Identification and regulation of plasma membrane sulfate transporters in Chlamydomonas.

Chlamydomonas (Chlamydomonas reinhardtii) exhibits several responses following exposure to sulfur (S)-deprivation conditions, including an increased efficiency of import and assimilation of the sulfate anion (SO(4)(2-)). Aspects of SO(4)(2-) transport during S-replete and S-depleted conditions were previously studied, although the transporters had not been functionally identified. We employed a reverse genetics approach to identify putative SO(4)(2-) transporters, examine their regulation, establish their biogenesis and subcellular locations, and explore their functionality. Upon S starvation of wild-type Chlamydomonas cells, the accumulation of transcripts encoding the putative SO(4)(2-) transporters SLT1 (for SAC1-like transporter 1), SLT2, and SULTR2 markedly increased, suggesting that these proteins function in high-affinity SO(4)(2-) transport. The Chlamydomonas sac1 and snrk2.1 mutants (defective for acclimation to S deprivation) exhibited much less of an increase in the levels of SLT1, SLT2, and SULTR2 transcripts and their encoded proteins in response to S deprivation compared with wild-type cells. All three transporters were localized to the plasma membrane, and their rates of turnover were significantly impacted by S availability; the turnover of SLT1 and SLT2 was proteasome dependent, while that of SULTR2 was proteasome independent. Finally, mutants identified for each of the S-deprivation-responsive transporters were used to establish their critical role in the transport of SO(4)(2-) into S-deprived cells.

Sulfur (S) is an essential element for all organisms and is present in proteins, lipids, carbohydrates, and several metabolites. Sulfate (SO 4 22 ) is the preferred S source for most organisms. In photosynthetic organisms, the reductive assimilation of SO 4 22 occurs in plastids, which means that this ion must traverse both the plasma membrane and the plastid envelope prior to reduction and incorporation into organic molecules. SO 4 22 is relatively inert and must be activated by the enzyme ATP sulfurylase before being reduced to sulfide and incorporated into the amino acids Cys and Met (Leustek et al., 2000), which can be used for the synthesis of proteins or converted into other metabolites, including glutathione and dimethyl sulfide.
Much of the SO 4 22 in the soil is not readily available to plants or microbes. The SO 4 22 anion can be adsorbed onto the surface of the soil particles, and a large proportion may be covalently bonded to organic molecules in the form of SO 4 22 esters and sulfonates. When experiencing low SO 4 22 availability, the unicellular, soil-dwelling alga Chlamydomonas (Chlamydomo-nas reinhardtii) exhibits a suite of responses that include the synthesis of extracellular arylsulfatases (ARS), elevated SO 4 22 transport activity, and an increase in the levels of transcripts encoding ATP sulfurylase and other enzymes associated with S assimilation. Many of these responses allow the alga to more efficiently scavenge and assimilate available SO 4 22 in the environment (Davies and Grossman, 1998;Grossman and Takahashi, 2001;Zhang et al., 2004;Gonzalez-Ballester et al., 2008, 2010Shibagaki and Grossman, 2008).
Specific polypeptides involved in regulating Chlamydomonas S-deprivation responses have been identified. The SAC1 (for sulfur acclimation 1) protein is required for many S-limitation-induced responses. Chlamydomonas mutants with lesions in the SAC1 gene exhibit abnormal SO 4 22 uptake, are unable to synthesize extracellular ARS, and show little increase in many S-deprivation-responsive transcripts, including those encoding ARS, ATP sulfurylase, Ser acetyltransferase, and the ferredoxin-dependent sulfite reductase. Furthermore, sac1 mutants cannot suppress photosynthetic electron transport activity and rapidly die when placed in S-deficient medium in the light. Even though the SAC1 gene encodes a protein similar to anion transporters from a number of different organisms, including the Na + /SO 4 22 transporter from mammals, the phenotypes of sac1 mutants strongly suggest that SAC1 functions in regulating cellular responses to S deprivation (Davies et al., 1996). A second polypeptide that plays a central role in the acclimation of Chlamydomonas to S deprivation is SNRK2.1, a member of the SNF1-related protein kinase 2 family. Like SAC1, SNRK2.1 is required for most responses associated with the acclimation of Chlamydomonas to S deprivation. A snrk2.1 mutant (initially designated ars11) is defective for the expression of several S-responsive genes, and the mutant bleaches and dies more rapidly than wild-type cells during S starvation; the responses of snrk2.1 to S deprivation are generally more severe than those of the sac1 mutant (Gonzalez-Ballester et al., 2008).
Arabidopsis has 12 putative SO 4 22 transporters that have been classified into four groups. The AtSULTR1 group represents high-affinity transporters responsible for the initial uptake of SO 4 22 into root cells, while AtSULTR2;1, AtSULTR2;2, and AtSULTR3;5 are lowaffinity transporters likely to play important roles in the translocation of SO 4 22 from roots to leaves (Takahashi et al., 2000;Hawkesford, 2003;Kataoka et al., 2004a). Furthermore, two isoforms of the AtSULTR4 group have been localized to the tonoplast and facilitate SO 4 22 efflux from vacuoles (Kataoka et al., 2004b). The plant-associated H + /SO 4 22 cotransporters usually contain 12 transmembrane domains followed by a linking region that connects to a C-terminal STAS (for SO 4 22 transporter and anti-sigma factor antagonist) domain. There is significant interest in the function of the STAS domain, since lesions in this domain can lead to serious human diseases (Everett and Green, 1999). While the precise function of the STAS domain is not clear, it is required for the activity and proper assembly of the SO 4 22 transporters Grossman, 2004, 2006;Rouached et al., 2005) and may help regulate interactions of the transporter with partner proteins that function downstream in the S assimilation pathway (Shibagaki and Grossman, 2010).
Features of SO 4 22 transport in Chlamydomonas during S-replete and S-depleted conditions have been reported, although there has been little molecular characterization of the genes and polypeptides encoding the putative SO 4 22 transporters. Initial studies demonstrated that both the V max and the substrate concentration at which SO 4 22 transport is at halfmaximum velocity (K 1/2 ) were altered in S-deprived cells. The V max increased approximately 10-fold and the K 1/2 decreased roughly 7-fold. The enhancement of SO 4 22 transport upon S deprivation is prevented by cycloheximide (CHX) but not by chloramphenicol, demonstrating that protein synthesis on 80S cytoplasmic ribosomes is required for the synthesis of the highaffinity transport systems .
Analysis of the Chlamydomonas genome sequence has led to the identification of several genes encoding proteins with high sequence similarity to known SO 4 22 transporters. Three of the putative SO 4 22 transporters (SULTR1 to SULTR3) are in the H + /SO 4 22 family (characteristic of vascular plants), while another three (SLT1 to SLT3; for SAC1-like transporter) are in the Na + /SO 4 22 transporter family that is present in bacteria, nonvascular plants, and mammals; these transporters have not been identified in vascular plants. Another SULTR-type transporter, originally designated SULTR4, was recently shown to encode a functional molybdate transporter, and the gene has been renamed MOT1 (Tejada-Jiménez et al., 2007). Transcripts from some of the transporter genes increase significantly during S deprivation; increased accumulation of SLT1 and SULTR2 transcripts during S starvation was previously reported (Gonzalez-Ballester et al., 2008). Chlamydomonas also possesses SO 4 22 permeases (SulP1 and SulP2) that resemble the permeases of bacteria (Laudenbach and Grossman, 1991;Chen et al., 2003;Lindberg and Melis, 2008). These transport proteins have been localized to the chloroplast envelope and function in the transport of SO 4 22 from the cytosol into chloroplasts, where reductive assimilation of the anion occurs.
Here, we identify, localize, and examine the regulation of Chlamydomonas SO 4 22 transporters. We define the kinetics of accumulation of both the RNA encoding the transporters and the transporter polypeptides following the imposition of S deprivation and monitor their decay after SO 4 22 is added back to starved cells. We also evaluate the impact of the regulatory elements SAC1 and SNRK2.1 on the accumulation of SULTR and SLT transcripts in S-replete and S-depleted cells and show that the proteasome is involved in the turnover of SLT but not SULTR transporters. Finally, various methods were used to localize the different transporters to specific cellular membranes and to identify and characterize mutants that were specifically defective for the function of the individual transporter polypeptides. These analyses provide a comprehensive view of the function, biogenesis, and regulation of the S-responsive SO 4 22 transporters in Chlamydomonas.
SO 4 22 transporters. These genes were characterized and used to identify potential SO 4 22 transporters from other species (Takahashi et al., 1997). Similarly, analysis of the entire Chlamydomonas genome sequence (Merchant et al., 2007) Fig. S1A , red bar) is present on various anion transporters and extends into the cytoplasm of the cell.
In contrast to the SULTR-type transporters, the deduced amino acid sequences of SLT1, SLT2, and SLT3 of Chlamydomonas exhibit strong sequence similarity to Na + /SO 4 22 transporters (SLC13 family). All three members of the SLT family in Chlamydomonas have 10 to 12 predicted transmembrane domains and an intracellular loop containing a TrkA-C domain (Supplemental Fig. S1B, orange bar). While the precise function of the TrkA-C domain is not known, it may be involved in the regulation of SO 4 22 transporter activity, potentially through interactions with partner proteins. The similarity among the Chlamydomonas SLTs and between the SLTs and the putative SO 4 22 transporter of the moss Physcomitrella patens is shown in Supplemental Figure S1B.
Induction of SO 4 22 Transport and Accumulation of SULTR2, SLT1, and SLT2 Transcripts during S Deprivation When organisms are deprived of S, they often exhibit increases in their affinity for SO 4 22 (decrease in the K 1/2 ) and in the maximal rate of SO 4 22 transport (elevated V max ). As shown in Figure 1, the maximal rate of SO 4 22 transport into wild-type Chlamydomonas cells (strain 21gr) increases approximately 13-fold after 24 h of S deprivation. To identify those transporters likely to be responsible for the change in kinetics of SO 4 22 transport by cells acclimated to S deprivation, we monitored changes in the abundance of transcripts encoding the putative SO 4 22 transporters of Chlamydomonas by real-time quantitative PCR (RT-qPCR). RNAs were isolated from wild-type cells placed in medium devoid of S, nitrogen (N), or phosphorus (P) for 12 h. The elevated levels (note the log scale) for transcripts from ARS2 (encoding an extracellular ARS) during S deprivation, from PHOX (encoding an alkaline phosphatase) during P deprivation, and from NIT1 (encoding a nitrate reductase) during N deprivation (Fig. 2) suggest that the cells were responding normally to each nutrient starvation. Transcripts encoding the SO 4 22 transporters showed remarkable changes in abundance when the cells were starved for S but not when starved for P or N (Fig. 2). SULTR2 and SLT2 transcript levels increased by more than 100-fold, while the level of the SLT1 transcript exhibited an increase of approximately 1,000-fold after 12 h of S deprivation. In contrast, SULTR1 and SLT3 transcript levels decreased roughly 1,000-and 10-fold, respectively, specifically during S deprivation, while the abundance of the SULTR3 transcript exhibited little change (Fig. 2). The accumulation of SULTR2, SLT1, and SLT2 transcripts following the imposition of S deprivation suggests that these three genes may encode the dominant, potentially high-affinity SO 4 22 transporters, while the decrease in SULTR1 and SLT3 transcript levels suggests that they may encode lowaffinity SO 4 22 transporters (not needed in an S-limited environment).
The induction of SO 4 22 transport activity observed in wild-type Chlamydomonas cells during S limitation is reduced in the sac1 mutant and completely abolished in the snrk2.1 mutant (Fig. 1). To understand the relationship between transporter activity and the levels of transcripts encoding the different transporters, we quantified SO 4 22 transporter transcripts in wild-type and mutant strains. While the level of the ARS2 transcript increased approximately 3 orders of magnitude in wild-type cells (21gr) within 2 h of S deprivation (Fig.  3A), the sac1 and snrk2.1 mutants showed marked reductions in the increase in ARS2 mRNA abundance relative to wild-type cells (Fig. 3, B and C, respectively). Furthermore, the elevation (SULTR2, SLT1, SLT2) and depression (SULTR1, SLT3) of levels of transcripts encoding the various putative SO 4 22 transporters following S deprivation were generally not nearly as marked in sac1, and especially in snrk2.1, relative to wild-type cells. The levels of accumulation of the SULTR2, SLT1, and SLT2 polypeptides were consistent with their transcript abundances during S starvation (Supplemental Fig. S2).
Interestingly, some transcripts encoding transporters were still elevated in the mutants following S deprivation; the most notable examples are the SLT1 transcript in the sac1 mutant and the SULTR2 transcript in the snrk2.1 mutant (Fig. 3, B and C); the SULTR2 transcript is also elevated to some extent under S-replete conditions in the sac1 mutant. Furthermore, while the SULTR2 transcript still increases significantly in snrk2.1, the absolute level of the transcript is much lower in the mutant than in the wildtype strain (2-3 orders of magnitude) under both S-replete and S-deprivation conditions. Also, the increase in the SLT1 transcript in the sac1 mutant is only observed after 24 h of S deprivation (it is observed 2 h after wild-type cells are exposed to S deprivation). The SLT1 transcript in the snrk2.1 strain is essentially not detected under S-replete or S-deprivation conditions. Finally, while the level of the SULTR1 transcript markedly declines in wild-type cells, its decline is diminished in the sac1 mutant, and it actually increases by 5-to 10-fold in the snrk2.1 strain. The SULTR3 transcript, which also decreases to some extent in the wildtype strain during S deprivation, may increase slightly in the snrk2.1 mutant (Fig. 3, compare A and C). These results suggest that there are some complex relationships linking the mutant phenotypes to transcript levels, although regulation in the mutant strains is aberrant for essentially all transporter transcripts that were examined; we conclude that the SAC1 and SNRK2.1 regulators are required for both increases and decreases in transcript abundance that are observed following the transfer of Chlamydomonas cells to S-deprivation conditions. Furthermore, SNRK2.1 appears to be important for maintaining moderate levels of SLT1 and SULTR2 transcripts when the cells are growing in nutrient-replete medium. Finally, some of the transcripts encoding transporters that normally sharply decline in wild-type cells during S deprivation can become elevated in the mutant strains (SULTR1 in the snrk2.1 mutant). These findings suggest that there may be both direct and indirect consequences of elimination of the SAC1 and/or SNRK2.1 regulatory proteins during S deprivation on the level of transporter transcripts; the increase in the SULTR3 transcript in snrk2.1 might reflect the inability of the strain to suppress expression of the SULTR3 gene, which might be the result of a compensatory effect elicited in cells severely compromised in their ability to scavenge S from external and probably internal sources.

SULTR2, SLT1, and SLT2 Protein Levels
We analyzed wild-type cells for levels of SULTR2, SLT1, and SLT2 transcripts and polypeptides, both transporters following imposition of S deprivation. Levels of transcripts encoding ARS2, SULTR2, SLT1, SLT2, SULTR1, SLT3, and SULTR3 were measured by RT-qPCR for wild-type cells (A) and the sac1 (B) and snrk2.1 (C) mutants. RNA samples were isolated from cells 0, 2, 6, and 24 h after they were transferred from TAP to TAP2S medium. The increase in ARS2 transcript abundance was used as a positive control to demonstrate proper S-starvation responses in wild-type cells. before and during S deprivation and following the addition of SO 4 22 back to starved cells. Figure 4A shows the levels of these transcripts in cells collected at various times from 1 to 32 h after removal of S from the medium and at 1, 2, 4, and 8 h after the addition of SO 4 22 to cells that had experienced 24 h of S starvation. The same samples that were used to quantify transcript abundance by RT-qPCR were also used for microsomal membrane preparations to monitor the levels of the transporter proteins by western-blot analyses, as shown in Figure 4B.
As presented in Figure 4A, levels of SULTR2, SLT1, and SLT2 transcripts are high after 1 h of S deprivation and remain at approximately the same level after 24 h of S deprivation. However, these mRNAs decline at different rates following the addition of SO 4 22 to the medium. The SLT1 transcript declined most rapidly, although none of the transcripts dropped to the +S levels, even at 8 h following the readdition of SO 4 22 to the medium.
For monitoring transporter protein abundance, polyclonal antibodies were raised against a peptide unique to SLT2 (Supplemental Fig. S1B, magenta bar), a peptide unique for SULTR2 (Supplemental Fig. S1A, blue bar), and a peptide present in both SLT1 and SLT2 (SLT "general" antibodies; Supplemental Fig. S1B, green bar; see "Materials and Methods"). The SLT general antibodies (SLT in Fig. 4B) detected a specific band of approximately 95 kD, which corresponds to the predicted molecular mass of SLT1 and SLT2 (93 and 96 kD, respectively). The specific SULTR2 antibodies recognized a polypeptide of approximately 80 kD, which corresponds to the predicted molecular mass of SULTR2 (approximately 84 kD). Generally, all of the transporter polypeptides appear to migrate as doublets (the most well-resolved doublet is for SULTR2), which could be a consequence of protein modification, proteolytic cleavage of a short N-or C-terminal region, and/or the recognition of more than one polypeptide by the antibodies (although the two proteins would have to show the same kinetic changes with respect to the S status of the medium). The marked, rapid accumulation of the transcript for the SLT and SULTR2 transporters (after 1 h of S deprivation; Fig. 4A) is followed by a dramatic increase in the levels of the transporter polypeptides, which is first detected 2 h after the initiation of S deprivation (Fig.  4B). Following administration of SO 4 22 to S-starved cultures, the levels of SULTR2, SLT2, and probably SLT1 (detected by the general SLT antibody and confirmed by analysis of transporter polypeptides in strains defective for specific SO 4 22 transporters; see Fig. 9 below) polypeptides declined and became barely detectable 8 h after the SO 4 22 content of the medium had been replenished (Fig. 4B). We also observed some decline in polypeptide levels at 32 h relative to 24 h of S deprivation, although the decline was much less severe than in cultures administered S.

Turnover of SULTR2, SLT1, and SLT2 Polypeptides
We analyzed whether the decrease in abundance of the SO 4 22 transporters following the addition of SO 4 22 was a consequence of enhanced degradation. Cells were placed in medium devoid of S for 10 h to allow the accumulation of SO 4 22 transporter polypeptides prior to the administration of CHX, an inhibitor of protein synthesis on 80S ribosomes, with or without the addition of 1 mM SO 4

22
. The concentration of CHX used in these experiments was previously shown to specifically block cytoplasmic protein synthesis (Kawazoe et al., 2000); it also completely blocked accumulation of the SO 4 22 transporter proteins in cells transferred to medium lacking S (Supplemental Fig.   Figure 4. Accumulation of SULTR2, SLT1, and SLT2 transcripts and proteins and their decay following the addition of SO 4 22 to S-deprived cells. Time course of accumulation of SULTR2, SLT1, and SLT2 mRNAs (A) and polypeptides (B) upon transfer of cells from S-replete to TAP2S medium is shown. After 24 h in TAP2S, 1 mM MgSO 4 was added to starved culture. Samples were taken at 1, 2, 4, 8, 24, and 32 h after the starvation (from +S to 2S) and 1, 2, 4, and 8 h after the addition of SO 4 22 back to cultures that had been S starved for 24 h. RNA and protein isolations are described in "Materials and Methods." A ferroxidase, FOX1, served as a loading control (accumulation of FOX1 is S independent). S3). The levels of the SO 4 22 transporters were monitored 20, 60, and 240 min following the addition of the inhibitor and/or SO 4 22 to the cultures. As expected, the SO 4 22 transporters accumulated after 10 h of S starvation (Fig. 5A, lane 2). Interestingly, these polypeptides were turned over significantly more rapidly in cultures that received SO 4 22 along with CHX ( Fig. 5A, lanes 10-12) compared with cultures that were only administered CHX but remained S starved (Fig. 5A, lanes 4-6). The band intensity was quantified, normalized to the level of the FOX1 protein, and the half-life of each polypeptide was derived by fitting the values to the decaying exponential equation. These results indicate that administration of SO 4 22 enhances degradation of the SULTR2, SLT1, and SLT2 polypeptides, although the kinetics of the turnover varies among the transporters. Since the SLT general antibodies recognize both SLT1 and SLT2 polypeptides, the half-life estimated from the signals detected with these antibodies reflects the turnover rate of both SLT proteins. Interestingly, when SO 4 22 was added back to the cultures, the addition of CHX made little difference to the kinetics of the loss of SLT and SULTR2 polypeptides (Fig. 5A, compare lanes 7-9 with 10-12), raising the possibility that the addition of SO 4 22 to starved cultures both rapidly blocked the synthesis of these proteins (which in part also declines because of transcript degradation) and stimulated their degradation.
To determine whether the proteasome was involved in degradation of the SO 4 22 transporters, we performed a similar experiment in which cells were S starved for 10 h prior to the addition of CHX and SO 4 22 to the cultures in the presence and absence of the proteasome inhibitor MG132; samples were collected 30, 90, and 240 min following the addition of the inhibitors plus SO 4

22
. Figure 5B shows that degradation of SLT1 and SLT2 polypeptides was completely blocked by MG132, suggesting that proteasome activity is required for the turnover of these proteins. Intriguingly, SULTR2 turnover was not affected when proteasome function was inhibited (repeated on three separate samples with similar results), suggesting that there are different mechanisms for eliminating the transporter proteins once S becomes sufficient in the environment.

Transporter Localization
To localize the SO 4 22 transporters, total membranes were prepared from S-starved cells and separated by a two-phase aqueous polymer system into a plasma membrane (polyethylene glycol) fraction and a fraction containing the rest of the cellular membranes. The polypeptides from the plasma membrane and "other membrane" (non-plasma membrane) fractions were resolved by SDS-PAGE, and immunoblot analyses were performed using antibodies to proteins of known localization. As shown in Figure 6A, the plasma membrane fraction was largely free of contamination from the thylakoid membrane and chloroplast envelope, based on the distribution of CRD1 (Allen et al., 2008), and from the mitochondrial membrane, based on the distribution of COX2b (Page et al., 2009). As expected, two plasma membrane markers, H + -ATPase (Norling et al., 1996) and FOX1 ferroxidase (Herbik et al., 2002), partitioned preferentially to the phase with the plasma membrane (polyethylene glycol). SLT2 and SULTR2 were detected almost exclusively in the plasma membrane fraction; however, a weak signal with the general SLT antibodies was also observed in the nonplasma membrane fraction. This latter partitioning pattern is similar to that of a tonoplast V-ATPase. In order to clearly establish whether or not these SO 4 22 transporters are localized to the plasma membrane, proteins from the polyethylene glycol phase were separated on a 15% to 45% Suc gradient (Fig. 6B). While the tonoplast V-ATPase was mainly in fractions 7 to 9, SLT1, SLT2, and SULTR2 migrated to the bottom of the gradient and were clearly localized with the plasma membrane marker proteins FOX1 ferroxidase and the H + -ATPase (fractions 1-3). proteins and the role of the proteasome in degradation. A, Chlamydomonas cells were grown in TAP and transferred to TAP2S. After 10 h of starvation, the culture was split into three aliquots: CHX was added to aliquots 1 and 3, and 1 mM SO 4 22 was added to aliquots 2 and 3. Samples were taken prior to starvation (lane 1), after 10 and 14 h of starvation (lanes 2 and 3, respectively), and 20, 60, and 240 min after the addition of CHX (lanes 4-6), SO 4 22 (lanes 7-9), and SO 4 22 + CHX (lanes 10-12). B, Identical cultures were used for this analysis, except that the cells starved for S for 10 h were administered SO 4 22 + CHX (lanes 4-6) or SO 4 22 + CHX + MG132 (lanes 7-9). The experiment was repeated three independent times, yielding essentially identical results, with representative data shown. The signal strength of each band was quantified by ImageJ software (Abramoff et al., 2004) and fitted to the exponential decay equation to obtain half-lives (t 1/2 ). A cytochrome c oxidase, COX2b, served as a loading control.

Expression in Heterologous Systems
To demonstrate that SULTR2, SLT1, and SLT2 encode functional SO 4 22 transporters, their cDNA clones were used to functionally complement the S. cerevisiae (yeast) uracil auxotrophy strain CP60-1C, which harbors mutations in the SUL1 and SUL2 SO 4 22 transporter genes. CP60-1C grows very slowly on medium containing 100 mM SO 4 22 (or less) as a sole S source (Cherest et al., 1997). Supplemental Table S1 shows that expression of Chlamydomonas SULTR2, SLT1, and SLT2 did not rescue the slow-growth phenotype of the CP60-1C strain in liquid medium containing 100 mM SO 4 22 as a sole S source; the same growth rate was observed for yeast cells carrying the vector alone and those carrying plasmids with Chlamydomonas SO 4 22 transporter genes. Both the GFP-tagged and untagged forms of transporters yielded the same results (failure to complement the yeast mutant). In contrast, CP60-1C cells transformed with Arabidopsis SULTR1;2 cDNA exhibited a significantly shorter doubling time. AtSULTR1;2 has previously been shown to rescue the yeast SO 4 22 transporter mutant (Shibagaki et al., 2002).

Generation of an Insertional Mutant Library
Our inability to confirm the functionality of the putative SO 4 22 transporters using the heterologous system prompted us to screen for Chlamydomonas strains bearing mutations in those transporter genes. Insertional mutagenesis has been widely used in Chlamydomonas to generate mutant strains affecting various biological processes. To obtain mutants with insertions in the putative SO 4 22 transporter genes, we generated a library containing approximately 52,000 insertional mutants using D66 as the parental strain. Cells were transformed with a 1.7-kb PCR fragment containing the AphVIII gene, which confers resistance to the antibiotic paromomycin, and transformants were plated on selective medium. The steps in the construction of the mutant library are described in "Materials and Methods." Genomic DNA isolated from transformants was screened by PCR to identify specific gene disruptions. Mutants with an insertion in the SO 4 22 transporter genes were identified by PCR using a target gene-specific primer and a primer specific to the inserted DNA (Supplemental Table S2 shows the list of primer pairs that were used). PCR products were sequenced to confirm that the disruption was in the targeted transporter gene. The specific transformant harboring the insertion was then identified, backcrossed (with strain 21gr) to generate a homogeneous genetic background, and then more fully characterized. Figure 7A shows the positions of the AphVIII gene in the strains disrupted for SLT1, SLT2, and SULTR2. The Southern-blot analyses illustrated that slt1, slt2, and sultr2 mutants all carry a single AphVIII insertion (Fig.  7B). These results were confirmed by genetic crosses of each single mutant with the wild-type strain, which demonstrated a 1:1 segregation ratio of paromomycinresistant to paromomycin-sensitive progeny (data not shown). In the slt1 mutant, the AphVIII construct was inserted in the eighth exon, interrupting the translation of a full-length product by introducing an incorrect splice site that caused a frame shift and the appearance of a premature stop codon (Supplemental Fig. S5A shows the alignment of the wild-type and Figure 6. Localization of the Chlamydomonas SLT1, SLT2, and SULTR2 transport proteins. A, Chlamydomonas strain D66 was grown in TAP medium to midlogarithmic phase and then grown in TAP2S for an additional 24 h. Total membranes (TM) were isolated and partitioned into plasma membrane fraction (polyethylene glycol fraction; PM) and a fraction containing all other membranes (dextran fraction; OM). A total of 20 mg of protein from each fraction was resolved by SDS-PAGE and transferred to a polyvinylidene difluoride membrane for immunoblot analyses using antibodies against SLT, SLT2, SULTR2, plasma membrane H + -ATPase, FOX1 ferroxidase (plasma membrane), CRD1 (thylakoid membrane and chloroplast envelope), COX2b (mitochondrial inner membrane), and V-ATPase (tonoplast). B, To separate the plasma membrane from the tonoplast, 600 mg of protein from the polyethylene glycol phase was loaded onto a 15% to 45% Suc gradient and separated into 11 fractions of equal volume (fraction 1 is at the bottom [B] of the gradient and fraction 11 is at the top [T] of the gradient). The proteins of each fraction were resolved by SDS-PAGE and transferred to a polyvinylidene difluoride membrane for immunoblotting, as described in "Materials and Methods." truncated SLT1 gene products). The aberrant SLT1 transcript in the slt1 mutant does not accumulate to the same extent as does the normal transcript in wild-type cells starved for S (the accumulation in the mutant is approximately 1 order of magnitude less than in the parental strain; Fig. 8A), suggesting that the mutation may also cause transcript instability. The lesion also abolished SLT1 protein accumulation during S deprivation, although a signal is observed when the general SLT antibody is used (Fig. 9A). Since the general SLT antibody recognizes both SLT1 and SLT2 polypeptides, the signal observed in the lane with polypeptides from the slt1 mutant is likely to result from the presence of SLT2, as there is essentially no detectable SLT signal in a slt1slt2 double mutant (Fig. 9B).
The integration site of the AphVIII gene in the slt2 mutant is in the 5# untranslated region (Fig. 7), 38 bp upstream of the translation start codon. Even though this disruption is not in the coding sequence, the accumulation of both the SLT2 transcript and polypeptide are severely impacted by the mutation (Figs. 8A and 9A): essentially no SLT2 protein is detected in the mutant. We also observed a novel phenotype of the slt2 mutant: the SLT3 transcript, which decreased during S-limited growth of wild-type cells and the slt1 and sultr2 mutants, consistently increased in the slt2 mutant (Fig. 8A).
In the sultr2 mutant, a fragment of the AphVIII construct was inserted in the ninth exon of SULTR2, resulting in the generation of a chimeric transcript with a premature stop codon. The SULTR2-AphVIII transcript would be translated into a truncated polypeptide missing the last 115 amino acids of SULTR2 (Supplemental Fig. S5B). This truncated polypeptide appears to be unstable and does not accumulate when the sultr2 mutant is starved for S (Fig. 9A).

SO 4 22 Uptake by the Mutant Strains
Prior to performing phenotypic analyses, each of the single mutants was backcrossed four times to the wildtype 21gr strain to eliminate other background mutations. The single mutants were then crossed to each other to generate double and triple mutants. The transcript and protein analyses for the strains with the multiple mutations are shown in Figures 8 and 9. There did not appear to be significant compensation for the absence of two S-deprivation-induced transporters in the double mutants; the level of transcript and polypeptide accumulation for the third transporter was similar to that observed in wild-type cells (Figs. 8B and 9B), based on normalization to the level of the CBLP transcript and COX2b protein, respectively. In order to assess whether the absence of specific SO 4 22 transporters impacted the rate of SO 4 22 transport, we measured the uptake rate of SO 4 22 over a range of external SO 4 22 concentrations (0.02-200 mM) in the wild type and the various mutant strains following S deprivation. As shown in Figure 10, the rates of SO 4 22 uptake in the slt1, slt2, and sultr2 single mutants were not markedly depressed relative to that of wild-type cells. In contrast, the three double mutants all showed a significant decline (by approximately 50%) in their rate of SO 4 22 transport relative to wild-type cells. Most extreme was the phenotype of the triple slt1slt2sultr2 mutant. This strain, which lacks all of the inducible transporters, exhibited a drastic decrease in the rate of SO 4 22 uptake. Indeed, the triple mutant exhibited essentially no increase in its SO 4 22 uptake capacity when deprived of S; the rate of SO 4 22 transport was the same in starved and unstarved cells (Fig. 10). Furthermore, wild-type cells and the slt1slt2sultr2 triple mutant showed a similar rate of SO 4 22 transport under nutrient-replete conditions. The K 1/2 was also calculated from the initial rates of SO 4 22 uptake for single, double, and triple mutants; we were unable to detect a significant difference in the transport affinity , and SULTR2 genes. Black bars, gray bars, and white bars represent exons, introns, and untranslated regions, respectively. B, DNA gel-blot analysis of wild-type, slt1, slt2, and sultr2 strains. Genomic DNA was digested with PstI, separated by agarose gel electrophoresis, transferred to a nitrocellulose membrane, and hybridized with the 1.7-kb PSAD::AphVIII PCR fragment. Asterisks designate the newly introduced PSAD::AphVIII gene sequences. Size standards (in kb) are shown to the left. of wild-type cells relative to any of the mutant strains (Supplemental Table S3). This may not be too surprising for the single and double mutants, but the triple mutant also showed a decrease in the K 1/2 that closely resembled that of wild-type cells. This might arise from the fact that the slt2 mutation only affects transcript accumulation during S deprivation and does not disrupt the coding region; a small amount of SLT2 protein can still be translated from the basal level of SLT2 transcript. In fact, a very weak signal using the SLT antibodies was detected in the slt1slt2sultr2 mutant in S-depleted medium, which is likely from a low level of SLT2 polypeptide (Fig. 9B). Alternatively, posttranslational modifications of the transporters that function during S-replete growth may alter their kinetic characteristics, although they remain at low levels in the plasma membrane (so the V max for uptake remains very low).
In order to confirm that the phenotype observed in the double and triple mutants is caused by the lesions in the SO 4 22 transporter genes, wild-type alleles of each transporter gene were introduced into the slt1slt2sultr2 mutant background by a genetic cross to a wild-type strain (D66). The SO 4 22 uptake rates of the progeny correlated perfectly with the number of SO 4 22 transporter genes present in each strain (Supplemental Fig. S6, A and B). Note that the relative uptake rates of the single and double mutants are similar to the results presented in Figure 10, even though the absolute transport rates are lower in this experiment (due to the background difference between the 21gr and D66 strains). In sum, these results demonstrate that SLT1, SLT2, and SULTR2 are the prominent, high-affinity transporters involved in SO 4 22 uptake when the cells are deprived of S, while other transporters function during S-replete growth. Furthermore, there appears to be some functional redundancy among the activities of these transporters.

SO 4 22 Transporters in Chlamydomonas
One of the early responses of most organisms to S deprivation is the generation of high-affinity SO 4 22 Figure 8. RT-qPCR for analysis of transcript accumulation in wild-type and mutant strains. The accumulation of ARS2, SLT1, SLT2, SLT3, SULTR1, SULTR2, and SULTR3 transcripts is shown in wild-type cells and in the single (slt1, slt2, sultr2; A), double (slt1slt2, slt1sultr2, slt2sultr2), and triple (slt1slt2sultr2) mutants (B). RNA samples were collected prior to, and 4 and 24 h after, the cells were transferred from TAP to TAP2S medium. Levels of individual transcripts are given as relative abundance with respect to that of the housekeeping control gene (CBLP ). transport systems. In this study, we identified genes encoding plasma membrane SO 4 22 transporters and demonstrated that the activation of these genes occurred specifically during S deprivation. The elevated transcript accumulation was detected 1 to 2 h after transferring cells to medium devoid of S (Fig. 2). It is noteworthy that the transcripts encoding the high-affinity SO 4 22 transporters accumulate significantly more rapidly than transcripts encoding other S-responsive genes, such as ARS2 and ECP76, following the transfer of cells to -S medium . These results suggest that there is a tiered regulation of the S-responsive genes and that the induction of the high-affinity SO 4 22 transport system is one of the earliest responses to S starvation. A distinct permease involved in translocating SO 4 22 from the cytoplasm into chloroplasts is composed of the subunits SulP1, SulP2, Sabc, and Sbp (Melis and Chen, 2005). Expression of all four genes encoding the components of this permease is induced upon S deprivation; levels of SulP and Sabc polypeptides also increase when the cells become S limited (Lindberg and Melis, 2008).
In addition, S-starved cells exhibit increased SO 4 22 uptake activity within 1 h of the onset of S deprivation . This initial increase in uptake rate precedes the accumulation of newly synthesized SO 4 22 transporters (Fig. 4B), suggesting that there might also be posttranslational modifications of the existing SO 4 22 transporters. This suggestion is in accord with the finding that the K 1/2 for SO 4 22 transport in the slt1slt2sultr2 triple mutant still rapidly changes following the imposition of S deprivation; the V max remains low because the three S-responsive transporters are not synthesized. Finally, transcripts for two putative SO 4 22 transporters, SULTR1 and SLT3, decline during S deprivation. The encoded proteins may have low-affinity transporter activity that is likely to primarily function in the uptake of SO transporters may be primarily used when the pH of the milieu around the cell is low, while the Na + /SO 4 22 cotransporters may be responsible for the majority of SO 4 22 uptake when the pH is high and when it is more efficient to use Na + as a counter ion.
The SLT2 and SLT3 genes are tandemly arranged on chromosome 10 in a head-to-tail orientation; the 3# untranslated region of SLT2 overlaps with the 5# untranslated region and the first exon of SLT3. In wild-type cells, SLT2 is heavily transcribed during S deprivation, and this likely interferes with the transcription of the downstream SLT3 gene, resulting in a decrease in SLT3 mRNA abundance (Fig. 8). In contrast, transcriptional activity of SLT2 in the slt2 mutant is decreased due to the insertion in the 5# untranslated region, which probably accounts for an unusually high rate of SLT3 transcription during S deprivation. Interestingly, SLT2 and SLT3 (76% identity, 85% similarity to each other) have a similar exonintron organization, suggesting that they arose from a gene duplication, but based on their unique expression patterns, the genes have subsequently specialized (Fig.  3A). Finally, the SLT and SULTR protein families each likely contains both high-and low-affinity transporters, allowing Chlamydomonas to tailor uptake processes to the features of the environment under both high-and low-S conditions.

Transcriptional Regulation of SO 4 22 Transporters
Several isoforms of SO 4 22 transporters play important roles in facilitating SO 4 22 uptake during S deprivation in Arabidopsis. The S-responsive induction of these transporter genes (AtSULTR1;1, AtSULTR1;2, AtSULTR4;2) has been shown to be under the control of SLIM1, the ETHYLENE-INSENSITIVE-LIKE3 transcriptional regulator (Maruyama-Nakashita et al., 2006). As already mentioned, a number of factors associated with S-responsive regulation in Chlamydomonas have also been identified; these include SNRK2.1, SNRK2.2, and SAC1. Mutants devoid of these regulatory elements show aberrant responses to S deprivation . We examined the accumulation of SO 4 22 transporter transcripts in wild-type cells and the sac1 and snrk2.1 mutants; these mutant strains do not elevate SO 4 22 transport activity nearly as much as wild-type cells when they are deprived of S (Fig. 1). In contrast to wild-type cells, the sac1 mutant only accumulates SLT1, SLT2, and SULTR2 transcripts 24 h after the imposition of S starvation. This is consistent with the model proposed by Moseley et al. (2009), in which SAC1 is a sensor that resides on the plasma membrane and acts as a negative regulator of the SNRK2.2 repressor protein kinase when environmental SO 4 22 levels fall below a certain threshold. This causes derepression of the S-responsive genes, including those encoding SO 4 22 transporters. After a prolonged period of S starvation, intracellular SO 4 22 levels also decline and further stimulate the transcription of S-responsive genes as a consequence of activation of the SNRK2.1 activator protein kinase; such activation can occur in the sac1 mutant, leading to elevated accumulation of SO 4 22 transporter mRNA 24 h after the onset of S starvation (Fig. 3B), which then leads to increases in the levels of transporter protein (Supplemental Fig. S2) and uptake activity (Fig. 1), although both are still lower than in wild-type cells.
Similarly, the snrk2.1 mutant is severely compromised in its ability to increase the rate of SO 4 22 uptake during S starvation (Fig. 1); its phenotype is significantly more severe than that of the sac1 mutant. The abundance of SLT1, SLT2, and SULTR2 transcripts in the snrk2.1 mutant, under both S-replete and S-depleted conditions, is much lower than the levels observed in wild-type cells (Fig. 3). For SULTR2, expression is induced in the snrk2.1 mutant during S deprivation, but the basal and fully induced transcript levels are both much lower than observed in wild-type cells. The SLT1 transcript does not accumulate in this mutant. These results suggest that SNRK2.1 is critical for maintaining both basal-level expression and the accumulation of SLT1 and SULTR2 transcripts during acclimation of the cells to S deprivation. Similar to the situation in the sac1 mutant, the inability of snrk2.1 to induce SO 4 22 transport activity during S deprivation is associated with its failure to accumulate SO 4 22 transporter transcripts. The capacity for SO 4 22 transport in wild-type, sac1, and snrk2.1 cells starved for S generally correlates with the levels of transcript and transporter polypeptide accumulation (Figs. 1 and 4; Supplemental Fig. S2). Together, these results indicate that SLT1, SLT2, and SULTR2 encode high-affinity SO 4 22 transporters that are controlled by the S status of the environment at the level of transcript abundance and that are responsible for the majority of SO 4 22 uptake when the cells become S deprived.
In contrast, the SULTR1 and SLT3 genes appear to be down-regulated during S starvation (the transcripts are reduced by 1,000-and 10-fold, respectively; Fig.  3A), suggesting that the proteins encoded by these genes are low-affinity SO 4 22 transporters that function primarily under S-sufficient conditions. Furthermore, S deprivation does not cause a significant reduction in SULTR1 and SLT3 transcript abundance in the sac1 mutant, while these transcripts are elevated to some extent in the snrk2.1 strain (Fig. 3, B and C). Hence, SAC1 and SNRK2.1 are important not only for stimulating the activity of the SLT1, SLT2, and SULTR2 genes but are also involved in depressing the activities of the SULTR1 and SLT3 genes during S limitation.
Even though the SULTR3 transcript does not appear to be regulated by the S status of the cells, its level accumulates during S starvation in a snrk2.1 background (Fig. 3, A and C). The inability of the snrk2.1 mutant to up-regulate expression of the high-affinity transporters and other genes involved in the acclimation program may result in severe internal S starvation that elicits secondary effects, including an increase in SULTR3 transcript accumulation. Alternatively, SNRK2.1 may play a direct role in regulating SULTR3 expression.
Accumulation and Turnover of SO 4 22 Transporters SLT1, SLT2, and SULTR2 polypeptides accumulate in S-starved cells 1 to 2 h after the transcripts peak (Fig.  4). These proteins are synthesized de novo upon imposition of S deprivation, since administration of CHX at the time of transferring cells from S-replete to S-depleted medium blocked their accumulation (Supplemental Fig. S3). Moreover, these SO 4 22 transporters are turned over rapidly when S becomes available; they are almost undetectable 8 h after SO 4 22 is added to S-starved cultures (Fig. 4B). The half-lives of SLT1, SLT2, and SULTR2 are two to three times longer in the absence of SO 4 22 ; the rate of turnover increases when the cells are placed in S-rich medium (Fig. 5). These data strongly suggest that de novo synthesis of SLT1, SLT2, and SULTR2 specifically facilitates SO 4 22 transport when cells experience S starvation. Once starvation conditions are relieved, these high-affinity transporters are rapidly degraded, with S (either SO 4 22 or a reduced S metabolite) stimulating the degradation process. The proteasome is involved in degrading SLT1 and SLT2, although it is not known whether the involvement is direct or indirect. Surprisingly, proteasome activity does not appear to be required for SULTR2 degradation. There are several examples of the degradation of plasma membrane transporter polypeptides by monoubiquitination and subsequent proteasome-independent proteolysis in vacuoles (Eguez et al., 2004;Wolf, 2004). Tight regulation of high-affinity transporter synthesis and turnover that depends on S availability would allow for an economy of energy utilization by the cells, balancing the transport process with intracellular demand, and might also help control the uptake of selenate, a toxic analog of SO 4

22
, under all environmental conditions. In addition, the high-affinity SO 4 22 transporters may have a lower capacity for SO 4 22 uptake than the low-affinity transporters (SULTR1 and SLT3), making them less suitable for SO 4 22 uptake when there is an abundance of the anion in the environment. To facilitate SO 4 22 uptake from the environment, algal and plant cells must have the capacity to synthesize high-affinity SO 4 22 transporters, some of which must reside on the plasma membrane. Using twophase aqueous polymer separation coupled with Suc density gradient centrifugation, we demonstrated that all three of the S-deprivation-induced transporters, SLT1, SLT2, and SULTR2, reside in the plasma membrane (Fig. 6). Expression of the cDNAs encoding these transporters in the yeast SO 4 22 transporter mutant CP60-1C failed to rescue its growth phenotype (Supplemental Table S1). While it was possible that the difference in Chlamydomonas and yeast codon usage could result in heterologous expression problems, this is probably not the case, since, as shown in Supplemental Figure S4A, the Chlamydomonas SO 4 22 transporter-GFP fusion proteins accumulated in the yeast transformants. However, the GFP fluorescence in yeast cells expressing the Chlamydomonas SULTR2-, SLT1-, and SLT2-GFP fusion proteins accumulates within intracellular membranes that probably originate from the endoplasmic reticulum (Supplemental Fig. S4B, I, II, and III). In contrast, in CP60-1C cells harboring AtSULTR1;2 fused to GFP, much of the fluorescence signal is distributed over the cell surface (Supplemen-tal Fig. S4B, IV). These results suggest that our inability to complement the yeast mutant is a consequence of inefficient localization of the Chlamydomonas transporters to the yeast plasma membrane. Perhaps plasma membrane localization of Chlamydomonas SO 4 22 transporters requires a posttranslational modification (e.g. phosphorylation) that does not occur in yeast cells. Recently, phosphorylation of the yeast nitrate transporter Ynt1 has been shown to be essential for its delivery to the plasma membrane during N limitation (Navarro et al., 2008). Similarly, phosphorylation of the yeast plasma membrane ATPase is critical for its proper maturation and routing to the cell surface (DeWitt et al., 1998). Alternatively, Chlamydomonas transporters may require an accessory protein in the secretory pathway for proper targeting. The PHOSPHATE TRANSPORTER TRAFFIC FACIL-ITATOR1, a plant-specific SEC12-related protein, has been demonstrated to play a critical role in facilitating endoplasmic reticulum exit of a high-affinity phosphate transporter in Arabidopsis .
We also attempted to express Chlamydomonas SO 4 22 transporters in Xenopus oocytes but never succeeded in obtaining oocytes with an increased rate of SO 4 22 uptake (data not shown). These results, along with those of the experiments described for heterologous expression of the Chlamydomonas transporters in the yeast system, suggested that there are biogenesisspecific processes that are required for proper maturation/localization of Chlamydomonas transporters. Intriguingly, while none of the high-affinity SO 4 22 transporters from Chlamydomonas was able to complement the CP60-1C strain, SO 4 22 transporters of Arabidopsis (SULTR1;1, SULTR1;2, and SULTR1;3) do rescue the yeast mutant phenotype (Supplemental Table S1; Takahashi et al., 2000;Shibagaki et al., 2002;Kataoka et al., 2004a). These results suggest that Chlamydomonas and Arabidopsis, which diverged over 1 billion years ago (Merchant et al., 2007), have different posttranslational modifications, stabilities, and/or mechanisms by which the transporters are localized to the plasma membrane.
Since we were unable to confirm the functionality of SULTR2, SLT1, and SLT2 in the heterologous systems, Chlamydomonas mutants with insertions in the SO 4 22 transporter genes were identified by a PCR-based screen (Fig. 7) and characterized at the physiological level by measuring SO 4 22 uptake rates in the various strains (Fig. 10). The results clearly show that the transporters are critical for the uptake and assimilation of SO 4 22 in S-deprived cells and that there is some functional redundancy among the transporters. Interestingly, in the S-starved sultr2 mutant, the uptake rates at low external SO 4 22 concentrations (0.02-2 mM) are comparable to those of wild-type cells, while the transport rates at higher SO 4 22 concentrations (20-200 mM) are much lower than the rates in the wild-type strain, suggesting that SULTR2 may be responsible for much of the SO 4 22 uptake into S-deprived cells.
Remarkably, increased uptake capacity normally observed during S deprivation is completely abolished in the triple mutant (Fig. 10B); the maximal SO 4 22 uptake rates of the slt1slt2sultr2 mutant during S-replete and S-depleted conditions are essentially the same. These data suggest that SLT1, SLT2, and SULTR2 are responsible for essentially all induced SO 4 22 uptake that is associated with S deprivation.

Isolation of SLT1, SLT2, and SULTR2 cDNAs
Total RNA was isolated from 12-h S-starved 21gr (wild-type) cells using a standard phenol-chloroform extraction protocol (Sambrook et al., 1989). Reverse transcription and PCR were performed using the Sensiscript RT Kit (Qiagen) and Pfu Turbo DNA Polymerase (Stratagene). Amplified PCR products were cloned into pENTR-D-TOPO (Invitrogen) and sequenced.

SO 4 22 Uptake Assays
SO 4 22 transport assays were performed as described previously . The rate of SO 4 22 uptake was measured at the indicated external concentrations of the anion over a 2-min time series.

Peptide Synthesis and Antibody Production
Specific antibodies against FOX1, CRD1, and COX2b were kind gifts from Dr. Sabeeha Merchant (University of California, Los Angeles). The antibodies to SULTR2 and SLT2 were prepared by Agrisera, while the SLT general antibody was prepared by Covance Research Products. Monospecific antibodies for SLT2 were generated in rabbits against a peptide region (DGKYLSKPDPNW) unique to this transport protein (Supplemental Fig. S1B). This peptide was conjugated to keyhole limpet hemocyanin via a terminal Cys and injected into rabbits. Resulting antibodies were affinity purified by column chromatography using an UltraLink Iodoacetyl Gel (Pierce) covalently bound to the unique SLT2 peptide used for antibody production. SLT2 peptide synthesis, peptide conjugation, antibody production, and antibody purification were all performed by Agrisera. To generate polyclonal antibodies that specifically recognize SULTR2, a cDNA encoding a unique region in the SULTR2 protein (Glu-677 to Gln-764; Supplemental Fig. S1A) was cloned in frame into the 3# terminus of the 103Hisencoding sequence of the expression vector pET-16b (Novagen) and introduced into Escherichia coli for isopropylthio-b-galactoside-inducible expression. E. coli cells expressing the fusion protein were grown to an A 600 of 0.5 and then induced with isopropylthio-b-galactoside for 4 h at 37°C. The 103His-SULTR2 protein was purified using a nickel-nitrilotriacetic acid agarose Superflow column (Qiagen) and then subjected to preparative SDS-PAGE (15% polyacrylamide gel). The region of the gel containing the fusion protein was excised and used for antiserum production in rabbits. To generate general antibodies that would react with both SLT1 and SLT2 (and potentially with SLT3), cDNAs encoding the region of SLT1 and SLT2 from Thr-337 to Val-584 (Supplemental Fig. S1B) were cloned into pET-16b (the SLT1 and SLT2 amino acid sequences from this region are 91% identical and 97% similar). These expressed 103His-SLT1-and 103His-SLT2-tagged polypeptides were resolved by preparative SDS-PAGE, as described above, and regions of the gel containing the fusion proteins were excised and used as antigens for antibody production.

Plasma Membrane Isolation
The D66 strain was used for plasma membrane isolations. Ten liters of cells were grown to midlogarithmic phase (3 3 10 6 cells mL 21 ) in a stirred bottle and starved for S for 24 h. Batches of resuspended cells (from 10 g pellet wet weight) were sonicated on ice using a Fisher Scientific Sonic Dismembrator model 550 with a microtip probe, power setting 3.5, and 15 cycles of 1 min of sonication, each followed by 1 min of cooling. The total microsomal fraction was separated into plasma membranes and a fraction containing the rest of the membranes using a two-phase aqueous polymer system described by Herbik et al. (2002). To fractionate membrane protein from the top (polyethylene glycol) phase of the two-phase aqueous polymer system, 600 mg of protein was loaded onto a 4.5-mL, 15% to 45% Suc gradient and centrifuged at 4°C for 16 h at 31,000 rpm in SW60Ti rotor (Beckman Coulter). The gradient was separated into 11 fractions of equal volume, diluted 5-fold in a dilution buffer (5 mM MOPS, pH 7.0, 2 mM dithiothreitol, 1 mM EDTA, and 1 mM phenylmethylsulfonyl fluoride), and centrifuged at 4°C for 1 h at 50,000 rpm in the TLA100.3 rotor (Beckman Coulter). Pellets were resuspended in 50 mL of the dilution buffer, and an equal volume of the loading buffer was added to the samples prior to an incubation at 42°C for 15 min. SDS-PAGE and immunoblot analyses were performed as described above.

Generation of Chlamydomonas Mutants
To generate a library of insertional mutants, a 1.7-kb PCR fragment containing the selectable marker gene AphVIII (conferring resistance to paromomycin; Sizova et al., 2001) under the control of the PSAD promoter was used for transformation. The cell wall-less strain D66 was transformed by electroporation (Shimogawara et al., 1998) using a modified procedure reported by Colombo et al. (2002). After transformation, cells were incubated in TAP medium for 16 to 18 h to allow for the accumulation of the AphVIII protein. Transformants were subsequently selected on TAP medium supplemented with 5 mg mL 21 paromomycin.

Generation and Screening of the Mutant Library
The mutant screen was performed according to Gonzalez-Ballester (2005). In brief, transformants were generated as described above, transferred to 96well microtiter plates containing approximately 100 mL of liquid TAP medium, grown for 5 d before they were used for genomic DNA isolation; the plates were wrapped in Parafilm to minimize evaporation. Each microtiter plate represents a pool of 96 insertional mutants. Genomic DNA was isolated from a pool of 96 transformants using a standard phenol-chloroform protocol (Sambrook et al., 1989). An equal concentration of genomic DNA from 10 pools was combined to create a "superpool" of DNA at a concentration of 100 ng mL 21 . To screen for mutants of interest, the superpool DNA was used as a template for PCR. Each PCR was performed using one gene-specific primer (the gene for which a disruption is sought; Supplemental Table S2) and one AphVIII construct-specific primer (RB2, 5#-TACCGGCTGTTGGAC-GAGTTCTTCTG-3#). Multiple gene-specific primers were paired with the AphVIII primer to increase our chances of identifying a disrupted target gene. PCRs were performed in a final volume of 25 mL containing 0.2 mL of Taq DNA Polymerase (Qiagen), 2.5 mL of 103 PCR buffer, 2 mL of deoxyribonucleotide triphosphates (2.5 mM each), 4 mL of Q solution (Qiagen), 1 mL of dimethyl sulfoxide (Sigma), 1 mL of template, 400 nM of each primer, and distilled water to make up the remainder of the 25-mL volume. Conditions used for amplification in the thermocycler were as follows: preincubation at 95°C for 5 min, followed by 35 cycles of sequential denaturation at 95°C for 30 s, annealing at 60°C for 30 s, and amplification at 72°C for 120 s. PCR products were separated by gel electrophoresis using 0.8% agarose gels. To identify DNA regions adjacent to the right border of the AphVIII construct (downstream from the PSAD promoter), the PCR product amplified with the gene-specific primer and RB2 was excised from the gel, purified using the QIAquick PCR Purification Kit (Qiagen), and sequenced. Primers used for this screening are listed in Supplemental Table S2.

Southern-Blot Analyses
Genomic DNA was isolated from 50-mL liquid cultures of the wild-type strain 21gr and slt1, slt2, and sultr2 mutants using a standard phenol-chloroform extraction protocol (Sambrook et al., 1989). Ten micrograms of genomic DNA was digested for 20 h with 10 units of restriction endonucleases (PstI; New England Biolabs). The fragments were separated by agarose (0.8%) gel electrophoresis and blotted overnight in 203 SSC onto nylon membranes (GeneScreen; DuPont-New England Nuclear), and the transferred DNA was cross-linked to the membrane by UV illumination. An alkaline phosphataselabeled probe was synthesized by chemical cross-linking of a thermostable alkaline phosphatase to the nucleic acid template. Probe synthesis and hybridization were performed using the AlkPhos Direct Labeling and Detection Systems following the manufacturer's protocol (Amersham Biosciences).

Rescue of the Mutant Phenotype
The phenotype of the slt1slt2sultr2 mutant was rescued by introducing wild-type alleles encoding the individual transporters into the triple mutant background. The slt1slt2sultr2 strain was crossed to the D66 wild-type strain, and the SO 4 22 transport rate of the progeny containing one (slt1, slt2, or sultr2) or two (slt1slt2, slt1sultr2, or slt2sultr2) mutations was determined by SO 4 22 uptake assays.

Supplemental Data
The following materials are available in the online version of this article.
Supplemental Figure S1. Amino acid sequence alignment of SO 4 22 transporter proteins.
Supplemental Figure S3. CHX inhibition of accumulation of the SO 4 22 transporter protein during S deprivation.
Supplemental Figure S4. Expression of SO 4 22 transporter-GFP fusion proteins in S. cerevisiae cells.
Supplemental Figure S5. Amino acid sequence alignments of SLT1 and SULTR2 wild-type and mutant gene products.
Supplemental Figure S6. Characteristics of SO 4 22 transport in wild-type cells and single, double, and triple SO 4 22 transporter mutants (progeny of a cross between a wild-type strain and a slt1slt2sultr2 triple mutant) deprived of S for 24 h.
Supplemental Table S2. List of SULTR2-, SLT1-, and SLT2-specific primers used for PCR screening of the insertion library.
Supplemental Table S3. Characteristics of SO 4 22 transport in wild-type cells and the single, double, and triple SO 4 22 transporter mutants after 24 h of S deprivation.
Supplemental Materials and Methods S1.