Elevation of cellular Mg2+ levels by the Mg2+ transporter, Alr1, supports growth of polyamine-deficient Saccharomyces cerevisiae cells

The polyamines putrescine, spermidine, and spermine are required for normal eukaryotic cellular functions. However, the minimum requirement for polyamines varies widely, ranging from very high concentrations (mm) in mammalian cells to extremely low in the yeast Saccharomyces cerevisiae. Yeast strains deficient in polyamine biosynthesis (spe1Δ, lacking ornithine decarboxylase, and spe2Δ, lacking SAM decarboxylase) require externally supplied polyamines, but supplementation with as little as 10−8 m spermidine restores their growth. Here, we report that culturing a spe1Δ mutant or a spe2Δ mutant in a standard polyamine-free minimal medium (SDC) leads to marked increases in cellular Mg2+ content. To determine which yeast Mg2+ transporter mediated this increase, we generated mutant strains with a deletion of SPE1 or SPE2 combined with a deletion of one of the three Mg2+ transporter genes, ALR1, ALR2, and MNR2, known to maintain cytosolic Mg2+ concentration. Neither Alr2 nor Mnr2 was required for increased Mg2+ accumulation, as all four double mutants (spe1Δ alr2Δ, spe2Δ alr2Δ, spe1Δ mnr2Δ, and spe2Δ mnr2Δ) exhibited significant Mg2+ accumulation upon polyamine depletion. In contrast, a spe2Δ alr1Δ double mutant cultured in SDC exhibited little increase in Mg2+ content and displayed severe growth defects compared with single mutants alr1Δ and spe2Δ under polyamine-deficient conditions. These findings indicate that Alr1 is required for the up-regulation of the Mg2+ content in polyamine-depleted cells and suggest that elevated Mg2+ can support growth of polyamine-deficient S. cerevisiae mutants. Up-regulation of cellular polyamine content in a Mg2+-deficient alr1Δ mutant provided further evidence for a cross-talk between Mg2+ and polyamine metabolism.

(CH 2 ) 4 NH(CH 2 ) 3 ) are ubiquitous in living cells and organisms and are normally present at high concentrations (mM) (1)(2)(3). In the yeast Saccharomyces cerevisiae, putrescine is produced from ornithine by ornithine decarboxylase (Spe1; Scheme 1A). Spermidine is produced from putrescine, and spermine is produced from spermidine by addition of an aminopropyl moiety from decarboxylated SAM, which is produced by SAM decarboxylase (Spe2; Scheme 1A). As the primary and secondary amino groups of polyamines are protonated at physiological pH, polyamines interact with negatively charged molecules such as nucleic acids, proteins, and phospholipids and influence their conformation, stability, and activity (4). The polyamines regulate a large number of cellular processes, including the enhancement of the efficiency and fidelity of translation (5). They are vital for survival of eukaryotes and are intimately involved in the regulation of eukaryotic cell growth. However, the precise modes of their action in supporting many cellular functions are not fully understood.
A high level (mM) of cellular polyamines is required for mammalian cell proliferation, and polyamine homeostasis is tightly regulated by intricate mechanisms (6). Numerous studies have demonstrated the antiproliferative effects of various inhibitors of polyamine biosynthesis or polyamine analogs that cause depletion of polyamines in mammals. The high polyamine requirement of mammalian cells is the rational basis of targeting the polyamine pathways in cancer chemoprevention and chemotherapies (1). Depletion of spermidine and spermine mediated by overexpression of the polyamine catabolic enzyme spermidine/spermine N 1 -acetyltransferase 1 also caused a dramatic inhibition of protein synthesis and cell growth (7). In contrast to mammalian cells, the polyamine requirement is extremely low in the yeast S. cerevisiae. A spe2⌬ mutant strain, which is unable to synthesize spermidine and spermine, grew at a nearly normal rate in medium containing 10 Ϫ8 M spermidine, a condition in which the cellular spermidine content dropped as low as 0.2% of WT levels (8).
One clearly defined function of the polyamine spermidine in eukaryotes is its role as precursor of the unusual amino acid hypusine (N ⑀ -4-amino-2-hydroxybutyl(lysine)) (9), which is formed by the post-translational modification of the eukaryotic translation factor eIF5A. The aminobutyl moiety of spermidine is conjugated to a specific lysine residue to form deoxyhypusine, which is subsequently hydroxylated to hypusine. Hypusine/de-oxyhypusine is essential for the activity of eIF5A. Eukaryotic cell proliferation and animal development depend on hypusinated eIF5A (10,11). Normally, Ͻ2% of cellular spermidine is used for hypusine synthesis in eukaryotic cells. However, in a yeast spe2⌬ mutant cultured in a medium containing Ͻ10 Ϫ8 M spermidine, spermidine became severely limiting, and as much as ϳ50% of the total cellular spermidine was mobilized for hypusine synthesis. These findings suggest that hypusination of eIF5A is the most critical function of polyamines in yeast, and that, unlike mammalian cells, yeast do not require a high intracellular concentration of polyamines for growth.
The apparent discrepancy in the minimal polyamine requirement between yeast and mammalian cells may be due to differences in the intrinsic functional roles of polyamines, or it may indicate that yeast possess a unique mechanism to compensate for polyamine deficiency. One such mechanism could be the compensatory accumulation of another cation such as magnesium (Mg 2ϩ ), the most abundant divalent cation in cells (12,13). Mg 2ϩ is a critical cofactor for over 300 enzymes, and it also serves a structural role by stabilizing protein domains. Like polyamines, the majority of cellular Mg 2ϩ is bound to negatively charged ligands such as ATP, RNA, DNA, or phospholipids (12). Functional overlap between polyamines and Mg 2ϩ has been suggested by several in vitro studies. Polyamines stimulated translation in cell-free lysates when Mg 2ϩ concentration was suboptimal (4,14), suggesting that polyamines and Mg 2ϩ can partially substitute for each other in protein synthesis. However, there is little information on the in vivo functional interaction between polyamines and Mg 2ϩ in the regulation of eukaryotic cell growth.
The cellular content of Mg 2ϩ is tightly controlled in yeast cells and remains constant over a range of 1-100 mM external Mg 2ϩ (15,16). However, when cells are cultured in low Mg 2ϩ medium (Ͻ100 M), intracellular Mg 2ϩ content is substantially reduced, and growth is limited (15). Regulation of cellular Mg 2ϩ is likely achieved by control of uptake systems, efflux from the cell, and sequestration within organelles (12,13,15) (Scheme 1A).
To mediate this regulation, yeast express five known Mg 2ϩ transporters, all related to the bacterial plasma membrane Mg 2ϩ transporter CorA (Scheme 1B): Alr1/Alr2 of plasma membrane, Mnr2 of the vacuolar membrane, and Mrs2/Lpe10 of the mitochondrial membrane (15,17). The CorA superfamily of Mg 2ϩ transporters share certain structural features, including two adjacent transmembrane domains near the C terminus that are connected by a short loop (Scheme 1B), although the amino acid sequences of the CorA/Mrs2/Alr1 superfamily have diverged significantly, and Alr1/Alr2 and Mnr2 have long N-domains compared with CorA and Mrs2/Lpe10. The universally conserved signature motif, GMN, near the end of the first transmembrane domain is important for the selectivity of Mg 2ϩ uptake (18,19). Alr1 was first identified as a factor whose overexpression confers resistance to aluminum, an inhibitor of CorA proteins (20). Loss of Alr1 function reduced cellular Mg 2ϩ content, and mutants lacking Alr1 displayed a severe growth defect. Alr2 is closely related to Alr1 in sequence but contributes little to Mg 2ϩ homeostasis and cell growth except in the absence of Alr1 (15,(20)(21)(22). Mnr2 is a vacuolar membrane protein required for release of Mg 2ϩ from storage vacuoles to the cytoplasm (15). Mrs2 and Lpe10 are two related mitochondrial membrane proteins required for the entry of Mg 2ϩ into the mitochondrial matrix (23). Both are required for mitochondrial function but do not influence whole-cell Mg 2ϩ accumulation nor do they contribute to the intracellular storage of excess Mg 2ϩ (15).
To investigate the relationship between polyamines and Mg 2ϩ in yeast cells, we first examined the effect of the spe1⌬ Scheme 1. The polyamine pathway and the Mg 2؉ transporters in S. cerevisiae and the inverse relation between cellular polyamines and Mg 2؉ . A, the polyamine biosynthesis pathway is shown on the left side. Deletion of SPE1 causes loss of all three polyamines, and deletion of SPE2 causes depletion of spermidine and spermine with increased accumulation of putrescine. Five S. cerevisiae Mg 2ϩ transporters are depicted on the right side. Of these, Alr1 is mainly responsible for the elevation of Mg 2ϩ content upon depletion of polyamines in S. cerevisiae. B, diagrammatic representation of five S. cerevisiae Mg 2ϩ transporters in relation to the bacterial Mg 2ϩ transporter CorA. Of these, Alr1, Alr2, and Mnr2 are important in the maintenance of cytoplasmic Mg 2ϩ concentration. Each Mg 2ϩ transporter contains two transmembrane domains near the C terminus that are connected by a short periplasmic loop. Each turn in the peptide chain represents 50 amino acids. The conserved GMN signature sequence critical for selective recognition of Mg 2ϩ is indicated. PA, polyamines; ODC, ornithine decarboxylase; SAMDC, SAM decarboxylase; DeSAM, decarboxylated SAM.

Interplay between polyamines and Mg in yeast cell growth
mutation (which blocks synthesis of all polyamines) or the spe2⌬ mutation (which blocks synthesis of spermine and spermidine) on cellular Mg 2ϩ content. Interestingly, cellular Mg 2ϩ content increased in response to polyamine depletion. Examination of double mutant strains lacking one polyamine biosynthesis gene (SPE1 or SPE2) and one Mg 2ϩ transporter gene (ALR1, ALR2, or MNR2) indicated that Alr1 alone was required for this elevated Mg 2ϩ accumulation. Consistent with this observation, Alr1 was found to be essential for the survival and growth of polyamine-deficient spe1⌬ and spe2⌬ cells. These findings provide strong evidence that yeast can specifically compensate for polyamine deficiency by up-regulating the accumulation of Mg 2ϩ ions. The elevation of the cellular polyamine levels in the Mg 2ϩ -deficient alr1⌬ mutant cultured in YPD 2 further suggests an interaction between polyamine and Mg 2ϩ metabolism.

Depletion of cellular polyamines leads to an elevation of Mg 2؉ content in S. cerevisiae
We compared the growth, polyamine content, and Mg 2ϩ content of WT yeast (Y534; BY4741) and the two polyamine biosynthesis mutants Y535 (spe1⌬) and Y536 (spe2⌬) ( Table 1 and Fig. 1). In Fig. 1 (and Figs. 3 and 4), all data are color-coded (parental strains (brown), spe1⌬ mutant (pink), and spe2⌬ mutant (blue)). For comparison of growth, cells were initially inoculated at a very low density (0.0003), and the optical density was followed for 72 h. Cultures were regularly diluted into fresh medium to maintain log phase. In YPD rich in polyamines, the mutations had no effect on growth (Fig. 1A), as the mutants utilized polyamines supplied from the medium. Growing cultures for long periods in polyamine-free SDC led to the depletion of initial polyamine stores in spe1⌬ and spe2⌬ mutants.
The growth of Y535 (spe1⌬) and Y536 (spe2⌬) in SDC declined with time and stalled after 20 -40 h (Fig. 1B), and the growth defects were magnified with prolonged incubation. The content of spermidine and spermine in the mutants grown in YPD was only slightly less than that in the WT (Fig. 1C). A higher level of putrescine was observed in spe2⌬, resulting from the blockage of conversion of putrescine to spermidine and spermidine to spermine in the absence of SAM decarboxylase (Spe2) (Scheme 1A). The spe1⌬ cells cultured in SDC did not contain any detectable polyamines, and spe2⌬ cells contained a highly elevated level of putrescine but no spermidine or spermine ( Fig. 1D) as expected. The Y536 (spe2⌬) strain grew better than Y535 (spe1⌬) in SDC, suggesting that the high level of putrescine partially fulfilled the polyamine requirement.
To determine the effect of these changes in polyamine content on Mg 2ϩ homeostasis, parallel samples of Y534 (WT), Y535 (spe1⌬), and Y536 (spe2⌬) cells were also taken for analysis of their elemental content. There was little or no difference in Mg 2ϩ content among the three strains after growth in YPD (3.1-3.3 mg of Mg 2ϩ /g dry weight) (Fig. 1E). However, after 24-h culture in SDC, the Mg 2ϩ content of the polyamine-deficient spe1⌬ and spe2⌬ cells increased substantially (by ϳ2.5and ϳ1.6-fold, respectively) (Fig. 1F). The increase was consistently more pronounced in spe1⌬ than spe2⌬ cells, suggesting that the degree of Mg 2ϩ accumulation was responsive to the severity of polyamine deficiency. To verify that this change in Mg 2ϩ content was a consequence of polyamine deficiency and not the absence of some other component of YPD in SDC, Mg 2ϩ content was also measured in cells cultured in SDC supplemented with 10 Ϫ8 or 10 Ϫ6 M spermidine (Fig. 1, G and H). The elevation in cellular Mg 2ϩ content of the mutants was reduced by spermidine supplementation. When cultured in SDC containing 10 Ϫ8 M spermidine, Mg 2ϩ content increased ϳ2.1-fold in spe1⌬ and ϳ1.4-fold in spe2⌬ (Fig. 1G). We chose this concentration of spermidine, as the spe2⌬ mutants can grow nearly normally in SDC containing 10 Ϫ8 M spermidine when cellular polyamines were limiting (0.2% of normal level) (8). A strong increase in the Mg 2ϩ content under this condition

Interplay between polyamines and Mg in yeast cell growth
suggests that Mg 2ϩ elevation contributes to the nearly normal growth of the mutant. When cells were supplemented with a much higher level of spermidine (10 Ϫ6 M), only small increases in Mg 2ϩ content were observed (13.2 and 6.8% in spe1⌬ and spe2⌬, respectively) ( Fig. 1H), confirming an inverse relationship between spermidine supply and Mg 2ϩ accumulation by the mutant strains.
To determine whether this effect of polyamine deficiency was specific to Mg 2ϩ or reflected a more general effect on nutrient accumulation, we examined the content of potassium (K), manganese (Mn), zinc (Zn), and phosphorus (P) ( Table 2). Only small variations (Ͻ13%) in potassium content were observed in the three strains cultured in YPD, SDC, and SDC supplemented with spermidine. No consistent negative or positive effect of polyamine depletion was observed on the content of potassium, indicating that polyamine deficiency does not cause accumulation of cations in general. Like the Mg 2ϩ content, the Zn 2ϩ content of Y535 and Y536 cells was increased after culture in SDC (ϳ3-and ϳ1.4-fold, respectively), and the elevated Zn content was reduced in cells cultured in SDC supplemented with spermidine. In contrast, the Mn 2ϩ content was not elevated in response to polyamine depletion. These results suggest that Mg 2ϩ and Zn 2ϩ are transported by the same transporter, Alr1, but Mn 2ϩ is not. A moderate increase in phosphorus content was observed in spe1⌬ cells when polyamines were depleted and cellular Mg 2ϩ was substantially increased. This effect is consistent with previous reports showing a close relationship between Mg 2ϩ content and phosphate accumulation by yeast (15,24). Overall, these data indicate that a major effect of polyamine depletion is elevation of the Mg 2ϩ content, perhaps mediated by a change in transporter activity.

Generation of mutant strains with a combined deletion of a polyamine biosynthesis gene and a Mg 2؉ transporter gene
As yeast responded to polyamine deficiency by increasing Mg 2ϩ accumulation, we suspected that this increase was essential to maintaining viability and growth. If so, inactivation of one of the required Mg 2ϩ transporters might prevent Mg 2ϩ accumulation and compromise the growth of polyamine-deficient cells. To examine this possibility, we constructed a set of double mutant strains combining the spe1⌬ or spe2⌬ mutation with a mutation in one of the three Mg 2ϩ transporters (Alr1, Alr2, or Mnr2) (Scheme 1, A and B), known to be important in the regulation of cytosolic Mg 2ϩ concentration (15). Double Interplay between polyamines and Mg in yeast cell growth mutant strains were constructed by deletion of SPE1 or SPE2 in the Mg 2ϩ transport mutant strains CM200 (alr1⌬), NP27 (alr2⌬), and NP180 (mnr2⌬) by transformation with a PCRamplified spe1⌬::KanMX4 or spe2⌬::KanMX4 marker (25) and selection of G418-resistant clones. Transformation of two isogenic parental strains (DY1457 for CM200 and NP27, and NP174 for NP180) was also performed in parallel. New spe1⌬ and spe2⌬ strains representing all combinations of mutations were isolated successfully (AH1-AH9; Table 1) with the exception of the spe1⌬ alr1⌬ strain. The inability to isolate this mutant suggests that the combination of the two mutations was lethal. The knockout status of the five genes, SPE1, SPE2, ALR1, ALR2, and MNR2, in the parental strains and the nine new knockout mutants (AH1-AH9) was confirmed by PCR (Fig. 2) using an ORF primer set and a knockout primer set (Table 3). For each spe1⌬ and spe2⌬ mutant, a PCR product with an expected size was detected with a knockout primer set but not with an ORF primer set.

Efficient Mg 2؉ uptake by Alr1 is essential for survival and growth of polyamine-deficient cells
To determine the effect of lack of a Mg 2ϩ transporter in polyamine deficiency, we first compared the growth of each of the spe1⌬ and spe2⌬ mutants derived from the WT DY1457 and the two mutants, CM200 (alr1⌬) and NP27 (alr2⌬), in standard SDC (Fig. 3). SDC is polyamine-free and was chosen to maximally display the growth defect resulting from deletion of SPE1 or SPE2. In addition, the Mg 2ϩ concentration of SDC (4 mM) is lower than that required for optimum growth of alr1⌬ strains (15,16) while still allowing measurable growth. Thus, SDC should reveal any growth defects resulting from novel synthetic interactions. As previously observed for the Y534, Y535, and Y536 series of strains (Fig. 1), loss of SPE1 in each case caused a more pronounced growth defect than loss of SPE2 (Fig. 3, A and C). Each spe1⌬ mutant lacked all polyamines, and each spe2⌬ mutant contained only putrescine at a highly elevated level, confirming the knockout of SPE1 or SPE2, respectively (Fig. 3, D-F).
The synthetic growth defect of AH3 (alr1⌬ spe2⌬) was associated with the loss of the ability to accumulate Mg 2ϩ upon polyamine depletion. There was little or no increase in the Mg 2ϩ content of AH3 (alr1⌬ spe2⌬) over that of its parental strain CM200 (alr1⌬) (Fig. 3H), whereas all other spe1⌬ and spe2⌬ mutants derived from Alr1-expressing strains (AH1, AH2, AH4, and AH5) showed considerable increases in Mg 2ϩ content (ϳ2-and ϳ1.7-fold for spe1⌬ and spe2⌬ strains, respectively) (Fig. 3, G and I). Although the polyamine patterns were consistent with deletion of SPE1 or  The status of SPE1 and SPE2 genes was determined in six sets of isogenic strains marked by bars above the strain names. Knockout of the SPE1 or SPE2 gene was confirmed by the presence of a PCR product with a knockout primer set and the absence of a PCR product using an ORF primer set. From each of the four sets of PCR reactions (four panels), one main product (either a knockout PCR product or an ORF PCR product) was generated, consistent with each genotype. The positions of 1-kb ladder DNA standards flanking the PCR product are marked on the right side of each panel.

Interplay between polyamines and Mg in yeast cell growth
SPE2 in each of these strains, putrescine was much higher in AH3 (Fig. 3E) than in other spe2⌬ mutants. The inability of AH3 (alr1⌬ spe2⌬) to enhance Mg 2ϩ levels might have caused a compensatory overaccumulation of putrescine. Expression of SPE2 in AH3 using pFL38/SPE2 restored polyamine content (Fig. 3E, tan bars) and cell growth (Fig. 3B, tan line) and complemented the Mg 2ϩ accumulation defect, indicating that these phenotypes were a specific consequence of the spe2⌬ mutation.
We also examined the effects of deletion of SPE1 or SPE2 in an mnr2⌬ background (Fig. 4). As Mnr2 is required for the release of vacuolar Mg 2ϩ stores under Mg 2ϩ -deficient conditions (15), we suspected that this transporter might release Mg 2ϩ from storage vacuoles to increase cytosolic Mg 2ϩ concentration in response to polyamine deficiency and enhance growth of spe1⌬ or spe2⌬ mutants. However, no notable differences in the growth (Fig. 4, A-B) and the polyamine patterns (Fig. 4, C-D) were observed between the polyamine synthesisdeficient mutants derived from WT(NP174) and the mnr2⌬ mutant (NP180). The growth and the polyamine patterns of AH8 (mnr2⌬ spe1⌬) and AH9 (mnr2⌬ spe2⌬) were similar to those of AH6 (spe1⌬) and AH7 (spe2⌬), respectively (Fig. 4, A  and B), an indication that Mnr2 is not required for tolerance to polyamine deficiency. Interestingly, the mnr2 mutation did substantially increase total cellular Mg 2ϩ content over the levels observed in spe1⌬ or spe2⌬ single mutants (Fig. 4, E and F, compare AH8 with AH6 and AH9 with AH7). Although direct measurement of the vacuolar Mg 2ϩ content has not been made, this increase is probably due to elevated uptake of Mg 2ϩ by Alr1 in the spe1⌬ or spe2⌬ mutants and the increased sequestration of Mg 2ϩ in the vacuoles of polyamine-depleted mnr2⌬ cells (AH8 and AH9) (15). The observation that loss of Mnr2 function did not inhibit growth of polyamine-deficient strains argues that this store is not normally utilized to compensate for polyamine deficiency. Taken together, these results indicate that Alr1, but not Alr2 or Mnr2, is required for the elevated accumulation of Mg 2ϩ in polyamine-deficient cells and that this response is essential for survival.

Effects of Mg 2؉ supply on polyamine metabolism
Because the above data revealed an inverse relationship between polyamine and Mg 2ϩ content in polyamine-depleted S. cerevisiae cells, we wondered whether Mg 2ϩ deficiency would lead to an increased polyamine accumulation. To address this question, we compared growth and Mg 2ϩ and polyamine content of DY1457 (WT) and CM200 (alr1⌬) strains cultured in YPD supplemented with different concentrations of Mg 2ϩ (0 -200 mM). Mg 2ϩ supplementation was necessary for normal growth of the alr1⌬ mutant, as the Mg 2ϩ concentration of YPD is quite low (ϳ500 M) (26). The growth of DY1457 was not enhanced by Mg 2ϩ supplementation and was slightly inhibited with increasing concentration of Mg 2ϩ (Ͼ30 mM) (Fig. 5A). The exogenously supplied Mg 2ϩ had little influence on cellular Mg 2ϩ content of WT cells (Fig. 5B, brown bars). In contrast, the growth of the alr1⌬ mutant was severely inhibited in YPD but increased with supplemented Mg 2ϩ and was restored to the WT level at 200 mM Mg 2ϩ (Fig. 5A, red line). The Mg 2ϩ content of the alr1⌬ mutant cultured in standard YPD was ϳ1.43 mg/g dry weight (46% of WT) (Fig. 5B). Mg 2ϩ supplementation increased the Mg 2ϩ content of CM200 (alr1⌬) to 2.1 mg/g dry weight at 200 mM Mg 2ϩ (Fig. 5B, red bars) but did not restore it to WT levels. Inability of the alr1⌬ to fully restore the Mg 2ϩ content is probably due to inefficient Mg 2ϩ uptake preventing the refill of all the intracellular stores, which represent up to 80% of the total Mg 2ϩ content (15). However, our results indicate that the minimal Mg 2ϩ requirement for normal growth was met in the alr1⌬ cells at this level by supplementation with 200 mM Mg 2ϩ .

Interplay between polyamines and Mg in yeast cell growth
Strikingly, the total polyamine content of Mg 2ϩ -deficient CM200 (alr1⌬) cells cultured in standard YPD was ϳ67% higher than that of WT DY1457 cells (Fig. 5C). This heightened level decreased as the Mg 2ϩ concentration in the medium increased and returned to the WT level at 100 mM Mg 2ϩ , a level that almost completely suppressed the alr1⌬ growth defect  (Fig. 5A). Thus, yeast responded to Mg 2ϩ deficiency by increasing polyamine content. The elevation in polyamine content in the Mg 2ϩ -deficient alr1⌬ cells suggests an interrelationship between Mg 2ϩ and polyamine metabolism. A slight decrease in total polyamines was also detected in DY1457 with increasing Mg 2ϩ supplementation. There was also a differential decline in the relative levels of spermine in both DY1457 and CM200 as the Mg 2ϩ level in the medium increased, suggesting effects of Mg 2ϩ on cellular polyamine metabolism.

Discussion
In this study, we present strong evidence that yeast cells accumulate excess Mg 2ϩ to maintain viability upon depletion of cellular polyamines (Scheme 1A). This increase in the cellular Mg 2ϩ content of polyamine-deficient spe1⌬ and spe2⌬ cells was reversed by the addition of spermidine in the medium (Fig.  1), suggesting a direct response to polyamine availability. Given that the cellular Mg 2ϩ content is normally maintained within a narrow range in yeast (15), even in medium containing high Mg 2ϩ (100 -200 mM), the observed increases in cellular Mg 2ϩ content (1.6 -2.5-fold) in polyamine-deficient cells represent a marked deviation with likely functional significance.   (n ϭ 4). B and C, approximately 10 OD 600 units of cells were harvested (at OD 600 of ϳ1 or less) for analysis of Mg 2ϩ content (B) and polyamine content (C) as described in Fig. 1. B, the values for the Mg 2ϩ content are indicated as black dots, and each bar indicates the mean and the error bars represent S.D. (n ϭ 4). C, the values for the total cellular polyamine content are indicated by a merged bar of three segments representing spermine, spermidine, and putrescine (top to bottom). Each segment represents the mean Ϯ S.D. (n ϭ 4). The p values were calculated for the Mg 2ϩ content by ANOVA. p values less than 0.05 were considered statistically significant: *, p Յ 0.05; **, p Յ 0.01 compared with the control.

Interplay between polyamines and Mg in yeast cell growth
The vital importance of elevated Mg 2ϩ accumulation for polyamine-deficient yeast cells was indicated by the severe synthetic growth defect of an alr1⌬ spe2⌬ mutant, which was not able to elevate cellular Mg 2ϩ content (Fig. 3) in response to polyamine deficiency. Our data further demonstrate that the polyamine depletion-induced Mg 2ϩ accumulation is mainly dependent on the plasma membrane Mg 2ϩ transporter Alr1 and not on its isoform Alr2 or the vacuolar Mg 2ϩ transporter Mnr2. This study presents firm in vivo evidence for the functional interaction between Mg 2ϩ and polyamines in the regulation of S. cerevisiae growth.
The elevation of cellular Mg 2ϩ in polyamine-deficient cells is not due to a general increase in cation uptake, as other cations such as K ϩ or Mn 2ϩ did not increase in response to polyamine depletion. Although the Zn 2ϩ content was also increased in response to polyamine depletion (Table 2), the increased Zn 2ϩ content was far too low to compensate for the cation loss from polyamine depletion. The elevated Mg 2ϩ would be sufficient to maintain the charge balance of polyamine-deficient cells. Furthermore, there are no reports suggesting that Zn 2ϩ can substitute for polyamines in macromolecular synthesis. These findings corroborate the specificity of the Mg 2ϩ function in supporting the growth of polyamine-depleted S. cerevisiae cells.
The specificity of Alr1 toward other divalent cations is not well-understood. Direct measurement of competitive inhibition of Mg 2ϩ transport by other metals has not been conducted in yeast as no radioactive Mg 2ϩ isotope is commercially available. The increased sensitivity of an Alr1-overexpressing strain toward other metals (La 3ϩ , Co 2ϩ , Mn 2ϩ , Ni 2ϩ , and Zn 2ϩ ) suggested a possibility of transport of these other metals by Alr1 (20). However, our metal content analysis data ( Table 2) imply that Mg 2ϩ and Zn 2ϩ , but not Mn 2ϩ , are transported by Alr1. Further studies are needed using the Alr1-overexpressing strain as well-as the alr1⌬ strain to clarify the substrate specificity of Alr1.
In addition to the apparent regulatory effect of cellular polyamines on Mg 2ϩ , an inverse relationship between polyamines and Mg 2ϩ was also observed in the Mg 2ϩ transport-deficient alr1⌬ cells cultured in YPD (Fig. 5). In these cells, the total polyamine content increased in response to Mg 2ϩ deficiency. These findings suggest that the polyamine pathway enzymes and/or polyamine transporters are regulated by cellular or exogenous Mg 2ϩ concentration. Future investigations to identify and elucidate the specific polyamine pathways regulated by the Mg 2ϩ level are warranted.
Despite the vital importance of Mg 2ϩ in cellular physiology, the mechanism of Mg 2ϩ homeostasis in eukaryotic cells, not to mention its functional interactions with polyamine homeostasis, is poorly understood. Although the yeast CorA family Mg 2ϩ transporters have diverged considerably from the bacterial transporter CorA, they share common structural features, including two transmembrane domains connected by a short loop near the C terminus and a signature motif, GMN, that is important for Mg 2ϩ selectivity (Scheme 1B). A cryo-EM structure of the Thermotoga maritima CorA suggested a mode of its regulation by intracellular Mg 2ϩ concentration (27); at high Mg 2ϩ concentration, Mg 2ϩ is bound to cytoplasmic N-terminal domains of the CorA homopentamer, and this binding induces a closed conformation, whereas loss of Mg 2ϩ binding in a low-Mg 2ϩ environment reverses it to an open channel. The acidic residues involved in binding cytosolic Mg 2ϩ , located at the subunit interfaces of the pentamer, were identified (28). Chemical cross-linking and the split-ubiquitin assay data (22) suggest that Alr1 and Alr2 also form homo-or heterooligomers and that, like CorA, they function as pentamers. Residues that participate in Mg 2ϩ binding at the regulatory sites are conserved in prokaryotic CorA proteins (28) and in Alr1 (29), suggesting conservation of this mechanism. The CorA and Alr1/ Alr2 proteins are functionally interchangeable, as overexpression of CorA partially restored the growth of an alr1⌬ strain (16).
Our observation that polyamine deficiency increases Mg 2ϩ accumulation via Alr1 suggests that this protein may be regulated directly or indirectly by cytosolic polyamine concentration. At least three possible models might explain this effect. First, polyamines might bind directly to Alr1 and inhibit its activity in a manner similar to cytosolic Mg 2ϩ (27). Polyamines might interact directly with the Mg 2ϩ -binding sites to mimic the effect of Mg 2ϩ , or they may bind to other sites within the Alr1 cytoplasmic domains. Second, polyamines might bind directly to the Mg 2ϩ channel pore itself to block it, as has been observed for some potassium channels (28). A third mechanism for increased Mg 2ϩ accumulation might be the induction of expression of Alr1 mRNA and/or protein by a polyamine-dependent regulatory mechanism. Comparison of Alr1 mRNA levels in polyamine-replete and polyamine-deficient S. cerevisiae cells by quantitative RT-PCR did not reveal up-regulation of Alr1 expression upon depletion of polyamines (data not shown). This mechanism is also less likely than the others, considering that Alr1 transport activity was enhanced under Mg 2ϩ deficient conditions, but the induction of the Alr1 protein or its mRNA was not observed (29).
A functional interaction between polyamines and Mg 2ϩ has been suggested by several previous studies. In a reticulocyte lysate freed of polyamines by gel filtration, polyamines enhanced translation at suboptimal concentrations of Mg 2ϩ and beyond the level achieved by high Mg 2ϩ alone, suggesting that polyamines can at least partly substitute for Mg 2ϩ (4,14). Evidence for their functional interaction in vivo was also reported in S. cerevisiae (30) and mammalian cells (31) in which polyamine overloading caused cellular toxicities by displacement of cellular Mg 2ϩ . The growth inhibition was attributed to excessive accumulation of polyamines leading to a concomitant decrease in cellular Mg 2ϩ . It was suggested that replacement of ribosome-bound Mg 2ϩ by accumulated polyamines inactivated ribosomes, leading to inhibition of protein synthesis and cell growth, but no underlying molecular mechanism was identified. Our study reveals an essential functional requirement for Mg 2ϩ content elevation in polyamine-deficient cells, a biological situation opposite to that of cellular toxicity caused by excess polyamines (30). Although the two studies complement each other in supporting the concept of functional interplay between polyamines and Mg 2ϩ , our study uniquely reveals a mechanism of up-regulation of cellular Mg 2ϩ content involving Alr1 upon depletion of polyamines and offers an explanation for the extremely low polyamine requirement for S. cerevisiae

Interplay between polyamines and Mg in yeast cell growth
growth. In addition, this study provides new evidence that polyamine pathways are regulated by Mg 2ϩ supply (Fig. 5).
The observations that spe1⌬ and spe2⌬ strains can survive and grow with very low amounts of cellular polyamines and that cellular Mg 2ϩ accumulation is up-regulated in polyamine-deficient cells support the notion that much of the polycationic polyamine requirement of yeast can be fulfilled by Mg 2ϩ . In SDC free of polyamines, the growth of spe1⌬ and spe2⌬ strains was arrested upon prolonged incubation in SDC despite increased Mg 2ϩ content (Figs. 1, 3, and 4). Obviously, Mg 2ϩ cannot totally substitute for polyamines for polyamine-specific functions, including the fine-tuning of translation (5) and the eIF5A hypusine modification. Thus, growth of spe1⌬ and spe2⌬ strains in SDC arrests when cellular polyamines and hypusinated eIF5A fall below the minimum threshold levels. Accumulation of reactive oxygen species and various morphological changes, including large cell size, large vacuoles, and apoptotic cell death, were reported in a S. cerevisiae spe2⌬ strain under conditions of an extreme depletion of polyamines (32,33). However, we found no/little changes in cell sizes and viability of the spe1⌬ and spe2⌬ mutants after 24-h culture in SDC, when cells were harvested for the analysis of Mg 2ϩ and polyamines. This suggests that elevation of Mg 2ϩ in our mutants was not associated with the loss of viability and other cellular changes that accompany extreme polyamine depletion. Instead, the elevated Mg 2ϩ sustains the growth and viability of the spe1⌬ and spe2⌬ mutants in medium containing very low spermidine (10 Ϫ8 M).
The fact that mammalian cells, unlike yeast, require a high level of polyamines for growth suggests either that there are mammalian-specific processes intrinsically dependent on a high polyamine concentration or that mammalian cells lack a mechanism for Mg 2ϩ accumulation in response to polyamine depletion. Mg 2ϩ homeostasis is poorly understood in mammalian cells, although a number of putative Mg 2ϩ transporters have been described (34). In addition to the human mitochondrial CorA protein Mrs2, these include novel families unrelated to CorA, such as SLC41, TRPM6/7, MagT, NIPA, MMgT, and HIP14 families. From a screening of a Jurkat cell library, MagT1 and its homolog TUSC3 were identified as cell-surface Mg 2ϩ transporters that complement the arl1⌬ mutation (35). As their structures are unrelated to the CorA superfamily, they are not likely to respond to polyamine depletion in a similar manner as does Alr1. Although intricate mechanisms of polyamine homeostasis have been established in mammalian cells, there is poor understanding as to how polyamine pathways interact with other cellular pathways such as metal metabolism and homeostasis. Parenthetically, cellular iron levels were reported to regulate many polyamine pathway proteins in several mammalian cells (36). Future investigations will be directed toward further exploration of the interplay between Mg 2ϩ and polyamine homeostasis in S. cerevisiae and in mammalian cells.

Yeast strains and cell growth assay
The S. cerevisiae strain WT Y534 (BY4741) and the two knockout strains Y535 (spe1⌬) and Y536 (spe2⌬) generated by a S. cerevisiae gene deletion project (25) were kindly provided by Dr. Herbert Tabor (NIDDK, National Institutes of Health). Other strains with a deletion of Mg 2ϩ transport-related genes (alr1⌬, alr2⌬, and mnr2⌬) and new strains generated in this study are listed in Table 1. Cells were routinely cultured in YPD or SDC with or without Mg 2ϩ , G418, or spermidine as indicated in specific experiments. YPD is rich in polyamines (putrescine, spermidine, and spermine, ϳ0.3, 2.2, and 0.4 mM, respectively) and low in Mg 2ϩ (26). Standard SDC contains 4 mM Mg 2ϩ but no polyamines. For the selection of Ura ϩ transformants, uracil-dropout medium (SDϪUra) was used. Because the polyamine requirement for optimal growth of spe1⌬ and spe2⌬ strains is extremely low, ultrapure water was used, and special care was taken to avoid any environmental contamination of polyamines in SDC. To compare the growth of different strains in SDC, each strain was patched on YPD plates, and freshly grown patches of cells were inoculated in SDC at 0.001 or 0.0003 OD 600 and cultured at 30°C with shaking. Cell density was measured using a spectrophotometer at 600 nm every 2 h during the day. To follow growth over a 72-h period, cells were repeatedly diluted in the same fresh medium to OD 600 0.001 or 0.003 when the density reached ϳ1 OD 600 .

Generation of spe1⌬ and spe2⌬ strains
The SPE1 and SPE2 knockout cassettes were prepared using genomic DNA of Y535 (spe1⌬) and Y536 (spe2⌬), respectively, as templates. The primers used for amplifying the spe1⌬::KanMX4 cassette were CATTTCTCTCCTTGTCTG-TGCT and TGGCCTGTGTTGAAGTATGGT. The primer sets for amplifying the spe2⌬::KanMX4 cassette were CCAGA-GATATGTAGCCTTCCATC and GGGCATAAACCTTTG-AGCATCATC. PCR was performed using the Easy-A 2ϫ Master Mix (Agilent) with the following program: 94°C for 5 min, denaturation at 94°C for 0.5 min, annealing at 58°C for 1 min, an extension reaction at 72°C for 2.5 min for 30 cycles, and a final extension reaction at 72°C for 5 min. Yeast transformation with the purified PCR products (ϳ2.51 and ϳ2.429 kb, respectively for spe1⌬::KanMX4 and spe2⌬::KanMX4) was carried out with the Yeast Transformation kit (Sigma-Aldrich), according to the manufacturer's instructions, using 1 g of linear DNA. The transformed clones were isolated on YPD plates containing 500 g/ml G418.

Isolation of genomic DNA and PCR confirmation of genotypes of parental and spe1⌬ and spe2⌬ mutant strains
Cells were cultured in YPD to 1 OD at 600 nm and harvested. Genomic DNA was isolated from the cell pellets following a published protocol (37). Briefly, the cell pellets (10 OD 600 units) were resuspended in 0.2 ml of 200 mM LiOAc, 1% SDS solution and heated at 70°C for 5 min. 600 l of 100% EtOH was added, and DNA and cell debris were spun down at 15,000 ϫ g for 3 min. The pellets were washed with 70% EtOH. After removal of all EtOH, DNA was extracted by resuspension in 0.1 ml of Tris-EDTA buffer, and the DNA concentration was measured using a NanoDrop ND-100 spectrophotometer. PCR was performed as described above using 0.1 g of genomic DNA, a knockout primer set or an open reading primer set (Table 3), and Jump-

Interplay between polyamines and Mg in yeast cell growth
Start REDTaq ReadyMix Reaction Mix (Sigma-Aldrich) according to the manufacturer's instructions.

Construction of pFL38/SPE1 and pFL38/SPE2 plasmids
To check the reversal of the phenotypes of the spe1⌬ and spe2⌬ strains, pFL38 plasmids encoding SPE1 and SPE2 were reintroduced into the corresponding null strains. The recombinant plasmids were constructed by GenScript USA Inc. by synthesis of SPE1 and SPE2 genes (each ORF with 200 bp 5Ј-UTR and 3Ј-UTR) and subcloning into pFL38, and the transformants were selected on SDϪUra plates.

Analysis of cellular Mg 2؉ content by inductively coupled plasma optical emission spectrometry (ICP-OES)
WT and most mutant cells were inoculated at a density of 0.001 in 10 -20 ml of the indicated medium, and exponentially growing cells (1 or less than 1 OD 600 ) were harvested at ϳ24 h after inoculation. For slow-growing mutant strains, a higher inoculum density and a larger culture volume were used to obtain ϳ10 -20 OD 600 units of cells in 24 h. Cells were harvested by centrifugation, transferred to preweighed Eppendorf tubes, and washed once with 1 mM EDTA and twice with ultrapure water, and the washed cell pellets were frozen on dry ice. The total Mg 2ϩ content was measured by ICP-OES as described previously (38).

Determination of yeast polyamine content
S. cerevisiae cells were cultured, harvested, and frozen, in parallel or similarly to those for Mg 2ϩ analysis, as described above. Polyamines were extracted from the cell pellets by resuspension in 1.2 N perchloric acid, repeated vortexing with glass beads, and incubation on ice. The extracted amines and the internal standard (1,7diaminoheptane) were derivatized using dansyl chloride, and the dansylated polyamine derivatives were analyzed in duplicate by reverse-phase HPLC as described previously (39). The polyamine content was normalized against cell proteins determined using the Pierce BCA protein assay dye reagent after dissolving the perchloric acid precipitates in 0.1 N NaOH.

Statistical analysis
All values are presented as means of four or more biological replicates (n) as indicated in the legends. The contents of Mg 2ϩ and polyamines are expressed as mean Ϯ S.D., and the optical densities of cells are expressed as mean Ϯ S.E. Differences between samples were compared using analysis of variance (ANOVA) and were considered statistically significant at p Ͻ 0.05. All statistical analyses were conducted using GraphPad Prism 8 (GraphPad Software, San Diego, CA).