To investigate the uptake and long-distance translocation of sulphate in plants, we have characterized three cell-type-specific sulphate transporters, Sultr1;1, Sultr2;1 and Sultr2;2 in Arabidopsis thaliana. Heterologous expression in the yeast sulphate transporter mutant indicated that Sultr1;1 encodes a high-affinity sulphate transporter (Km for sulphate 3.6 ± 0.6 μm), whereas Sultr2;1 and Sultr2;2 encode low-affinity sulphate transporters (Km for sulphate 0.41 ± 0.07 m m and ≥ 1.2 m m, respectively). In Arabidopsis plants expressing the fusion gene construct of the Sultr1;1 promoter and green fluorescent protein (GFP), GFP was localized in the lateral root cap, root hairs, epidermis and cortex of roots. β-glucuronidase (GUS) expressed with the Sultr2;1 promoter was specifically accumulated in the xylem parenchyma cells of roots and leaves, and in the root pericycles and leaf phloem. Expression of the Sultr2;2 promoter–GFP fusion gene showed specific localization of GFP in the root phloem and leaf vascular bundle sheath cells. Plants continuously grown with low sulphate concentrations accumulated high levels of Sultr1;1 and Sultr2;1 mRNA in roots and Sultr2;2 mRNA in leaves. The abundance of Sultr1;1 and Sultr2;1 mRNA was increased remarkably in roots by short-term stress caused by withdrawal of sulphate. Addition of selenate in the sulphate-sufficient medium increased the sulphate uptake capacity, tissue sulphate content and the abundance of Sultr1;1 and Sultr2;1 mRNA in roots. Concomitant decrease of the tissue thiol content after selenate treatment was consistent with the suggested role of glutathione (GSH) as a repressive effector for the expression of sulphate transporter genes.
Sulphur is an essential nutrient for plants. Uptake of sulphate in roots is the initial and crucial step to obtain sulphur for the synthesis of sulphur-containing organic compounds ( Leustek & Saito, 1999). Physiological studies have suggested that two or three sulphate transport systems having different affinities for sulphate are involved in sulphate uptake and translocation ( Kreuzwieser et al. 1996 ; Lass & Ullrich-Eberius, 1984; Vange et al. 1974 ). It has been postulated that various steps in these processes are mediated by transporter proteins expressed in different cell types ( Clarkson et al. 1993 ). Genes encoding several different isoforms of sulphate transporter that are likely to be involved in uptake and translocation have been cloned from Arabidopsis ( Takahashi et al. 1996 ; Takahashi et al. 1997 ; Takahashi et al. 1999a ; Takahashi et al. 1999b ; Yamaguchi et al. 1997 ). Smith et al. (1995 , 1997) have shown that expression of root-specific sulphate transporter genes is induced by sulphate starvation in Stylosanthes hamata and barley. However, cell-type-specific expression of these inducible transporters was not reported in those studies. We have previously shown that the sulphate transporter gene, Sultr2;1 (originally called AST68 by the clone name), is specifically expressed in the central cylinder of roots and in the vascular tissues of leaves in Arabidopsis ( Takahashi et al. 1997 ). The mRNA expression level of this gene in roots is also up-regulated during sulphate starvation, suggesting that Sultr2;1 makes a significant contribution to enhancement of the sulphate transport capacity in the vascular tissues of roots. In the present study, we have characterized the kinetic properties, cell-type specific expression and regulation of three functional sulphate transporters, Sultr1;1, Sultr2;1 and Sultr2;2 (originally called AST56 by the clone name) ( Takahashi et al. 1996 ), and used this information to ascribe possible roles to these transporters. We provide evidence that accumulation of Sultr1;1 high-affinity transporters in the root surface and Sultr2;1 low-affinity transporters in the root stele enhances the sulphate uptake capacity in Arabidopsis.
Studies with barley ( Smith et al. 1997 ) have indicated that concentrations of sulphate, cysteine and glutathione (GSH) in plant tissues decline when the expression of the sulphate transporter gene HVST1 increases during sulphate starvation. These data suggest that the expression of HVST1 is induced when the sulphur status is generally low. Recent reports have indicated that accumulation of GSH may act negatively on the expression of the sulphate transporter and ATP sulphurylase genes in sulphate-starved plants ( Lappartient et al. 1999 ; Vidmar et al. 1999 ). Experiments with sulphate-starved seedlings of maize suggest that cysteine is also likely to be a repressor for the regulation of these genes ( Bolchi et al. 1999 ). In this study, we demonstrate that increased expression of Sultr1;1 and Sultr2;1 mRNA in sulphate-starved Arabidopsis roots can be mimicked by the addition of selenate to sulphate-replete medium. This positive regulation occurred in parallel with the decrease in tissue GSH content, suggesting a GSH-dependent de-repressive control for expression of these sulphate transporter isoforms responsible for the uptake and translocation of sulphate in Arabidopsis.
This is the first comprehensive analysis of sulphate transporter genes in Arabidopsis leading to characterization of their physiological functions based upon their kinetic properties, cell-type-specific expression and regulation.
cDNA and genomic sequences encoding sulphate transporters in Arabidopsis
Accumulation of nucleotide sequence information in the database enabled us to isolate cDNAs and genes encoding sulphate transporters from Arabidopsis. Accession numbers of the cDNA and genomic clones isolated from the libraries, and the corresponding BAC and P1 clones in the database are summarized in Table 1. We have designated the sulphate transporter genes as ARAth;Sultr or Sultr, according to the rules of Commission of Plant Gene Nomenclature ( http://mbclserver.rutgers.edu/CPGN/).
Table 1. Sulphate transporter genes in Arabidopsis
cDNA clones for Sultr2;1, Sultr2;2, Sultr3;1 and Sultr4;1 were isolated by screening a cDNA library with expressed sequence tag (EST) clones, 142F20T7, 130L5T7, 76E7T7 and H2H3T7, respectively ( Takahashi et al. 1996 ; Takahashi et al. 1997 ; Takahashi et al. 1999a ; Takahashi et al. 1999b ). EST clone G8B12T7 encoded the full-length cDNA clone, Sultr3;2 ( Takahashi et al. 1999b ). cDNA clones for Sultr1;1 and Sultr3;3 were isolated by RT–PCR using the nucleotide sequences of BAC clones T3F12 and T26J12, respectively. The protein sequences of all these Arabidopsis isoforms were predicted to contain 12 membrane-spanning domains resulting in similar topologies to those found for the barley and Stylosanthes transporters ( Smith et al. 1995 , 1997). Phylogenic relationships between the encoded protein sequences indicated that these seven isoforms could be classified into at least four groups ( Fig. 1). Sultr1;1 belongs to group 1 which contains the root-specific high-affinity subtypes, SHST1, SHST2 ( Smith et al. 1995 ) and HVST1 ( Smith et al. 1997 ). Sultr2;1 and Sultr2;2 fell into group 2 with the low-affinity transporter SHST3 ( Smith et al. 1995 ). Sultr3;1, Sultr3;2 and Sultr3;3 formed group 3, the members of which were only expressed in leaves ( Fig. 7). The roles of these transporters that make up group 3 are not well understood at present. Sultr4;1 belongs to group 4 which also contains the hypothetical sulphate transporter protein, slr1776, from the cyanobacterial genomic sequence. Members of group 4 have the least similarity to the other plant sulphate transporters. The Sultr4;1 protein was shown to be localized in chloroplasts in our previous study ( Takahashi et al. 1999a ). Prediction of subcellular localization by the computer program psort ( http://psort.ims.u-tokyo.ac.jp/) suggests that Sultr4;1 is probably located in the thylakoid membranes, but its function and role in the sulphur assimilation pathway is unresolved at present.
Genomic clones for Sultr2;1, Sultr2;2 and Sultr3;1 were isolated from a genomic library and sequenced as described previously ( Takahashi et al. 1997 ; Takahashi et al. 1999b ). Recently, genomic sequences for Sultr1;1, Sultr2;2, Sultr3;1, Sultr3;2, Sultr3;3 and Sultr4;1 were identified in the BAC and P1 clones released from the Arabidopsis genome sequencing project ( Table 1). To determine the complete exon/intron structure of Sultr4;1, a DNA fragment for the C-terminal region of Sultr4;1 (exon d3 to e3) was amplified from Arabidopsis genomic DNA by PCR. Map positions of Sultr2;1, Sultr3;1 and Sultr3;2 were separately determined by using recombinant inbred lines ( Lister & Dean, 1993). Comparison of the cDNA and corresponding genomic sequences revealed that insertion sites of introns in these seven sulphate transporter genes are fairly well conserved ( Fig. 2). Four sites were conserved in all transporters and the coding sequences were divided into five regions (a, b, c, d, e) by these four common insertion of introns. Sultr3;1 and Sultr3;2 had exactly the same exon/intron organization. Sultr2;1 and Sultr2;2 contained similar patterns. Similarities of genomic structures suggest that these genes may have closely related functions. Interestingly, classification of sulphate transporter genes with the genomic structure was generally consistent with similarities between the protein sequences ( Fig. 1), kinetic properties ( Fig. 3) and organ-specific expression of mRNA ( Fig. 7).
Expression of Arabidopsis sulphate transporters in a yeast mutant lacking the capacity to take up sulphate
To confirm the functions of the proteins encoded by the Arabidopsis sulphate transporter genes, cDNAs were expressed in the yeast mutant, CP154-7A ( Cherest et al. 1997 ). CP154-7A lacking both the SUL1 and SUL2 sulphate transporter genes has a very low sulphate uptake capacity and requires homocysteine as a sulphur source. cDNA fragments encoding the Arabidopsis transporters were cloned into the expression vector, pYE22m ( Ashikari et al. 1989 ). pYE22m contains a promoter region of the yeast glyceraldehyde-3-phosphate dehydrogenase gene which enables constitutive expression of the inserted gene product. Expression of Sultr1;1, Sultr2;1 and Sultr2;2 rescued the growth of CP154-7A on a medium containing 0.1 m m of sulphate as a sole sulphur source, suggesting that these cDNA clones encode sulphate transporter proteins that are functional in yeast plasma membranes ( Fig. 3a). The kinetics of sulphate uptake by these yeast transformants was determined by short-term sulphate uptake studies ( Fig. 3b,c). These revealed that Sultr1;1 is a high-affinity sulphate transporter (Km for sulphate 3.6 ± 0.6 μm) and Sultr2;1 is a low-affinity sulphate transporter (Km for sulphate 0.41 ± 0.07 m m). Similar short-term uptake experiments for Sultr2;2 have not yielded a reliable Km for this transporter due to its low Km and very low uptake rates. However, those experiments did indicate that Sultr2;2 is a low-affinity sulphate transporter with a Km > 1.2 m m. Sultr3;1, Sultr3;2, Sultr3;3 and Sultr4;1 when expressed in CP154-7A in the same way, failed to complement the phenotype of the mutant when supplied with 0.1–1 m m sulphate in the medium (data not shown). Inability of these transporters to complement the mutant is presumably due to even lower uptake rates and/or lower affinities for sulphate than those for Sultr2;2. An alternative possibility is that they may not have been correctly recognized as plasma membrane-localizing proteins in yeast cells. Sultr4;1, the chloroplast-localizing isoform, may have similar problems when expressed in the heterologous yeast system.
Cell type-specific expression of Sultr1;1, Sultr2;1 and Sultr2;2 in Arabidopsis
The cell type-specificity of Sultr1;1, Sultr2;1 and Sultr2;2 was studied in transgenic Arabidopsis plants expressing promoter–reporter fusion gene constructs. For each construct, more than 10 independent transgenic lines were analysed to confirm the specificity of the expression.
Figure 4 shows the fluorescence of GFP in transgenic Arabidopsis plants expressing the Sultr1;1 promoter–GFP fusion gene. A DNA fragment flanking 1944 bp upstream of the translational initiation site of Sultr1;1 was amplified by PCR and fused to the coding sequence of GFP ( Chiu et al. 1996 ). This fusion construct was introduced into Arabidopsis plants by Agrobacterium-mediated transformation ( Bechtold et al. 1993 ). Localization of GFP was described in relation to the anatomy of the developing roots of Arabidopsis ( Dolan et al. 1993 ). GFP was found in the lateral root cap ( Fig. 4A) and epidermal cell layers of the trichoblast (root hair) differentiation zone ( Fig. 4B) in the root tips. In mature roots in which the vascular system had fully developed, GFP was also expressed in cells of the cortex ( Fig. 4C). GFP was also localized in root hairs ( Fig. 4D), but no signals were found in the endodermis and central cylinder in any of the root sections examined. Very weak signals were found in the hydathode of cotyledons ( Fig. 4E). When plants were grown on low external sulphate concentration, fluorescence of GFP occurred in auxiliary buds of leaves ( Fig. 4F). These signals in the above-ground tissues were weaker than those in roots. Spatial expression patterns of GFP suggested that the high-affinity sulphate transporter Sultr1;1 is responsible for the uptake of sulphate from the environment at the outermost cell layers in roots.
Cell-type-specific expression of Sultr2;1 was studied with two promoter–reporter constructs using the uidA gene ( Jefferson et al. 1987 ) as a reporter. First, an XhoI–HindIII fragment that covers the region from positions −2990 to −45 of the translational initiation site of Sultr2;1 was fused to the coding sequence of the uidA gene. The second construct contained an XhoI–NcoI fragment covering the region from −2990 to +90 of Sultr2;1 fused to the uidA gene. Accumulation of the uidA gene product, GUS, was observed by staining the plants with 5-bromo-4-chloro-3-indolyl-β- d-glucuronide (X-gluc) ( Fig. 5). Transgenic plants with the two different constructs showed exactly the same tissue-specific accumulation patterns of GUS. Figure 5 shows the results of the first fusion construct. The results from these promoter–GUS fusion experiments were identical with those from the previously reported in situ hybridization studies using Sultr2;1 mRNA as a riboprobe ( Takahashi et al. 1997 ). GUS protein was accumulated in the vascular tissues in both leaves and roots ( Fig. 5A,B). In leaves, they were located in the xylem parenchyma and phloem cells ( Fig. 5C), but in roots they were located in the xylem parenchyma and pericycle cells ( Fig. 5D). Xylem was not stained in either leaves or roots. No staining was observed in the epidermis and endodermis of roots. These data, together with the in situ hybridization data ( Takahashi et al. 1997 ), suggest that the Sultr2;1 protein is responsible for the uptake of sulphate form the apoplastic solution within the vascular bundle and is possibly involved in root-to-shoot transportation of sulphate in Arabidopsis.
Transgenic plants expressing the Sultr2;2 promoter–GFP fusion gene were constructed in a similar manner to those for Sultr1;1. A DNA fragment covering the region from position −3384 to +66 of the translational initiation site of Sultr2;2 was fused to the coding sequence of GFP ( Chiu et al. 1996 ). GFP driven by the Sultr2;2 promoter was specifically expressed in the phloem in roots ( Fig. 6A,B), but no fluorescence was observed in the phloem of vascular bundles in leaves. Uptake of sulphate into phloem is suggested to involve two different organ-specific transporters, Sultr2;1 in leaves ( Fig. 5C) and Sultr2;2 in roots ( Fig. 6A,B). In leaf tissues, GFP fluorescence was localized in the vascular bundle sheath cells ( Fig. 6C), suggesting that the Sultr2;2 low-affinity sulphate transporter is involved in the distribution of sulphate from the vascular bundle to the palisade cells.
Regulation of Arabidopsis sulphate transporter genes by sulphur supply
To investigate the effects of sulphate supply on the regulation of genes encoding sulphate transporters, steady-state mRNA levels were estimated by RT–PCR ( Fig. 7a,b). Plants were grown vertically for 2 weeks on agarose plates containing 10 μm, 100 μm or 1.7 m m of sulphate ( Fig. 7a). Plants on 10 μm sulphate media were stunted and chlorotic. Short-term response to sulphate starvation was examined by transferring 3-week-old plants onto agarose medium containing no sulphate ( Fig. 7b).
Sultr1;1 mRNA was over 20-fold more abundant in roots grown on 10 or 100 μm sulphate than it was in the control plants grown on 1.7 m m sulphate ( Fig. 7a). There was a lower increase of Sultr1;1 mRNA in the leaves of these plants than in roots. When plants were transferred to medium containing no sulphate, the transcript level of Sultr1;1 in roots increased rapidly within 24 h ( Fig. 7b). The high abundance and rapid response of expression of Sultr1;1 mRNA to the absence of sulphate in the medium suggests that this high-affinity transporter is regulated at the transcriptional level in a demand-derived manner by the sulphur status of the plant.
The amounts of Sultr2;1 mRNA in roots were estimated to be over fivefold more abundant under 10 or 100 μm sulphate conditions than in roots of control plants ( Fig. 7a). When plants were subjected to a short-term sulphate starvation stress ( Fig. 7b), steady-state levels of Sultr2;1 mRNA increased in the roots. Expression of Sultr2;1 in the xylem parenchyma cells of the root stele ( Fig. 5) suggests that this functional low-affinity transporter is involved in the internal translocation of sulphate from roots to leaves. In contrast to the high level of induction in roots, steady-state levels of Sultr2;1 mRNA in leaves were lower in plants supplied with 10 or 100 μm sulphate ( Fig. 7a), or remained unchanged by a short-term sulphate starvation ( Fig. 7b). The mRNA levels of Sultr2;1 in roots and leaves were inversely regulated in response to the decrease of sulphate concentration in the medium ( Fig. 7a). A similar response to sulphate deprivation was reported for the expression of the SHST3 sulphate transporter in roots and leaves ( Smith et al. 1995 ).
Sultr2;2 mRNA accumulated abundantly in the leaves of plants continuously grown on 10 or 100 μm sulphate ( Fig. 7a). During a short-term sulphate starvation stress, the steady-state level of Sultr2;2 mRNA did not change ( Fig. 7b). The expression of the Sultr4;1 responded similarly to the sulphur status of the plant. There were high levels of Sultr4;1 mRNA in the leaves of plants continuously grown on 10 and 100 μm sulphate ( Fig. 7a) but there was no response to short-term sulphate starvation ( Fig. 7a). Sultr2;2 in the vascular bundle sheath cells ( Fig. 6) and Sultr4;1 in chloroplasts ( Takahashi et al. 1999a ) may play important roles in maintaining sulphate status when plants are grown continuously under sulphur-less conditions. mRNA levels corresponding to the other three leaf-specific isoforms did not change markedly in response to sulphur deficiency ( Fig. 7a,b).
Uptake of sulphate correlates with the increase in mRNA levels of Sultr1;1 and Sultr2;1 in roots
When Arabidopsis plants were grown on 1.7 m m sulphate medium containing 0.1 m m selenate, the mRNA levels of Sultr1;1 and Sultr2;1 in roots were markedly increased ( Fig. 7c). The uptake rate of 35S-labelled sulphate was increased twofold ( Table 2) by this treatment, which coincided with the increase in the mRNA levels of these sulphate transporter genes in roots. These responses were similar to those elicited by a short-term sulphate starvation ( Fig. 7b; Smith et al. 1997 ). As a result, the expression of Sultr1;1 and Sultr2;1 was up-regulated by the addition of selenate and this led to increased rates of sulphate uptake.
Table 2. Effects of selenate on sulphate uptake rates, and sulphate and thiol contents
Cys (pmol mg−1 FW)
γ-GluCys (pmol mg−1 FW)
GSH (pmol mg−1 FW)
SO42– (nmol mg−1 FW)
Uptake of 35SO42– (pmol mg−1 FW h−1)
Thiol and sulphate contents were determined in 3-week-old plants transferred and grown for 48 h on the control GM medium or on the GM medium containing 0.1 m m of selenate (+ Se) as in Fig. 7(c) (n = 3). Uptake rates of 35S-labelled sulphate were measured in 8-day-old plants transferred and grown for 48 h on the control GM medium or on the GM medium containing 0.1 m m of selenate (+ Se). The radioactivity of incorporated 35S was measured in whole plants (n = 4).
17.8 ± 2.2
6.7 ± 3.3
700.4 ± 78.6
9.5 ± 0.2
24.8 ± 4.6
9.4 ± 0.9
4.2 ± 1.0
247.7 ± 4.1
9.4 ± 0.6
10.8 ± 0.6
2.9 ± 1.2
435.0 ± 38.0
33.9 ± 1.9
53.1 ± 4.9
9.6 ± 0.6
1.5 ± 0.4
158.7 ± 21.4
4.9 ± 0.1
In selenate-treated plants, the total sulphate pool increased about twofold by this treatment ( Table 2). More precisely, sulphate contents increased 3.5-fold in leaves and decreased 50% in roots by this treatment. These data suggest that internal movement of sulphate from roots to leaves occurred in addition to the overall increase in the sulphate uptake. This movement may be related to the function of Sultr2;1 over-expressed in roots.
Selenate is a toxic analogue of sulphate, and is assumed to be assimilated to selenium-containing analogues of sulphur metabolites. Selenate treatment generally decreased the levels of the sulphur assimilation metabolites, cysteine, γ-glutamylcysteine and GSH in both leaves and roots ( Table 2). These data suggest that addition of selenate negatively affects the synthesis of sulphur assimilatory metabolites as in the sulphate-starved plants. It appears likely that exposure to selenium reduces the level of GSH, which is proposed to be a key sulphur-containing compound involved in the repression of sulphate transporter genes ( Lappartient et al. 1999 ; Vidmar et al. 1999 ).
Sulphate is the form of sulphur taken up from the soil by roots and distributed to the above-ground tissues ( Leustek & Saito, 1999). A number of distinct steps are involved in the process of sulphate uptake by roots and translocation to leaf tissues. These steps include uptake into the root symplast from the soil solution, radial transport across the root to the central stele, unloading into xylem vessels for translocation to the shoot, discharge from xylem vessels into the apoplast of the stelar cells in leaves, and uptake into the symplast of leaf cells. Many of these steps require the transport of sulphate across plasma membranes. Phylogenetic analyses based upon sequence comparisons suggest that the plant sulphate transporters may be grouped into at least four groups that appear to have similar functional characteristics and tissue specificity. This paper reports data on the phylogenetic grouping, localization of expression and the regulation of genes encoding the Sultr1;1, Sultr2;1 and Sultr2;2 sulphate transporters from Arabidopsis and the kinetics of their protein products. These data enable these transporters to be allocated to specific steps in the uptake and translocation process and likely roles to be ascribed to them.
Initial uptake of sulphate from the soil solution is mediated by the high-affinity sulphate transporter, Sultr1;1
Uptake of sulphate from the soil solution into plant roots requires transport of sulphate from the apoplast surrounding the root epidermal and cortical cells across the plasma membrane into the symplast. Data on the cell-type-specific localization of expression of Sultr1;1 ( Fig. 4) indicate that it is ideally located in the root tips and external cell layers of roots for uptake of sulphate from the soil solution. Root tips continuously grow into soil that is not depleted of sulphate, and the long root hairs associated with epidermal cells considerably expand the surface area of the root and the volume of soil from which the root can draw sulphate. Where there is no suberized exodermis, the solution in the apoplast can move radially through the cell walls and intercellular spaces of epidermal and cortical cells as far as the endodermis where entry to the central stele is restricted by the suberized barrier of the Casparian strip. However, there are reports of apoplastic movement being restricted in young tissues ( Enstone & Peterson, 1992). The expression of Sultr1;1 appears to coincide with the root tissues that are exposed to solution in the apoplast. In young root tissues, Sultr1;1 is expressed in root tips and epidermal cells, whereas in more mature root tissues it is also expressed in cortical cells that are presumably surrounded by apoplastic solution. Similar kinds of developmentally controlled expression have been reported for nitrate ( Huang et al. 1996 ) and potassium ( Lagarde et al. 1996 ) transporters in roots.
Expression of high-affinity sulphate transporters in group 1 has been shown to be similarly regulated by sulphate starvation ( Fig. 7a,b) ( Bolchi et al. 1999 ; Smith et al. 1995 , 1997; Vidmar et al. 1999 ). Data from the study by Smith et al. (1997) indicate that either sulphate or the products of sulphate assimilation such as cysteine or GSH may be involved in the feedback regulation of the HVST1 high-affinity sulphate transporter gene. The end result of this regulation is that the number of high-affinity sulphate transporters in the plasma membranes of the root cells in close contact with the soil solution are considerably increased when the general sulphur status is declined. This would increase the opportunity for the root to capture any sulphate ions that did become available in the depleted soil solution. The corollary to this is that such feedback regulation also reduces the number of transporters available for sulphate uptake when the plant contains adequate sulphur. This would prevent accumulation of unnecessary sulphate by the plant thus maintaining sulphate chemostasis in the plant tissues and conserving energy otherwise expended on active sulphate uptake. More recently, data have been presented indicating that GSH may be the negative effector metabolite that represses the expression of the sulphate transporter, Sultr2;1, and ATP sulphurylase, APS1, in Arabidopsis ( Lappartient et al. 1999 ). The barley sulphate transporter, HVST1, is also similarly regulated by GSH in sulphate-starved roots ( Vidmar et al. 1999 ). Data on the regulation of Sultr1;1 by selenate suggest that the expression of this gene is induced by decrease of the tissue GSH level in sulphate-replete plants ( Fig. 7c, Table 2).
All of these regulatory properties of the members of group 1 are consistent with those expected for primary sulphate transporters involved in the initial uptake of sulphate from the soil solution and are supported by physiological data on the effects of plant sulphur status on sulphate uptake rates ( Clarkson et al. 1983 ; Smith et al. 1997 ). Taken together, the kinetic data ( Fig. 3b), the localization data ( Fig. 4) and the data on regulation of the gene encoding Sultr1;1 ( Fig. 7a,b) provide compelling evidence that Sultr1;1 is the primary high-affinity transporter responsible for uptake of sulphate from the external soil solution into the root symplast. Phylogenetic relationships and the information that is available on the kinetics and regulation of members of group 1 ( Fig. 1) suggest that all members of this group are likely to be high-affinity root sulphate transporters that play a similar role to Sultr1;1.
Translocation of sulphate from roots to leaves is mediated by the two low-affinity sulphate transporters, Sultr2;1 and Sultr2;2
Once inside the root symplast, sulphate can move radially to the central stele through the cortical and endodermal cells via plasmodesmata connections without traversing cell membranes. From within the stele, sulphate is then loaded into xylem vessels for transfer to the shoots. This process involves efflux through plasma membranes of millimolar concentrations of sulphate from active, living cells of the vascular tissues. Channels or low-affinity transporters are likely to be involved in this process. It is possible that Sultr2;1 mediates this transfer. It is well placed to serve this function being expressed in cells of the xylem parenchyma immediately surrounding xylem vessels and in cells of the pericycle ( Fig. 5). Its expression in roots is up-regulated during sulphate deficiency ( Fig. 7a,b) and it has a Km for sulphate of 0.41 m m. However, heterologous expression in yeast indicates that Sultr2;1 moves sulphate into yeast cells ( Fig. 3c). This poses the question of a possible role for a low-affinity transporter involved in the influx of sulphate into xylem parenchyma and pericycle cells of vascular tissues. One possibility is that Sultr2;1 serves as a scavenger re-absorbing sulphate from millimolar concentrations leaked into intercellular spaces and cell walls within the stele. Such a function could be particularly important in optimizing the amount of sulphate transferred to shoots during sulphate deficiency and thus provide a reason for the observed up-regulation of Sultr2;1 in roots during sulphur deprivation ( Fig. 7a,b). This function is indicated by an apparent movement of sulphate from roots to leaves ( Table 2) with increase of the mRNA level of Sultr2;1 in roots ( Fig. 7c) by the selenate treatment. The data from the selenate treatment also suggest that expression of Sultr2;1 in roots is possibly regulated in a GSH-dependent manner as in the case of Sultr1;1. For efflux of sulphate from cells of the xylem parenchyma, a different transport system may be responsible. Studies on potassium release into xylem vessels of Arabidopsis indicate that this is accomplished by SKOR, a potassium-selective voltage-regulated outward-rectifying channel ( Gaymard et al. 1998 ). Like Sultr2;1, SKOR is expressed in cells of the xylem parenchyma and pericycle, but the SKOR protein mediates potassium efflux from these cells.
Sultr2;1 in the above-ground tissues could play a similar role, ensuring that any excess sulphate is transported back into xylem parenchyma cells for efflux into xylem vessels. However, there would need to be a balance in leaves between the amount of sulphate used for assimilation and that re-adsorbed for further transport. Sulphate discharged from xylem vessels in the vascular bundles of minor veins in leaves would need to be transferred past the xylem parenchyma cells so that it could be used for assimilation in leaf palisade and mesophyll cells. It might be expected that this balance would be particularly important during periods of sulphur stress and could be achieved by the observed down-regulation of the expression of the Sultr2;1 and SHST3 genes of group 3 ( Fig. 1) in leaves during sulphur deficiency ( Fig. 7a, Smith et al. 1995 ). The other tissue in which Sultr2;1 is expressed is leaf phloem ( Fig. 5). This suggests a role in phloem loading from millimolar concentrations of sulphate within the leaf vascular bundles for transfer to other organs. One of the characteristic symptoms of sulphur deficiency is an apparent restricted mobility of internal sulphur compounds that results in chlorotic yellowing of the younger leaves whilst the older leaves retain their green coloration. Reduced phloem loading in leaves, due to down-regulation of the expression of genes such as Sultr2;1 or SHST3 during sulphur deficiency, would be consistent with such symptoms.
Sultr2;2 appears to have a lower affinity for sulphate than Sultr2;1. Localization data indicated that Sultr2;2 plays a role in the transport of sulphate via root phloem. In leaves, however, Sultr2;2 was expressed in the bundle sheath cells surrounding the vascular tissues of leaf veins ( Fig. 6). This suggests a role in the uptake of sulphate released from xylem vessels at millimolar concentrations for transfer to the primary sites of assimilation in leaf pallisade and mesophyll cells. The balance suggested above between allocation of sulphate discharged in leaf vascular tissues and uptake by xylem parenchyma cells and bundle sheath cells could be accomplished by the observed regulation of genes encoding Sultr2;1 and Sultr2;2 ( Fig. 7a). During sulphur stress, this balance would move towards uptake by bundle sheath cells in order to maintain a supply of sulphate to leaf cells for assimilation. The observed up-regulation of the expression of Sultr2;2 in bundle sheath cells and down-regulation of the expression of Sultr2;1 in leaf xylem parenchyma cells during sulphur stress could achieve such a balance.
Genes encoding sulphate transporters Sultr3;1, Sultr3;2 and Sultr3;3 from group 3 ( Fig. 1) appear to be exclusively expressed in leaves and their expression does not appear to be significantly modulated by the sulphur status of the plant ( Fig. 7). We have no definitive information on the roles of these transporters in translocation of sulphate in leaves at present. Sultr4;1 from group 4 ( Fig. 1) is expressed in chloroplasts in leaves ( Takahashi et al. 1999a ) and is likely to be involved in the transport of sulphate to the initial site of reduction. The role of Sultr4;1 in roots ( Fig. 7) has not yet been elucidated.
Growth of Arabidopsis
Arabidopsis thaliana ecotype Columbia was used for all experiments. Wild-type and transgenic plants were grown on GM solid medium ( Valvekens et al. 1988 ) at 22°C under 16 h/8 h light and dark cycles. To avoid contamination of sulphate from agarose, electrophoretic-grade agarose LO3 (Takara) was used at 0.8% w/v. Sulphate-deficient medium was prepared by replacing sulphate salts contained in Murashige–Skoog medium ( Murashige & Skoog, 1962) with equivalent chloride salts. For long-term sulphate starvation treatment ( Fig. 7a), seeds were sown on solid medium containing 10 or 100 μm of sulphate as a sulphur source, and grown vertically for 2 weeks. For short-term sulphate starvation ( Fig. 7b) and selenate treatments ( Fig. 7c), 3-week-old plants vertically grown on GM media were transferred either to the sulphur-deficient GM medium with no sulphate or GM medium containing 0.1 m m of selenate.
Isolation of cDNA and genomic clones
cDNA clones for Sultr3;1, Sultr2;2, Sultr2;1 and Sultr4;1 were isolated by screening a λgt11 cDNA library with expressed sequence tag (EST) clones, 76E7T7, 130L5T7, 142F20T7 and H2H3T7, as described previously ( Takahashi et al. 1996 ; Takahashi et al. 1997 ; Takahashi et al. 1999a ; Takahashi et al. 1999b ). EST clone G8B12T7 was fully sequenced and designated Sultr3;2 ( Takahashi et al. 1999b ). cDNAs for Sultr3;3 and Sultr1;1 were isolated by RT–PCR according to the nucleotide sequences of BAC clones, T26J12 (accession no. AC002311) and T3F12 (accession no. AC002983), respectively. For RT–PCR, total RNA was extracted from leaves and roots of Arabidopsis by a phenol/SDS method and precipitated by LiCl following the standard protocol ( Ausubel et al. 1994 ). Reverse transcription was carried out at 42°C in a 10 μl solution containing 1 μg of root total RNA, 10 pmol of SMART cDNA synthesis primer (Clontech), 1 m m of dNTP and 20 units of Moloney murine leukaemia virus (M-MLV) reverse transcriptase (Promega). The coding sequences of Sultr1;1 and Sultr3;3 cDNAs were amplified by PCR from the first-strand cDNA using oligonucleotide primers as follows: Sultr3;3-FE (5′-CTGTTGAATTCTTAGGAAAACCAATTAATG-3′) and Sultr3;3-RE (5′-CCGGAATTCTTAAACGTTACTGAGAGATGG-3′) for Sultr3;3; Sultr1;1-FE (5′-CCGGAATTCATGTCCGGGACTATTAATC-CC-3′) and Sultr1;1-RE (5′-CCGGAATTCTTAAGTTTGTTGCTCAGCC-3<normal′) for Sultr1;1. PCR was carried out using ExTaq DNA polymerase (Takara). The amplified EcoRI-ended fragments were cloned into the EcoRI site of pBluescript II SK– (Stratagene) and fully sequenced on both strands.
Genomic clones for Sultr3;1, Sultr2;2 and Sultr2;1 were isolated by screening a λEMBL3 genomic library (Clontech) with corresponding cDNA clones as described previously ( Takahashi et al. 1997 ; Takahashi et al. 1999b ). Genomic sequences for Sultr3;1, Sultr2;2, Sultr3;2, Sultr4;1, Sultr3;3 and Sultr1;1 were identified in the BAC and P1 clones from the database ( Table 1). The exon/intron structure of the C-terminal region of Sultr4;1 (exon d3 to e3) was determined in a DNA fragment amplified from Arabidopsis genomic DNA by PCR using oligonucleotide primers Sultr4;1G-FE (5′-CCTGAATTCTTTTCTGCATACCCAGCTACA-3′) and Sultr4;1-RE (5′-ATTGAATTCCCTTTATTTCTCCACTGATAA-3′). PCR was carried out using ExTaq DNA polymerase (Takara). The amplified EcoRI-ended fragments were cloned into the EcoRI site of pBluescript II SK– (Stratagene) and fully sequenced on both strands.
Expression and sulphate uptake measurements in yeast
The coding regions of the Sultr2;2 and Sultr2;1 cDNAs were amplified by PCR as BamHI-ended fragments using oligonucleotide primers as follows: Sultr2;2-FB (5′-CTTGGATCCATGGGCAT-AGAGCTTCAGAAT-3′) and Sultr2;2-RB (5′-ATTGGATCCATATCA-ACACAACTCGTGGGA-3′) for Sultr2;2; Sultr2;1-FB (5′-CTTGGATC-CATGAAAGAGAGAGATTCAGAG-3′) and Sultr2;1-RB (5′-ATT-GGATCCTGGTCCTTTGAAAACTGTTTC-3′) for Sultr2;1. These BamHI-ended cDNA fragments and the EcoRI-ended fragment of Sultr1;1 isolated by RT–PCR described above were cloned into the BamHI and EcoRI sites of the expression vector, pYE22m ( Ashikari et al. 1989 ). The resulting plasmids were transferred into the Saccharomyces cerevisiae mutant strain, CP154-7A (Mataa, his3, leu2, ura3, ade2, trp1, sul1::LEU2, sul2::URA3) ( Cherest et al. 1997 ) by the lithium acetate method ( Gietz et al. 1992 ). Trp+ transformants were selected on SD minimal media ( Sherman, 1991) containing 20 g l−1 of glucose, 0.25 m m of homocysteine and required amino acids. Uptake of 35S-labelled sulphate was measured by the method described by Smith et al. (1995) .
Construction of transgenic Arabidopsis expressing promoter–reporter fusion genes
The promoter region of Sultr1;1 flanking 1944 bp upstream of the translational initiation site was amplified by PCR as a BamHI–NcoI fragment from Arabidopsis genomic DNA using oligonucleotide primers Sultr1;1G-FB (5′-CGCGGATCCATTTTCATGTCCAAAG-TTTCC-3′) and Sultr1;1G-RN (5′-GCATACCATGGTTGCTGAAAA-TTAGTTGCC-3′). The promoter region of Sultr2;1 is derived from the genomic clone isolated from the λEMBL3 library. XhoI–HindIII and XhoI–NcoI fragments, which contain genomic regions for Sultr2;1 from −2990 to −45 and from −2990 to +90, respectively, were used for the preparation of two fusion gene constructs. The numbers indicate the nucleotide positions from the translational initiation site of Sultr2;1. The promoter region of Sultr2;2 is derived from the genomic clone isolated from the λEMBL3 library. The 5′-end of the Sultr2;2 genomic clone which covers the region from positions −3384 to +66 was amplified by PCR as an XhoI–BamHI fragment using a primer for the λEMBL3 vector arm sequence and a reverse primer specific for Sultr2;2 as follows: EMBL3-XbT7 (5′-GCTCTAGATAATACGACTCACTATAGGGAG-3′) and Sultr2;2G-RB (5′-ACGCGGATCCCATGGGCTCTTCTGCAGG-AC-3′). The amplified DNA fragments were fully sequenced to confirm their identity.
The BamHI–NcoI fragment of Sultr1;1 was cloned in the place of the cauliflower mosaic virus 35S (CaMV 35S) promoter sequence of the GFP expression vector, pTH2 ( Chiu et al. 1996 ). For construction of Sultr2;1–uidA cassettes, the coding sequence of uidA ( Jefferson et al. 1987 ) was amplified by PCR as HindIII–BamHI or NcoI–BamHI fragments using GUS-FH (5′-GCCAAGCTTATGTTACGTCCTGTAGAAACC-3′) or GUS-FN (5′-CAGTACCATGGTACGTCCTGTAGAAACCCC-3′) as forward primers, and GUS-RB (5′-CCTGGATCCGATTCATTGTTTGCCTCCC-TG-3′) as a reverse primer. The XhoI–HindIII and XhoI–NcoI fragments of Sultr2;1 were fused with these uidA fragments. The CaMV 35S promoter sequence in the expression vector pHTT202 ( Teeri et al. 1989 ) was replaced with the Sultr1;1–GFP and Sultr2;1–uidA fusion cassettes. For construction of the Sultr2;2–GFP fusion gene, the coding sequence of uidA in pBI101 (Clontech) was first replaced with the coding sequence of GFP. The XhoI–BamHI fragment of Sultr2;2 was cloned between the SalI and BamHI sites of this vector.
The intermediate plasmids of the Sultr1;1–GFP and Sultr2;1–uidA fusion genes were introduced in the Ti plasmid pGV3850 of Agrobacterium tumefaciens C58C1 ( Zambryski et al. 1983 ) by electrotransformation. The binary plasmid of the Sultr2;2–GFP fusion gene was transferred in Agrobacterium tumefaciens GV3101 (pMP90) ( Koncz & Schell, 1986) by the freeze–thaw method. Arabidopsis plants were transformed with the in planta transformation method ( Bechtold et al. 1993 ). Transgenic plants were selected on GM solid media ( Valvekens et al. 1988 ) containing 50 mg l−1 of kanamycin. Kanamycin-resistant progenies were analysed. For preparation of cross-sections, tissues were embedded in 5% agar. The embedded tissues were cut into 200 μm sections with a microslicer (DTK-1000, Dosaka). GUS proteins were stained with 5-bromo-4-chloro-3-indolyl glucuronide (X-gluc, Sigma) ( Jefferson et al. 1987 ). Plants and sections were observed under a fluorescent microscope, BX50-FLA (Olympus). GFP was visualized with the Chroma Dual Band Filter (Olympus) which provides excitation at 475–490 and 545–565 nm, and emission at 510–530 and 585–620 nm.
Extraction of total RNA and reverse transcription were carried out as described above in the methods for cDNA isolation. First-strand cDNA that derives from 25 ng of total RNA was used for each PCR amplification. PCR was carried out by ExTaq DNA polymerase (Takara) using gene-specific primers for sulphate transporters and α-tubulin ( Ludwig et al. 1987 ) as follows: Sultr3;1-FB (5′-TTAGGATCCATGGGCACGGAGGACTACACA-3′) and Sultr3;1-RB (5′-ATAGGATCCCTATACGTTGTTCCAAGGCTC-3′) for Sultr3;1; Sultr2;2-FB (5′-CTTGGATCCATGGGCATAGAGCT-TCAGAAT-3′) and Sultr2;2-RB (5′-ATTGGATCCATATCAACACAA-CTCGTGGGA-3′) for Sultr2;2; Sultr2;1-5F (5′-TTCATAGTTAAACT-TCCACACAACG-3′) and Sultr2;1–3R (5′-TAAATTCATCACATGCA-ATAACCCG-3′) for Sultr2;1; Sultr3;2-FB (5′-TTAGGATCCATGAGC-AGCAAAAGGGCATCA-3′) and Sultr3;2-RB (5′-ATAGGATCCTTAA-ACGTTGTTAAATTCCGG-3′) for Sultr3;2; Sultr4;1-FE (5′-CCTGA-ATTCATGTCCTACGCATCTCTCAGC-3′) and Sultr4;1-RE (5′-ATTG-AATTCCCTTTATTTCTCCACTGATAA-3′) for Sultr4;1; Sultr3;3-FE (5′-CTGTTGAATTCTTAGGAAAACCAATTAATG-3′) and Sultr3;3-RE (5′-CCGGAATTCTTAAACGTTACTGAGAGATGG-3′) for Sultr3;3; Sultr1;1-FE (5′-CCGGAATTCATGTCCGGGACTATTAATC-CC-3′) and Sultr1;1-RE (5′-CCGGAATTCTTAAGTTTGTTGCTCAG-CC-3′) for Sultr1;1; TUB-F (5′-AAATTAGGGTTTCTACTGAGA-GAAG-3′) and TUB-R (5′-ACGAATATTTTACAGGATTTAAACA-3′) for α-tubulin. Primers were designed on both ends of the coding sequence. PCR was carried out for 20 cycles where the cDNAs were exponentially amplified. PCR products were separated in agarose gels and were transferred to Hybond N+ membranes (Amersham) by capillary blotting with 0.4 m NaOH. Membranes were hybridized with 32P-labelled probes synthesized with the random primer DNA labelling kit (Takara). Hybridization was carried out at 65°C in a buffer containing 5 × SSPE (0.9 m NaCl, 0.05 m sodium phosphate pH 7.7, 5 m m EDTA), 0.5% SDS, 5 × Denhardt's solution, and 20 μg ml−1 of salmon sperm DNA. Membranes were washed at 65°C sequentially with 2 × SSPE buffer containing 0.1% SDS, 1 × SSPE buffer containing 0.1% SDS, and finally with 0.1 × SSPE buffer containing 0.1% SDS ( Sambrook et al. 1989 ). Hybridization signals were detected and quantified with an image analyser BAS- 2000 (Fujifilm).
Sulphate and thiol analyses
For measurements of uptake of 35S-labelled sulphate in Arabidopsis, plants were germinated and grown vertically on GM solid media for 8 days. Plants were transferred on the GM media with or without 0.1 m m of selenate, and grown for 2 days. The uptake rate was measured in a buffer solution containing 2 m m of 2-morpholinoethanesulphonic acid (MES) (pH 6.0), 0.1 m m of CaCl2, 0.1 m m of MgSO4 and 2.5 kBq ml−1 of 35S-labelled Na2SO4 (NEN) as described in Honda et al. (1998) . After weighing the fresh weights, plants were placed in the scintillation vials with 1 ml of scintillation fluid, and the radioactivity was measured by a liquid scintillation counter.
For estimation of total sulphate contents, sulphate was extracted in boiling hot de-ionized water from the freeze-dried tissues. Quantification was carried out by ion chromatography ( Smith et al. 1997 ). For estimation of thiol compounds, plants were ground in liquid nitrogen and extracted in 0.1 m HCl. Thiol compounds were labelled with monobromobimane (Molecular Probes) and separated by a reverse-phase HPLC column (Inertsil ODS-80A, GL Sciences) ( Fahey & Newton, 1987). N-acetylcysteine was used as an internal standard.
We thank Dr Y. Surdin-Kerjan (Centre National de la Recherche Scientifique, France) for the yeast mutant strain, CP154-7A; Dr Y. Tanaka (Suntory Ltd, Japan) for the expression vector, pYE22m; Dr Y. Niwa (University of Shizuoka, Japan) for the GFP expression vector, pTH2; and Dr T. Teeri (University of Helsinki, Finland) for the plant expression vector, pHTT202. This work was supported in part by Grants-in-Aid for Scientific Research from the Ministry of Education, Science, Sports and Culture, Japan, by Research for the Future Program (96I00302) from the Japan Society for the Promotion of Science, and by Asahi Glass Foundation. H.T. is a postdoctoral research fellow of Japan Society for the Promotion of Science. The Institute of Arable Crop Research receives grant-aided support from the Biotechnology and Biological Science Research Council of the UK. M.B.K. is supported by a grant from the Home-Grown Cereals Authority of the UK.
GenBank/EMBL/DDBJ accession numbers AB018695(Sultr1;1 cDNA), AB003591(Sultr2;1 cDNA) and D85416(Sultr2;2 cDNA).