(III) from DMA to NA would appear to be possible. However the kinetics of the movement of Fe(III) from one hexadentate siderophore to another is exceedingly slow (Hider, 1984) and, therefore, as previously suggested by von Wirén et al. (1999) and Sugiura & Nomoto (1984), reduction to Fe(II) before transfer to NA would appear to be a likely mechanism. The important point that emerges from this debate is that both von Wirén et al. (1999) and Reichman & Parker (2002) agree that NA form stable Fe(II) complexes. To avoid further speculation on metal speciation in plant extracts, future investigations should definitely focus on the direct chemical analysis of metal chelates. Such an analysis has recently been performed with yeast subjected to Ni toxicity, where Ni(II)-NA complexes were detected (Vacchina et al., 2003). As demonstrated recently, NA complexes with Ni(II), Fe(II) and Fe(III) form suitable substrates for membrane transport mediated by the metal phytosiderophore transporter ZmYS1 (Schaaf et al., 2004). Thus apparent Fe deficiency symptoms as observed in the NA-deficient tomato mutant chloronerva (Ling et al., 1999) are likely to be related with an intracellular maldistribution of Fe (Becker et al., 1995). Such a role is further supported by an immunochemical localization of NA in vacuoles (Pich et al., 2001). Another sharp contrast to the conclusion by Reichman & Parker (2002) that Fe-NA is unlikely to be employed for long-distance Fe transport, is the finding that NA deficiency in transgenic tobacco plants resulted in malnutrition of flowers with metal micronutrients including Fe. Again, this was best explained by an involvement of NA in intra- and intercellular Fe trafficking (Takahashi et al., 2003). Furthermore, recent studies on a ZmYS1 homolog from Arabidopsis, AtYSL2, indicated that nicotianamine-chelated Fe(II) is a suitable substrate for AtYSL2. Since AtYSL2-GFP fusion proteins are mainly located in the plasma membrane of vascular cells in roots (DiDonato et al., 2004), there is a high probability for Fe(II)-NA acting as an endogenous mobile Fe binding form. Taken together, in addition to the physico-chemical characterization of Fe-NA chelates, there is plenty of evidence from physiological and genetic studies that further indicate a role of NA in Fe trafficking, also in graminaceous species where DMA and NA occur in parallel (Becker et al., 1995; Pich et al., 2001; Bereczky et al., 2003; Takahashi et al., 2003; Schaaf et al., 2004). NA chelation and subsequent transport or compartmentation offers an attractive opportunity for plant cells for short- and long-distance transport of Fe and other essential metals and to decrease the immediate toxicity of Fenton-reactive free metals or simply to support their tissue and cellular distribution. In view of all this experimental evidence, Reichman & Parker's (2002) criticisms would appear to be unfounded. We thank Prof Dr T Kitahara and Prof Dr S Mori, (University of Tokyo, Japan) for the supply of NA and DMA. We thank the DFG Bonn for financial support to NVW (Grant WI 1728/6-1); the BBSRC for financial support to RCH; and CREST (Care Research for Evolutional Science and Technology) of Japan Science and Technology Corporation to EY.