The Saccharomyces cerevisiae High Affinity Phosphate Transporter Encoded by PHO84 Also Functions in Manganese Homeostasis*

In the bakers’ yeast Saccharomyces cerevisiae , high affinity manganese uptake and intracellular distribution involve two members of the Nramp family of genes, SMF1 and SMF2 . In a search for other genes involved in manganese homeostasis, PHO84 was identified. The PHO84 gene encodes a high affinity inorganic phosphate transporter, and we find that its disruption results in a manganese-resistant phenotype. Resistance to zinc, cobalt, and copper ions was also demonstrated for pho84 (cid:1) yeast. When challenged with high concentrations of metals, pho84 (cid:1) yeast have reduced metal ion accumulation, suggesting that resistance is due to reduced uptake of metal ions. Pho84p accounted for virtually all the manganese accumulated under metal surplus conditions, demonstrating that this transporter is the major source of excess manganese accumulation. The manganese taken in via Pho84p is indeed biologically active and can not only cause toxicity but can also be incorporated into manganese-requiring enzymes. Pho84p is essential for activating manganese enzymes in smf2 (cid:1) mutants that rely on low affinity manganese transport systems. A role for Pho84p in manganese accumulation was also identified in a standard laboratory growth medium when high affinity manganese uptake is active. Under these conditions, cells lacking both Pho84p and the high affinity Smf1p transporter that similar systems for and which contain both high (24–26) and low affinity 28, 30). Although Pho84p seems to primarily under manganese surplus or low affinity conditions, our results indicate that PHO84 as well as SMF1 to the acquisition of manganese under conditions of moderate metal exposure ( in a standard laboratory medium). both SMF1 and PHO84 accumulated low levels of manganese, although no major on was observed in a standard laboratory medium. Our results suggest that under standard conditions the of is also in pho84 (cid:1) yeast, indicating that, like Pho84p contributes to accumulation under a wide range of environmental metal concentrations. Yet, this is not the case for all metals, because with copper and zinc the pho84 (cid:1) mutations only affected uptake with metal surplus and not under standard laboratory conditions. The different levels of PHO84 dependent metal accumulation observed with manganese, cobalt, zinc, and copper may be the result of metal availability in the medium or the competing influence of other specific transporters for these

In the bakers' yeast Saccharomyces cerevisiae, high affinity manganese uptake and intracellular distribution involve two members of the Nramp family of genes, SMF1 and SMF2. In a search for other genes involved in manganese homeostasis, PHO84 was identified. The PHO84 gene encodes a high affinity inorganic phosphate transporter, and we find that its disruption results in a manganese-resistant phenotype. Resistance to zinc, cobalt, and copper ions was also demonstrated for pho84⌬ yeast. When challenged with high concentrations of metals, pho84⌬ yeast have reduced metal ion accumulation, suggesting that resistance is due to reduced uptake of metal ions. Pho84p accounted for virtually all the manganese accumulated under metal surplus conditions, demonstrating that this transporter is the major source of excess manganese accumulation. The manganese taken in via Pho84p is indeed biologically active and can not only cause toxicity but can also be incorporated into manganese-requiring enzymes. Pho84p is essential for activating manganese enzymes in smf2⌬ mutants that rely on low affinity manganese transport systems. A role for Pho84p in manganese accumulation was also identified in a standard laboratory growth medium when high affinity manganese uptake is active. Under these conditions, cells lacking both Pho84p and the high affinity Smf1p transporter accumulated low levels of manganese, although there was no major effect on activity of manganese-requiring enzymes. We conclude that Pho84p plays a role in manganese homeostasis predominantly under manganese surplus conditions and appears to be functioning as a low affinity metal transporter.
Manganese is a biologically important metal that is required by many enzymes for activity. The enzymes that rely on manganese for activity range from carboxylases and phosphatases in the cytosol (1)(2)(3) to sugar transferases in the Golgi (4 -7) to a manganese-containing superoxide dismutase in the mitochondria (8 -10). However, this same metal can have deleterious effects in biological systems. Excessive accumulation of manganese can interfere with calcium metabolism (11) and increase errors during replication of mitochondrial DNA (12,13). In humans manganese is a potent neurotoxin, and indus-trial uses of manganese have led to cases of "manganism," which is characterized by disturbances in mental processes and symptoms similar to those of Parkinson's disease (14 -16). It is therefore critical that cells maintain manganese under tight homeostatic control.
The bakers' yeast Saccharomyces cerevisiae has served as an excellent model system in which to study the homeostasis of manganese. Two of the manganese transporters identified in this organism are Smf1p and Smf2p (17)(18)(19), members of the Nramp family of divalent metal transporters that are conserved from bacteria to humans (20). Smf1p is localized to the plasma membrane when manganese is limiting (21) and has the capacity for high affinity uptake of manganese (22). However, when cells are not starved for manganese, as is the case in standard laboratory growth medium, Smf1p does not appear to be essential for manganese accumulation. In smf1⌬ cells lacking this transporter there is no major deficiency in the uptake of manganese, and the activity of manganese-requiring enzymes is not significantly decreased, indicating that other cell surface manganese transporters must be active (23). However, the identity of such transporter(s) is currently unknown. The other Nramp manganese transporter of S. cerevisiae, Smf2p, does not directly function in the uptake of extracellular manganese but is localized to intracellular vesicles (19) and appears to function in the distribution of manganese within the cell (23). Yeast lacking SMF2 display a severe reduction in whole cell manganese and activity of manganese-requiring enzymes. The reduction in whole cell manganese of smf2⌬ mutants has been proposed to result from a feedback inhibition of cell surface manganese uptake (23).
A separate set of manganese transporter(s) may become active when cells accumulate elevated or toxic levels of the metal ("manganese surplus" state). This is certainly the case for zinc, copper, and iron uptake in yeast. Although high affinity transporters contribute to zinc, copper, and iron uptake when these metals are present at moderate levels (24 -26), upon exposure to elevated concentrations of these metals the low affinity transporter(s) become the primary source of metal for the cell (27)(28)(29)(30). In the current study we provide evidence that the apparent low affinity uptake of manganese in yeast is accomplished by the action of Pho84p, a cell surface transporter of phosphate (31,32). Yeast lacking PHO84 are resistant to high concentrations of manganese and have altered metal ion accumulation. Our results indicate that Pho84p works with the high affinity transporter Smf1p to help maintain cellular manganese levels. This is the first example of an eukaryotic phosphate transporter functioning in metal ion transport.
Yeast transformations were performed using the lithium acetate procedure (33). Cells were propagated without shaking at 30°C either in an enriched yeast extract, peptone-based medium supplemented with 2% dextrose (YPD) 1 or in minimal synthetic dextrose medium (34). Cultures for invertase activity assays were grown in a yeast extract, peptone-based medium with 2% lactate, and 0.1% glucose (YPLG).
The PHO84 disruption plasmid was generated by PCR amplifying the upstream (from Ϫ837 to Ϫ15) and downstream sequences (from 2008 to 2672) introducing BamHI and NotI or XhoI and BamHI sites, respectively. The PHO84 PCR products were digested with the indicated enzymes and ligated in a trimolecular reaction into pRS305 (LEU2) digested with XhoI and NotI, resulting in pLJ246. Transformation of yeast strains with pLJ246 digested with BamHI resulted in deletion of PHO84 sequences from Ϫ14 to 2007.
Metal Measurements-Yeast cells were grown to an A 600 of 2 in YPD medium or the same medium supplemented with the indicated metal ions. The cultures were harvested and washed with TE (10 mM Tris-Cl, and 1 mM EDTA, pH 8), then deionized water, and resuspended in deionized water as described (35). Metal analysis of whole yeast cells was carried out on a PerkinElmer Life Sciences AAnalyst 600 graphite furnace atomic absorption spectrometer according to the manufacturer's specifications.
Biochemical Assays-Superoxide dismutase enzymatic activity was assayed in cells grown in YPD medium to an A 600 of 2. For invertase activity gels, cells were grown in YPD to an A 600 of 1, washed with deionized water, resuspended in YPLG medium at an A 600 of 0.3, and grown for 5 h. To prepare cell lysates for electrophoresis on both denaturing and non-denaturing gels, cells were harvested, and spheroplasts were generated as described previously (35). Spheroplasts were lysed with 30 strokes of a microcentrifuge tube pestle in 10 mM HEPES, 1 mM EDTA, and 0.1% Tween 20, pH 7.5, containing protease inhibitors. Extracts were filtered through a 0.45-m membrane, and protein content was quantitated by the method of Bradford (36). Analysis of superoxide dismutase activity by non-denaturing gel electrophoresis and staining with nitro blue tetrazolium and immunoblot analysis of Sod2p levels were performed as described previously (23,37). Invertase activity gels were analyzed by non-denaturing gel electrophoresis and staining with tetrazolium red as described previously (23,38).

Disruption of PHO84 but Not Other Phosphate Transporters
Genes Causes Resistance to Manganese-Through an analysis of the Research Genetics collection of yeast deletion mutants, we noted that yeast lacking PHO84 were resistant to high concentrations of manganese. PHO84 encodes an inorganic phosphate transporter that localizes to the cell surface of S. cerevisiae and appears to be a major source of cellular phosphate (31,32). However, pho84⌬ mutants are viable because these mutants can obtain phosphate from other transporters. S. cerevisiae is known to express at least six phosphate transporters that contribute to cellular phosphate under various conditions (39). To determine whether manganese resistance was associated with the deletion of other phosphate transporters, we tested all six corresponding mutants in a manganese toxicity growth test. As seen in Fig. 1, of the six potential phosphate transporters only disruption of PHO84 resulted in increased resistance to manganese. The identical manganese-resistant phenotype of the pho84⌬ strain was observed in two independent genetic backgrounds, i.e. the wild type strain used in the Research Genetics collection (BY4741) (Fig. 1) and an unrelated wild type strain AA255 (not shown). This confirmed that the manganese resistance was in fact due to loss of PHO84.
Yeast Lacking PHO84 Are Resistant to Other Metal Ions-We tested whether pho84⌬ cells were resistant to other metal ions.
As seen in Fig. 2, deletion of PHO84 was not specific to manganese and conferred resistance to zinc and cobalt and a slight resistance to copper. As one possible explanation for the metal resistance, pho84⌬ mutants might accumulate lowered concentrations of these metals. We therefore monitored the total cellular concentration of manganese, zinc, cobalt, and copper in cells grown under conditions of moderate metal exposure (e.g. standard laboratory YPD growth medium not supplemented with metals) or exposed to concentrations of these metals that inhibited wild type cell growth by Ϸ50% (e.g. metal surplus conditions). When grown in a standard laboratory medium without metal supplementation, the cellular levels of zinc and copper were not affected by disruption of PHO84; however, we did observe nearly a 2-fold decrease in manganese and a 4-fold decrease in cobalt levels in the pho84⌬ strain relative to wild type (Fig. 3A). Under metal surplus conditions, accumulations of manganese, zinc, cobalt, and copper in the pho84⌬ strain were all reduced compared with the wild type (Fig. 3B). The largest effect was observed with manganese in which an 80-fold decrease in metal accumulation was obtained with pho84⌬ mutants. Metal accumulation was down 4-to 5-fold for both zinc and cobalt in the pho84⌬ strain, and copper showed the smallest difference with a ϳ2-fold decrease in pho84⌬ yeast.
These results indicate that reduced metal accumulation and presumably reduced metal ion uptake is the mechanism of metal ion resistance in pho84⌬ yeast. However, Pho84p may also be contributing to the accumulation of manganese and cobalt in a standard laboratory medium in which these metals are present at moderate concentrations. PHO84 Is Responsible for Increased Manganese Accumulation When Cells Are Exposed to Elevated Concentrations of the Metal-Using wild type and pho84⌬ strains, we sought to determine at what concentration of manganese in the growth medium does Pho84p contribute to manganese accumulation. Both strains were grown in YPD medium containing increasing concentrations of MnCl 2 , and whole cell manganese was measured by atomic absorption spectrophotometry. The steady state accumulation of manganese under these conditions is shown in Fig. 4. Interestingly, intracellular manganese remained relatively constant until the medium manganese concentration reached Ϸ5 M. At this manganese concentration a manganese surplus state ensued in which the wild type strain began to accumulate excess manganese at a level that was proportional to the level of extracellular metal. This increase in cellular manganese was PHO84-dependent, as manganese levels in the pho84⌬ strain were unchanged with concentrations of up to 25 M MnCl 2 (Fig. 4). The results clearly show that, under condi- tions of manganese surplus, virtually all the over-accumulation of manganese is dependent on PHO84.
PHO84 and SMF1 Both Contribute to Manganese Acquisition-To date, the only reported high affinity manganese transporter at the cell surface in yeast is Smf1p (22). However, yeast lacking SMF1 only displays a mild reduction in manganese accumulation and does not display symptoms of manganese starvation such as lowered activity in the manganese-containing enzymes Mn-Sod2p and the manganese-requiring mannosyltransferases (23). Clearly other routes of manganese acquisition are present beyond this high affinity transporter. To test the role of Pho84p in these other pathways of manganese acquisition, we created a strain lacking both SMF1 and PHO84. Under standard laboratory conditions (YPD medium not supplemented with manganese), a single disruptant of either SMF1 or PHO84 was associated with similar reductions in manganese content (Ϸ30 or Ϸ40% decrease, respectively). Disruption of both SMF1 and PHO84 was found to be additive and further decreased manganese accumulation Ϸ80% compared with wild type (Fig 5A). This additivity demonstrates that Smf1p and Pho84p are functioning in separate pathways for the accumulation of manganese.
The in vivo manganese status of cells grown under standard laboratory conditions was monitored using the activity of two distinct manganese-requiring enzymes, namely Mn-Sod2p in the mitochondria and mannosyltransferases in the Golgi. The activity of manganese-dependent mannosyltransferases was measured indirectly by observing the level of invertase glycosylation, which can be measured by electrophoretic shifts on non-denaturing gels (38). For example, smf2⌬ mutants, which accumulate very low levels of manganese, show an increased mobility of invertase, indicative of poor glycosylation (Fig. 5C) and (23). We noted that even with the dramatic decrease in manganese content, the smf1⌬ pho84⌬ strain did not exhibit symptoms of manganese starvation. As is seen in Fig. 5, B and C, smf1⌬ pho84⌬ cells grown in a standard laboratory medium contained near wild type levels of Mn-Sod2p activity, and the ability of mannosyltransferases to modify invertase was not impaired. In fact, invertase from cells disrupted for PHO84 showed slower mobility, indicating that invertase from these strains was modified to a somewhat greater extent than that from the wild type sample; however, the cause of this mobility difference is not clear. Therefore, even though the smf1⌬ pho84⌬ strains accumulate low levels of manganese in standard laboratory medium, there is no obvious impairment in the delivery of manganese to enzymes in the mitochondria or the Golgi. By comparison, smf2⌬ strains. which also accumulate low manganese, exhibit a noticeable deficiency in manganeserequiring enzymes (Fig. 5, B and C) and (23). Although the rationale for this difference is unknown, it may reflect differential compartmentalization and availability of the manganese in the two types of mutants.
Rescue of Mn-Sod2p Activity by Manganese Is Impaired in smf2⌬ pho84⌬ Yeast-The aforementioned experiments demonstrate that Pho84p is not critical for activation of manganese enzymes in a standard laboratory medium when high affinity manganese uptake systems are operative. However, the situation may be different when cells rely on more low affinity transport systems for manganese acquisition (e.g. manganese FIG. 4. PHO84 is responsible for increased manganese accumulation when cells are exposed to elevated concentrations of the metal. Strains AA255 (WT) and LJ328 (pho84⌬) were grown to mid-log phase in YPD medium supplemented with increasing concentrations of MnCl 2 , and whole cell metal content was measured as described in the Fig. 3 legend. surplus conditions). Low affinity transport systems are certainly important in the case of smf2 mutants of yeast, as these cells show severe symptoms of manganese starvation unless the medium is supplemented with elevated manganese (23). The addition of excess manganese to the growth medium, in this case 5-25 M manganese, corrects the smf2⌬ defect and completely restores the activity of Mn-Sod2p (Fig. 6A). Whole cell manganese content was also returned to wild type levels in the smf2⌬ strain with the addition of 5 M manganese, and excess manganese accumulation was observed upon exposure to higher manganese concentrations (Fig. 6B). This important low affinity pathway does not involve Smf1p, because deletion of SMF1 did not block the rescue of the manganese starvation phenotype by excess manganese in smf2⌬ yeast (Fig. 6A, right  panel). Thus, this manganese uptake must be occurring through an alternative manganese transporter, perhaps Pho84p. To determine whether this was the case, a strain containing a double disruption of SMF2 and PHO84 was generated. Measuring Mn-Sod2p activity and whole cell manganese content of smf2⌬ pho84⌬ yeast revealed that PHO84 is critical in alleviation of the manganese starvation phenotype when excess manganese is added to the growth medium (Fig.  6A). Whereas smf2⌬ cells require the addition of 5 M MnCl 2 to restore activity of Mn-Sod2p, yeast containing double disruptions of SMF2 and PHO84 require 50 M MnCl 2 to achieve similar levels of Sod2p activity. Even with 50 M manganese added, the whole cell manganese content of smf2⌬ pho84⌬ is still only ϳ60% that of untreated wild type yeast (Fig. 6B). Taken together, these results clearly demonstrate that Pho84p is a major component of the low affinity manganese transport system. This manganese taken up via Pho84p is indeed biologically active. The metal not only contributes to cell toxicity but can enter manganese utilization pathways and be incorporated into manganese-requiring enzymes such as Sod2p.

DISCUSSION
The results presented here provide strong evidence that PHO84 encoding an inorganic phosphate transporter also has a role in metal ion transport. Unlike dedicated metal transporters whose activity is tightly regulated by metal concentration (21, 24 -26, 40 -43), the uptake of excess metals by Pho84p appears to escape this regulation, perhaps because of its primary function in phosphate transport. As we observed, even in the presence of toxic concentrations of metal ions, PHO84-dependent metal transport continues. Disruption of PHO84 has the strongest effect on manganese accumulation under conditions of manganese surplus, and Pho84p appears to be the primary low affinity transporter of manganese. We find that the manganese transported by Pho84p is biologically active and can both cause toxicity and be incorporated into manganese-requiring enzymes.
The intracellular manganese content of yeast is remarkably constant over a certain range of extracellular manganese levels, e.g. up to 5 M in the case of enriched YPD medium. This buffering of intracellular manganese levels is likely maintained, at least in part, through the action of Pmr1p, a manganese-transporting P-type ATPase (44) that facilitates manganese export from the cell through the secretory pathway (45,46). It seems that when the external manganese concentration reaches a certain threshold (e.g. 5 M in YPD medium), Pho84p (pho84⌬ smf1⌬), and XL117 (smf2⌬) were grown to mid-log phase in YPD medium, and whole cell manganese content was measured as described in the Fig. 3 legend. B, lysates were generated from the strains shown in panel A; superoxide dismutase activity was assayed by native gel electrophoresis and staining with nitro blue tetrazolium (upper panel), and Mn-Sod2p protein levels were assayed by immunostaining (lower panel). C, the same strains shown in panel A were grown in YPLG medium, and the activity of mannosyltransferases was monitored by the mobility of invertase by native gel electrophoresis and staining with tetrazolium red. becomes effective in causing manganese accumulation, and the export of manganese through Pmr1p is not sufficient to offset the Pho84p-transported manganese.
We suggest that Smf1p and Pho84p together comprise a manganese uptake system that functions over a wide array of external metal concentrations. This is similar to what has been described for the metal transport systems for iron, copper, and zinc, which contain both high (24 -26) and low affinity transporters (27,28,30). Although Pho84p seems to function primarily under manganese surplus or low affinity conditions, our results indicate that PHO84 as well as SMF1 contribute to the acquisition of manganese under conditions of moderate metal exposure (e.g. growth in a standard laboratory medium). Yeast lacking both SMF1 and PHO84 accumulated low levels of manganese, although no major effect on activity of manganese-requiring enzymes was observed in a standard laboratory medium. Our results suggest that under standard conditions the transport of cobalt is also reduced in pho84⌬ yeast, indicating that, like manganese, Pho84p contributes to cobalt accumulation under a wide range of environmental metal concentrations. Yet, this is not the case for all metals, because with copper and zinc the pho84⌬ mutations only affected uptake with metal surplus and not under standard laboratory conditions. The different levels of PHO84dependent metal accumulation observed with manganese, cobalt, zinc, and copper may be the result of metal availability in the medium or the competing influence of other specific transporters for these metals.
Presumably, Pho84p is co-transporting divalent metal ions and phosphate in vivo in the form of metal-phosphate complexes. Evidence for such a transport mechanism has come from studies of recombinant Pho84p in reconstituted proteoliposomes (47). Recombinant Pho84p was found to require the presence of metal ions such as manganese and cobalt for phosphate transport, and a MeHPO 4 metal-phosphate complex was the proposed substrate. Not all metal phosphate complexes are good substrates for Pho84p, as the addition of magnesium was shown to actually have an inhibitory effect on phosphate transport with recombinant Pho84p (47). Consistent with this, we found an inhibitory effect of magnesium on the Pho84p-dependent uptake of manganese in vivo (data not shown). The preference of Pho84p for metal-phosphate complexes containing manganese (and also cobalt) is consistent with our proposal that Pho84p functions as a low affinity transporter for these metals in vivo.
A phosphate transporter that utilizes a metal-phosphate complex as a substrate is not unique to yeast. Bacterial phosphate transporters from Escherichia coli and Acinetobacter johnsonii have been described that require the presence of divalent cations for phosphate transport in reconstituted proteoliposomes (48,49). These bacterial phosphate transporters show a striking biochemical similarity to Pho84p. The substrate for both Pho84p and the bacterial phosphate transporters are thought to be neutral metal-phosphate complex, and each of these transporters shows a preference for MnHPO 4 (47)(48)(49).
Pho84p is homologous to transporters from mammalian, plant, fungal, and bacterial sources and include transporters for both phosphate and sugars (50 -54). Although Pho84p is the first eukaryotic phosphate transporter described with a role in metal transport, it is likely that this type of system is present in higher organisms, including humans. Exposure to high concentrations of manganese has been shown to result in the development of manganism, a neurological disorder similar to Parkinson's Disease that is characterized by high concentrations of manganese in the brain (14 -16). It is possible that neuronally expressed homologues to Pho84p contribute to manganese accumulation under metal surplus conditions. The unexpected connection between phosphate and manganese me-tabolism revealed here with studies in yeast may have implications in the understanding of manganese toxicity and diseases of manganese exposure in humans.