INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Selenium (Se) is an essential element for many organisms, but it is also toxic at higher concentrations. Se is essential because seleno-Cys (Se-Cys) is in the active site of certain selenoproteins, several of which are involved in oxidative stress resistance (Stadtman, 1996). Se becomes toxic at higher levels due to incorporation of Se into sulfur (S)-containing molecules, especially the nonspecific replacement of Cys by Se-Cys in proteins (Ohlendorf et al., 1986; Anderson, 1993). This replacement of S by Se in molecules is due to the chemical similarity of these two elements; most enzymes involved in S metabolism can catalyze the analogous reaction with the corresponding Se substrates with similar affinity for both substrates (Stadtman, 1990; Anderson, 1993). On the other hand, Se-specific enzymes tend to have a much higher affinity for the Se substrate than for the S analog (Mihara et al., 2000).

The specific incorporation of Se into selenoproteins involves the translation of UGA opal (stop) codons in specific mRNAs encoding Se-Cys-containing proteins (Böck et al., 1991). Se-Cys-tRNA is formed from Ser-tRNA using selenophosphate as a Se substrate. Selenophosphate is formed by selenophosphate synthetase, using elemental Se (Se⁰) as a substrate (Lacourciere and Stadtman, 1998; Lacourciere et al., 2000). Se⁰ is released from Se-Cys by Se-Cys lyase (SL; a pyridoxal phosphate-dependent enzyme of the NifS family), also producing Ala. Se-Cys is produced from selenate by the sulfate assimilation pathway, both in Se-requiring and non-requiring organisms (Stadtman, 1990; Anderson, 1993; Pilon-Smits et al., 1999). In plants, this pathway is localized mainly in the chloroplast (Leustek and Saito, 1999).

At this point, it is not clear whether Se is essential for plants. There are no confirmed reports of selenoenzymes in higher plants, only in the unicellular alga Chlamydomonas reinhardtii (Fu et al., 2002; Novoselov et al., 2002). However, the machinery for incorporation of Se into selenoproteins may be present in plants, because a Se-Cys-tRNA recognizing the UGA anticodon was found in Beta vulgaris (Hatfield et al., 1992). Also, a gene encoding a plastidic NifS-like protein with SL activity was recently cloned from Arabidopsis, but the role of this enzyme in plant Se and/or S metabolism has yet to be elucidated (Pilon-Smits et al., 2002).

Se toxicity occurs when organisms are exposed to sediments that are naturally rich in Se (shale rock) or to wastewater derived from these sediments or from Se-rich coal or oil (Ohlendorf et al., 1986). Different Se tolerance mechanisms have evolved, many of which are based on prevention of incorporation of Se-Cys into proteins. For instance, Se can be volatilized as low-toxic dimethyl(di)selenide (Ganther et al., 1966; Lewis et al., 1966), or accumulated in the form of nonprotein seleno-amino acids (e.g. methyl-Se-Cys and methyl-Se-Met; Neuhierl et al., 1999). Also, some organisms accumulate Se⁰, which is insoluble and has a relatively low toxicity (Wilber, 1980). The Se⁰ may either be formed via reduction of selenite (Tomei et al., 1995; Kessi et al., 1999) or released from Se-Cys by SL activity (Mihara et al., 2000).

Because plants are efficient at accumulating and volatilizing Se from soil or water (Hansen et al., 1998; Zayed et al., 2000), they may be used in various ways for cleanup of Se-polluted sites. Plant genetic engineering offers a tool to further enhance the efficiency of Se phytoremediation and to obtain a better understanding of the factors controlling Se tolerance, accumulation, and volatilization. In a previous study, overexpression of the key enzyme of the sulfate assimilation pathway, ATP sulfurylase, was shown to lead to enhanced selenate reduction and higher Se tolerance and accumulation (Pilon-Smits et al., 1999).

In the present study, we wanted to test the hypothesis that (over) expression of SL in plants reduces nonspecific incorporation of Se into proteins and, thus, leads to enhanced Se tolerance and perhaps accumulation. This is based on the assumption that in transgenic SL plants, the flow of Se will be redirected toward accumulation as relatively low-toxic Se⁰, rather than into proteins. To test this hypothesis, a mouse (Mus musculus) SL enzyme was expressed in Arabidopsis as a model plant species. The mouse enzyme was chosen because of its very high activity toward Se-Cys but negligible activity toward Cys (Mihara et al., 2000). Thus, introducing the mouse SL gene in plants was expected to alter Se fluxes without disrupting S metabolism. The mouse SL enzyme was targeted to either the cytosol (a site of translation) or the chloroplast (another site of translation and the main site of Se-Cys formation). The transgenic SL plants were analyzed with respect to Se incorporation into proteins and Se tolerance and accumulation.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Transgenic Arabidopsis plants were obtained using the two gene constructs for cytosolic (Fig. 1A) and chloroplastic (Fig. 1B) expression, respectively. Ten transgenic lines per construct were identified using PCR and propagated to homozygosity. No phenotypic differences were observed between the transgenics and wild-type plants when grown under standard conditions on agar or soil. Four lines per construct were selected for further analysis, two with high and two with low expression levels, as judged from immunoblotting (Fig. 2A) and SL enzyme activity measurements (Fig. 2B). The transgenic SL plants contained a protein that reacted with antiserum raised against the mouse SL enzyme (Fig. 2A), which was of the expected size (43 kD) and not present in wild-type plants. Therefore, it was concluded that the mouse SL protein is expressed in the transgenic plants, and that the transit sequence is cleaved off in the chloroplastic lines. The mouse SL also appears to be active in the plants, because the transgenics showed SL activities up to 2-fold higher in cytosolic lines and up to 6-fold higher in chloroplastic lines, compared with wild type (Fig. 2B). The endogenous SL activity can be attributed to NifS-like proteins, like the recently characterized Arabidopsis NifS-like chloroplast protein with SL activity (Pilon-Smits et al., 2002). The SL activity in roots was lower than in shoots for both the transgenic and wild-type plants (Fig. 2B, left versus right). The SL activity levels in the transgenic lines relative to each other and to wild type showed a similar pattern in roots and shoots, except that the chloroplastic lines with the highest expression levels, cpSL1 and 9, showed a less dramatic increase in SL activity in roots compared with shoots. A relatively low activity in roots compared with shoots has also been found for other proteins targeted to the plastids (Pilon-Smits et al., 1999, 2000); this phenomenon may be due to lower plastid import capacity or cofactor availability in root tissue.

View larger version (17K): [in this window] [in a new window]   Figure 1.   Representation of the cytosolic (A) and chloroplastic (B) SL gene constructs, used for the transformation of Arabidopsis. LB, Left border; RB, right border; nptII, coding sequence of the neomycin phosphotransferase gene; mouse SL, coding sequence of the mouse SL gene; TrFd, chloroplast transit sequence of ferredoxin; 35SCAMV, cauliflower mosaic virus promoter; 5'NOS, promoter of the nopaline synthetase gene; 3'NOS, terminator of the nopaline synthetase gene; AMV, translational enhancer sequence. Forward (a) and reverse (b) primers used for PCR are shown with arrows.

View larger version (43K): [in this window] [in a new window]   Figure 2.   Expression analysis of the mouse SL gene in the Arabidopsis SL transgenics. Left, Shoot tissue; right, root tissue. A, Immunoblot, using antiserum raised against the mouse SL protein. WT, Arabidopsis wild type. cytSL 1, 8, 12, and 13, Transgenic Arabidopsis lines expressing the cytosolic SL construct. cpSL 1, 2, 8, and 9, Transgenic Arabidopsis lines expressing the chloroplastic SL construct. Extract of Escherichia coli expressing the mouse SL was used as a positive control. Protein was extracted from shoots of 2-week-old plants, and equal amounts of total protein (20 µg) were loaded in each lane. B, SL enzyme-specific activity in wild-type (WT) and transgenic Arabidopsis seedlings grown for 2 weeks on 0.5-strength Murashige and Skoog agar medium. Shown are the means and SEs of three measurements, using extracts made from 10 plants each. SEs not shown were too small to be visualized by the graphing program. Please note the scale difference in Figure 2B between the shoot and root tissues.

The intracellular location of the mouse SL protein in the transgenic plants was investigated. Protein extracts from total shoot tissue and from isolated chloroplasts were analyzed for the presence and activity of the mouse SL protein. Homogenate, total chloroplast, and stromal fractions from transgenic line cpSL1 all contained a 43-kD band that reacted with the anti-SL antiserum with similar intensity (Fig. 3A). The cytosolic SL13 fractions only showed this band in the homogenate but not in the chloroplasts. Similarly, SL enzyme activity in chloroplast stroma of cpSL1 plants was 80% of that in homogenate, whereas in cytSL13 plants, the SL activity in stroma was much lower than in homogenate and comparable with wild type (Fig. 3B). Therefore, we conclude that the cytosolic SL transgenics most likely express the mouse SL in their cytosol, and that the cpSL transgenics target the mouse SL enzyme to the chloroplast stroma.

View larger version (36K): [in this window] [in a new window]   Figure 3.   Localization of the mouse SL protein in Arabidopsis SL transgenic lines, comparing shoot homogenate and chloroplasts of 3-week-old soil-grown plants. A, Western blot immunostained with mouse SL-specific antibodies, showing homogenate (H), intact chloroplasts (CI), protease-treated chloroplasts (CP), and chloroplast stroma (S). Protein amounts reflecting equal amounts of chlorophyll were loaded in each lane. B, Enzyme activities in shoot homogenate (H) and chloroplast stroma (S) of wild-type (WT) and transgenic SL plants. Shown are the means and SEs of three measurements. The band indicated by an asterisk is unrelated. Note that the reduced mobility of SL in the homogenate fractions is a result of sample preparation (trichloroacetic acid precipitation of homogenate in bovine serum albumin-rich chloroplast grinding buffer).

Both the cytosolic and chloroplastic SL transgenics contained significantly less Se in protein compared with wild type (Fig. 4; P < 0.05 for all lines except cytSL8 and 12), with an average of 42% reduction. Therefore, it can be concluded that expression of the mouse SL enzyme in either the cytosol or the chloroplast can significantly prevent the nonspecific translational incorporation of Se-Cys into proteins.

View larger version (41K): [in this window] [in a new window]   Figure 4.   Se incorporation into proteins of wild-type (WT) and SL transgenic Arabidopsis seedlings grown for 14 d on 0.5-strength Murashige and Skoog agar medium supplied with 5 µM selenite. Shown are the means and SEs of five replicates from the shoots of 25 plants each.

The prevention of nonspecific incorporation of Se into protein would be expected to result in enhanced Se tolerance. Initial experiments on horizontal agar plates with Se indicated that the cytosolic SL plants were more tolerant to Se than wild type, but surprisingly, the chloroplastic SL plants appeared less tolerant, as judged from shoot size and degree of chlorosis. Se tolerance was quantitatively analyzed by measuring root length after 2 weeks of growth on vertical agar plates containing various forms of Se (Fig. 5). The cytosolic SL plants grew significantly better on agar medium supplied with either selenate, selenite, or Se-Cys (P < 0.05). The biggest difference was observed on Se-Cys, where the cytSL plants developed 4-fold longer roots than wild type. The increase in Se tolerance was correlated with SL expression level, with the highest expressors showing the biggest increase in tolerance. The chloroplastic SL plants showed significantly reduced Se tolerance to all three forms of Se (P < 0.05, Fig. 5). The degree of Se tolerance was inversely correlated with the SL expression level, ranging from slightly reduced tolerance for low cpSL expressors to 3-fold shorter roots for high expressors. On control medium, there were no differences in growth among the plant types (Fig. 5).

View larger version (81K): [in this window] [in a new window]   Figure 5.   Tolerance of wild-type (WT) and transgenic SL Arabidopsis seedlings to 50 µM selenate, 25 µM selenite, or 50 µM Se-Cys. Tolerance was measured as root length. Shown on the left are the means and SEs of 24 seedlings per treatment per line. The picture on the right shows the growth of representative cytSL and cpSL lines in comparison with WT.

Both the cytSL and cpSL plants showed a general trend of having higher levels of Se in their shoots, at least when supplied with selenite (up to 50% higher, Fig. 6). There were no significant differences in Se concentrations when the plants were supplied with selenate (results not shown). Among the selenite-treated plants, all four cpSL lines and two of the cytSL lines tested contained significantly higher Se levels than wild type (P < 0.05). There was no apparent correlation between SL expression level and tissue Se concentration.

View larger version (66K): [in this window] [in a new window]   Figure 6.   Shoot Se concentrations in wild-type (WT) and transgenic SL Arabidopsis seedlings, grown for 14d on 0.5-strength Murashige and Skoog agar medium supplied with 25 µM selenite. Shown are the means and SEs of four samples from six plants each per treatment per line.

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

The transgenic Arabidopsis SL plants showed enhanced leaf SL enzyme activity levels, up to 6-fold for chloroplastic SL lines and up to 2-fold for cytosolic expressors (Fig. 2). The endogenous SL activity found in wild-type Arabidopsis plants (Figs. 2 and 3) may be explained by NifS-like proteins. Around 30% of this endogenous SL activity was chloroplastic (Fig. 3), and probably due to the plastidic NifS-like protein AtCpNifS (Pilon-Smits et al., 2002). The remainder of the endogenous SL activity may be mitochondrial, because a search of the Arabidopsis genomic database revealed the presence of two NifS-like proteins, one chloroplastic (At1g08490) and the other predicted to be mitochondrial (At5g65720; Kushnir et al., 2001). If these are the only two NifS-like proteins in Arabidopsis, then the cytSL transgenics likely contain a novel SL enzyme activity in their cytosol, whereas the cpSL transgenics contain an increase in SL activity of up to approximately 15-fold in their chloroplasts (wild-type chloroplast SL activity was approximately 30% of 0.05 units mg¹ protein, i.e. 0.015 units mg¹ protein, and cpSL1 SL activity was 0.25 units mg¹ protein). When stroma SL activity levels were compared between wild type and cpSL1 (Fig. 3), there was an approximately 15-fold increase in SL-specific activity.

Both the cytSL and cpSL transgenics incorporated approximately 50% less Se into proteins. This is in agreement with the hypothesis that the conversion of Se-Cys to Se⁰ and Ala will prevent nonspecific incorporation of Se-Cys into proteins. These results demonstrate that it is possible to prevent Se incorporation into proteins via genetic engineering and to redirect Se fluxes in the plant to a novel metabolic sink. This redirection of Se, away from incorporation into protein, had a profound effect on Se tolerance. The cytosolic SL transgenics were more Se tolerant, probably because replacement of Cys by Se-Cys in proteins is toxic, but Se⁰ is relatively inert. In contrast, the transgenic cpSL plants showed reduced Se tolerance. Thus, the intracellular location of the mouse SL enzyme mattered greatly, with cytosolic and plastidic locations leading to opposite effects on Se tolerance.

Why are the transgenic cpSL plants less tolerant to Se? Because the cpSL plants showed a normal phenotype in the absence of Se, the adverse effect of expression of the SL enzyme in the chloroplast appears to be related to SL activity on Se-Cys, presumably to the produced Se⁰. Perhaps Se⁰ is toxic to plastids because it interferes with aspects of plastid S metabolism, such as the biosynthesis of iron (Fe)-S clusters and thiamin (Li et al., 1990; Belanger et al., 1995). Because enzymes involved in S metabolism generally do not discriminate between analogous S and Se substrates, it is feasible that accumulated elemental Se can effectively compete with S⁰ for the plastidic enzymes involved in Fe-S cluster or thiamin synthesis.

The various transgenic SL lines showed an expression-related increase or decrease in Se tolerance. The negative correlation between Se tolerance and SL activity in the cpSL plants may reflect the degree of disruption of S metabolism by Se⁰. The positive correlation between Se tolerance and SL activity in the cytSL plants may reflect the degree of prevention of Se incorporation into protein. Although the cytSL plants showed enhanced Se tolerance, their growth was still significantly reduced by Se. They still showed incorporation of Se into protein, although it was 2-fold lower than in the wild type. Perhaps the amount of SL protein in the cytosol was not enough to completely prevent translational incorporation of Se into protein in the cytosol, and higher SL expression levels could further enhance Se tolerance. Alternatively, the Se toxicity experienced by the cytSL plants may be due to translational incorporation of Se into mitochondrial and plastidic proteins (e.g. Rubisco large subunit, a significant fraction of total plant protein, is plastid encoded).

When supplied with selenite, the cytSL and cpSL transgenics both showed enhanced shoot Se concentrations. The degree of Se accumulation was similar for most transgenic lines and not related with the SL expression level. Therefore, it appears that the introduction of a new Se sink triggers Se uptake by the plant, even at low levels and regardless of the intracellular compartment of the new Se sink. Although the SL transgenics accumulated significantly more Se from selenite compared with wild type, there were no differences when supplied with selenate. This difference in the effect of SL expression on shoot Se levels between selenate- and selenite-supplied plants may be explained by a slower conversion rate of selenate to Se-Cys relative to selenite-Se-Cys. In earlier studies with Brassica juncea plants, the reduction of selenate to selenite appeared to be a rate-limiting step for Se assimilation into organic compounds (de Souza et al., 1998; Pilon-Smits et al., 1999). If, as in B. juncea, selenate-supplied Arabidopsis plants accumulate approximately 95% selenate, whereas plants supplied with selenite accumulate approximately 100% organic Se, then expression of SL is not expected to be able to influence Se accumulation from selenate very much, because it is located downstream from the rate-limiting reaction. This may also explain why the differences in tolerance between SL and wild-type plants were more pronounced for Se-Cys and selenite than for selenate (Fig. 5).

There was a significant difference in tolerance between SL and wild-type plants, even to selenate. This indicates that the rate of production of Se-Cys from selenate by the sulfate assimilation pathway was high enough for the introduced SL activity to have a significant effect. This may be explained by the earlier observation that selenate is taken up and translocated by plants at a much faster rate than selenite (de Souza et al., 1998). Thus, it is feasible that even if only a small fraction of the shoot selenate is assimilated to Se-Cys, the actual amount of Se-Cys in the shoot is comparable with the amount of Se-Cys produced from selenite, because a much smaller fraction of the supplied selenite is taken up and translocated.

How useful would (over) expression of SL be for enhancing Se phytoremediation efficiency? The cytSL transgenics show promise in this respect, because they display both higher Se tolerance and higher shoot Se concentrations. The cpSL low expressor lines may also be useful because they contained the highest Se concentrations of all lines tested, without much reduction in Se tolerance. Of course, Arabidopsis would not be a suitable species for phytoremediation, but the same strategy may be used to create SL transgenics of other plant species. For this purpose, the creation and characterization of B. juncea SL transgenics are under way. In addition to their possible use for phytoremediation of Se-contaminated soil or water, these SL plants may also have enhanced nutritional value, if they contain elevated levels of the essential element Se.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Arabidopsis var. Columbia 0 seed was obtained from Dr. June Medford (Colorado State University, Fort Collins).

Plasmid Construction and Plant Transformation

The mouse (Mus musculus) SL cDNA was obtained as a NdeI-HindIII fragment in vector pET21d (Mihara et al., 2000). This vector was digested with HindIII and a linker with the sequence 5'AGCTGGATCC was inserted to replace the HindIII site with a BamHI site. Subsequently, the NdeI site was removed and replaced by NcoI using an oligonucleotide linker with the sequence 5' TAGCCATGGC to generate vector pESL-n, or by BstEII, using the oligonucleotide linkers 5' TAGGTTACCCT and 5' TAAGGGTAACC to generate vector pESL-b. For cytosolic expression, the SL sequence from pESL-n was cloned as an NcoI/BamH fragment under the control of a 35S-CAMV promoter into cloning vector pMOG18 (Sijmons et al., 1990) generating pSLY. In pSLY, the sequence, as confirmed by dideoxy sequencing, at the upstream fusion point, starting with the six bases for the NcoI recognition site, was: 5' ... . CC-ATG-GCt-atg-gac-gcg ... . 3' (the newly introduced start codon is underlined, the linker-derived sequence is in capital letters, and the original SL sequence in lowercase letters). This produces the amino acid sequence Met-Ala-Met-Asp-Ala-etc. The two underlined amino acids (Met and Ala) were added to the N terminus of SL to accommodate cloning in the expression vector. The sequence Met-Asp-Ala is the start of the mouse SL sequence. For chloroplastic expression, the SL sequence was subcloned as a BstEII/BamHI fragment in the pMOG18-derived vector pWA1 (Rensink et al., 1998) generating pSLC. In pSLC the sequence, as confirmed by dideoxy sequencing, at the upstream fusion point was 5'... gca atg gcc aca tac aag GTT ACC CTt atg gac ... 3'. The linker-derived sequence is in capitals, and the ferredoxin precursor-derived sequence is underlined. In the pSLC construct, SL is also under control of the CAMV35S promoter, but the protein is fused with the 48-amino acid-long transit sequence and the first seven amino acids of mature ferredoxin from Silene pratensis (Rensink et al., 1998). The protein sequence predicted from the DNA sequence would be: Met-Ala-44 amino acids-Ala-Met-Ala-Thr-Tyr-Lys-Val-Thr-Leu-Met-Asp-Ala etc. The ferredoxin precursor sequence is underlined, the ferredoxin transit sequence is given in italics, and the seven amino acids that will remain added to the SL sequence after cleavage of the transit sequence are given in bold face. The sequence Met-Asp-Ala is the start of the mouse SL sequence. Both the cytosolic and chloroplastic expression constructs have the NOS transcription terminator sequence. The gene constructs were cloned as EcoRI/HindIII fragments into plant binary vector pMOG23 (Sijmons et al., 1990) containing the nptII kanamycin resistance marker and the T-DNA border sequences generating vectors pSCY and pSCH, respectively. The sequences of the inserts were verified by dideoxy sequencing. The two constructs (Fig. 1) were transferred to Agrobacterium tumefaciens C58C1, which was used to transform Arabidopsis using the flower dip method (Clough and Bent, 1998).

Kanamycin-resistant lines were selected on 0.5-strength Murashige and Skoog agar medium containing 1% (w/v) Suc and supplemented with 50 µg mL¹ kanamycin and 50 µg mL¹ cefotaxime. Kanamycin-resistant seedlings were transplanted to soil and PCR was used to confirm the presence of the transgenes. Ten kanamycin-resistant lines per construct were confirmed by PCR to be transgenic SL lines. The following primers were used. The forward primer was directed against the 35S promoter, with sequence 5' CCT TCG CAA GAC CCT TCC TC 3'. The reverse primer was directed against the mouse SL gene and had sequence 5' TGA TCT CGG AGA CAG GCA TGA 3'. The plants were selfed, and homozygous seed batches from the T₂ generation were used for further experiments.

Enzyme Activity Assays

For measurement of SL enzyme activity, entire shoots were collected from 2-week-old seedlings grown on one-half-strength Murashige and Skoog agar medium (Murashige and Skoog, 1962). The samples were ground in liquid nitrogen and extracted with 1 mL g¹ fresh weight of extraction buffer containing 50 mM Tris (pH 8), 20% (w/v) glycerol, 2 mM EDTA, 0.5% (v/v) Triton X-100, 1 mM dithiothreitol (DTT), and 0.1 mM phenylmethylsulfonyl fluoride. SL enzyme activity was measured at 37°C in 0.12 M Tricine-NaOH (pH 7.9), 10 mM L-Se-Cys, 50 mM DTT, and 0.2 mM pyridoxal phosphate. The Se⁰ produced was measured with lead acetate as described by Esaki et al. (1982). Specific activity is expressed in units per milligram protein, with one unit of enzyme defined as the amount that catalyzes the formation of 1 µmol of product in 1 min.

Chloroplast Isolations and Immunoblotting

Chloroplast isolations were performed according to Rensink et al. (1998) using 3-week-old soil-grown plants. The purity of the chloroplast fractions from cytosolic contamination was confirmed by measuring phosphoenolpyruvate carboxylase activities in plant homogenate and chloroplast fractions, as described by Pilon-Smits et al. (1990). Proteins from plant homogenate (collected during chloroplast isolation) and isolated chloroplast fractions were separated by SDS-PAGE and transferred to nitrocellulose membranes. Immunostaining was used to visualize the mouse SL protein, using polyclonal antibodies raised against the mouse SL protein (Mihara et al., 2000).

Protein Se Content

Se content in protein was determined in shoots of 2-week-old seedlings grown on one-half-strength Murashige and Skoog medium with 5 µM sodium selenite. The seedlings were harvested, washed, and ground in liquid nitrogen. Aliquots of 1 g fresh weight were extracted in 5 mL g¹ buffer containing 100 mM NaCl, 50 mM Tris/HCl (pH 7.5), 0.5% (v/v) Triton X-100, 1 mM DTT, and 0.1 mM phenylmethylsulfonyl fluoride. The homogenate was cleared by centrifugation (7,500g for 10 min). A small sample was taken for protein determination, and the volume of the extract was measured. The proteins in the extract were precipitated by adding trichloroacetic acid to a final concentration of 15% (w/v). The mixture was incubated on ice for 30 min and then centrifugated for 20 min at 7,000g at 4°C in a glass tube. The pellet was washed with ice-cold acetone, dried, and dissolved in 1 mL of concentrated nitric acid. After acid digestion, the Se concentrations in these samples were determined by inductively coupled plasma atomic emission spectrometer (Pilon-Smits et al., 1999).

Tolerance Measurements

To determine Se tolerance, seedlings were grown for 2 weeks on vertical plates containing one-half-strength Murashige and Skoog agar medium with or without Se (at the concentrations indicated), and root length was measured. For measurement of Se accumulation, seedlings were grown for 2 weeks on horizontal 0.5-strength Murashige and Skoog agar plates, harvested, and Se concentration in shoot tissue was analyzed by inductively coupled plasma atomic emission spectrometer after acid digestion, as described by Pilon-Smits et al. (1999).

Data Analysis

The software package JMP-IN was used for statistical analyses (SAS Institute, Cary, NC). Pairs of means were compared statistically using Student's t tests. The statistically significant differences ( = 0.05) are indicated in the text.

Distribution of Materials

Upon request, all novel materials described in this publication will be made available in a timely manner for noncommercial research purposes, subject to the requisite permission from any third party owners of all or parts of the material. Obtaining any permissions will be the responsibility of the requestor.