ITCB-EDTA, cadmium chloride (CdCl₂), bovine serum albumin (BSA), anti-mouse IgG-gold antibody, and LR white resin were purchased from Sigma-Aldrich (St. Louis, MO). The FITC-conjugated goat anti-mouse secondary antibody was from Zhongshan Golden bridge Biotechnology Co., LTD (Beijing, China). All other chemicals were of analytical grade and obtained from Beijing Chemical Reagents Co. (Beijing).

Buffers and solutions used include a formaldehyde-acetic acid-alcohol (FAA) solution (50 mL of formaldehyde 40%+ 50 mL of acetic acid + 900 mL of 50% alcohol); phosphate-buffered saline (PBS) (0.1 M phosphate buffer containing 0.9% sodium chloride, pH 7.5); PBS with 0.1% (v/v) Tween-20 (PBST); and PBST containing 0.5% (w/v) gelatin (PBSTG).

The soil was a mixture of purchased nutrient soil and vermiculite (1:1, v/v), air dried in sunlight and passed through a 5-mm mesh polyethylene sieve to remove large stones and grass debris. Seven treatments (one control and 6 experimental) were tested in triplicate in a randomized block design. The control was not supplemented with Cd²⁺; and the others were fortified with different concentrations of CdCl₂ solutions followed by a 7-day incubation at ambient temperature as described in Table 1. Wheat seeds were sown and grown in a greenhouse under natural light conditions with day/night temperatures of 25/15°C. Pots (14 cm x 14 cm) were watered once every three days at 250 mL of tap water per time per pot, which was sufficient to saturate the soil to field water capacity without leaching. No additional fertilizers were applied to the soil during the experiment.

The collected wheat plant samples were carefully washed with deionized water, and the roots and leaves of the wheat plants were separated with plastic scissors; the plant samples were dried at 80°C to a constant weight. The Cd²⁺ content in all samples was then determined according to the ELISA procedures described previously [30]. The C1, C3 and C5 samples were simultaneously determined with graphite furnace atomic absorption spectroscopy (GFAAS).

After the treatment period, selected roots were quickly washed with deionized water and cut transversely with a razor blade into 3 cm-long segment in the root hair zone. The segments were immediately and completely embedded in OCT embedding compound and then frozen at -60°C. The sample segments were sectioned crosswise at 10 μm thick using a Microm HM525 cryostat from Microm Instruments, Inc. (San Macros, CA) and mounted on microscopic glass slides. Sections were air dried for 30 min at room temperature to prevent them from falling off the slides during antibody incubations. The tissue sections were prefixed in 5-mM ITCB-EDTA solution for 10 min at 4°C, then fixed in a FAA solution for another 10 min. After the FAA solution was discarded, the sections were washed gently 3 times with PBS.

The slides were incubated in PBS containing 1% BSA for 10 min at room temperature to block non-specific binding sites. After the BSA/PBS solution was washed off, mMb4F₃B₆D₉A₁ in PBS buffer was placed on the tissue followed by incubation for 1 h at room temperature. After removal of the primary antibody mMb4F₃B₆D₉A₁, the FITC-labeled anti-mouse IgG secondary antibody in PBS was added and then incubated for 1 h at room temperature. The optimum working dilutions for the primary antibody and FITC-labeled secondary antibody were 1:500 and 1:50, respectively. For the control test, the tissue sections were treated in the same manner as the treatment described above, except without mMb4F₃B₆D₉A₁. The stained sections were visualized on a Nikon C1 Plus LCSM microscope equipped with a 488-nm laser for excitation.

The selected roots were quickly rinsed with deionized water and cross-sectioned with a razor blade into 1 mm tissue blocks. The tissue sections were immediately fixed in 53.1 mM ITCB-EDTA solution at 4°C for 15 to 30 min followed by a mixture of 10% paraformaldehyde and 1% glutaraldehyde in 0.1 M PBS. They were entirely infiltrated with these solutions under brief vacuum. After fixation for approximately 8 h at room temperature, the samples were washed in 0.1 M PBS and dehydrated sequentially in a series of ethanol solutions (30%-100% in a 10% increment). The tissue sections were infiltrated and embedded in LR White resin. Ultrathin sections (70–80 nm) of the treated samples were obtained using a Leica EM UC6 Ultra-S microtome and then mounted onto nickel grids (50 mesh).

Immunoelectronic tissue staining was all performed at ambient temperature. Ultrathin tissue sections were treated sequentially with the following solutions (each 30 μL): 50 mM glycine in 0.1 M PBS buffer for 5 min; 1% BSA in PBSTG for 5 min; mMb4F₃B₆D₉A₁ in PBSGT (containing 1% BSA) for 60 min; rinse with PBS 3 times; 10 nm colloidal gold-labeled anti-mouse IgG secondary antibody in PBSTG for 45 min; rinse with ultrapure water 3 times; 1% aqueous uranyl acetate for 30 min; rinse with ultrapure water3 times. For immunoelectronic staining, the optimum working dilutions for the primary antibody and colloidal gold-labeled anti-mouse IgG secondary antibody were 1:1000 and 1:100, respectively. All the specimens were finally observed under a JEM-1230 transmission electron microscope with an energy-dispersive X-ray (EDX) analyzer.

The sample preparation for the conventional Cd²⁺-S coprecipitation method was performed as described previously [31]. Tissue segments were washed with deionized water and then fixed in the presence of 1% Na₂S with 3% glutaraldehyde in 0.2 M phosphate buffer (pH 7). Post-fixation in 1% OsO₄ was performed before dehydration with ethanol series (50, 60, 70, 80, 90, and 100%). After dehydration, samples were treated with propylene oxide and infiltrated in Spurr resin. The segments were then cut into ultrathin slides (70–80 nm). Prior to final observation under the JEM-1230 transmission electron microscope, the specimens were stained with 1% aqueous uranyl acetate for 15 min followed by 1% lead citrate for 15 min.

After 28 days of exposure, toxicity symptoms such as quite noticeable browning, shortening and thickening of roots, were observed when the fortified Cd²⁺ content in the soil was greater than 50 μg g^-1. The concentrations of Cd²⁺ in the wheat plants were determined with icELISA and confirmed with GFAAS. Cd²⁺ accumulation in the roots and leaves of the wheat plants cultivated in soil fortified with different concentrations of Cd²⁺ is shown in Table 1. Cd²⁺ concentrations in the roots were higher than those in the leaves. The translocation factor (TF) is the ratio of metal concentration in other plant parts to that in plant roots [32] and is an important parameter in studies of metal accumulation and translocation. All the TF values of Cd²⁺ from roots to leaves were <1, which indicate a strong capability to hold Cd²⁺ in root cells against transport to the aboveground parts of wheat plants.

The major obstacles to developing an IHC method for metallic ions are preserving tissue structure integrity as close to the living situation as possible [33] and avoiding loss and redistribution of ions. To overcome this challenge, ITCB-EDTA was used to chelate the mobile Cd²⁺ in the plant tissue and fix to proteins at the site via the isothiocyano functional group prior to fixation of frozen tissue samples or fixation for transmission electron microscopy. Fig 1 shows the proposed mechanism of IHC Cd²⁺ prefixation in plant tissues. ITCB-EDTA solution was applied to bind the mobile Cd²⁺ to the surrounding proteins, while mAb4F₃B₆D₉A₁ recognized the EDTA-Cd²⁺ portion of the ITCB-EDTA-Cd²⁺ complex. For immunofluorescent staining, the optimum working dilutions for the primary antibody (mAb4F₃B₆D₉A₁) and FITC-labeled secondary antibody were 1:500 and 1:50, respectively. For immunoelectronic staining, the optimum working dilutions for the primary antibody and colloidal gold-labeled anti-mouse IgG secondary antibody were 1:1000 and 1:100, respectively.

Fig 2I A-F show fluorescent microscopic images of Cd²⁺ distribution in wheat exposed to different Cd²⁺ doses (0, 1, 5, 25, 50 and 200 μg g^-1) for 28 days, and Fig 2I G shows the image of no primary antibody control tests, while Fig 2I a-g show the corresponding bright field images. Except for the images of the Cd²⁺-negative tissue section (Fig 2I A) and control treatment (Fig 2I G), bright green fluorescence could be observed in all other images of Cd²⁺-treated wheat. There was a distinct enhancement of the root cross-sectional area and the green fluorescent signal intensity with increased concentrations of Cd²⁺ in the soil. For plants that grew in soil fortified with 1 and 5 μg g^-1 Cd²⁺, bright green fluorescence could be observed mainly in the endodermis, and dots of fluorescence signal could be also observed in the stele cells (Fig 2I B and C). When the Cd²⁺ concentrations in the soil were 25 and 50 μg g^-1, bright fluorescence could be observed in the stele (Fig 2I D and E). When the Cd²⁺ concentration was 200 μg g^-1, extremely bright fluorescence was observed throughout the whole root (Fig 2I F). The results indicated that at low concentrations, Cd²⁺ principally precipitated within and outside the endodermis. When the Cd²⁺ concentration in the soil was higher than 25 μg g^-1, excess Cd²⁺ passed through the endodermis and entered the stele. In unexposed plants and controls, a faint but not bright signal within the background range was still observed, probably as a result of autofluorescence produced by polyphenols and other compounds [14].

Leitenmaier and Küpper [18] reported that fluorescence microscopy cannot quantitatively detect target metal ions in cells because the dyes do not display high selectivity for the target metals, and the fluorescence response of the dye to metal concentrations is not linear. In this experiment, a standard curve relating the total FITC fluorescent value and the Cd²⁺concentration in the plant roots was developed (Fig 2II). The mean correlation coefficient of the standard curve was 0.9955, which demonstrated that the immunofluorescent method for Cd²⁺ was sensitive and specific enough for the quantitative imaging of Cd²⁺ in plant tissues. The fluorescent value was expressed as the product of the average fluorescent intensity and its area, which was obtained by Nikon EZ-C1 Free Viewer software.

In addition to the immunofluorescent method, a new immunoelectronic staining method was developed and confirmed with EDX analysis. To verify whether gold particles can represent the location of Cd²⁺ in the plant cell, gold particles from the colloidal gold-labeled anti-mouse IgG secondary antibody were analyzed randomly on an EDX analyzer. Fig 3A–3C shows the EDX spectra of the gold particles. The particles showed the specific Cd²⁺ location in the xylem and endodermis cells and curdled macromolecular substances of wheat plants exposed to 100 μg g^-1 Cd²⁺. Fig 3D shows the spectrum of a blank region without gold particles. The Cd peak was present in each gold particle EDX spectrum (Fig 3A–3C) but was not present in the spectrum of the blank region. The peaks labeled Al and Ni are the Al holder and Ni grid, respectively. The EDX spectra of the gold particles and blank region proved the reliability of the established immunoelectronic microscopic method.

The conventional histochemical method of Na₂S precipitation for heavy metal ions was also implemented for comparison with the IHC method developed in the present study. Fig 4 shows the Cd²⁺ distribution in the stele of Cd²⁺-negative (Fig 4A and 4C) and Cd²⁺-treated (Fig 4B and 4D) wheat plants. The dark deposits in the wheat plants were analyzed randomly on an EDX analyzer, and the corresponding spectra are displayed in Fig 5. The deposits in the Cd²⁺-negative wheat sample (Fig 4A and 4C) suggest that Na₂S precipitation method could produce false positive, and the corresponding spectra (Fig 5A and 5C) proved the standpoint, which had Cu and Mg peak, and no obvious Cd peak. The spectrum of the Cd²⁺-treated samples had an obvious Cd peak, Mg peak and Cu peak (Fig 5B), which also suggest that it is lack of specificity for Cd²⁺ using the Na₂S precipitation method for heavy metal ion detection, an inherent disadvantage of the method [12]. Note: the peaks labeled Al, Os, Pb and Cu indicate the Al holder, osmium tetroxide fixative, lead citrate staining solution and Cu grid, respectively.

The immunofluorescent method developed in the present study was used to study the dynamic process of Cd²⁺ absorption in wheat roots exposed to 50 ug g^-1Cd²⁺ (Fig 6). Samples were collected 2, 4 and 15 days after germination. The epidermis cells were the first binding sites for Cd²⁺. Bright green fluorescence could be observed only 2 days after seed germination (Fig 6A). As the exposure time increased, Cd²⁺ was observed in the xylem and then phloem; the xylem accumulated more Cd²⁺ than the primary phloem, as indicated by the fluorescence intensity (Fig 6B and 6C).

Fig 7 shows immunoelectronic microscopic images of wheat root transverse sections. Compared to the unexposed controls (Fig 7A), the cells of the Cd²⁺-treated wheat plants were smaller and more closely arranged, were more seriously lignified, and were structurally impaired, such as cell wall fracture [34] and coagulation of macromolecular substances in the vacuole.

Fig 8 shows the location of Cd²⁺ deposition in the cortex, endodermis and stele cells of control and Cd²⁺-treated wheat roots. In the Cd²⁺-treated plants (Fig 8B), distinct gold particles were observed in the cortical tissue of the middle lamella and the outer surface of the cell wall in the intercellular space bordered by three cortical cells, whereas few particles were observed in the control plants (Fig 8A).On the outer and inner tangential endodermal cell wall of Cd²⁺-treated plants (Fig 8C and 8E), there were no distinct colloidal gold particles. Fig 8D shows that the non-thickened outer tangential endodermal wall of the Cd²⁺-treated plant was one of the main Cd²⁺ deposition locations. However, only a few gold particles appeared on the membrane along the inner tangential cell wall and none on the inner tangential wall (Fig 8F). Fig 8H shows Cd²⁺-induced cell wall fracture and Cd²⁺ deposition in the cytoplasm near the cell wall, which were not observed in Cd²⁺ control plants (Fig 8G). To date, there is no direct evidence that Cd²⁺ can be transported through the symplast pathway. Gold particles were observed in plasmodesmata (PD) of Cd²⁺-treated plants (Fig 8J) but not in control plants (Fig 8I). There was notable difference in gold particle density on the curdled macromolecular substances in the vacuole between the control plant (Fig 8K) and the Cd²⁺-treated plant (Fig 8K). In addition, more membrane-bound organelles were found in some root phloem cells of Cd²⁺-positive plants compared to the control (Fig 8M), and numerous gold particles were observed (Fig 8N).