• The translocation of manganese (Mn), nickel (Ni), cobalt (Co), zinc (Zn) and cadmium (Cd) in white lupin (Lupinus albus cv. Amiga) was compared considering root-to-shoot transport, and redistribution in the root system and in the shoot, as well as the content at different stages of cluster roots and in other roots.
• To investigate the redistribution of these heavy metals, lupin plants were labelled via the root for 24 h with radionuclides and subsequently grown hydroponically for several weeks.
• 54Mn, 63Ni and 65Zn were transported via the xylem to the shoot. 63Ni and 65Zn were redistributed afterwards via the phloem from older to younger leaves, while 54Mn remained in the oldest leaves. A strong retention in the root was observed for 57Co and 109Cd.
• Cluster roots contained higher concentrations of all heavy metals than noncluster roots. Concentrations were generally higher at the beginning of cluster root development (juvenile and immature stages). Mature cluster roots also contained high levels of 54Mn and 57Co, but only reduced concentrations of 63Ni, 65Zn and 109Cd.
Plants have a natural ability to extract elements from soil and to distribute them between roots and shoot depending on the biological processes in which the element is involved (Ximénez-Embún et al., 2002). In addition to the uptake of nutrients, toxic compounds such as heavy metals can also be taken up by the plants. Heavy metals are defined as metals with a density > 5.0 g cm−3 (Seaward & Richardson, 1990). Some heavy metals also play a role in plant metabolism, and can be considered nutrients. This is the case for manganese (Mn), zinc (Zn) and nickel (Ni), which are involved in major functions (Welch, 1995). Manganese has been shown to play a role in enzyme activation, biological redox systems (e.g. electron transport reactions in photosynthesis) or detoxification of oxygen free radicals; while Zn is involved in membrane integrity, enzyme activation and gene expression. Nickel is needed for urea metabolism, iron absorption and nitrogen (N) fixation (Welch, 1995). There is no evidence that cobalt (Co) has any direct role in the metabolism of higher plants in general (Marschner, 1995), but it has been demonstrated to be required for N2 fixation in legumes and in root nodules of nonlegumes (Ahmed & Evans, 1960; Marschner, 1995). In contrast, some heavy metals such as cadmium (Cd) are not needed by the plant, but are also taken up (Römer et al., 2000, 2002; Ximénez-Embún et al., 2002; Zornoza et al., 2002). Both kinds of heavy metal, nutrients and pollutants, can accumulate in excess in the plant to levels undesirably high for human or animal nutrition, and may even become toxic to plants at a certain concentration (William et al., 2000). Thus uptake of heavy metals by the roots, their transport in the different parts of the root system, their release to the shoot and further redistribution within the shoot are important processes for (i) redistribution of heavy metals in the plant/soil system; (ii) supply of shoot parts with nutrients; and (iii) quality of harvested plant parts.
The transport and mobility of heavy metals have been studied in plants including bean and wheat. In bean, Bukovac & Wittwer (1957) determined the absorption, transport and mobility of radionuclides applied on the leaves. They observed that these radioactive isotopes had a different mobility in the phloem. Schmidke & Stephan (1995) concluded that the metal micronutrients Mn, Zn, Co and iron (Fe) were supplied by phloem transport, and that this transport was linked to the presence of nicotianamine. Takahashi et al. (2003) reported that nicotianamine is important for the regulation of metal transfer within cells, in addition to its role in long-distance transport. More information concerning the distribution of Zn, Mn, Ni, Co and Cd is available for wheat. For example, Zn can be transported rapidly in the phloem (Herren & Feller, 1994, 1996; Pearson et al., 1995; Haslett et al., 2001; Erenoglu et al., 2002; Page & Feller, 2005; Riesen & Feller, 2005), while Cd, which is chemically very similar to Zn (Chesworth, 1991) and thus may be transported in plants by similar pathways (Grant et al., 1998), has a mobility different from Zn in wheat (Page & Feller, 2005; Riesen & Feller, 2005). Varieties of crop plants may differ considerably in the symplastic redistribution of Zn (Hajiboland et al., 2001; Erenoglu et al., 2002). Comparing two rice varieties differing in Zn efficiency, the transport of Zn from older (source) to younger (sink) leaves has been found to be more rapid under Zn deficiency than under adequate Zn supply, and higher in the Zn-efficient than in the Zn-inefficient genotype (Hajiboland et al., 2001). Erenoglu et al. (2002) reported that differences in the expression of Zn efficiency were not related to the redistribution of foliar-applied Zn within wheat plants. Therefore caution is recommended when generalizing results obtained with a particular species or genotype. Manganese may be translocated as a free divalent cation in the xylem from roots to shoot (Marschner, 1995). On the other hand, redistribution of Mn in the phloem is very limited (El-Baz et al., 1990; Page & Feller, 2005; Riesen & Feller, 2005), and may depend on the plant species and developmental stage (Herren & Feller, 1994). The distribution of heavy metals may depend on the availability of other elements. An accumulation of Mn at the penetration position of Erysyphe graminis on wheat leaves was observed only in plants supplied with silicate, but not in comparable plants grown without silicate (Leusch & Buchenauer, 1988). These results suggest that, at least under certain conditions, silicon influences the redistribution of Mn in wheat. Nickel and Co were considered to have an intermediate phloem mobility (Marschner, 1995), but recent studies in wheat demonstrated that the mobility of Ni in the phloem is rather high compared with Co (Page & Feller, 2005; Riesen & Feller, 2005). In legumes, Ni and Mo are redistributed efficiently from vegetative plant parts to maturing seeds, while the accumulation of these two micronutrients in the ears of cereals is less pronounced (Horak, 1985).
In the studies cited above, it was demonstrated that heavy metals have a different mobility in plants. It remains an interesting challenge to elucidate in more detail the acquisition and redistribution of various heavy metals in a plant forming cluster roots (e.g. white lupin). Lupin differs in its root anatomy from cereals and other plants mentioned above. Four recent reports on Cd uptake by white lupin are available in the literature (Römer et al., 2000, 2002; Ximénez-Embún et al., 2002; Zornoza et al., 2002). Ximénez-Embún et al. (2002) investigated the tolerance, uptake and accumulation of Cd in white lupin grown on contaminated sand, and found that Cd concentration was significantly higher in roots than in shoots. Zornoza et al. (2002) demonstrated that lupin roots have a high capacity of Cd retention by cell walls and complexation by thiol groups. Römer et al. (2000, 2002) compared the mobilization and uptake of Cd by white lupin, blue lupin and ryegrass on two soil types. These authors observed that the Cd content in shoots was 28 times lower for white lupin than for ryegrass, while blue lupin had an intermediate Cd content in the shoot. They suggested that the higher secretion of carboxylates into the rhizosphere by lupins might modify Cd speciation in the soil solution, and that the consequence was a lower Cd uptake, observed in both types of soils. The secretion of carboxylates may contribute to the solubilization of iron and Mn in the soil and, as a consequence, lead to higher contents of these metals in lupin plants (Dinkelaker et al., 1995 and references therein).
Apart from the results mentioned above, only limited information is available on heavy metal translocation in white lupin, despite the fact that this plant is often used as a model for the study of nutrient uptake because of its capacity to produce cluster roots and acquire phosphorus (P) from soils where it is sparsely available (Purnell, 1960; Dinkelaker et al., 1995; Neumann & Martinoia, 2002; Lamont, 2003). These cluster roots are the main site of phosphate acquisition in the root system of white lupin, thanks to their large surface and their secretion activity. White lupin cluster roots secrete large amounts of carboxylates, mainly citrate and malate, as well as protons, phenolic compounds and phosphatases. In white lupin, the carboxylate secretion and acidification of the rhizosphere have been studied intensively for many years (Gardner et al., 1982a, 1982b, 1983; Johnson et al., 1994; Neumann et al., 1999, 2000; Massonneau et al., 2001; Penaloza et al., 2002; Shane et al., 2003), and it has been shown that these secretion processes are related to the developmental stages of cluster roots. Massonneau et al. (2001) defined four developmental stages of white lupin cluster roots. At the juvenile stage, roots are still growing and small amount of malate is secreted. Several days later, roots reach the immature stage, characterized by fully grown roots secreting reduced amounts of carboxylates and high quantities of phenolic compounds (Weisskopf et al., 2006). The mature stage is when most carboxylate secretion occurs and acidification takes place. At the senescent stage, secretion of carboxylates and phenolic compounds is much reduced. It is very likely that these induced changes, mediated by the massive secretion of protons and carboxylates into the rhizosphere of cluster roots and especially into the rhizosphere of mature cluster roots, will affect not only the availability of phosphate, but also the solubility and uptake of heavy metals (Dinkelaker et al., 1995). Moreover, white lupin is the only cluster-rooted species of agricultural importance. Because of its very efficient P-acquisition strategy, as well as its ability to fix N in symbiotic association with Bradyrhizobiae (Raza et al., 2001), white lupin is a very promising crop in a world of nonrenewable resources, as it is able to acquire both N and phosphate with reduced application of external fertilizers. Thus more precise knowledge about the heavy metal uptake and transfer to the plant parts used as fodder or human food would be very useful from the agronomic point of view.
The secretion of citrate from mature cluster roots is important for the solubilization of phosphate and heavy metals in soil. Such solubilization processes are not relevant in hydroponic culture. However, such an experimental setup can be used to test whether mature cluster roots are also characterized by high acquisition rates for solubilized heavy metals from the medium, and particularly by a higher retention of heavy metals in vacuoles of cluster roots, thus resulting in a low root-to-shoot translocation.
The aim of the present study was to characterize the distribution of five heavy metals in the roots and shoots of white lupin after labelling a single root in hydroponic culture. The acquisition of heavy metals in the different developmental stages of white lupin cluster roots was also analysed. We investigated three essential heavy metals (Mn, Ni and Zn); the conditionally (for N2 fixation) required Co; and a nonessential heavy metal, Cd, which is a pollutant found in many contaminated soils (Sauvéet al., 2000; Lugon-Moulin et al., 2004). White lupin was chosen for this study because of the well known physiology of its roots and the potential agronomic application of this research.
Materials and Methods
Transport of heavy metals in white lupins
Seeds of white lupin (Lupinus albus L. cv. Amiga, Suedwestdeutsche Saatzucht, Rastatt, Germany) were washed in 1% NaOCl for 15 min, rinsed three times with deionized water, incubated overnight in aerated deionized water, and germinated in the dark (for 4 d) and in the light (for 1 d) in a quadratic Petri dish prepared with five sheets of filter paper (Whatman 3MM) moistened with 0.2 mm CaSO4. Lupin seeds were placed in between the filter papers (four papers below and one above the seeds). The Petri dish was arranged with a slope of 30°. After 5 d germination the root was placed for 24 h on a 10-ml tube containing 10 ml nutrient solution with radioactive heavy metals. The nutrient solution used in these experiments contained 1.5 mm KH2PO4, 0.76 mm MgSO4, 0.34 mm Ca(NO3)2, 0.22 mm KNO3, 7.5 µm Fe (added as Sequestren), 0.25 µm MnCl2, 1.23 µm H3BO3, 0.04 µm ZnSO4, 0.05 µm Na2MoO4, 0.012 µm Ni(NO3)2, 0.025 µm CuSO4 (Hildbrand et al., 1994, four times diluted). For preparation of the nutrient solution without P, KH2PO4 was eliminated from the nutrient solution. To investigate the distribution of radioactive heavy metals in whole plants by gamma or beta spectrometry, the plants were labelled with nutrient solution containing a mixture of 54Mn (6.0 kBq l−1), 63Ni (739.4 kBq l−1) and 57Co (2.9 kBq l−1) (first set of plants); or a mixture of 65Zn (24.4 kBq l−1) and 109Cd (79.2 kBq l−1) (second set of plants). With gamma spectrometry it is possible to measure more than one isotope at the same time: 65Zn and 109Cd, which have different gamma energies, can be measured at the same time. To diminish the number of samples, 63Ni (beta energy), which is not detected by the gamma counter, was added to the couple 54Mn and 57Co (which also have different gamma energies). Afterwards, 63Ni was detected using a beta counter. For autoradiography (separate experiments), plants were labelled separately either with 54Mn (274.1 kBq l−1), 63Ni (261.1 kBq l−1), 57Co (274.1 kBq l−1), 65Zn (274.1 kBq l−1) or 109Cd (274.1 kBq l−1). After labelling, roots were washed (dipped three times sequentially in 100 ml nutrient solution) to remove radioactive solutes from the root surface, and placed for 2 h in nutrient solution with 0.1% Congo red to allow identification of the root part initially labelled with heavy metals after further elongation of the root. Then the roots were washed three times to remove excessive dye and incubated for 1 h on nutrient solution. Thereafter the seedlings were transferred to aerated nutrient solution. Four labelled plants and two identically treated but initially unlabelled control plants were placed together on pots with 1 l nutrient solution. The four replicate plants from the same pot were sampled and analysed separately. Plants were grown at a temperature of 21°C during the night and 25°C during the day, 65% humidity. The photoperiod was 14 h light (100 µmol photons m−2 s−1 from four Philips TLD 36 W/25 and two Osram Fluora L 36 W/77 fluorescent lamps, measured 20 cm above the culture pot); 10 h night. The nutrient solution was changed every week.
For analysis of the distribution of heavy metals in white lupins labelled with radionuclides at the seedling stage (after 5 d germination), plants were harvested at different time points after the labelling phase (0, 1, 4, 8, 12, 20 and 28 d after labelling, four plants per time point). Then the plants were dissected into the labelled part of the main root (stained red); lateral roots outgrowing from the labelled part of the main root; the apical part of the main root including its lateral roots; the hypocotyl; the cotyledons; the stem; and leaf 1 (oldest) to leaf 12 (youngest). The different plant parts were dried at room temperature and then weighed. Plant parts were analysed simultaneously at the end of the experiment. The radioactivity of 54Mn, 57Co, 65Zn and 109Cd was detected in a gamma counter (1480 Wizard 3′, Wallac Oy, Turku, Finland). For 63Ni measurement, plant parts were ashed afterwards at 550°C for 8 h. After cooling, the ash was dissolved in 1 ml 20 mm citric acid, mixed, and 200 µl were transferred to Ready Caps· (Beckman Instruments, Fullerton, CA, USA) and dried at < 65°C for 4 h. The radioactivity of 63Ni was measured in a liquid scintillation counter (beta counter, Betamatic V, Kontron Instruments, Zurich, Switzerland). Interferences from other radionuclides presented in the samples (54Mn and 57Co) were corrected. For autoradiography, the harvested plants were placed on paper, protected with baking paper and three sheets of typewriter paper, and dried by placing a metal plate heated to 220°C on the top for 30 s. At the end of the experiment, all plants were exposed simultaneously to X-ray film (Fuji medical X-ray film, super RX) for 3 months (54Mn, 57Co, 65Zn and 109Cd); for 63Ni a longer exposure time of 6 months was needed.
The dry matter values for the different plant parts are shown for one set of plants (labelled with 54Mn, 63Ni and 57Co) in Fig. 1. These values were very similar to the other set of plants (data not shown). Comparing the dry matter values of plant parts and their radionuclide content allows us to distinguish the absence of heavy metals because an organ was not yet present; and the absence of heavy metals because they were not transported into an existing organ.
To investigate the transport of heavy metals in older plants, 6-wk-old plants (germinated and grown following the method mentioned above, but not yet labelled with radioactive heavy metals) were placed for 24 h on radiolabelled nutrient solution containing a mixture of 54Mn (0.17 kBq l−1), 63Ni (0.016 kBq l−1) and 57Co (0.61 kBq l−1). One plant was placed on each pot with 1 l radiolabelled nutrient solution. Four plants were collected immediately after the labelling phase (day 0), and four other plants were transferred to nonlabelled nutrient solution for seven other days (day 7) and harvested afterwards. For this experiment, plants were dissected into root system, hypocotyl, stem and leaf 1 (oldest) to leaf 30 (youngest). On plants of this age, cotyledons and leaves 1–3 were often missing as a consequence of senescence and abscission. At the end of the experiment, the radioactivity present in the plant parts was analysed using a gamma or a beta counter as described above.
Heavy metal contents in cortex, vascular cylinder and different cluster root stages of 6-wk-old plants
Plants were germinated following the method mentioned above, but not yet labelled with radioactive heavy metals. Then plants were grown on nutrient solution without phosphate to induce the formation of cluster roots. One plant was placed on each pot with 1 l nutrient solution. After 5 wk, cluster roots appeared in the roots. When the plants were 6 wk old they were placed on nutrient solution without phosphate containing a mixture of 54Mn (0.16 kBq l−1), 63Ni (0.46 kBq l−1) and 57Co (0.70 kBq l−1); or a mixture of 65Zn (0.017 kBq l−1) and 109Cd (0.13 kBq l−1) for labelling. One plant was placed on each pot with 1 l radiolabelled nutrient solution. Four plants were labelled with 54Mn, 57Co and 63Ni, while four other plants were labelled with 109Cd and 65Zn. After labelling for 24 h, the root system was dipped three times sequentially in 1 l nutrient solution to remove radioactive solutes from the root surface, then immersed in a solution of bromocresol purple (0.04% w/v) to differentiate the developmental stages of cluster roots. The pH of the solution was adjusted to 6.5 so that both acidification (yellow) and alkalinization (purple) could be observed. After 20 min contact with the root system, colour changes became visible. The root system was dissected into different parts, such as apex and juvenile, immature, mature, senescent and noncluster roots according to the different cluster root developmental stages described by Massonneau et al. (2001). The shoot was also collected. A small section (1 cm) of the main root was collected (immediately above the first lateral roots) and dissected afterwards. After a longitudinal cut, the vascular cylinder and cortex could easily be separated with surgical tools. The plant parts were weighed and radioactivity was counted using a gamma or a beta counter.
Solubility of radionuclides accumulated in roots
The solubility of heavy metals within plant tissues was addressed by a sequential extraction experiment. Lupin plants were germinated and seedlings (5 d old) were labelled for 24 h with a mixture of 54Mn (0.29 kBq l−1) and 57Co (1.33 kBq l−1), or with 109Cd (0.12 kBq l−1) alone. After labelling with radionuclides, roots were dipped three times sequentially in 100 ml nutrient solution to remove the radioactive solutes from their surface, and incubated 10 min in nutrient solution. Then plants were collected and a 1-cm section was cut in the main root and dissected into cortex and vascular cylinder. The two root parts were placed overnight in water for initial counting in a gamma counter. Then a sequential extraction with two different buffer systems was performed to test solubility at pH 5.4 (20 mm sodium acetate buffer) and at pH 8 (20 mm Tris–HCl). In a first step, plant parts were placed in tubes containing the buffer, homogenized with a Polytron mixer (20 s middle speed; 5 s high speed) and incubated for at least 30 min. After centrifugation (5 min, 16 000 g), pellets were extracted in the original buffer supplemented with 500 mm NaCl. After centrifugation, pellets were extracted again in the original buffer supplemented with 10 mm Na2EDTA. The supernatants of each step and the pellets of the final step (pellet 3) were analysed by gamma spectrometry. The radionuclide content in the water initially used was added to that of supernatant 1 (first column in Table 1). Three replicates were analysed for each buffer and each radionuclide.
Table 1. Solubility of heavy metals in cortex and vascular cylinder of white lupin (Lupinus albus)
Buffer Supernatant 1
Buffer + 500 mm NaCl Supernatant 2
Buffer + 10 mm Na2EDTA
A 1-cm segment of the main root of white lupins was collected and dissected into cortex and vascular cylinder. Solubility was tested at pH 5.4 (20 mm sodium acetate buffer) and pH 8 (20 mm Tris–HCl buffer). In the first step, the heavy metals were extracted in the original buffer; after centrifugation, pellets were extracted in the original buffer supplemented with 500 mm NaCl. After centrifugation, pellets were extracted again in the original buffer supplemented with 10 mm Na2EDTA. Supernatants of each step and pellets of the final step (pellet 3) were analysed by gamma spectrometry. Means and standard errors of three replicates are shown for relative radionuclide contents in the percentage of total label in the sample (100% = supernatant 1 + supernatant 2 + supernatant 3 + pellet 3).
Sodium acetate buffer (20 mm, pH 5.4)
45 ± 23
44 ± 19
3 ± 3
8 ± 5
94 ± 1
2 ± 2
2 ± 2
2 ± 2
69 ± 9
16 ± 3
4 ± 2
11 ± 5
43 ± 10
12 ± 7
21 ± 7
23 ± 5
40 ± 5
36 ± 15
14 ± 7
10 ± 9
38 ± 18
19 ± 9
34 ± 17
9 ± 9
Tris–HCl buffer (20 mm, pH 8)
41 ± 9
54 ± 5
2 ± 2
3 ± 2
55 ± 9
24 ± 17
21 ± 14
0 ± 0
69 ± 11
7 ± 4
5 ± 4
19 ± 4
44 ± 5
34 ± 17
6 ± 3
16 ± 12
45 ± 4
33 ± 7
9 ± 5
13 ± 9
56 ± 23
24 ± 24
20 ± 18
0 ± 0
Values of heavy metal content are means of three or four replicates. Differences in heavy metal content between cortex and vascular tissues were tested for statistical significance using Student's t-test (P < 0.05, n = 3 or 4). For the different parts of cluster roots, anova was performed and least significant differences were calculated (statistix for Windows ver. 1.0, Analytical Software, Tallahassee FL, USA; rejection level 0.05, n = 3 or 4).
Time courses for distribution of heavy metals in whole plants
The time courses for 54Mn, 63Ni, 57Co, 65Zn and 109Cd differed in whole plants labelled via the main root at the seedling stage (Fig. 1). The 54Mn and 63Ni contents decreased rapidly in the labelled part of the main root, while 65Zn, 109Cd and 57Co were released more slowly. At the end of the experiment (day 28), only small amounts of 54Mn (6.5% of total) and 63Ni (18.6%) could still be found in the initially labelled part of the root, while higher percentages were found for 109Cd (75%), 57Co (55%) and 65Zn (33%).
A release from the labelled root zone to other parts of the root system occurred for some heavy metals (Fig. 1). For instance, 15% of the total content of 57Co was found in the lateral roots of the labelled part of the main root. In contrast, 54Mn and 63Ni were transferred more rapidly to the shoot than 57Co. 54Mn transiently accumulated in the hypocotyl (day 1) before moving into leaves 1–3 (day 4). A small quantity of 54Mn also reached leaves 4–6. Afterwards, no major changes in 54Mn distribution were detected until the final harvest (day 28). The transient accumulation of 63Ni in the hypocotyl was less pronounced than for 54Mn, as well as accumulation in leaves 1–3. But in contrast to 54Mn, the content of 63Ni diminished in leaves 1–3 to reach leaves 4–6, and later also leaves 7–9 and 10–12. This indicates that 63Ni was retranslocated from the older leaves to newly expanding leaves. Similar translocation properties were observed for 65Zn, which decreased in the oldest leaves and increased in the upper (youngest) leaves.
The three radionuclides 63Ni (rapidly released from roots to shoot and retranslocated from older to younger leaves); 54Mn (rapidly released from roots to shoot but not retranslocated within the shoot); and 57Co (retained in roots) behaved very differently in young plants (Fig. 1). The transport of these radionuclides was also investigated in older plants with a more complex root system. Figure 2 shows the release of 54Mn, 63Ni and 57Co from roots to shoot in 6-wk-old lupin plants. Immediately after the labelling phase (day 0), 54Mn was present in a large amount in the root system, hypocotyl and stem. Seven days later (day 7) almost all 54Mn had moved to the youngest fully expanded leaves (leaves 16–18) as well as to the other leaves, to a lesser extent. A similar pattern was observed for 63Ni, but with slower transfer from roots to shoot (after 1 wk 36% remained in roots, compared with 3% for 54Mn) and with a more even distribution between the different leaves. Finally, 57Co showed extreme behaviour with practically no transfer to the shoot.
Distribution of heavy metals in whole plants of lupin as revealed by autoradiography
Autoradiography is a suitable technique for visualizing and localizing in detail the labelled heavy metals in different plant parts. Figure 3 shows autoradiographs of lupin seedlings 4 d after labelling (for 24 h) via the main root with 54Mn, 63Ni, 57Co, 65Zn or 109Cd. These pictures clearly show the different distribution of the five radionuclides analysed in whole lupin plants. While 54Mn and 63Ni were already transferred to the shoot, most of 57Co, 65Zn and 109Cd were still found in the root system 4 d after labelling. This method allowed us to determine more precisely where heavy metals were stored in the leaves. Interestingly, 54Mn moved to the shoot and accumulated at the very periphery of the oldest leaves (Fig. 3a, d). Autoradiography was also prepared for roots and shoot of older plants labelled with 54Mn, 63Ni, 57Co, 65Zn and 109Cd. The distribution of 54Mn, 57Co, 65Zn and 109Cd in roots and shoot of 4-wk-old white lupins (20 d after labelling phase) is shown in Fig. 4. For 63Ni, the signal was too low and the distribution could not be seen. Plants were labelled simultaneously and in the same stage as those used for the quantitative analyses (Fig. 1) and autoradiography analyses (Fig. 3). In agreement with the quantitative results presented in Fig. 1, at day 20 54Mn was mainly present in leaves 1 and 2 (the oldest leaves) and in smaller quantities in leaves 3 and 4, while no signal was detected in the youngest leaves (leaves 6–8). Again, 54Mn accumulated at the periphery of the oldest leaves. 57Co was also present in larger amounts in the oldest leaves (1 and 2) but, in contrast to 54Mn, 57Co was distributed more homogeneously in the leaf lamina and also reached the youngest leaves. The autoradiograph of lupin labelled with 65Zn showed that this heavy metal was present in higher quantities in the youngest leaves (leaves 5–8). There was a strong accumulation of 65Zn in the newly formed (expanding) leaf 8. In contrast to the three other radionuclides, the small amount of 109Cd that reached the shoot was distributed evenly among the different leaves. Moreover, the autoradiographs showed that 109Cd was concentrated in the central vein of the leaflets.
Distribution of heavy metals in cluster roots of white lupin
To investigate the accumulation of heavy metals in the root system more precisely, the distribution of heavy metals in the different stages of cluster roots was analysed. Figure 5 shows the partitioning of 54Mn, 63Ni, 57Co, 65Zn and 109Cd in the different cluster root stages, as percentages of radionuclides per plant part (Fig. 5a); and the content of these heavy metals in the different cluster root stages, measured in counts per minute per root FW (Fig. 5b). The content of heavy metals in different parts of the root system varied. For all heavy metals except 63Ni, the noncluster roots contained the highest part of the heavy metals absorbed, because of their dominance in the root system (Fig. 5a). While the contribution of cluster roots to heavy metal sequestration was very minor for 54Mn, it was more significant for 63Ni, 57Co, 65Zn and 109Cd. Most of the heavy metals were found in senescent cluster roots, with a minor contribution of mature cluster roots in the case of 63Ni and 109Cd. When taking into account the differences in abundance of each root type within the entire root system, and thus considering the amounts of heavy metals per FW root, the situation was quite different (Fig. 5b): for all radionuclides, cluster roots showed a much higher content of heavy metals than noncluster roots. Highest heavy metal concentrations were found in young and immature cluster roots, while senescent cluster roots behaved like noncluster roots and accumulated only small amounts of heavy metals. Mature cluster roots did not accumulate all heavy metals tested in the same way: while 54Mn and 57Co were present in high amounts (similar to the levels observed in young and immature cluster roots), 63Ni, 65Zn and 109Cd levels in mature cluster roots were not significantly higher than in senescent or noncluster roots.
Localization of heavy metals within roots
The results presented above and in Figs 1–4 indicated that a high percentage of some radionuclides was retained for several weeks in the labelled part of the main root (57Co, 109Cd), while other radionuclides were released more rapidly to the shoot (54Mn, 63Ni). The question of localization of heavy metals staying in the root at the tissue level was addressed by separating root cortex from vascular tissue in a 1-cm segment of the main root, segment collected immediately above the first lateral roots (Fig. 6). More than 80% of 57Co and of 109Cd was present in the cortex, while 54Mn and 65Zn were present in similar quantities in the two parts of the root. A high percentage (> 80%) of 63Ni was detected in the vascular cylinder.
Solubility of radionuclides
As the solubility of heavy metals is a key factor in their movement and distribution within plant tissues, this aspect was also analysed. Three heavy metals were tested: Mn (rapidly released to the first leaves); and Cd and Co (retained in the main root). In general, a high percentage of all radionuclides was soluble in the two buffers used (Table 1). Most of the heavy metals were present in the supernatants after the first and second steps of solubilization. Surprisingly, 109Cd was also very soluble in both buffers and < 15% was present in the final pellet. 57Co showed the strongest insolubility, as revealed by the highest percentage in the last pellet, after the three subsequent extraction steps. These measurements refer to the solubility of the heavy metals regardless of their subcellular distribution in intact roots, as compartmentation was completely destroyed during extraction. Accumulation in certain compartments may be important for the low root-to-shoot translocation of 57Co and 109Cd.
Distribution of heavy metals in white lupin
In this study, the redistribution of five heavy metals in the roots and shoot of white lupin was characterized. 54Mn, 63Ni, 57Co, 65Zn and 109Cd behaved differently with regard to root-to-shoot transfer and redistribution within the shoot and root system. 65Zn and 63Ni were probably transported via the xylem to transpiring leaves and then retranslocated to newly formed leaves. These findings suggest that 65Zn and 63Ni were transported via the phloem from the older leaves to younger, expanding leaves. As Zn and Ni are micronutrients (Welch, 1995), it makes sense that they move from roots to shoot in the xylem to reach the leaves. The rapid retranslocation of 65Zn and 63Ni from old to young leaves via the phloem contributes to the supply of rapidly growing plant parts with these micronutrients. In contrast to Zn and Ni, Mn (which is also a micronutrient; Welch, 1995) was transported only from the labelled part of the main root to the first leaves, and accumulated in these leaves. This distribution pattern suggests that 54Mn is rapidly released from roots into the xylem and reaches photosynthetically active leaves via the transpiration stream, and that this radionuclide is not (or only in minor quantities) redistributed afterwards via the phloem to other leaves. This is in agreement with experiments performed with other plant species, such as maize and horse bean (El-Baz et al., 1990); and wheat (Page & Feller, 2005; Riesen & Feller, 2005). Solubilization of Mn in the soil by the secretion of carboxylates from cluster roots, and the rapid translocation of Mn from roots to shoot, may cause an accumulation of Mn in lupin leaves under certain conditions (Dinkelaker et al., 1995). The low mobility of 54Mn within the shoot may be explained by a restricted loading of soluble Mn into the phloem, or by insolubilization in the leaves. These possibilities remain to be investigated in future studies. It must be also considered whether Mn may be translocated to some extent from stems and petioles via the phloem to maturing seeds (Hannam et al., 1985).
Cadmium showed a specific behaviour in the sense that most of it stayed in the root system. This is in agreement with the results of Römer et al. (2000, 2002); Ximénez-Embún et al. (2002); Zornoza et al. (2002). One can imagine that Cd, a nonessential and pollutant heavy metal found in different concentrations in soils (Sauvéet al., 2000; Lugon-Moulin et al., 2004), is recognized as a toxic compound at the root level and, as it is not needed in the shoot, the plant sequesters it in the roots to avoid damage to the shoot. Cadmium is known to have several toxic effects on plants (Sanità di Toppi & Gabbrielli, 1999 and references therein). Cadmium sequestration in the cell wall, as already shown by Zornoza et al. (2002), or in the vacuoles of root cells may explain the high quantity of this radionuclide found in the root throughout this experiment. As judged from the solubility of 109Cd in extracts of cortex and vascular cylinder (Table 1), it can be supposed that 109Cd is sequestered in a soluble form (e.g. in the vacuole of root cells). The root-to-shoot long-distance transport of 109Cd in white lupin was not similar to that observed in wheat (Page & Feller, 2005): 109Cd was shown to be transported from roots to shoot in wheat, and was also found in the grain (Herren & Feller, 1997; Riesen & Feller, 2005). By contrast, as 109Cd stayed in the root system of lupin and, as reported by Grant et al. (1998), restricting Cd transport from roots to shoot reduces the Cd concentration in grains much more than in leaves, lupin seeds should contain only very low levels of Cd, which is an advantage for seed consumption.
The behaviour of 57Co was very similar to that of 109Cd. Cobalt is also known to have toxic effects in plants (Palit et al., 1994), and it is suggested that Co is also recognized as a toxic compound at the root level and sequestered in roots to avoid damage to the shoot. Moreover, Co is required by legumes and nodulating nonlegumes for N2 fixation (Palit et al., 1994; Marschner, 1995). Even if, in our experiments, N fixation did not take place because of a sufficient nitrate supply, Co may be recognized by lupin as a micronutrient useful in the roots, where N fixation is normally taking place, and thus stay in the root system (labelled part and distribution in lateral roots and apical part as shown in Fig. 1) rather than being translocated to shoots as 54Mn, 65Zn or 63Ni.
The results obtained with autoradiographs confirmed the results of gamma and beta counting (Figs 3, 4) and provided more detail on the levels of heavy metals in the different parts of individual leaves. 54Mn accumulated strongly at the periphery of older leaves (fully expanded with a high transpiration rate). Interestingly, a similar pattern of heavy metals at the edge of the leaf was observed for 109Cd in Thlaspi caerulescens (Cosio et al., 2005). These authors asked whether 109Cd allocation at the edge of older leaves might be explained by transport of 109Cd with the transpiration stream and excretion of 109Cd in excess with guttation. The guttation fluid may serve to excrete various elements, such as potassium, magnesium and calcium, as shown in sunflower (Tanner & Beevers, 2001). The pattern of guttation drops can be found on the entire leaf surface, as in tobacco, potato and bean; or at the edge or tips of the leaf, as in mustard, barley and cucumber (Kormarnytsky et al., 2000). It is therefore possible that guttation could be a way for white lupin to excrete excess Mn, but this question requires further investigation. In contrast to the findings of Cosio et al. (2005) in T. caerulescens, no Cd accumulation was observed in our experiments at the periphery of the leaves. Very little 109Cd was detected in the shoot compared with the root. Nevertheless, this small quantity of 109Cd was present in all parts of the leaf area. This homogeneous distribution may be explained by the uptake of 109Cd into mesophyll or epidermal cells. A large amount of 109Cd was also detected in the major leaf veins (leaf 4, Fig. 4g). It remains open whether the 109Cd detected in the veins was loaded into the phloem or was present in other tissues.
Distribution of heavy metals in cluster and noncluster roots
Considering the root system as a whole, because of the large majority of noncluster roots over cluster roots at the time of the experiment, noncluster roots had the largest contribution to heavy metal root content (Fig. 5a). When looking at the specific acquisition rates of heavy metals by the different root types and stages, it was clear that cluster roots were more efficient than noncluster roots in terms of heavy metal accumulation per FW unit (Fig. 5b). This may be caused by the very densely branched and ramified structure of cluster roots, and the consequently higher uptake surface. Secretion of carboxylates and protons did not appear to be correlated with heavy metal uptake, as mature cluster roots, where secretion of the highest amounts of carboxylates as well as acidification occur, did not show higher uptake than the other stages of cluster roots. This is consistent with previous results of Römer et al. (2002), who showed that plant uptake was not enhanced despite a higher excretion of citrate in P deficiency causing an increased Cd concentration in the soil solution. This suggests that, even if citrate and acidification affect Cd solubility, some mechanisms must enable the plant to limit root uptake of this toxic heavy metal. Interestingly, high heavy metal acquisition coincided with the stages where the highest secretion of phenolics takes place (juvenile and immature stages; Weisskopf et al., 2006). Previous reports have studied the impact of phenolics on heavy metal tolerance: for aluminium in maize (Kidd et al., 2001); and copper in alfalfa (Parry & Edwards, 1994) and white lupin (Jung et al., 2003). However, little information is available about the possible role of phenolics in tolerance to the heavy metals used in our study.
The results reported here were obtained from hydroponic cultures allowing direct access to the various parts of the root system. However, this system differs considerably from soil cultures (high phosphate concentration in the nutrient solution; no solubilization of heavy metals from soil particles). Root activities may affect soil properties and the solubility of heavy metals locally. The release of chelators, carboxylates or protons, as well as an increased reductive potential, may affect the solubilization and acquisition of these elements (Moraghan, 1979; Dinkelaker et al., 1995). Strong gradients can be generated by the local activities of cluster roots in soil, but not in hydroponic culture.
It is possible that a reallocation of the heavy metals within the different stages of cluster roots and within the different plant parts occurred during the time of treatment (24 h). The distribution of heavy metals in old lupin plants collected immediately after a 24-h labelling period support this assumption (Fig. 2). These results, showing the presence of 54Mn and 63Ni in the shoot, lead to the conclusion that some of these heavy metals might be reallocated within the labelling phase.
Localization and solubility of heavy metals in main root
After a 24-h labelling period, a high percentage of 63Ni was present in the vascular cylinder of the main root, while < 20% was in the cortex (Fig. 6). This distribution is in agreement with the good mobility of 63Ni in xylem and phloem. In contrast to 63Ni, 80% of 57Co and 109Cd was located in the cortex. The poor transfer of these radionuclides to the vascular cylinder may partially explain the very low transfer to the shoot. These heavy metals might be insolubilized (e.g. in cell wall components) or accumulated in soluble form in special compartments (e.g. vacuoles). The solubility of 57Co and 109Cd in extracts of cortex and vascular cylinder (Table 1) shows that most of these heavy metals were present in the supernatants after the first and second steps of solubilization, indicating a good solubility of these elements. Thus it appears likely that the compartmentalization of soluble 57Co and 109Cd, and not the formation of insoluble forms, was the primary cause for retention in the root system. The root-to-shoot translocation of Cd can be more rapid in other plants species (Page & Feller, 2005). Therefore caution is recommended when generalizing these results obtained with white lupin.
In conclusion, the study presented here showed that the five heavy metals tested differed in their distribution patterns and mobility in white lupin. While most of the 57Co and 109Cd remained in the root system, the other radionuclides were transported from roots to shoot. 63Ni and 65Zn were redistributed via the phloem from the oldest to the youngest leaves. 54Mn accumulated in the first leaves, especially in the edge of the leaflets. Regarding the acquisition of heavy metals by the different stages of cluster roots, the levels were highest in young and immature cluster roots, where large amounts of phenolics are secreted; mature cluster roots accumulated less heavy metal, despite the high secretion of carboxylates and protons. Our results showed that white lupin may absorb heavy metals in the roots and transfer some of them to the shoot. This knowledge may be useful in agriculture in the context of the quality of harvested plant parts. As most Cd stays in the root, we can assume that, in general, it will not accumulate to high levels in the seeds – the part of white lupin used for animal and human nutrition. Hydroponic culture proved a useful system for precise monitoring of heavy metal distribution within the plant. However, the situation may be quite different when studying soil-grown plants, where heavy metals are attached to soil particles and might be less available than in a nutrient solution. A study in microcosms with soil-grown plants would help to come closer to natural conditions and would constitute the next step towards a better understanding of the real field situation.
This project was founded by the National Centre of Competence in Research (NCCR) Plant Survival research programme of the Swiss National Science Foundation.