Tobacco (zhongyan 100) seeds were donated from Institute of Tobacco Research, Chinese Academy of Agricultural Sciences. The seeds were propagated in the test base (Danzhou, Hainan province, China) of Haikou Experimental Station, Chinese Academy of Tropical Agricultural Sciences.

Tobacco seeds were immersed in 70% ethanol for 30 seconds. And then, the ethanol was removed and the seeds were soaked with H₂O₂ for 10 minutes. After the H₂O₂ was removed, the seeds were washed with sterilized water for five times, one minute each time. Finally, the seeds were cultured on MS medium with 0.3% (w/v) phytogel (pH 5.7). Plates were incubated in dark at 4˚C for three days and then were moved to a growth chamber with controlled temperature (22˚C - 24˚C), relative humidity (75% - 90%), light (750 µEm⁻²) and photoperiod (16-h day/8-h night) for one month. The fresh leaves were cut from the seedlings in superclean bench and cultured on MAI solid medium (MS, 10 mg/ml maltose, 1 mg/ml 2,4-D, 1 mg/ml IAA, 0.1 mg/ml NAA, 1 mg/ml biotin, 50 mg/ml glutamine, 30 g/l sucrose, 1.3 g/l phytogel, Ph = 5.8). MAI solid medium was placed in dark with controlled temperature (22˚C - 24˚C), relative humidity (75% - 90%). After two months, calli appeared. And then, the calli was cut from the explants and cultured in MAI liquid medium (MS, 10 mg/ml maltose, 1 mg/ml 2,4-D, 1 mg/ml IAA, 0.1 mg/ml NAA, 1 mg/ml biotin, 50 mg/ml glutamine, 30 g/l sucrose, Ph = 5.8) under 28˚C for three days. The debris in MAI liquid medium was filtered and discarded. And then, tobacco suspension cells were gotten.

Dye reaction buffer was prepared as the following: 0.1 mmol/L CaCl₂, 10 mmol/L KCl, 0.5 mol/L sorbitol, 0.05% PVP40, 0.2% BSA, 5 mmol/L MES, Ph = 5.5. 1 ml of suspension cells were transferred from triangular flask into 1.5 ml tube. 1 mg of chemosensor 1 [18] was dissolved with dye reaction buffer. 10 µl of dye was added into the suspension cell mixture. The tube was placed under 28˚C for ten minutes. And then, 100 µl of the suspension cell mixtures were placed on the glass slide and observed using Laser scanning Confocal Microscopy (Olympus, Japan). The wavelength of exciting light was 420 nm. The wavelength of emitted light was 597 nm.

Differential centrifugation was performed according to the papers published and modified [19] [20] . Tobacco seeds were cultured on MS plate for one month. And then, the seedlings were cultured in MS liquid medium. After ten days, 10 µmol/L HgCl₂ were added into the medium for 2 days. The roots of the seedlings were immersed in 20 mmol/L of Na₂-EDTA for 3 hours. And then, the roots were washed with distilled water for three times, 10 minutes each time. The roots were cut and collected. The roots samples were stored in −70˚C freezer until use. Buffer I was prepared as the following: 0.2 mol/L Tris-HCl, 1 mmol/L PMSF, 1 mmol/L EDTA, 1 mmol/L SDS, 20 mmol/L mercaptoethanol, ph = 8.0. 12 g of the roots sample were ground into powder in liquid nitrogen. The powder was mixed with 100 ml buffer I. The mixture was sonicated for 30 minutes. And then, the mixture was filtered with gauze. The debris was collected and named as D0. The filtered mixture was centrifugated at 500× g. The supernatant was collected and named as S1. The debris was collected and named as D1. D0 and D1 were mixed and ground in liquid nitrogen once again. 100 ml buffer I was added into the powder. The mixture was sonicated for 30 minutes once again. The mixture was filtered with gauze. The debris was collected and named as D01. The filtered mixture was centrifugated at 500× g. The supernatant was collected and named as S11. The debris was collected and named as D11. D01 and D11 were mixed and treated once again following the procedure above. The debris collected from gauze was named as D02. The debris collected from centrifugation was named as D12. The supernatant was named as S12. S1, S11 and S12 were mixed and centrifugated with 10,000× g at 4 oc. The debris was collected and named as F1. The supernatant was centrifugated with 18,000× g at 4 oc. The debris was collected and named as F2. The supernatant was named as F3.

Mercury content in different tissues was measured according to the paper published [21] . After the seedlings were cultured for 15 days, they were harvested, washed thoroughly in distilled water, shoots and roots were separated, and immersed directly in liquid nitrogen. The frozen plants were dried using a freezer-dryer, and the dry weight was determined. After being ground to a fine powder using liquid nitrogen, samples (three replicates of each treatment) were acid-digested by stepwise additions of 70% (v/v) nitric acid, 30% (v/v) hydrogen peroxide, and concentrated HCl. For measuring mercury content of the whole seedling, the procedure was as the same as described above, except that shoots and roots were not separated.

The Hg content in debris was measured following a previously published method [22] . The debris was washed with fresh LB liquid medium three times. And then, the pellets were acid-digested in 70% (v/v) nitric acid, 30% (v/v) hydrogen peroxide and concentrated HCl at 95˚C. The supernatant was collected and put in a tube. 70% (v/v) nitric acid, 30% (v/v) hydrogen peroxide, and concentrated HCl were added into the tube incrementally. The tube was kept at 95˚C for 2 hours. The samples were then analyzed in an AANALYST 200 Perkin Elmer Spectrometer with an MHS-15 Mercury-Hydride System. The Hg content was calculated in µmol/L as the paper published previously [22] . All samples were measured three times, and the average values were used.

After tobacco seedlings were cultured in MS liquid medium containing 5 µmol/L mercuric chloride for two days, the mercury content in root and shoot was measured respectively. Results showed that the mercury content in shoot of the seedling was 12 µg/g, while the value in the root of the seedling cultured in the same medium was about 44 µg/g (Figure 1, Figure 2). The mercury content in roots was about three times higher than that in the shoots. Similar results were also found in seedlings cultured in liquid MS medium containing 2.5 µmol/L mercuric chloride (Figure 1, Figure 2). These demonstrated that after mercury was absorbed by plants, most of the mercury accumulated in roots. Only a small section of the mercury in plants was transferred into the above tissue.

To analyze the mercury distribution in plant root further, tobacco roots were ground in liquid nitrogen and different cell fractions were separated. After the mercury content in different fractions were measured, it was found that in the seedlings cultured in MS liquid medium containing 5 µmol/L mercuric chloride, the mercury content in cell wall was 6.6 µg/g (Figure 3). The mercury content in cell membrane fraction was 2.1 µg/g (Figure 3). The mercury content in

supernatant fraction was 35.1 µg/g (Figure 3). After the mercury contents of fractions from seedlings cultured in MS liquid medium containing 2.5 µmol/L mercuric chloride were measured, it was found that the mercury content in cell wall, cell membrane and supernatant was 3.6 µg/g, 1.2 µg/g, 19.2 µg/g, respectively (Figure 3). These demonstrated that in roots, most of the mercury accumulated in liquid fraction, about 15% of the mercury was attached by cell wall. Only a small part assembled in cell membrane.

For further identify where the mercury in liquid fraction was distributed, tobacco suspension cell was prepared and mercury was labeled as described in

materials and methods. After the suspension cells were treated with 0.05 µmol/L mercuric chloride for two hours, the fluorescence indicator was added and cultured in 28˚C for 20 minutes. And then the cells were observed under confocol. It was found that most of fluorescence located in vacuole and cell wall (Figure 4). This was consistent with the results gotten from differential centrifugation (Figure 3).