Mössbauer and LC-ICP-MS investigation of iron trafficking between vacuoles and mitochondria in vma2ΔSaccharomyces cerevisiae

Vacuoles are acidic organelles that store FeIII polyphosphate, participate in iron homeostasis, and have been proposed to deliver iron to mitochondria for iron–sulfur cluster (ISC) and heme biosynthesis. Vma2Δ cells have dysfunctional V-ATPases, rendering their vacuoles nonacidic. These cells have mitochondria that are iron-dysregulated, suggesting disruption of a putative vacuole-to-mitochondria iron trafficking pathway. To investigate this potential pathway, we examined the iron content of a vma2Δ mutant derived from W303 cells using Mössbauer and EPR spectroscopies and liquid chromatography interfaced with inductively-coupled-plasma mass spectrometry. Relative to WT cells, vma2Δ cells contained WT concentrations of iron but nonheme FeII dominated the iron content of fermenting and respiring vma2Δ cells, indicating that the vacuolar FeIII ions present in WT cells had been reduced. However, vma2Δ cells synthesized WT levels of ISCs/hemes and had normal aconitase activity. The iron content of vma2Δ mitochondria was similar to WT, all suggesting that iron delivery to mitochondria was not disrupted. Chromatograms of cytosolic flow–through solutions exhibited iron species with apparent masses of 600 and 800 Da for WT and vma2∆, respectively. Mutant cells contained high copper concentrations and high concentrations of a species assigned to metallothionein, indicating copper dysregulation. vma2Δ cells from previously studied strain BY4741 exhibited iron-associated properties more consistent with prior studies, suggesting subtle strain differences. Vacuoles with functional V-ATPases appear unnecessary in W303 cells for iron to enter mitochondria and be used in ISC/heme biosynthesis; thus, there appears to be no direct or dedicated vacuole-to-mitochondria iron trafficking pathway. The vma2Δ phenotype may arise from alterations in trafficking of iron directly from cytosol to mitochondria.

Vacuoles are acidic organelles that store Fe III polyphosphate, participate in iron homeostasis, and have been proposed to deliver iron to mitochondria for iron-sulfur cluster (ISC) and heme biosynthesis. Vma2Δ cells have dysfunctional V-ATPases, rendering their vacuoles nonacidic. These cells have mitochondria that are iron-dysregulated, suggesting disruption of a putative vacuole-to-mitochondria iron trafficking pathway. To investigate this potential pathway, we examined the iron content of a vma2Δ mutant derived from W303 cells using Mössbauer and EPR spectroscopies and liquid chromatography interfaced with inductively-coupled-plasma mass spectrometry. Relative to WT cells, vma2Δ cells contained WT concentrations of iron but nonheme Fe II dominated the iron content of fermenting and respiring vma2Δ cells, indicating that the vacuolar Fe III ions present in WT cells had been reduced. However, vma2Δ cells synthesized WT levels of ISCs/hemes and had normal aconitase activity. The iron content of vma2Δ mitochondria was similar to WT, all suggesting that iron delivery to mitochondria was not disrupted. Chromatograms of cytosolic flow-through solutions exhibited iron species with apparent masses of 600 and 800 Da for WT and vma2Δ, respectively. Mutant cells contained high copper concentrations and high concentrations of a species assigned to metallothionein, indicating copper dysregulation. vma2Δ cells from previously studied strain BY4741 exhibited iron-associated properties more consistent with prior studies, suggesting subtle strain differences. Vacuoles with functional V-ATPases appear unnecessary in W303 cells for iron to enter mitochondria and be used in ISC/heme biosynthesis; thus, there appears to be no direct or dedicated vacuole-to-mitochondria iron trafficking pathway. The vma2Δ phenotype may arise from alterations in trafficking of iron directly from cytosol to mitochondria.
Vacuoles are acidic organelles in fungi and plants that are evolutionarily connected to endosomes and lysosomes in humans. Their acidity is essential for sequestering and storing iron and for participating in cellular iron homeostasis (1,2). The vacuolar ATPase (V-ATPase) is a membrane-bound multisubunit complex that pumps protons from the cytosol into the vacuole. This results in an acidic vacuolar lumen (pH 5.6-6.1) and a slightly basic cytosol (pH 7.4) (3)(4)(5)(6)(7). vmamutant cells lack functional V-ATPases and grow well only in media buffered between pH 3 and 6, a narrower and more acidic range than is normal for WT cells (7)(8)(9). The pH of the growth medium influences the pH of vacuoles (8). When vmamutant cells were grown at pH 5.5, their vacuolar pH was 5.9 (4,6,7). When the same cells were transferred to pH 7.5 medium, vacuolar pH shifted to 7.05. In contrast, vacuoles of WT cells treated equivalently remained at pH 5.9.
V-ATPases consist of two subcomplexes including a peripheral membrane subcomplex V 1 and an integral membrane complex V o . In yeast, V-ATPases are found in vacuolar membranes or vacuole-mitochondria contact sites. Inactivation or deletion of any component subunit leads to a similar mutant phenotype characterized by reduced replicative lifespan, decreased inorganic polyphosphates, sensitivity to oxidative stress, and iron dysregulation. VMA2 encodes subunit B of the V 1 subcomplex; the knockout is viable although it lacks V-ATPase activity.
Iron-replete WT vacuoles contain high (mM) concentrations of polyphosphate that coordinate Fe III ions (10)(11)(12). Polyphosphate chains are synthesized by the VTC complex on the vacuolar membrane (13,14), a process that depends on the proton gradient generated by V-ATPase. Without this gradient, vma mutants are deficient in polyphosphate ions and are less able to store iron (15,16). Thus, polyphosphate ions and protons both impact vacuolar iron metabolism.
Iron homeostasis in yeast involves about two dozen genes called the iron regulon (17). These genes are activated when mitochondrial iron-sulfur cluster (ISC) assembly is compromised (18). For example, this condition is evident in strains that are deficient in yeast frataxin homolog 1 (Yfh1) (19), yeast adrenodoxin homolog 1 (Yah1) (20), and the inner membrane transporter Atm1 (21). These cells respond to iron regulon activation by importing excess iron and mobilizing vacuolar iron stores.
The iron regulon is activated in vma2Δ cells grown in ironreplete (μM range) media even though their mitochondria are not deficient in proteins involved in ISC assembly (22,23). This implies that mitochondria cannot assemble sufficient ISCs when the vacuolar lumen is insufficiently acidic, and it highlights an intriguing connection between the two organelles. vma2Δ cells have reduced activities of mitochondrial ISC-containing proteins that are involved in respiration, including aconitase, succinate dehydrogenase, and the Rieske iron-sulfur protein (23). vma-mutant cells grow poorly on respiring media (3,22,24,25). However, supplementing growth media with mM concentrations of iron a) rescues this phenotype, b) deactivates the iron regulon (22,23), c) increases the activity of ISC-containing enzymes, d) increases the rate of O 2 consumption, and e) stimulates respiration (24)(25)(26).
Excess supplemented iron provides the additional iron needed for the biosynthesis of iron-rich mitochondria which are in turn needed for respiration. This requirement is uncompromising, even in WT cells, such that cells grown on iron-deficient minimal medium (MM) succumb to autophagy so that they can recycle sufficient cellular iron for use in mitochondriogenesis (27). During autophagy, Smf3 and Fet5/ Fth1 iron exporters on the vacuolar membrane pump iron into the cytosol to satisfy mitochondrial demand. Iron-deficient vma mutants have difficulty undergoing autophagy which hinders respiration (27).
When the iron regulon is activated in ISC-mutant cells, the rate of cellular iron import increases dramatically owing to expression of a high-affinity iron importer (composed of Fet3 and Ftr1) on the plasma membrane. Expression of the highaffinity vacuolar iron exporter (composed of Fet5 and Fth1) also increases. Fet3 and Fet5 are paralogous multicopper oxidases controlled by the iron regulon. As a result, the vacuoles of ISC-mutant cells are nearly devoid of iron, and much of the excess iron that flows into the cytosol ultimately accumulates in mitochondria as Fe III oxyhydroxide phosphate-associated nanoparticles (19)(20)(21).
vma-mutant cells have difficulty importing extracellular iron because copper-containing Fet3 has difficulty maturing (i.e., metallating with copper) in the Golgi (1,28). Whether these cells also have difficulty exporting vacuolar iron is unknown, but this seems likely, given the similarities between Fet3 and Fet5. In any event, the concentration of iron in vmamutant cells is not excessive relative to WT cells even though the iron regulon is activated. vma2Δ cells reportedly contain only 30% more iron than WT cells (23). In contrast, ISC mutants contain an order-of-magnitude more iron than WT cells.
The mechanism by which a decline in the acidity of vacuoles causes a decline of mitochondrial function, and how excessive iron rescues this decline, has been investigated by Kane (23) and more recently by Hughes (26). Both hypothesize that the flow of iron from vacuoles to mitochondria is disrupted in vma-mutant cells and that excess media iron re-establishes this pathway. Besides sequestering iron, vacuoles store amino acids, including cysteine (29). Hughes et al. (26) found that loss of vacuolar acidity increases nonvacuolar pools of cysteine and other amino acids. One intriguing possibility is that excess cysteine in the cytosol binds iron, thereby blocking import into mitochondria.
vma mutants generate high levels of reactive oxygen species (ROS) and are hypersensitive to oxidative stress (5,22,30,31). They are more easily intoxicated by high levels of transition metals than WT cells. This likely occurs because vma cells have difficulty storing metals in their vacuoles, which otherwise imparts metal resistance to the cell (17,22,24,32,33). Surprisingly, mitochondrial respiration is not the major source of ROS in vma2Δ cells (22). Rather, excessive ROS is likely associated with iron dyshomeostasis and increased Fenton chemistry (5,23).
Also surprising is that the relationship between vacuolar acidity, cellular iron, mitochondrial function, and ROS damage is related to human aging and longevity. Hughes and Gottschling found that a decline of vacuolar acidity is associated with age-related deterioration of mitochondria (34). Overexpressing V-ATPase subunits suppresses mitochondrial dysfunction in aging cells by hyperacidifying vacuoles. Young yeast cells treated with a V-ATPase inhibitor and mutant cells lacking V-ATPase subunits both have impaired mitochondrial morphology similar to the organelle in aged cells (34).
The objective of the current study was to better understand the connection between vacuolar acidity, mitochondrial function, and iron homeostasis by using powerful biophysical and bioanalytical probes, namely Mössbauer (MB) and EPR spectroscopies and liquid chromatography interfaced with inductively-coupled-plasma mass spectrometry (LC-ICP-MS). Mitochondria, vacuoles, and the cytosol of vma2Δ cells were isolated and investigated. Our results confirm and extend some but not all relevant previous results, with differences probably due to the background strains and media/conditions used. New insights into the connection between vacuoles and mitochondria are presented, highlighting the potential importance of low-molecular-mass (LMM) iron, copper, and sulfur species in the cytosol of these yeast cells.

Results
A haploid VMA2 deletion strain was generated from a W303 parent to allow for direct comparison to previous biophysical studies that used the same genetic background (10,11,20,21,(35)(36)(37)(38)(39). We refer to W303 and the mutant strain as WT and vma2Δ W , respectively. We initially investigated whether the iron content of these two strains differed and whether the pH of the medium influenced that content. WT and vma2Δ W cells were grown in MM containing 2% glucose as the carbon source (fermentation) and supplemented with iron (typically 40 μM 57 Fe III citrate). Medium pH was buffered using citric acid and sodium citrate. Under these conditions, vma2Δ W cells grew well at pH values 3, 4, and 5 although growth was inhibited when the medium pH was 6 or 7; in contrast, WT cells grew well in media buffered from pH 3 to 7.
Whole-cell iron concentrations (Table 1) revealed no obvious pH dependence; however, some strain-dependent trends were evident. vma2Δ W cells contained 30% less iron, 60% less phosphorus, 6× less manganese, 4× less zinc, and 2.5× more copper than WT cells. The increased copper when considered with previous reports that Fet3 in vma2Δ cells is not metallated in the Golgi (1,28) suggested that copper was dysregulated in these mutant cells and that vma2Δ cells compensated by importing excessive copper.
MB spectra of both strains were collected to investigate the type of iron present. WT cells exhibited qualitatively similar spectra regardless of pH ( Fig. 1), but some subtle pHdependent trends were evident. At low pH, over 60% of spectral intensity was a magnetic feature originating from nonheme high-spin (NHHS) Fe III . The solid purple line in Figure 1 is a simulation using parameters in Table S1. The Fe III ions are coordinated to anionic polyphosphate chains in vacuoles (10,11). With increasing pH, the percentage of cellular iron in this form declined to 30% at pH 7. We conclude that WT vacuoles store less iron as medium pH increases.
Also evident in WT MB spectra was the central doublet (CD) with isomer shift and quadrupole splitting parameters typical of [Fe 4 S 4 ] 2+ clusters and low-spin Fe II hemes (simulated by the green line in Fig. 1). MB spectra of isolated mitochondria are dominated by the CD (36)(37)(38), which originates mainly from iron-rich respiratory complexes and respiration-related iron-containing proteins, e.g., aconitase. A substantial portion of the CD intensity in whole cells is due to cytosolic [Fe 4 S 4 ] 2+ -containing proteins. Percentagewise, the spectral intensity of the CD in whole-cell WT spectra increased from 15% at low pH values to 30% at pH 7. This increase likely arose from the percentagewise decrease of vacuolar iron, such that the absolute concentration of the [Fe 4 S 4 ] 2+ clusters and LS Fe II hemes that give rise to this spectral feature were fairly constant with changes in pH.
Another 20% of the intensity of WT cell spectra was a quadrupole doublet (simulated by the blue line in Fig. 1) with parameters typical of NHHS Fe II complexes coordinated by 5 to 6 O/N donors (δ = 1.27 mm/s; ΔE Q = 2.9 mm/s). Such species are probably located in the cytosol and mitochondria, among other regions of the cell. Another 10% of spectral intensity consisted of poorly resolved material near the center of the spectrum (at 0 velocity). In Figure 1, this material was simulated (orange line) assuming parameters of Fe III phosphate-associated nanoparticles. Apart from the modest decline of vacuolar iron with increased pH, the iron in WT cells generally remained invariant over four orders-of-magnitude of acidity in the growth media. This implies that intracellular pH of WT cells is well regulated against extracellular perturbations, probably due to the activities of proton pumps on the plasma and vacuolar membranes.  The WT EPR spectrum of whole WT cells cultured at pH 5 with 40 μM supplemented Fe III citrate ( Fig. 2A) exhibited signals similar to those observed previously (11,38). Vacuolar Fe III exhibited a prominent g = 4.3 signal. Features at g = 6.5 and 5.5 are typical of high-spin Fe III hemes. A hyperfine-split signal arising from Mn II ions was observed at g = 2.00, as was an isotropic radical signal. No signals attributable to Cu II ions were observed.
MB spectra of vma2Δ W cells grown in MM at pH 3, 4, and 5 were dominated by an Fe II quadrupole doublet (Fig. 3) that accounted for 60% of spectral intensity. The high-energy line of the CD was also evident as a partially resolved shoulder on the low-energy line of the dominating Fe II doublet at + 1 mm/s. Like WT cells, the iron content of vma2Δ W cells was generally invariant in media in which acidity varied by three orders of magnitude. We had predicted that the magnetic feature due to vacuolar Fe III in MB spectra of WT cells would also be present in spectra of vma2Δ W cells grown at pH 3 or 4, given that the acidity of vacuoles lacking functional V-ATPase is affected by the pH of the growth media. This prediction was not realized, possibly for reasons discussed below.
The CD represented 20% of the spectral intensity for vma2Δ W cells at each pH within this range. Assuming an average cellular iron concentration of 180 μM (Table 1) suggests that 40 μM iron in these cells arose from [Fe 4 S 4 ] 2+ clusters and LS Fe II hemes. For WT cells, with an average iron concentration of 250 μM and 20% CD intensity, the corresponding concentration would be 50 μM. The two concentrations are the same within our uncertainties. We had expected that the concentration of such centers in vma2Δ W cells would have been much lower and that most of the iron in these cells would have been Fe III phosphate-associated nanoparticles as is observed in other ISC-mutant cells in which the iron regulon is activated. However, MB spectra of vma2Δ W cells did not exhibit intense features typical of nanoparticles nor did they provide clear evidence that the iron regulon was activated.
EPR spectra of vma2Δ W cells at pH 5 ( Fig. 2, B-C) supported the MB spectra and provided new insights. The intensity of the g = 4.3 signal due to vacuolar Fe III was strongly reduced relative to in WT spectra, consistent with the absence of the magnetic feature due to vacuolar Fe III in corresponding MB spectra. Minor features attributed to high-spin hemes at g = 6.5 and 5.5 were comparable in intensity to those of WT, suggesting normal levels of cellular hemes in the mutant cells. The hyperfine-split signal due to Mn II ions was dramatically reduced in vma2Δ W EPR spectra, consistent with the 6-fold decline of Mn concentration relative to in WT cells (Table 1). No signals attributable to S = ½ Cu II ions were observed, despite the 2.5-fold increase in Cu concentration in  the mutant cells. This suggests that the additional Cu ions in vma2Δ W cells were in the reduced diamagnetic Cu I state.
We assessed whether the dominating Fe II species in vma2Δ W cells arose from an accumulation of LMM iron complexes in the cytosol. The cytosol was isolated from WT and vma2Δ W cells, and the concentration of iron in these cytosol isolates was determined ( Table 1). The iron concentration of cytosol from both strains was high (200, 270 μM), similar to values reported previously (38). Cytosol solutions were passed through a 10-kDa cutoff membrane, and the metal contents of flow-through solutions (FTSs) were also determined. We previously found that LMM iron species constituted ca. 70% of cytosolic iron at 40 μM iron supplementation (38). Here, we observed that 30% of cytosolic iron from cells grown with 1 μM Fe was present in an LMM form ( Table 1). The concentration of LMM iron in the cytosol was higher than expected by an order-of-magnitude. We suspected that ironrich vacuoles may have burst during cytosol isolation and metals from the ruptured organelles leached into the cytosol. However, vacuoles contain little iron at 1 μM iron supplementation (37,38), yet the iron concentrations of cytosol isolated from 1 μM and 40 μM iron-supplemented WT cells were similar. Thus, vacuole-bursting seems unable to explain the high concentrations.
FTSs were subjected to LC-ICP-MS chromatography. The resulting traces (Fig. 4) revealed unresolved LMM irondetected peaks for each strain. The dominant LMM iron species in WT cytosol FTS migrated with an apparent mass of 600 Da (all masses quoted in the text are apparent), whereas the dominant LMM iron species in vma2Δ W cytosol FTS migrated with a mass of 800 Da. Both species were present in the FTS of both strains, and their collective intensities were similar. We conclude that the LMM iron in the cytosolic FTS of vma2Δ W and WT cells are similar though not identical in terms of species and concentration. We caution readers that about half of the LMM iron in the cytosolic fraction adsorbed onto the column.
We considered that much of the NHHS Fe II in vma2Δ W cells was located in the mitochondria, and so this organelle was isolated from both WT and vma2Δ W strains (grown on fermenting MM at pH 5). MB spectra of isolated mitochondria from each strain (Fig. 5) were dominated by the CD (ca. 80% intensity). A minor NHHS Fe II doublet (15%-20%) was also evident. Aconitase activities in these mitochondrial lysates were similar (WT = 2.6 ± 1.2 units/mg protein, vma2Δ W = 2.3 ± 1.2 units/mg protein; n = 2). Little if any Fe III oxyhydroxide nanoparticles were present in vma2Δ W mitochondria, unlike mitochondria from strains in which ISC assembly is defective. In summary, the iron contents of mitochondria isolated from WT and vma2Δ W cells, as evaluated by MB spectroscopy and by the activity of an ISC-containing enzyme, were similar. The low-intensity NHHS Fe II doublet observed in WT and vma2Δ W mitochondria indicate that mitochondrial Fe II alone cannot account for the dominating Fe II intensities observed in vma2Δ W whole-cell MB spectra.
We wondered whether the metabolic state of vma2Δ W cells (fermenting versus respiring) affected the iron content of the cells. To address this, WT and vma2Δ W cells were grown on a respiring carbon source (glycerol/ethanol) in MM buffered at pH 5 and supplemented with 40 μM 57 Fe III citrate. In this case, vma2Δ W cells required prior growth in glucose-based media before switching to respiring MM. About 65% of the intensity of the resulting MB spectrum of WT cells (Fig. 6B) arose from vacuolar high-spin Fe III (10,11); another ca. 25% arose from the CD. The higher intensity observed for the CD confirms that respiration-related ISC proteins are upregulated during respiration (36). In contrast, the MB spectrum of respiring vma2Δ W cells (Fig. 6D) lacked the magnetic feature arising from vacuolar Fe III but exhibited an Fe II doublet representing 40% of spectral intensity. Over half of the spectral intensity  Mössbauer and LC-ICP-MS of vma2Δ yeast cells arose from the CD combined with unresolved iron in the middle of the spectrum which fit to a doublet with the parameters of nanoparticles. The cellular concentrations of [Fe 4 S 4 ]-and LS Fe II heme-containing species in respiring WT and vma2Δ W cells were again similar, as observed for fermenting cells. In summary, the iron content of respiring vma2Δ W cells was similar to that of WT cells, though with a dominance of Fe II instead of vacuolar high-spin Fe III . The simplest interpretation is that vacuolar Fe III in WT cells are converted to NHHS mononuclear Fe II and Fe III nanoparticles in fermenting and respiring vma2Δ W cells.
We considered that some of the dominant NHHS Fe II species in iron-replete vma2Δ W cells were present in vacuoles, and so two batches of vacuoles from vma2Δ W cells were isolated and FTSs were prepared. LC traces of these solutions exhibited LMM iron peaks that were either of similar intensity as peaks from WT vacuolesor less intense (data not shown). However, these studies used a column that was not treated such that adsorption of iron was considerable. We considered that vma2Δ W vacuoles contained additional iron that adsorbed onto the column, but the iron concentration of those vacuoles was not unusually high.
Cells grown in media containing low concentrations of iron do not store much of the metal in their vacuoles. Thus, the contribution of vacuolar iron to MB spectra can be largely eliminated by growing cells under these conditions. The MB spectrum of respiring WT cells grown in MM supplemented with 1 μM 57 Fe citrate (Fig. 6A) was noisy because the cells contained only 180 μM Fe. However, the S/N was sufficient to conclude that the vacuolar Fe III feature was essentially absent (<15% of spectral intensity). The corresponding spectrum of vma2Δ W cells, containing 220 μM Fe (Fig. 6C), also lacked the vacuolar Fe III feature and also exhibited an intense CD (even stronger than for WT cells). This indicates that vma2Δ W mitochondria were fully able to generate [Fe 4 S 4 ] clusters (and LS Fe II hemes). We conclude that vma2Δ W mitochondria were able to import cytosolic iron and were able to use it for ISC assembly even though vacuoles were largely devoid of iron. This behavior was unexpected, given previous results which indicated a strong need for functional acidic vacuoles to deliver iron to mitochondria during mitochondriogenesis associated with respiration, as well as with the reported rescue of growth of vma-mutant cells with high (mM) concentrations of nutrient iron (22,24,26). In summary, these experiments performed with only 1 μM added iron gave essentially the same result in vma2Δ cells, showing a normal ISC pool in cells where vacuoles did not store significant iron.
To further gauge the vacuolar iron content of WT and vma2Δ W cells, we cultured such cells in high (mM) concentrations of nutrient iron and monitored growth rates. As yeast lack a means of exporting iron, vacuolar iron storage typically provides resistance to excessive iron levels. WT cells grew rapidly until 20 to 30 mM Fe III citrate had been added to the growth medium, beyond which the growth rate declined (Fig. 7, triangles). vma2Δ W cells grew slower than WT in medium containing up to 10 mM iron and grew poorly in media containing higher iron concentrations (Fig. 7, circles). We suspect that under these high-iron conditions, vma2Δ W cells were unable to sequester toxic iron in their vacuoles in a nontoxic form.
The shapes of the plots of Figure 6 suggested that we could simulate growth rates using Equation 1: which assumes two contributions to the growth rate, one that is sensitive to excessive iron in the media (α cyt ) and another that is not (α vac ). The observed experimental growth rate α was the slope of the ln(OD600) versus time in the exponential growth region. This empirical equation may only apply to the range of media iron concentrations investigated. It implies that  excess cytosolic iron is toxic, whereas iron sequestered in the vacuole is not, as least in WT cells. The red line in Figure 7 simulates the growth rate of WT cells in which vacuoles sequester iron from the cytosol. The blue line simulates the growth of vma2Δ W cells in which vacuoles are presumed unable to sequester iron in a benign form such as Fe III polyphosphate. vma2Δ W cells grew more slowly than WT cells, even in media that was not supplemented with iron at high concentrations. Our MB studies indicate that these cells accumulated Fe II ions. We considered that the Fe II species in vma2Δ W cells might be susceptible to Fenton chemistry and that this caused slow growth. We investigated this by treating two of the 57 Feenriched samples used to generate spectra in Figure 2 with H 2 O 2 . In both experiments, the resulting MB spectra did not change noticeablythe NHHS Fe II doublet continued to dominate (data not shown).
We also probed the effect of varying phosphate supplementation on the WT and vma2Δ W strains, the latter of which contains little vacuolar polyphosphate (16). We grew both WT and vma2Δ W cells in media containing low (0.5 mM) and high (21 mM) supplemented phosphate. This treatment did not have an effect on the oxidation state of iron in vma2Δ W cells (Fig. S1). Both high-and low-phosphorus vma2Δ W spectra were dominated by the Fe II doublet. Curiously, the spectrum of high-phosphate cells exhibited a quadrupole doublet with parameters typical of Fe III oxyhydroxide phosphate-associated nanoparticles. The same doublet was present in the lowphosphate spectrum, albeit at lower intensity. This result suggests that in vma2Δ W cells, the formation of nanoparticles may have been limited by phosphate rather than iron. In contrast, the presence of high or low phosphate in the growth media had no noticeable effect on iron in WT cells.
Although iron was our major focus, ICP-MS provides information on other elements, and so we also monitored Cu, Mn, Zn, S, and P. Copper was especially interesting because vacuoles also function in copper homeostasis. FTS from vma2Δ W cytosol exhibited an intense copper-detected peak with an apparent mass of 4000 Da (Fig. 8). This peak was tentatively assigned to metallothionein Cup1 (38). The greater intensity of this peak in vma2Δ W cytosol FTS relative to WT suggested greater expression of Cup1 and/or more extensive copper binding to this protein in vma2Δ W cells. This makes sense because vma2Δ W cells contained 2.5× more copper than WT cells (Table 1) and Cup1 sequesters copper. The same trend was reflected in the copper concentration of isolated cytosol ( Table 1). The lack of a Cu II -based EPR signal in vma2Δ W cells is consistent with diamagnetic Cu I binding to metallothioneins. Also present were ca. 2 low-intensity Cu peaks in the range of 500 to 700 Da; their intensities were similar for both vma2Δ W and WT FTSs (Fig. 8, insets).
LMM sulfur-based LC peaks were also detected in cytosolic FTSs from both WT and vma2Δ W cells (Fig. 9). The dominant sulfur peak in the FTS from both cells (V e = 20 ml) originated from the sulfur-containing MES buffer used early on in isolating the cytosol. The right shoulder of this peak coeluted with numerous sulfur standards including but not limited to cysteine (Fig. 9F). In vma2Δ W cytosol FTS, the intensity of the shoulder was greater relative to that of WT FTS. The sulfur peak at an apparent mass of 100 Da. originated from the PMSF added to the buffer during isolation; the sulfur peak at an apparent mass of 50 Da originated from the DTT used earlier during isolation.
Hughes et al. (26) recently suggested that increased concentrations of cytosolic cysteine in vma2-mutant cells inhibited the vacuole-to-mitochondria iron trafficking pathway. To examine this, we grew vma2Δ W cells on fermenting MM (pH 5) supplemented with 1 mM cysteine. The MB spectra of such cells were unaffected by cysteine supplementation (Fig. 3B) except for a modest decline in the CD (22% in the spectrum of unsupplemented vma2Δ W cells down to 14% for the same cells supplemented with cysteine). Hughes et al. (26) suggested that cysteine supplementation caused "bioavailable" iron to decline and nonvacuolar cysteine to accumulate. The MB spectrum of the cysteine-supplemented vma2Δ W sample did not show an increase in Fe III nanoparticles, a form of iron considered to be bio unavailable.
Mn and Zn chromatograms of vma2Δ W and WT cytosol (Fig. 10, A-B) were like those described previously (38), including a single Zn peak at 700 Da and a single Mn peak at 200 Da. Phosphorus chromatograms (Fig. 10, C-F) exhibited a major peak at 400 Da and a shoulder at 600 Da. The intensities of Mn, Zn, and P peaks were significantly lower in vma2Δ W traces, consistent with the lower concentration of these  elements in mutant cells and isolated cytosol (Table 1). Notably, only the P peaks at 400 Da declined in intensity for vma2Δ W cytosol; the peak at 600 Da remained the same. The peak at 70 Da comigrated with an AMP standard.
We were puzzled that many of the results obtained using vma2Δ W cells differed from published results on what was ostensibly the same mutant. However, the strain used to generate the mutant in those reports was BY4741, whereas ours was W303. We obtained the mutant in the BY4741 strain (to be called vma2Δ B ) and collected MB spectra of those cells grown at different pHs in fermenting MM supplemented with 40 uM 57 Fe citrate. vma2Δ B cells grown at low pH exhibited spectra that were virtually indistinguishable from WT cells, including an intense magnetic feature due to vacuolar Fe III (Fig. 11E versus Fig. 11B). At higher pH, vma2Δ B cells exhibited spectra dominated by a broad quadrupole doublet due to Fe III nanoparticles (Fig. 11C-D). Also, the percent absorbance was higher than in WT spectra, similar to that observed for ISC-mutant cells with an activated iron regulon (19)(20)(21). The spectra of vma2Δ B cells grown at intermediate pH contained a stronger-than-WT NHHS Fe II doublet (Fig. 11D), with parameters similar to those used to simulate the Fe II doublet in vma2Δ W cells.
We attributed the differences between our results and those in the literature to differences between the W303 and BY4741 background strains. Notably, our experiments had not indicated an iron trafficking defect in the vma2Δ W cells. Prompted by a reviewer's comment, we sought to determine by Western blot of whole-cell extracts whether the iron regulon in vma2Δ W cells was activated by monitoring the relative abundance of the high-affinity iron importer Fet3. Cytoplasmic Pgk1 was a positive control. We did not observe a significant difference in Fet3 abundance between WT W W303 and vma2Δ W cells (Fig. 12, top row, B and C respectively), indicating that the iron regulon is not activated. In contrast, Fet3 was more abundant in vma2Δ B versus WT BY4741 cells (Fig. 12, top row, E versus D respectively), supporting previous studies indicating that the iron regulon is activated in vma2Δ B cells grown in yeast extract peptone adenine dextrose medium (YPAD) media (22,23).

Discussion
As described in the Introduction, there is significant evidence for an iron trafficking pathway from vacuoles to mitochondria in yeast. Vacuoles store iron that is needed for mitochondriogenesis especially when cells shift from fermentation to respiration. Nonacidic vacuoles with defects in V-ATPase have been reported to generate mitochondrial defects in ISC synthesis and iron homeostasis. Our initial objective was to probe this relationship and defects in vma2Δ W cells by

MB and EPR spectroscopies and LC-ICP-MS chromatography.
However, analyzing our results was more challenging than anticipated.
One result was clear: most of the iron in fermenting vma2Δ W cells was nonheme high-spin Fe II . In W303 cells, approximately the same percentage of iron was HS Fe III , previously assigned to vacuolar iron. The overall iron concentrations in WT and vma2Δ W cells and the spectral contributions of the CD (due to [Fe 4 S 4 ] 2+ clusters and LS Fe II hemes) were also similar. We conclude that vacuolar Fe III ions in WT cells are reduced to the Fe II state in vma2Δ W cells. We suspect but are not certain that the Fe II remains in the organelle. We have observed a similar Fe III → Fe II reduction of vacuolar iron in yeast cells grown on adenine-deficient media in which metabolism has been altered (39). Some vacuolar iron in iron-replete respiring vma2Δ W cells was reduced to Fe II (and some may have formed Fe III nanoparticles).
The reduction of Fe III to Fe II as the acidity of vacuoles declines can be partially or wholly explained by the conclusions of Raguzzi et al. (40) and Singh et al. (41) using insights from Schafer and Buettner (42). These researchers hypothesized that the glutathione redox couple fGSSG þ2H þ þ2e − %2GSHg controls the redox status of vacuoles in accordance with the Nernst equation: The standard half-cell reduction potential at pH 7 is E 0(pH7) GSH = −240 mV, whereas at pH 6, it is E 0(pH6) GSH = −180 mV (i.e., more oxidizing under acidic conditions). On this basis, they correctly predicted that the iron in WT vacuoles should be Fe III . For each pH increase of 1 (a 10fold drop in acidity), the effective reducing power of glutathione increases by 59 mV. Assuming pH 6 to 7 in vma2Δ vacuoles, the thermodynamic reduction potential for the Fe III /Fe II couple in vacuoles should be in the range of the values listed above.
Interestingly, the same analysis predicts that the Fe II should reoxidize to Fe III in vma2Δ W cells grown at lower pH, which was not observed; the dominant iron in these cells remained Fe II for pH values as low as 3. Perhaps the pH of vma2Δ vacuoles is less sensitive to media pH than we have assumed. Other factors may also be involved (43). The redox state of vacuolar iron is likely influenced by the potential ligands in the organelle. For instance, vacuolar iron is coordinated by polyphosphate anions, and the concentration of these chains is diminished in vma2Δ vacuoles (16).
Diab and Kane predicted that vma2Δ cells should contain high levels of Fe II ions (23). These researchers suggested that this form of iron is toxic to the cell because it participates in Fenton chemistry {Fe II + H 2 O 2 → Fe III + OH − + OH⋅}. However, we treated vma2Δ W cells with H 2 O 2 but found no evidence using MB spectroscopy that the Fe II was altered in peroxidetreated cells. Reducing equivalents in the cell may have quickly reduced Fe III back to Fe II such that Fe II remained dominant in the treated cells despite being used in Fenton chemistry. Upregulation of the cytosolic peroxidase Tsa2 in vma2Δ cells may have also diminished the effects of H 2 O 2 (22).
Many of our results differed from expectations based on previous studies of vma2Δ B cells. To examine whether there might be a strain dependence, we collected MB spectra of vma2Δ B cells grown at different media pH values. Resulting spectra were closer to our expectations, with WT-looking vacuolar Fe III in vma2Δ B cells grown at low pH and NHHS Fe II and nanoparticles for cells grown at high pH. BY4741 and W303 cells have different plasma membrane potentials and different sensitivity to alkali-metal salts (44). W303 cells are larger than BY4741 cells, their overall protein concentration is lower, and their vacuoles occupy a greater fractional cell volume (45). We found that vma2Δ W cells were less sensitive to media pH and exhibited milder phenotypes associated with nonacidic vacuoles such that mitochondria were largely unaffected under the conditions used. We speculate that this reduced sensitivity might reflect larger vacuoles (a lower surface-to-volume ratio and therefore a smaller interface with the cytosol) in W303 cells, but further studies are required to verify this.
We have compared our results mainly to those of Kane and Hughes, but other studies report results regarding vma mutants that are more similar to ours. Ohya et al. (46) found that O 2 consumption rates in vma-mutant cells (from parent strains other than W303 or BY4741) grown on glucose were "identical" to those of WT cells. They also reported no significant change in F-ATPase and succinate dehydrogenase activities, consistent with our results and with our suggestion of strain-dependent differences in vma-mutant phenotypes.
Hughes et al. (26) suggested that cytosolic accumulation of cysteine drives the low iron and oxidative stress response in V-ATPase-deficient cells. We observed a LMM sulfur species in isolated cytosol from WT and vma2Δ W cells and found that its intensity increased in Vma2Δ w FTS relative to that of traces of WT FTSs. This shoulder to the MES buffer peak approximately comigrated with cysteine and other LMM sulfur species, including GSH and methionine. The increased shoulder intensity in vma2Δ W isolated cytosol is consistent with the vma2Δ W cells grew slower than WT cells and were more easily intoxicated by high concentrations of iron in the medium. We simulated these differences by assuming that vma2Δ W cells are unable to sequester iron (or unable to sequester it in a benign form). However, the situation is undoubtedly more complicated. The toxicity of imported iron may also be associated with the oxidation state of that iron, with Fe II being more toxic than Fe III .
vma2Δ W cells and cytosol contained less Mn than in WT cells and cytosol; a similar decline has been observed in PPN1-UP cells (47) which have elevated levels of the polyphosphatase Ppn1. In both strains, cells contain low levels of polyphosphate relative to WT. Mn binds vacuolar polyphosphate (12), suggesting that these mutant cells may have responded to a decreased capability to store Mn in vacuoles by slowing Mn import.
The decline of phosphorus in vma2Δ W cells and cytosol may be related to changes in the activity/expression levels of Pho84 and Pma1, both of which are located on the plasma membrane (14). Pho84 is a symporter that imports protons and phosphate ions into the cytosol; Pma1 exports protons from the cytosol. The activity of Pma1 decreases in vma mutants (7), which along with the loss of V-ATPase activity increases the acidity of the cytosol. We speculate that this might decrease the activity of Pho84 and thus reduce the import of phosphate into vma cells. Pho84 may also be involved in manganese and zinc import (48), and a decline in activity could have contributed to the observed decline of Mn and Zn concentrations in vma2Δ W cells.
The increased concentration of Cu in our vma2Δ W cells and cytosol confirms earlier reports that vma-mutant cells contain high levels of copper and that copper is dysregulated in these cells (17,24). A similar accumulation may be driving the reported decline in Ctr1 expression in vma2Δ B cells (22). Most LMM Cu in both WT and vma2Δ W cytosol is probably bound to metallothionein Cup1. Higher copper concentrations in vma2Δ W cells likely afforded increased levels of Cu-bound Cup1. The inability of nonacidic vacuoles to sequester copper may have increased the concentration of cytosolic Cu, which in turn induced expression of CUP1. Another possibility is that Cu-bound Cup1 is mobilized to vacuoles for degradation and that this process is inhibited in nonacidic vacuoles, thereby causing a buildup of Cu-bound Cup1 in the cytosol. A similar peak was tentatively assigned to Cup1 in chromatograms of vacuolar lysates (12).
Does iron trafficking in Saccharomyces cerevisiae include a direct or dedicated vacuole-to-mitochondria pathway? A direct pathway would involve contact of the two organelles (kiss-andrun) and would be independent of cytosol. A dedicated pathway would involve the synchronized export of a particular iron species from the vacuole into the cytosol, and as a second step, the import of that species exclusively into mitochondria. If the iron exported from vacuoles entered a cytosolic pool of iron that had various trafficking destinations, including but not limited to mitochondria, we would not regard that pathway as being direct or dedicated.
We have not systematically distinguished these possibilities, but our results disfavor direct or dedicated pathways. The vacuoles are nearly empty in cells grown in iron-deficient or iron-limited media (e.g., 1 μM 57 Fe citrate), yet MB spectra of such vma2Δ W cells afforded a CD as intense as for equivalent WT cells. This suggests that vacuoles are unnecessary for iron to enter mitochondria and be used to generate ISCs. We tentatively conclude, pending further investigations, that cytosolic iron can enter mitochondria directly without first entering and exiting vacuoles and that there is no direct/ dedicated vacuole-to-mitochondria trafficking pathway as defined above.
So how does the nonacidity of vacuoles negatively impact ISC metabolism in mitochondria? We suggest, following Hughes (26), that the effect indirectly involves vacuoles and directly involves the cytosol. Cysteine or a derivative thereof may be blocked from entering nonacidic vacuoles, and its buildup in the cytosol may inhibit iron traffic into mitochondria. Likewise, a hyperacidic cytosol, caused by the absence of V-ATPase to pump cytosolic protons into vacuoles and/or a decline in Pma1 activity, may also hinder iron trafficking. In support of this, we observed a shift in the LMM iron species in vma2Δ W cytosol, from dominance of a species with an apparent mass of 600 Da to one with an apparent mass of 800 Da.
Finally, we have had difficulty throughout this study reconciling the milder phenotype of the vma2Δ W strain with the more dramatic phenotype of vma2Δ B . Both mutant cells grow slowly and with difficulty under respiring conditions, but vma2Δ W mitochondria are indistinguishable (by our methods) from WT mitochondria. In contrast, vma2Δ B mitochondria show difficulty generating ISCs, the iron regulon is clearly activated, and nanoparticles and ROS are likely generated. Those cells recover only under conditions in which the media is supplemented with excessive iron.
We suggest that vma2Δ W mitochondria suffer from the same originating problem as vma2Δ B but to a milder degree. We hypothesize that the fundamental problem is a slow rate of iron import from the cytosol to mitochondria, due to the hyperacidity of this cellular region and/or to the inhibitory effect of high cysteine concentrations. The effect of slow iron import into mitochondria on the growth of vma2Δ W cells would not be evident from our studies because we patiently waited for them to grow to sufficient absorbance at 600 nm prior to harvesting them for MB spectroscopy. By doing this, the ultimate iron content is expected to be like that of WT cells (apart from the reduction of vacuolar Fe III → Fe II ). For vma2Δ B cells, the rate of iron traffic from the cytosol to the mitochondria might be so slow that the respiratory shield weakens and a vicious cycle ensues, leading to nanoparticles, ROS, and an activated iron regulon, similar to that observed in Mrs3/4ΔΔ cells, which can only import iron slowly into mitochondria under iron-limited growth conditions (49,50). In vma2Δ W cells, this effect might be muted.
A PCR was performed on template plasmid pF6A-His3MX6 (51) carrying the S. pombe his5+ gene, and a 1403-bp DNA fragment was amplified as expected. The DNA product (about 5 μg) was used to transform W303 yeast using the lithium acetate procedure (52), and several colonies were selected for histidine prototrophy and carried forward. The correctness of the VMA2 deletion was verified as follows. Genomic DNA was extracted from the transformants, and a PCR, using forward primer I (ctcatgaccgatggtacg) and reverse primer N (ATGT-GATGTGAGAACTGTATC), was performed. The I primer resides in the 5' region of VMA2 gene outside the region of recombination. The N primer resides in the TEF terminator region of the pF6A-His3MX6 plasmid. The derived PCR product was approximately 688 bp in size and was present in the knockouts but not in the controls (Fig. S2), confirming that the VMA2 flanking region and His3MX6 module including the S. pombe his5+ gene were correctly juxtaposed in the genome of the putative knockouts. The knockout strains had the expected auxotrophies and prototrophies and grew more slowly than the parent on rich or defined media.
BY4741 cells for Western Blot were cultured in YPAD media (20 g/l glucose, 10 g/l yeast extract, 20 g/l peptone, 100 mg/l adenine). Cell cultures were incubated at 30 C and 180 rpm shaker speed. Absorbance was monitored at 600 nm using a Genesys6 spectrophotometer (Thermofisher). Fifty to 100 ml of precultures was transferred to 1-to 2-l media when absorbance at 600 nm was 0.7 ± 0.1. Cells were harvested from 1 to 2 l when culture was at exponential phase (absorbance at 600 nm = 0.8 ± 0.2). Cells were pelleted via centrifugation at 5000g for 5 min using a Sorvall LYNX6000 superspeed centrifuge (Thermofisher). Cells were washed twice with 1 mM EDTA and then twice with deionized water (5 ml of wash per gram wet cell pellet). After each wash, cells were pelleted as above and the supernatant was discarded. Finally, pellets were suspended in a minimal volume of double-distilled water and pelleted into MB cups by centrifugation at 6400 RPM for 10 min using a Beckman Coulter Optima L-90K ultracentrifuge (SW32Ti rotor); samples were frozen in liquid N 2 for later analysis. EPR samples were pelleted after washing by centrifugation at 5000g for 5 min into EPR tubes.
For H 2 O 2 studies, whole-cell MB samples were thawed and incubated in 1 ml 5% H 2 O 2 for 15 min or in 200 μl of 0.0034% H 2 O 2 (initial concentrations) overnight. Cell pellets were reacquired for MB spectroscopy via centrifugation at 5000g for 5 min.

Respiring cell growth
Cell cultures were initially grown in 50 mL of fermenting MM as above. Once absorbance at 600 nm reached 0.6, the cells were pelleted and resuspended in 1.00 L of respiring MM (same as fermenting MM pH 5 except that 3% (v/v) glycerol and 1.5% (v/v) ethanol replaced glucose). Cells were harvested as described above.

Varying phosphate cell growth
Cell cultures were grown in fermenting CSM media at 30 C and 180 RPM. CSM media consisted of 20 g/l glucose, 1.6 g/l CSM, 5.6 g/l yeast nitrogen base minus phosphates (MP Bio), 100 mg/l NaCl, 60 mg/l adenine, 100 mg/l leucine, 48 mg/l tryptophan, 10 μM copper sulfate, and 40 μM ferric citrate. The medium was supplemented with 0.5 and 21 mM monopotassium phosphate. All such studies were conducted at pH 5 using a 0.05 M citric acid/sodium citrate buffer. Cells were harvested as above.

Mössbauer and EPR spectroscopies
Mössbauer spectra were collected at 5 K and 0.05 T on a MS4 WRC spectrometer (SEE Co, Minneapolis, MN) calibrated at RT using an α-iron foil. The applied field was parallel to the gamma radiation. EPR spectra were collected in an X-band Elexsys EPR spectrometer (Bruker) using the following conditions: 9.3789 GHz microwave frequency, 25 dB attenuation, 0.6325 mW power, 100 kHz modulation, 10G modulation amplitude, 5000G sweep width, 120 s sweep time, 5 to 10 scans.

Mitochondria studies
Mitochondria were isolated anaerobically from cells harvested from 24 L of fermenting cell cultures (pH 5), as described (35). Mitochondrial protein concentration was determined using the BCA assay. Mitochondria were pelleted via centrifugation at 12,000g for 10 min. Aconitase was assayed as described (53,54) but with minor changes. Mitochondria were handled within an anaerobic glove box (MBraun Labmaster 130) containing c.a. 5 ppm O 2 . A pellet of mitochondria containing 80 mg of protein was diluted to 250 μl with lysis buffer (degassed 4.7 mM Triton X-100, 50 mM NaCl, and 50 mM Tris-HCl). Two hundred microliters of the resulting solution was mixed with 200 μl of 50 mM NaCl, 50 mM Tris-HCl, and 50 μl of double-distilled water in a 1-mm pathlength quartz cuvette. The sample was sealed with a rubber septum and removed from the box. Fifty microliters of 10 mM cis-aconitic acid (Sigma-Aldrich) was added through the septum, and the sample was mixed and immediately monitored at 240 nm for 360 s using a Hitachi U-3310 spectrophotometer. The molar absorptivity of cis-aconitic acid at 240 nm (ε 240 ) was 4.88 mM −1 cm −1 (54).

Cytosol and vacuole isolation
The cytosol was isolated from fermenting cells (pH 5) and filtered through a 10-kDa cutoff membrane as described (55). Filtered FTS was injected into a treated size-exclusion (SEC) Superdex Peptide 10/300 Gl column (GE Life Sciences) connected to an Agilent 1260 Bioinert quaternary pump (G5611A) and Agilent 7700x ICP-MS. The mobile phase was 20 mM ammonium acetate pH 6.5 flowing at 0.6 ml/min. Vacuoles were isolated as described (12).

Western blot
Western blot was completed using the whole-cell extract as described (12,55). An anti-Fet3 antibody from rabbit was used in a 1:1000 dilution. The antibody was generated as described (28).

Elemental analysis
Isolated cytosol samples (100 μl) or double-distilled waterwashed whole cells (ca. 100 mg) suspended in fresh doubledistilled water were digested with 500 μl of undiluted (70%) trace metal-grade HNO 3 (Fisher) at 70 C for ca. 16 h in plastic falcon tubes sealed using electrical tape. Samples were cooled to RT, and 250 μl of 35% hydrogen peroxide was added. Samples were resealed and incubated at 65 C for 90 min. Samples were diluted to 10 ml with double-distilled water, affording c.a 3.5% and 0.86% (w/w) final concentrations of trace metal-grade nitric acid and hydrogen peroxide, respectively. A cell pellet packing efficiency of 0.7 and whole-cell pellet density of 1.1029 g/ml (56) were used in calculations.

Data availability
All data except those described as "Data not shown" in the text are contained within the manuscript and SI.