• Localization of cadmium (Cd) and other elements was studied in the leaves of the field-collected cadmium/zinc (Cd/Zn) hyperaccumulator Thlaspi praecox from an area polluted with heavy metals near a lead mine and smelter in Slovenia, using micro-PIXE (proton-induced X-ray emission).
• The samples were prepared using cryofixation. Quantitative elemental maps and average concentrations in whole-leaf cross-sections and selected tissues were obtained.
• Cd was preferentially localized in the lower epidermis (820 µg g−1 DW), vascular bundles and upper epidermis, whereas about twice the lower concentrations were found in the mesophyll.
• Taking into account the large volume of the mesophyll compared with the epidermis, the mesophyll is indicated as a relatively large pool of Cd, possibly involved in Cd detoxification/dilution at the tissue and cellular level.
Within the genus Thlaspi there are 23 species known to hyperaccumulate nickel (Ni), 10 that hyperaccumulate Zn (Brooks, 1998), while only three species (T. caerulescens, T. goesingense and T. praecox) hyperaccumulate Cd (Lombi et al., 2000; Vogel-Mikušet al., 2005) and one hyperaccumulates lead (Pb) (Reeves & Brooks, 1983). Most studies on metal uptake, transport and accumulation in hyperaccumulating plants were done on the model Zn/Cd hyperaccumulator T. caerulescens (Assunção et al., 2003). The ability of ecotypes to hyperaccumulate Cd differed significantly and was in most cases attributed to Zn transporters, whereas in the Ganges ecotype it was mainly attributed to the expression of high-affinity Cd transporters in the roots (Lombi et al., 2001). The recently discovered Cd/Zn hyperaccumulator Thlaspi praecox Wulfen (Brassicaceae) in the vicinity of a lead mine and smelter in Slovenia has been reported to accumulate up to 1.5% Zn, 0.6% Cd and 0.4% Pb (dry weight) at the most polluted locations (Vogel-Mikušet al., 2005), depending on the metal concentrations in soil and on the developmental cycle (Pongrac et al., 2007). This species was earlier reported as a Zn hyperaccumulator (Reeves & Brooks, 1983). It hyperaccumulates Cd in seeds, where it is primarily located in the epidermis of cotyledons of the embryo, as demonstrated using proton-induced X-ray emission (PIXE) (Vogel-Mikušet al., 2007). To date, no information concerning the physiology of metal (e.g. Cd) hyperaccumulation and tolerance in this newly discovered Cd-hyperaccumulating species is available.
Studies on the distribution of metals within plant tissues are necessary for understanding the physiology of metal hyperaccumulation and tolerance. Various methods offer different degrees of complexity and sensitivity, as well as spatial resolution. Qualitative evaluation of Cd distribution can be made rather easily using histochemical methods (Seregin & Ivanov, 1997; Baranowska-Morek & Wierzbicka, 2004; Pielichowska & Wierzbicka, 2004; Szarek-ukaszewska et al., 2004) or autoradiography (Cosio et al., 2005). Short-term desorption with radiotracers has limited spatial resolution, appropriate to identify compartments, but is not adequate for the precise location of metals (Lasat et al., 1996). Quantitative studies of the distribution of Cd are more difficult. Spatial localization of metals within biological tissues is most often done using microanalytical techniques based on the emission of characteristic X-rays, excited by charged particles (electrons, protons or heavier particles) or photons (micro-XRF, synchrotron radiation). Cd analyses have so far been restricted to plants kept in enriched substrates or hydroponic solutions (Khan et al., 1984; Vázquez et al., 1992a,b; Frey et al., 2000a; Küpper et al., 2000; Cosio et al., 2005; Wójcik et al., 2005). Quantitative results from scanning electron microscopy coupled with energy-dispersive X-ray microanalysis (SEM-EDX and STEM-EDX) were in most cases reported from point analysis. Cd images were most often qualitative, obtained by displaying the number of counts integrated over the energy window corresponding to the strongest Lα line of this element. Caution is required in interpreting these images since this line overlaps with the tail of the potassium Kα line. Other L lines of cadmium overlap with the Kα lines of potassium (K) and calcium, typically high-intensity lines in naturally growing plants. Paradoxically, losses of K during chemical fixation make possible Cd analysis with a smaller risk of overlap between the Cd-L X-ray lines and the Kα X-ray lines of this element. Such studies are, however, restricted to Cd and no comparison of its distribution with other elements is possible (Wójcik et al., 2005).
Synchrotron-based micro-X-ray fluorescence (µ-XRF) is a unique technique for in situ detection of chemical forms of Cd by micro-X-ray absorption near edge structure spectroscopy (µ-XANES). Using monochromatic radiation of energy below the absorption edge of K (3550 eV) avoids problems related to overlaps between X-ray lines of K and Cd (Isaure et al., 2006). A high-energy synchrotron radiation X-ray microbeam makes possible Cd mapping by using K-lines free of elemental overlaps (Hokura et al., 2006).
Proton-induced X-ray emission (PIXE) shows many similarities with SEM-EDX and has superior sensitivity (Legge & Cholewa, 1994). To date, reports on the use of micro-PIXE in studies of Cd localization and quantification in plants are scarce. In addition to its localization in seeds and seedlings of T. praecox (Vogel-Mikušet al., 2007), analyses done on the surface of Arabidopsis thaliana leaves showed intense Cd accumulation in trichomes (Ager et al., 2003), and on root cross-sections of Brassica juncea, demonstrating a clear decrease of this element from the outer cortex toward the stele (Schneider et al., 1999). In a study on biofiltering of potentially toxic elements by mycorrhizal fungi, high Cd enrichment was found in spores of Glomus intraradices cultivated with Pisum sativum (Przybyłowicz et al., 2004).
The aim of this study was to determine the sites of preferential accumulation of Cd and other mineral nutrients within leaf cross-sections of the recently discovered Cd/Zn hyperaccumulator T. praecox growing in an area polluted by heavy metals in the vicinity of a lead mine and smelter in Slovenia (Vogel-Mikušet al., 2005).
Materials and Methods
Sample collection and preparation
Plants of Thlaspi praecox Wulf. were collected in September 2004 (autumn, close to the end of the vegetation season) in an area polluted by Cd, Zn and Pb in Slovenia, at the plot corresponding to P3 from our previous studies (Regvar et al., 2006; Pongrac et al., 2007). The substrate was developing rendzina with partly decomposed organic matter and humus. The rhizosphere soil was highly enriched in total Cd, Zn and Pb (84.3 ± 7.6, 1249 ± 141 and 16 225 ± 1865 µg g−1, respectively). It contained 1.9 ± 0.3 µµg g−1 of available P, 18.4 ± 2.3% of organic matter and was of neutral pH (6.9 ± 0.1). Available metal fractions represented, on average, 28% of total soil Cd, 6% of total soil Zn and 10% of total soil Pb concentrations (Pongrac et al., 2007). Whole plants with roots and soil were collected in the field, transplanted into pots and transported to the laboratory. Selected leaves were quickly washed in deionized water, and cross-sections cut with a razor blade and immediately frozen by plunging in liquid propane cooled by liquid nitrogen using a Leica CFC Cryoworkstation. They were next freeze-dried in a Leica EM CFD Cryosorption Freeze Dryer. The freeze-drying process followed a 208 h programmed cycle starting at –80°C and eventually reaching ambient temperature, in a vacuum of the order of 10−3 mbar. Such a long cycle was applied to minimize shrinkage of specimens. Dry specimens were mounted between two layers of formvar film (1.5% formvar) coated with a thin carbon layer on the side of the incoming proton beam to prevent charge build-up during measurements.
Instrumentation and analytical methods
Microanalyses were performed using the nuclear microprobe at the Materials Research Group, iThemba LABS, South Africa. A proton beam of 3.0 MeV energy and 100–150 pA current was focused to a 3 × 3 µm2 spot and raster-scanned over the areas of interest, using square or rectangular scan patterns with a variable number of pixels (up to 128 × 128). PIXE and proton backscattering spectrometry (BS) were used simultaneously. PIXE spectra were registered in the energy-dispersive mode, using a LINK Si(Li) detector (active area 80 mm2, resolution c. 170 eV for Mn Kα line) positioned at a take-off angle of 135° and working distance of 25 mm. The energy range was set between 1 and 36 keV. An external absorber (125 µm Be) was positioned between the detector and the specimen in order to stop backscattered protons. BS spectra were detected with an annular Si surface barrier detector (100 µm thick) positioned at an average angle of 176°. Data were acquired in the event-by-event mode. The normalization of results was done using the integrated beam charge, collected simultaneously from a Faraday cup located behind the specimen and from the insulated specimen holder. A more detailed description of the nuclear microprobe setup at iThemba LABS can be found in Prozesky et al., (1995) and Przybyłowicz et al. (1999, 2005).
Elemental concentrations were obtained using GeoPIXE II software (Ryan, 2000). Quantitative analysis was standardless and has been described in detail (Ryan et al., 1990a,b). The error estimates were extracted from the error matrix generated in the fit (Ryan et al., 1990a), and the minimum detection limits (MDL) are calculated using the Currie formula (Currie, 1968). Elemental images were generated using the dynamic analysis method (Ryan & Jamieson, 1993; Ryan et al., 1995), which forms an integral part of GeoPIXE II. These images were free from the influence of background and interferences from overlapping X-ray lines, and were expressed as element concentration (µg g−1 or wt %). An identical matrix composition (cellulose C6H10O5) was assumed in each pixel of mapped areas. The image of each element was based on a group of X-ray lines (typically K or L lines). Images of P, S, Cl, K, Ca, Ti, Mn, Fe, Ni, Cu, Zn, As, Br, Rb and Sr were obtained from K lines, whereas L lines were used for Pb imaging. Cd was mapped using K and L groups of X-ray lines separately, and the results were compared.
Elemental images were constructed for cross-sections of five leaves from each of three plants. In addition, for two cross-sections, scans of elongated, larger areas were complemented by measurements of smaller areas within them, characterized by easily visible vascular bundles. Areas corresponding to the lower and upper epidermis, vascular bundle and palisade and spongy mesophyll were next marked in each cross-section. Contours of areas were drawn after examination of optical micrographs taken before and after measurements. PIXE and BS spectra extracted from these areas were employed to obtain average concentrations using a full nonlinear deconvolution procedure, thus complementing the results from maps. Matrix composition and areal density in each area were obtained from analysis of the corresponding BS spectra using a RUMP simulation package (Doolittle, 1986) and experimental, nonRutherford cross-sections for C and O at a laboratory angle of 170° (Amirikas et al., 1993). BS analyses revealed that all leaf cross-sections could be treated as infinitely thick specimens for the purpose of PIXE matrix corrections.
Leaves of T. praecox showed typical dorsiventral anatomy, with five distinct anatomical areas: the upper and lower epidermis, palisade and spongy mesophyll, and vascular strands (Fig. 1). Stomata were present in the upper and lower epidermis.
Cadmium was detected separately via K and L X-ray lines. Cd results obtained from PIXE spectra extracted from selected areas and from maps are discussed on the basis of K X-ray lines, which are ‘clean’, that is, without any interference and with very low background (Fig. 4a). Overlapping of the Cd L group of lines with the Kα lines of K and Ca, as well as the tail of the potassium Kα line presents a potential danger of over-estimating Cd concentrations (Fig. 4b). The agreement between Cd results obtained from K and L X-ray lines was in general acceptable for PIXE spectra extracted from selected areas within maps and processed using the full nonlinear deconvolution procedure (Table 1), However, Cd maps obtained on the basis of L X-ray lines (not shown) followed closely the distribution of K rather than Cd distribution, and the concentrations from maps in leaf tissues were about twice as high as the results obtained from corresponding, fully deconvoluted PIXE spectra. Therefore, despite the fact that the Cd-L maps showed much better counting statistics and visually looked more convincing than the Cd-K maps with relatively low counting statistics, only the latter are used in the discussion.
Table 1. Element concentrations of whole-leaf cross-section areas and selected leaf tissues of Thlaspi praecox obtained by proton induced X-ray emission (PIXE) microanalysis
No. of replications
Whole leaf area
Values are means ± SD; results are given in µg g−1 DW.
*Some results are below the limits of detection. nd, not detected.
The average concentration of Cd in the leaf cross-sections was 516 µg g−1 DW. The preferred accumulation sites were the lower epidermis, followed by vascular bundles and upper epidermis (Table 1, Fig. 2). The highest recorded value was 1078 µg g−1 DW (lower epidermis) and the lowest was 381 µg g−1 DW (palisade mesophyll).
The average concentration of Zn in the leaf cross-sections was 1170 µg g−1 DW. The preferred accumulation sites were the lower and upper epidermis (Fig. 2, Table 1), where about five times higher concentrations were observed compared with vascular bundles and the mesophylls. The highest recorded value was 3580 µg g−1 DW (average in lower epidermis) and the lowest was 180 µg g−1 DW (palisade mesophyll). Elemental maps showed that some cells in the epidermis were characterized by significantly higher Zn concentration than the neighbouring ones (Fig. 2). The image is intentionally shown with the concentration scale adjusted for better visibility and the concentration in white areas exceeds 0.5 wt %, typically reaching 1.0 wt % or slightly more. The highest concentration obtained from full deconvolution of PIXE spectra extracted from a single epidermal cell was 9610 µg g−1 DW.
The average concentration of Pb in the leaf cross-sections was 45 µg g−1 DW. The highest average concentrations were noted in the lower and upper epidermis (Table 1, Fig. 2), with the highest recorded value of 245 µg g−1 DW (lower epidermis). Values in the mesophylls were of the order of 30 µg g−1 DW, while in the vascular bundles, the concentrations usually varied from 40 to 80 µg g−1 DW, but were sometimes below the variable limits of detection (2.6–96 µg g−1 DW). This variability of the detection limits was caused by the sizes of the vascular bundles and the resulting variable counting statistics in the PIXE spectra.
Iron is another element showing a clear affinity to the epidermis (Fig. 2, Table 1), where concentrations significantly higher than the average concentrations in the leaf cross-sections were detected. Concentrations reaching 500 µg g−1 DW were noted in the upper epidermis, while in the mesophyll and vascular bundles, the concentrations were lower (Table 1).
Calcium was the only element with its most intense localization in the mesophyll, with a preference for palisade tissue (Fig. 3), which was also confirmed by its average concentrations (Table 1). Significantly less Ca was noted in the vascular tissue. Potassium showed the opposite pattern, with the highest concentrations in the vascular bundles and the lowest in the mesophylls. Enrichment in the epidermis was also noted.
Phosporus and S were preferentially localized in the vascular bundles (Fig. 3; Table 1), while Cl showed a more or less homogeneous distribution throughout leaf cross-sections, with a slight enrichment in the epidermis and vascular bundles. However, its concentration in the mesophylls was of the same order of magnitude (Fig. 3, Table 1).
Elemental maps of other elements – Mn, Ni, Rb, Br, Sr, Ti and Cu – are not shown. Their concentrations were too low for creation of reliable images due to low counting statistics. Their average concentrations showed either a rather uniform distribution or their they were too low to reveal distinct patterns in their localization preferences (Table 1).
The concentrations of the whole leaf area cross-sections obtained by micro-PIXE analysis confirmed hyperaccumulation of Cd (> 100 µg g−1 DW), while Zn and Pb concentrations did not meet hyperaccumulation criteria (Table 1). In the field-collected plants, metal concentrations largely depend on total and bioavailable soil metal concentrations, and a considerable sample variation between soil and plant metal concentrations from the individual sampling locations could be observed (Vogel-Mikušet al., 2005). In addition, the preferential localization of Cd at the edges of the leaves and at the points of higher concentrations spread over the whole limb surface was observed in T. caerulescens (Cosio et al., 2005), pointing to a high heterogeneity of metal distribution in the leaves of this species. High heterogeneity of Cd distribution in mesophyll cells of T. caerulescens was also observed in a recent study by Küpper et al. (2007), who found that differential Cd uptake in mesophyll cells during Cd stress acclimation represents a defence mechanism in which some mesophyll cells are sacrificed and used as a metal dump, while in the others photosynthesis can be performed without interruptions.
Micro-PIXE localization of Cd in T. praecox leaves showed that Cd was preferentially stored in the lower and upper epidermis (Table 1, Fig. 2), where it is delivered by the transpiration stream, as indicated by its intense localization in the vascular bundles. Taking into account the large volume of the mesophyll compared with the epidermis, however, the mesophyll is indicated as a relatively large pool of Cd, since half as much Cd was found in the mesophyll as in the epidermal cells. Similar results were also observed for Cd in T. caerulescens (Ma et al., 2005). This is also in line with our previous study, where it was demonstrated, using the same technique, that Cd in T. praecox seeds is concentrated in the epidermis of the cotyledons (Vogel-Mikušet al., 2007).
In mature leaves of T. caerulescens, Cd is mainly bound to oxygen ligands (presumably malic and citric acid) (Küpper et al., 2004; Ueno et al., 2005). Significant amounts were found in T. praecox seeds, indicating high Cd mobility within the plant tissues and an efficient phloem transport (Vogel-Mikušet al., 2007). A very strong accumulation of Cd was detected in the cytoplasm of the root phloem and companion cells in metal nontolerant A. thaliana, which was attributed to retranslocation of Cd from the shoots, therefore protecting the shoot from Cd toxicity (Van Belleghem et al., 2007). Similarly Reid et al. (2003) and Cakmak et al. (2000) showed that Cd can be rapidly distributed via phloem to all tissues in potato and wheat. To date, however, data on Cd ligands in phloem are scarce. Nicotianamine is the sole molecule that has been identified in the phloem sap as a potential metal transporter in association with Fe, Cu, Zn and Mn (Staphan & Scholz, 1993). The possible connection to anion(s) (e.g. P, S and Cl) showing similar localization patterns (e.g. epidermal and vascular) in T. praecox should, however, not be ruled out for Cd trafficking.
Significant concentrations of Ca (up to 3 wt %) were observed in palisade mesophyll cells with significantly lower concentrations noticed in vascular and epidermal tissues. Similar distribution pattern was also observed in other metal-hyperaccumulating plants, for example, Alyssum lesbiacum, A. bertolonii, T. goesingense and Arabidopsis halleri (Küpper et al., 2000, 2001). Compartmentation of Ca in mesophyll cells may play a significant role in Ca homeostasis, preventing interference with the Ca2+ signalling pathways and avoiding stomatal closure induced by extracellular Ca2+(De Silva et al., 1985). The high Ca mesophyll concentrations of metal-accumulating plants may also play a significant role in protecting photosynthetically active tissues from metal toxicity by preventing accumulation of large quantities of metals in photosynthetically active cells and possibly in chloroplasts (Antosiewicz, 2005), and against heavy metal inhibition (e.g. Pb) of the photosystem II (Rashid & Popovic, 1990; Antosiewicz, 2005). However, measurements of Cd and Ca should be performed at cellular level in order to confirm a protective role of Ca against Cd toxicity. Previously, positive correlations between Cd and Mg in A. halleri (Küpper et al., 2000) and Ni and Mg in T. goesingense (Küpper et al., 2001) were found at single cell level and concentration of Mg in mesophyll cells was interpreted as a defence against the substitution of Mg in chlorophyll by heavy metals.
The Ca distribution pattern was the opposite of that of K, which accumulated mainly in vascular tissues and to a lesser extent also in the epidermal tissues, indicating the differential accumulation capacity of different cell types for Ca and K. The same effect was also observed in citrus leaves (Storey & Leigh, 2004).
• Simultaneous use of PIXE and proton BS microanalyses enabled localization and elemental mapping of Cd, Zn and Pb, as well as the majority of nutrients within leaf cross-sections of field-collected T. praecox.
• Elemental mapping using micro-PIXE showed that metals and nutrients were asymmetrically distributed between and within different leaf tissues of T. praecox, indicating either differential uptake of ions from the apoplast as the transpiration stream passes different cell types or different apoplastic or symplastic pathways of different ions from the xylem to particular cell layers, in line with the conclusions of Karley et al. (2000).
• Cadmium was preferentially accumulated in metabolically less active epidermis and vascular bundles, but taking into account the large volume of the mesophyll compared with the epidermis, mesophyll was indicated as a relatively large pool of Cd as well. Intense vascular Cd localization indicates high Cd mobility with the transpiration stream.
• Lead showed preferential localization in the lower and upper epidermis, similar to the Cd localization pattern.
• Zinc and Fe showed high enrichment in the epidermis.
• A high preference of Zn and Fe for sequestration in the large vacuolated epidermal cells was observed.
• Preferential accumulation of Ca in the palisade mesophyll might additionally alleviate metal (especially Cd) toxicity in photosynthetically active tissues of the leaves.
The work was supported by the following projects: MSZS L1-5146-0481 Tolerance of Organisms in Stressed Ecosystems and Potential for Remediation, with the sponsorship of Rudnik Mežica MPI, Črna na Koroškem and Mobitel dd; MSZS PO-0212 Biology of Plants research programme and EU COST 859 entitled Phytotechnologies to Promote Sustainable Land Use and Improve Food Safety. Some financial support was provided by the International Atomic Energy Agency, Vienna.