Unlike Zn starvation, Fe starvation had a major effect on wheat growth (Table 1), Cu, Zn and Fe concentrations in wheat (Fig. 1 and Table 4) and significantly enhanced the release of PS (Fig. 3). It is well documented that, like other grasses, wheat responds to Fe deficiency by enhanced PS release which has been described as Strategy II of Fe acquisition (Marschner & Römheld, 1994; Marschner, 1995). There were only minor differences between the two bread wheat genotypes in the severity of Fe deficiency symptoms: ‘Aroona’ appeared less susceptible to Fe deficiency than ‘Songlen’. The growth of both wheat genotypes was considerably reduced under Fe deficiency: the dry biomass of Fe-deficient whole plants was 40% lower than that of adequately supplied plants (Table 1). However, Fe deficiency did not result in a significant change in Fe concentration in either roots or shoots of any of the two genotypes, except for hydroponically grown ‘Aroona’ shoots (Table 4). Conversely, Fe deficiency resulted in much increased concentrations of both Cu and Zn in the shoots and roots of both wheat genotypes (Fig. 1 and Table 4). This effect was significant in most cases and coincided with a much (significantly) increased release of PS in both genotypes at the end of the hydroponic stage of the experiment (Fig. 3). This suggested the possible implication of the release of PS in the acquisition of Cu (Fig. 1) and Zn (Table 4) for both hydroponically grown plants and, more interestingly, for plants grown on the calcareous soil as sole source of Cu and Zn. In the latter case, however, the relationship between Cu or Zn acquisition and PS release was not as clear when PS release at the end of the soil stage of the experiment was taken into account. The release of PS at the end of the hydroponic stage of the experiment (i.e. immediately before placing the plants in contact with the soil for the second stage of the experiment) increased two- and three-fold in response to Fe deficiency in ‘Songlen’ and ‘Aroona’, respectively (Fig. 3). In comparison, the acquisition of soil Cu increased three- and four-fold in response to Fe deficiency in ‘Songlen’ and ‘Aroona’, respectively (Table 3). It is noteworthy that the acquisition of Cu was only about 5% of total soil Cu for plants adequately supplied with Fe (and Zn), while it increased to about 20% for Fe-deficient plants (Table 3), which is considerable in the context of the fairly large total Cu content of the contaminated soil studied in the present experiment. In addition, the rate of PS release by Fe-deficient plants ranged between 0.1 µmol and 0.2 µmol per pot (for a 3-h period of measurement) prior and after the soil stage of the experiment (Fig. 3), suggesting that over the 8-d period of contact with the soil, the plants released at least c. 1–2 µmol PS per pot. This is an underestimate because it assumes that plants did not release any PS except during this peak period of the day, which does not hold true. In comparison, the amount of Cu acquired by the plants over this 8-d period was about 1 µmol Cu per pot (as deduced from Fig. 2). Meanwhile, the acquisition of Fe from soil was about 2–4 µmol Fe per pot and that of Zn was 0–0.5 µmol Zn per pot (as deduced from Table 4). Such results point to a potential role of PS release in the acquisition of not only Fe but also Cu from the calcareous, contaminated soil used in the present experiment. This is in line with the results of Treeby et al. (1989) who reported an increase in shoot Cu, but also Zn and Mn concentrations in Fe-deficient barley relative to Fe-sufficient barley. This was, however, found for an uncontaminated soil, which contained only 24 µg Cu g−1 soil, whereas the soil presently used contained 144 µg Cu g−1 soil, of which as much as about 20% was acquired by Fe-deficient wheat plants over the 8-d period of contact. As suggested by Treeby et al. (1989), PS do not only complex Fe; they can also complex other metals such as Cu, Zn and Mn, or Co and Ni, as mentioned by Mench and Fargues (1995). Murakami et al. (1989) measured the affinity of PS for Fe, various divalent metals and showed that they ranked as follows: Mn2+ < Fe2+ < Zn2+ < Ni2+ < Fe3+ < Cu2+ (Hinsinger, 2001). This suggests that the stability of the complex that PS form with divalent Cu is even greater than that of the complex formed with trivalent Fe. This provides a theoretical basis for the strong implication of PS in the acquisition of Cu by wheat, as suggested by the present results.
In the calcareous soil used in the present experiment, significant acquisition of soil Zn was only found for ‘Aroona’, which is the Zn-efficient genotype (tolerant to Zn deficiency), and Zn acquisition (Table 4) was well-positively correlated with PS release at the end of the hydroponic stage (Fig. 3a): both parameters increased about threefold in response to Fe-deficiency for the ‘Aroona’ genotype. This suggests that, as for soil Cu, there is some implication of PS in the acquisition of soil Zn by ‘Aroona’. If not related to the amounts of PS released, the greater efficiency of ‘Aroona’ in the use of soil Zn, compared with ‘Songlen’ (Table 4), might be related to a greater ability to take up PS–Zn complexes and make use of these as a source of Zn.
The present results do not provide proof of a direct causal relationship between the release of PS and the acquisition of soil Cu or Zn in wheat. They simply show that both the release of PS and the acquisition of Cu (and that of Zn in ‘Aroona’) are coincidentally enhanced under Fe deficiency. Previous works have shown that Fe deficiency can induce an increased uptake capacity (increased sink) not only of Fe, but also of other metals such as Cd, Zn and Cu (Cohen et al., 1998; Korshunova et al., 1999). This might occur as a consequence of an induction of the synthesis of metal transporters at the plasma membrane of root cells and of the poor selectivity of those transporters. In addition, although it has been used by many other authors, the method that we used for measuring the amount of PS released by wheat roots has several drawbacks. A major one is that it is not specific for PS (Gries et al., 1998): other complexing molecules among root exudates, such as organic and phenolic acids, amino acids (e.g. histidine) or proteins (e.g. nicotianamine), might also have contributed some proportion of the mobilization of Cu bound to the resin in the proposed method. Strictly speaking, it is thus more indicative of the complexing power of root exudates than of the amount of PS. A chemical analysis of the composition of root exudates would have helped clarify this point.