SULTR3;1 is a chloroplast-localized sulfate transporter in Arabidopsis thaliana

Authors


(e-mail xiangcb@ustc.edu.cn or xiangcb@gmail.com).

Summary

Plants play a prominent role as sulfur reducers in the global sulfur cycle. Sulfate, the major form of inorganic sulfur utilized by plants, is absorbed and transported by specific sulfate transporters into plastids, especially chloroplasts, where it is reduced and assimilated into cysteine before entering other metabolic processes. How sulfate is transported into the chloroplast, however, remains unresolved; no plastid-localized sulfate transporters have been previously identified in higher plants. Here we report that SULTR3;1 is localized in the chloroplast, which was demonstrated by SULTR3;1-GFP localization, Western blot analysis, protein import as well as comparative analysis of sulfate uptake by chloroplasts between knockout mutants, complemented transgenic plants, and the wild type. Loss of SULTR3;1 significantly decreases the sulfate uptake of the chloroplast. Complementation of the sultr3;1 mutant phenotypes by expression of a 35S-SULTR3;1 construct further confirms that SULTR3;1 is one of the transporters responsible for sulfate transport into chloroplasts.

Introduction

Sulfur is essential for plants because it participates in many biological processes including disease resistance, the biosynthesis of the two sulfur-containing amino acids, and the detoxification of reactive oxygen species (ROS), xenobiotics, and heavy metals (Leustek et al., 2000; Saito, 2000; Xiang et al., 2001; Takahashi et al., 2011; Alvarez et al., 2012). Sulfate is the major form of inorganic sulfur utilized by plants. It is absorbed and transported by specific sulfate transporters (Takahashi et al., 2011). Typical sulfate transporters are generally pH-dependent inline image co-transporters containing 10–12 membrane-spanning helices followed by a STAS domain (Smith et al., 1995; Sharma et al., 2011b).

In Arabidopsis thaliana, 12 sulfate transporters were initially identified and subdivided into four groups according to their phylogenic relationships and kinetic properties (Takahashi et al., 2000; Shibagaki et al., 2002; Yoshimoto et al., 2002). They are usually tissue-specific and function coordinately to acquire sulfate from soil and to transport it to plastids for reduction. Sulfate is taken up from the soil by two root-specific high-affinity transporters of group 1, SULTR1;1 and SULTR1;2 (Yoshimoto et al., 2002), and is transported to the xylem by two low-affinity transporters of group 2, SULTR2;1 and SULTR2;2 (Takahashi et al., 2000). SULTR1;3, a phloem-localized sulfate transporter of group 1, mediates the shoot-to-root translocation of sulfate (Yoshimoto et al., 2003). SULTR4;1 and SULTR4;2 are co-localized in the tonoplast membrane and facilitate the efflux of sulfate from the vacuole, which is the main site for storage of sulfate in plants (Kataoka et al., 2004b). A recent report showed that SULTR3 knockout mutants affected sulfate content and sulfur assimilation in seeds (Zuber et al., 2010). SULTR3;5 was previously reported to reinforce the capacity for root-to-shoot sulfate transport mediated by SULTR2;1 (Kataoka et al.,2004a). Interestingly, sulfate transporters SULTR3;1, SULTR3;2, and SULTR3;3, preferentially expressed in leaves, failed to complement the yeast sultr mutant CP154-7A grown on low-sulfur medium (Cherest et al., 1997), which might be due to their incorrect localization and malfunction in yeast (Takahashi et al., 2000).

As reducers of sulfate, plants play a prominent role in the global sulfur cycle. For most of the pivotal functions of sulfur in plants, the oxidized sulfur in sulfate must be reduced and assimilated to cysteine by plants before entering other metabolic processes. Sulfate uptake, reduction, and assimilation are highly regulated (Hell and Wirtz, 2011; Takahashi et al., 2011) and key regulators have been identified (Kawashima et al., 2011; Lee et al., 2011)

Plastids, and especially chloroplasts, are the main site of reductive assimilation of sulfate in plants (Takahashi et al., 2011). How sulfate is transported into chloroplasts remains unresolved so far. No sulfate transporter is represented in plastid proteomics data (Ferro et al., 2010; Friso et al., 2010). An ABC-type chloroplast-localized sulfate transporter was found in the unicellular green alga Chlamydomonas reinhardtii (Chen et al., 2003; Chen and Melis, 2004). This sulfate permease complex consists of two chloroplast envelope-targeted transmembrane proteins, an ATP-binding protein and a sulfate-binding protein. The complex is encoded by four nuclear genes, but none of their homologs are found in vascular plants (Buchner et al., 2004). Until now, no chloroplast-localized sulfate transporters have been identified despite their essential role in sulfur assimilation in higher plants.

Among the four subfamilies of SULTR family, SULTR3 remains the least known. Like the sulfate transporters of other subfamilies such as SULTR1;2, -2;1, and -4;1, SULTR 3;1 is predicted to encode a protein with 12 transmembrane helices and a STAS domain (Figure 1). The STAS domain is mandatory for a functional sulfate transporter (Shibagaki and Grossman,2004) and may function in sensing, metabolism, or transport of nutrients in other proteins (Sharma et al., 2011a). The threonine-587 that is essential for transport activity of SULTR1;2 (Rouached et al., 2005) is also conserved in the STAS domain of SULTR3;1 protein (Thr-578). The proteins SULTR5;2 and SULTR5;1 are missing the STAS domain; these were once considered homologous proteins of the sulfate transporter family but were later found to be high-affinity molybdate transporters and renamed MOT1 and MOT2, respectively (Tomatsu et al., 2007; Gasber et al., 2011). Based on structural similarities, it is likely that SULTR3;1 is a functional sulfate transporter. Likewise, these features are also conserved in the other four members of the SULTR3 subfamily.

Figure 1.

SULTR3;1 is a typical sulfate transporter. The online program smart (http://smart.embl-heidelberg.de/) was used to predict the conserved structural or functional domains of sulfate transporter proteins. Conserved transmembrane and STAS domains of SULTR 3;1 and other subfamily members are illustrated. The solid black box and ellipse represent the transmembrane helices and the STAS domain, respectively. SULTR 3;1 comprises a transmembrane domain with 12 transmembrane helices and a C-terminal STAS domain.

The leaf-preferential expression of the SULTR3 subfamily suggests they may function in the chloroplast, which prompted us to investigate whether the SULTR3 subfamily participates in sulfate transport into plastids. By using loss-of-function mutants and complementation, GFP tagging, chloroplast protein import, and in organello sulfate uptake by isolated chloroplasts, the current study demonstrates that SULTR3;1 (At3g51895) is localized in the chloroplast and contributes to the sulfate uptake of plastids. Knockout of SULTR3;1 significantly decreases the sulfate uptake of chloroplasts, which can be rescued by complementation. These results not only demonstrate that SULTR3;1 is a sulfate transporter in the chloroplast but also indicate that other sulfur transporters exist in chloroplasts. Indeed, loss of SULTR3;2, SULTR3;3, and SULTR3;4 also significantly reduced chloroplast sulfate uptake.

Results

SULTR3;1 is localized in the chloroplast

The subcellular localization of SULTR3;1 was examined using transgenic Arabidopsis plants expressing 35S-SULTR3;1-GFP (Figure S1A in Supporting Information). Confocal images show that the green fluorescence of the SULTR3;1-GFP fusion protein is localized in the chloroplasts of SULTR3;1-GFP transformants, while the chloroplasts in wild-type cells show only red fluorescence (Figure 2a). The green fluorescence of the SULTR3;1-GFP fusion protein was also found in plastids in hypocotyl cells (Figure 2b).

Figure 2.

Subcellular localization of SULTR3;1. (a) Confocal images of mesophyll protoplasts. The left panel shows the GFP fluorescence of SULTR3;1-GFP fusion protein (upper two rows). The middle panel shows the auto fluorescence of chlorophyll. The right panel shows the merged images of green and red fluorescence. Wild type protoplasts are used as the control (lower two rows). (b) Confocal images of hypocotyls cells. The upper left image shows the green channel of GFP fluorescence of SULTR3;1-GFP fusion protein. The upper right image shows the red channel of autofluorescence of chlorophyll. The bottom left image shows the bright field and the bottom right image is the merged image of green, red fluorescence, and bright field.

To examine where in the chloroplast SULTR3;1-GFP was localized, we analyzed chloroplast protein fractions using Western blot with anti-GFP antibody. Positive signals were found in total protein extract of SULTR4;1-GFP plants (as the GFP-positive control), the protein from whole chloroplasts and the chloroplast envelope membrane fraction but not in thylakoids of SULTR3;1-GFP plants (Figure 3a). LHCB2 protein, a thylakoid marker, was not detected in the envelope fraction (Figure 3a). Given that SULTR3;1 contains multiple membrane-spanning domains, the result suggests that SULTR3;1 is localized in the envelope membrane of Arabidopsis chloroplasts.

Figure 3.

Chloroplast envelope membrane localization of SULTR3;1. (a) Western blotting of GFP protein. From left to right: total proteins of the wild type Col-0, total proteins of SULTR4;1-GFP (serving as positive control for specific GFP signal), chloroplast proteins of SULTR3;1-GFP transformant (chloroplast), envelope membrane proteins (containing stroma proteins) of SULTR3;1-GFP transformant (envelope), and thylakoid proteins (containing lumen proteins) of SULTR3;1-GFP transformant (thylakoid). The LHCB2 was used as thylakoid-fraction marker. The loaded protein of the envelope and thylakoid fractions was normalized to fraction volume instead of protein amount. Coomassie brilliant blue stained SDS-PAGE was used as loading control. (b) Chloroplast protein import. The N-terminal 86 amino acid peptide of SULTR3;1 facilitates luciferase into the chloroplast. Luciferase (589 amino acids) and N-terminal fused luciferase (675 amino acids), expressed and labeled with biotin in vitro, were incubated with wild-type chloroplasts, which were then separated from supernatant for SDS-PAGE. The biotin-labeled proteins were detected with Streptavidin-HRP.

These experiments cannot entirely rule out that SULTR3;1-GFP fusion protein was mis-targeted to chloroplasts in SULTR3;1-GFP transgenic plants. A chloroplast protein import experiment was therefore carried out to verify the chloroplast localization of SULTR3;1. The N-terminal 86 amino acids of SULTR3;1 (from ATG to the beginning of the first transmembrane domain) were fused to the N-terminus of luciferase. The chloroplasts were incubated with the in vitro translated luciferase or the SULTR3;1 N-terminus-luciferase fusion protein. The fused luciferase was detected in both the chloroplasts and the supernatant, while luciferase was only detected in the supernatant (Figure 3b).

Sulfate uptake by purified chloroplasts

Sulfate uptake into chloroplasts by Sultr-type transport was functionally characterized using uptake experiments with labeled sulfate. Intact chloroplasts were isolated from the rosette leaves of 4-week-old wild-type plants. To verify the integrity of the isolated chloroplasts the Hill reaction was used to determine the activity of chloroplasts. 2,6-Dichlorophenol-indophenol was used as electron acceptor instead of NADP+. Ammonia was used as inhibitor to eliminate the proton gradient. The result shows that the reaction was light dependent, indicating the isolated chloroplasts were intact and active (Figure S2A).

Given that the SULTR family are proton-coupled transporters, the isolated chloroplasts should show a pH-dependent sulfate uptake. The pH dependence of sulfate uptake of the isolated chloroplasts was examined at pH 6.0 and 7.2, respectively. The sulfate uptake rate of isolated chloroplasts from wild-type plants was significantly higher at pH 6 in comparison with pH 7.2 (Figure S2B), which is in agreement with the proton-coupled symport of Sultr-type sulfate transporters.

In the uptake-assay, membrane-bound radioactive sulfate was negligible after washing of chloroplast with 10 mm non-radioactive Na2SO4 (Figure S2C).

Loss of SULTR3;1 reduces sulfate uptake by chloroplasts

The quantitative contribution of SULTR3;1 to the transport of sulfate into chloroplasts was determined by comparison of the uptake kinetics of radioactive 35S-labeled sulfate by isolated chloroplasts of the knockout mutants sultr3;1-2 and sultr3;1-4 (Figure S1B), the complementation transgenic line 35S-SULTR3;1 (Figure S1C, D), and the wild type plants. The knockout mutant line sultr3;1-2 was the same line used by Zuber et al. (2010).

The in organello sulfate uptake was assayed with nine exogenous sulfate concentrations from 0 to 5.12 mm which covered the range of cytosolic sulfate content (Buchner et al., 2004). The intact chloroplasts were adjusted to the same concentration for each line before the uptake experiments. According to the previous pH-dependent uptake assay, the reaction buffer was adjusted to pH 6.0 to provide a sufficient proton gradient. The kinetic results are shown in Figure 4-7ac. The sulfate uptake rates of the chloroplasts from SULTR3;1 knockout mutants were significantly decreased compared with that of the wild type, which could be reverted to the wild-type level when the mutant was complemented with 35S-SULTR3;1 (Figure 4-7ac). The two SULTR3;1 knockout mutants showed a lower Vmax but similar Km compared with the wild-type plants. The Vmax of the sultr3;1 mutant decreased by 29% compared with the wild type. The 29% decrease should be contributed by the SULTR3;1 knockout.

Figure 4.

Sulfate uptake by isolated chloroplasts of the SULTR3;1 knockout mutants, wild type, and 35S-SULTR3;1 complementation line. (a) Dose-dependent sulfate uptake kinetics curves of the SULTR3;1 knockout mutants, wild type and 35S-SULTR3;1 complementation line. The number of chloroplasts was adjusted based on chloroplast concentration with a hemacytometer. Nine sulfur concentrations (from 0 to 5.12 mm) were used in the experiment. Km, Vmax and R2 of each line were calculated using Michaelis–Menten fit of GraphPad prism (GraphPad Software Inc., http://www.graphpad.com/) and are shown in the table. Values are means ± SD (= three experiments for each line). The asterisks indicates statistical significance at: *< 0.05 and **< 0.01. (b) Lineweaver–Burk plot of the kinetic data in (a). (c) Enlarged Lineweaver–Burk plot of the kinetic data for the low sulfate concentration range.

Figure 5.

Sulfate uptake by isolated chloroplasts of the SULTR3;2, SULTR3;3, SULTR3;4, and SULTR3;5 knockout mutants and the wild type. (a) Dose-dependent sulfate uptake kinetics curves of the SULTR3 knockout mutants and the wild type. The number of chloroplasts was adjusted based on chloroplast concentration with a hemacytometer. Nine sulfur concentrations (from 0 to 5.12 mm) were used in the experiment. Km, Vmax and R2 of each line were calculated using the Michaelis–Menten fit of GraphPad prism (GraphPad Software Inc., http://www.graphpad.com/) and are shown in the table. Values are means ± SD (= three experiments for each line). The asterisks indicates statistical significance at: *< 0.05 and **< 0.01. (b) Lineweaver–Burk plot of the kinetic data in (a).

Figure 6.

Expression pattern of SULTR3;1 in Arabidopsis. (a)–(d) Showing GUS staining of 3- (a), 6- (b), 9- (c), and 12-day-old seedlings (d) expressing SULTR3;1 promoter-GUS reporter. (e) Rosette leaves. (f) Cauline leaf and inflorescence stem. (g) Inflorescence. (h) Siliques and seeds.

Figure 7.

Levels of l-cysteine or glutathione in sultr3;1 mutant versus the wild type. (a), (b) Thiol levels in leaves. Thiols were extracted from the different leaves of 4-week-old SULTR3;1 knockout mutants and the wild type and analyzed for l-cysteine (a) and glutathione (b) using HPLC. ‘1–2 leaves’ means the first two true leaves, ‘3–6’ means the third to sixth true leaves and ‘7–10’ means leaves merged after the seventh true leaves. Values are means ± SD (= 3 for each treatment). Asterisks indicate statistical significance at ***< 0.001. (c), (d) Thiol levels in seedlings under low sulfate conditions. Seven-day-old seedlings of sultr3;1-2, the wild type, and 35S-SULTR3;1 transferred to liquid 1/8 MS with 0 or 200 μm sulfate for another 7 days. Thiols were extracted and analyzed for l-cysteine (c) and glutathione (d) content using HPLC. Values are means ± SD (= 3 for each treatment). Asterisks indicate statistical significance at: *< 0.05 and **< 0.01.

We used the same in organello sulfate uptake assay to examine whether the other four members of SULTR3 subfamily are also involved in chloroplast sulfate transport. The kinetic results are shown in Figure 5 for the mutants and the wild type. The sulfate uptake rates of the chloroplasts from SULTR3 knockout mutants except SULTR3;5 were significantly decreased compared with that of the wild type. SULTR3;2, SULTR3;3, SULTR3;4 knockout mutants showed a lower Vmax than the wild-type plants, while in SULTR3;5 knockout mutant the Vmax was not significantly different from the wild type. The Km of the mutants did not show a statistically significant difference from that of the wild type. Based on Vmax values, the contribution of sulfate influx into chloroplasts by SULTR3;2, SULTR3;3, and SULTR3;4 was estimated at 74, 66, and 69% of the wild type, respectively.

Leaf-preferential expression pattern

Promoter analysis in transgenic plants expressing the SULTR3;1 promoter-GUS reporter construct (Figure S1E) displayed GUS staining in green tissues (cotyledons and leaves) as shown in Figure 6(a–f), which is consistent with the previously reported preferential expression of SULTR3;1 in leaves of seedlings and mature plants (Takahashi et al., 2000). In addition, flowers and seeds also stained positive for GUS (Figure 6g, h), suggesting that SULTR3;1 also functions in the plastids of flowers and seeds. No GUS staining, however, occurred in roots under normal conditions.

Levels of l-cysteine and glutathione in the mutant

To examine the effect of SULTR3;1 disruption on thiols, we measured the cysteine and glutathione contents in sultr3;1-2 leaves of different ages of plants grown in soil under normal conditions. No significant difference was found except for the younger leaves after the seventh true leaf in which the cysteine content was significantly lower than the wild type and the complementation line (Figure 7a). This result is consistent with a previous report about the cysteine level in sultr3;1 mature leaves (Zuber et al., 2010). The glutathione levels of the two SULTR3;1 knockout mutants were comparable to the wild type and complementation lines (Figure 7b), similar to the stable glutathione levels in both mutant seeds and mature leaves reported for the same sultr3;1 mutant (Zuber et al., 2010).

When sulfate availability was reduced to 200 μm, the cysteine level was significantly reduced in the mutant shoots compared with that in the wild type and the complementation line (Figure 7c). Under the same conditions, a reduced glutathione level did not show a statistically significant difference between the mutant and the wild type and the 35S-SULTR3;1 line (Figure 7d). When reducing sulfate to 0 μm the cysteine level was further decreased, but no significant difference was observed between the mutant and the wild type. However, the complementation line remained at a higher level, probably due to the overexpression of SULTR3;1 (Figure 7c). The glutathione level was not significantly different between the mutant and the wild type but levels in both were significantly lower than the complementation line (Figure 7d).

Discussion

How sulfate is transported into chloroplasts has long being debated. Transport studies showed that the outside-acidic pH gradient enabled the action of a proton/sulfate co-transporter for influx into the chloroplast (Buchner et al., 2004). Our results with chloroplast sulfate uptake confirm that sulfate transport across the chloroplast envelope membrane system is pH dependent (Figure S2B). Earlier studies with isolated chloroplasts showed saturable kinetics with Km values of 2.5–3 mm sulfate (Mourioux and Douce,1979; Gross et al.,1990), corroborating our observations (Figures 4 and 5). In these experiments sulfate uptake was competitively inhibited by phosphate. Hence the triose-phosphate/phosphate translocator was suggested to be responsible for sulfate uptake due to a side activity of the protein (Mourioux and Douce,1979; Gross et al.,1990), but no conclusive evidence for this hypothesis has been provided so far. Indeed, no transporters for sulfate influx into chloroplasts in vascular plants had been identified prior to this study.

In this study we have demonstrated the subcellular localization of SULTR3;1 in the chloroplast membrane system in combination with fluorescence and immunological detection of SULTR3;1 fused with GFP (Figures 2 and 3a). In addition, we have shown that the N-terminus of the SULTR3;1 protein is able to direct the luciferase protein to the chloroplast (Figure 3b). These results indicate that SULTR3;1 is localized to the chloroplast. Moreover, SULTR3;1 is preferentially expressed in leaves of seedlings and mature plants (Figure 6), which overlaps with the chloroplast localization and is consistent with the recently published results of Zuber et al. (2010), also supporting a function of SULTR3;1 in chloroplasts of leaves in which reduction of sulfate takes place.

In order to demonstrate the sulfate transport functionality of SULTR3;1 in chloroplasts, we exploited the knockout mutants for chloroplast sulfate uptake with the validated in organello assay (Figure S2). Comparison of dose-dependent uptake rates between the wild type and mutant demonstrated that SULTR3;1 is a functional sulfate transporter in chloroplasts (Figure 4a, b). Similar affinity of the chloroplast for exogenous sulfate despite the disruption of SULTR3;1 (Figure 4a) indicates the existence of other sulfate transporters in chloroplasts with comparable affinities. We also tried to resolve the time-dependent kinetics but failed because the equilibrium was reached too quickly.

The kinetic results indicates a saturable component of the chloroplast sulfate transport system, ruling out a diffusion-driven transport mechanism by an unspecific pore protein in the envelope. The decreased but not eliminated uptake of sulfate in the chloroplasts of the SULTR3;1 knockout mutants indicates that SULTR3;1 accounts for only part of the total chloroplast sulfate uptake. This result, together with the unaltered Km, indicates that there must be other sulfate transporters in chloroplasts responsible for the rest of the bulk sulfate transported into chloroplasts. Indeed, the SULTR3;1 mutant with no obvious phenotype also suggests functional redundancy.

To find out whether other SULTR3 subfamily members are also involved in chloroplast sulfate uptake, we obtained knockout mutants for the in organello sulfate uptake assays. Interestingly, the chloroplast sulfate uptake was also partially impaired in sultr3;2, sultr3;3, and sultr3;4 but not in sultr3;5 (Figure 5). The kinetics results suggest that four out of five members of the SULTR3 subfamily may be chloroplast sulfate transporters. SULTR3;1, SULTR3;2, SULTR3;3, and SULTR3;4 contributed 29, 74, 66, and 69%, respectively, to the total sulfate uptake. However, the sum of contributions by four individual sulfate transporters is much higher than the theoretical 100%. We speculate that this could occur when some of these transporters form heterologous multimeric complexes. Knockout of one transporter may cause dysfunction of the complexes, resulting in a much larger decrease in sulfate uptake as we observed. Our data do not support SULTR3;5 as a chloroplast sulfate transporter, which is consistent with the previous report that SULTR3;5 was an essential component of the root-to-shoot transport of sulfate in the vasculature (Kataoka et al., 2004a).

Our results also confirm the functional redundancy observed in the single knockout mutants. Further analysis with subcellular localization and multiple knockout mutants would eventually confirm the function of SULTR3 transporters.

Perhaps due to functional redundancy, single knockout mutation did not cause a significant difference in glutathione levels between the mutants and the wild type. Only cysteine levels were found to be significantly lower in young leaves of the sultr3;1 mutant grown in soil under normal conditions. This seems consistent with a sulfur deficiency symptom that generally arises initially in young leaves. Reducing external availability of sulfate usually lowers the cysteine and glutathione levels. Therefore we analyzed the seedlings grown under reduced sulfate (200 μm) and sulfate-free (0 μm) condition (Figure 7). The cysteine level was significantly lower in the mutant under 200 μm sulfate compared with the wild type. When sulfate availability was completely removed, the cysteine level dropped further. However, no difference in the cysteine level was observed between the wild type and mutant, which might indicate that severe sulfate depletion overshadowed the effect of the knockout mutation of SULTR3;1. Consistent with their results, Zuber et al. (2010) recently reported that the cysteine level was significantly decreased in seeds but not in mature leaves of the sultr3;1 mutant. Furthermore, none of the single knockout mutations in the SUTLR3 subfamily caused any growth phenotypes in the mutants that were obviously different from the wild type, indicating that the single mutation might be functionally compensated by other sulfate transporters in the subfamily that are functionally redundant. The generation and analysis of multiple mutants of the subfamily should be able to clarify the role of SULTR3 in chloroplast sulfate transport and their impact on growth.

Experimental Procedures

Plant materials and growth conditions

Arabidopsis thaliana (ecotype Columbia, Col-0) seedlings were grown on half-strength Murashige and Skoog (MS) solid medium containing 1% (w/v) sucrose at 22°C under 12-h light/12-h dark cycles. Sulfur-deficient medium was prepared by replacing sulfate salts and agar in the MS medium with equivalent chloride salts and agarose. The plants for chloroplast and protoplast isolation were grown in soil for 1 month.

Identification of the SULTR3;1 knockout mutants and complementation with 35S-SULTR3;1

The mutants sultr3;1-2 and sultr3;1-4 were both T-DNA insertion lines (SALK_023190 and SALK_127024) obtained from the Arabidopsis Biological Resource Center (ABRC). The homozygotes were identified by PCR as described (Alonso et al.,2003) using a common primer LBb1 (5′-GCGTGGACCGCTTGCTGCAACT-3′) and gene-specific primers 3;1-2-F (5′-TGAACTCCCTAGCACCGCGTA-3′) and 3;1-2-R (5′-AATCGGTGCTTTATCCTCTTC-3′) for sultr3;1-2 and 3;1-4-F (5′-TGGCATGTAAATACGAATGAGG-3′) and 3;1-4-R (5′-AAGGCTTAAATCACCAACACAC-3′) for sultr3;1-4. The homozygotes were confirmed by RT-PCR using primers 3;1 RT-F (5′-TGCGGTGAAGGGAAACATAC-3′) and 3;1 RT-R (5′-CTATACGTTGTTCCAAGGCT-3′).

For complementation analysis, the 35S-SULTR3;1 overexpression construct was made by inserting the coding region of SULTR3;1 (1977 bp) amplified by PCR using primer pair 3;1 attb1-F (5′-GGGGACAAGTTTGTACAAAAAAGCAGGCTGA ATGGGCACGGAGGACTACA-3′) and 3;1 attb2-R (5′-GGGGACCACTTTGTACAAGAAAGCTGGGTA CTATACGTTGTTCCAAGGCT-3′) into pCB2004 (Lei et al.,2007) via Gateway recombinant cloning (Invitrogen, http://www.invitrogen.com/) and transformed into sultr3;1-2 mutant using the floral-dip method as described (Bechtold et al., 1993; Bent, 2000). Transformants were selected for glufosinate resistance and confirmed by RT-PCR using primers 3;1RT-F and 3;1RT-R and β-tubulin8 as control with specific primers (5′-CTTAAGCTCACCACTCCAAGCT-3′and 5′-GCACTTCCACTTCGTCTTCTTC-3′).

Localization of SULTR3;1-GFP fusion protein and SULTR3;1 promoter analysis

The coding region of SULTR3;1 (1977 bp) was amplified by PCR using the primer pair 3;1 SpeI-F (5′-CACTAGT ATGGGCACGGAGGACTACA-3′) and 3;1 XbaI-R (5′-CGGTCTAGA TATACGTTGTTCCAAGGCTC-3′) to construct the GFP fusion vector in pCB2008E (Lei et al., 2007). The SULTR3;1-GFP construct was transferred into the wild type and the T2 generation was used for fluorescence imaging.

The promoter region of SULTR3;1 (−1862 to +79) was amplified by PCR using primer pair pSULTR3;1 attb1-F (5′-GGGGACAAGTTTGTACAAAAAAGCAGGCTGACCGGGTCAACAACCGTCTCT-3′) and pSULTR3;1 attb2-R (5′-GGGGACCACTTTGTACAAGAAAGCTGGGTACGACTGCTCGGGTAGGGTTT-3′) to construct pSULTR3;1-GUS in pCB308R for promoter analysis. The reporter construct pSULTR3;1-GUS was transferred into the wild type and the T2 population was used for GUS staining.

Preparation of mesophyll protoplasts and GFP imaging

Mesophyll protoplasts were prepared as described (Demidchik and Tester,2002). Fluorescence of GFP in the protoplasts of SULTR3;1-GFP transgenic plants was observed using a confocal microscope (Carl Zeiss LSM510, http://www.leica.com/) under 488 nm excitation. The emission wavelength was restricted to 530 nm for green fluorescence and 650 nm for red fluorescence of the chlorophyll background.

Isolation of intact chloroplasts

Plants were grown under day-neutral conditions at 22°C. Crude chloroplasts were obtained from the rosette leaves of 4-week-old plants using isolation buffer (pH 8.0) and purified using Percoll gradients as described (Kunst,1998).

Separation of the envelope from the thylakoid membrane for Western blot analysis

Purified intact chloroplasts from SULTR3;1-GFP transgenic plants were lysed in thylakoid isolation buffer and thylakoids were isolated as described (Peltier et al.,2002). Thylakoid and envelope membranes (including stroma contents) were fractionated and solubilized with protein extraction buffer (2% SDS, 1 × PBS, 1 mm EDTA, 0.1 mm phenylmethylsulfonyl fluoride, 1% cocktail inhibitor), while total proteins of the wild type and SULTR4;1-GFP transgenic plants were extracted with the same extraction buffer. Protein contents were estimated using the Bradford method. Acrylamide gels (12%) were used for SDS-PAGE analysis and LHCB2 was used as the marker of thylakoids. The protein blots were detected using 1:1000 diluted anti-GFP antibodies (Santa Cruz Biotechnology, http://www.scbt.com/) and anti-LHCB2 antibody (Agrisera, http://www.agrisera.com/).

The Hill reaction

The Hill reaction was carried out as previously described (Bregman, 1990) with some modifications to verify the chloroplast isolation protocol by determining the activity of chloroplasts. 2,6-Dichlorophenol-indophenol (DCIP) was used as the electron acceptor and ammonia as the inhibitor. The chloroplasts was isolated using 2-amino-2-(hydroxymethyl)-1,3-propane diol (TRIS)-sucrose buffer (0.3 m sucrose, 0.2 m TRIS-HCl, 5 mm MgSO4, adjust to pH 7.5) on ice. The TRIS-sucrose buffer was also used as the reaction buffer and was mixed with DCIP (and ammonia if needed) in advance before chloroplast suspensions were added. The reaction time was 10 min in light or the dark (as a negative control) and the absorbance under 600 nm was determined every 2 min using a spectrophotometer (UNICO UV-2101). Each assay was repeated three times.

Sulfate uptake by purified chloroplasts

The purified chloroplasts from the wild type, 35S-SULTR3;1, and the two knockout mutants (sultr3;1-2 and sultr3;1-4) were quantified by counting the number with a hemacytometer. The suspensions were adjusted to the same number concentration of chloroplasts with isolation buffer. Two hundred microliters of chloroplasts was used in each uptake reaction by mixing with the same volume of reaction buffer containing 0.3 m sorbitol, 50 mm 2-(N-morpholino)ethanesulfonic acid/KOH (adjust to pH 6.0), 10 mm NaHCO3, 5 mm MgCl2, 5 mm EDTA, 5 mm EGTA, corresponding Na2SO4 and 35S labeled Na2SO4 (10 mCi ml−1, GE Healthcare, http://www3.gehealthcare.com/). Eight exogenous sulfate concentrations (0, 0.04, 0.08, 0.16, 0.32, 0.64, 1.28, 2.56, and 5.12 mm) were used in the uptake assay. The pH-dependent assay was carried out using two different reaction buffers with 5.12 mm sulfate and pH 6.0 or 7.2, respectively. After 5 min incubation at room temperature, the mixture was immediately centrifuged at 2000g for 30 sec and the chloroplast pellet was washed with isolation buffer containing 10 mm non-radioactive Na2SO4 before being transferred into a 10-ml scintillation cocktail. The cocktail contained 4 g of 2,5-diphenyloxazole (PPO) and 0.01 g of 1,4-bis[2-(5-phenyl)-oxazolyl] benzene (POPOP) dissolved in 1 L of 1:1 (v/v) toluene:alcohol. The radioactivity was measured in a scintillation counter (Perkin Elmer Tri-carb 2910TR, http://www.perkinelmer.com/). A standard curve was created to determine the correlation between the radioactivity and quantity of Na235SO4.

Chloroplast protein import

The TNT coupled transcription/translation system and transcend biotin-lysyl-tRNA were used for in vitro protein expression (Promega, http://www.promega.com/). The N-terminal of SULTR3;1 protein containing 86 amino acids was amplified by PCR with primers 5′-ACGAAGCTTATGGGCACGGAGGACTACA-3′ and 5′-GTCGGATCCGATCTGATTTGAAGAACTTGAG-3′, and inserted into the luciferase T7 vector between HindIII and BamHI. Luciferase and N-terminal-fused luciferase were expressed and biotin-labeled in vitro following manufacturer's instructions. The chloroplast-import reaction was carried out as described (Forsman and Pilon,1995). Chloroplasts were isolated from wild-type leaves as described earlier for isolation of intact chloroplasts. The isolated chloroplasts were resuspended in import reaction buffer containing 50 mm HEPES (pH 8.0), 0.33 m sorbitol, 2 mm MgCl2, 0.5 mm DTT, 2 mm MgATP, and 2 μg/ml antipain. Proteins were incubated with chloroplasts (30 μg chlorophyll/150 μl import reaction buffer) for 20 min in light at 26°C. The chloroplasts were separated by centrifugation and Western blot was used to detect the biotin-labeled protein with streptavidin–horseradish peroxidase.

GUS staining assay

The GUS staining was conducted as described in Jane K (1988).

Thiol quantification

Thiols were quantified as described before (Xiang and Oliver, 1998).

Acknowledgement

This work was supported by grants from NNSFC (90917004, 30471038). The authors thank Dr Hideki Takahashi for providing SULTR4;1-GFP transgenic seeds and the ABRC for SULTR3 T-DNA insertional lines.

Ancillary