J Biol Chem, Vol. 274, Issue 39, 27590-27596, September 24, 1999

Oxidative Stress and Iron Are Implicated in Fragmenting Vacuoles of Saccharomyces cerevisiae Lacking Cu,Zn-Superoxide Dismutase^*

Laura B. Corson^§, Janet Folmer^¶, Jeffrey J. Strain^¶, Valeria C. Culotta^¶, and Don W. Cleveland^§**

From the  Predoctoral Program in Human Genetics, Johns Hopkins University, Baltimore, Maryland 21205, the ^§ Ludwig Institute for Cancer Research, La Jolla, California 92093, the ^¶ Department of Environmental Health Sciences, Johns Hopkins School of Hygiene and Public Health, Baltimore, Maryland 21205, and the  Departments of Medicine and Neuroscience, University of California at San Diego, La Jolla, California 92093

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The absence of the antioxidant enzyme Cu,Zn-superoxide dismutase (SOD1) is shown here to cause vacuolar fragmentation in Saccharomyces cerevisiae. Wild-type yeast have 1-3 large vacuoles whereas the sod1 yeast have as many as 50 smaller vacuoles. Evidence that this fragmentation is oxygen-mediated includes the findings that aerobically (but not anaerobically) grown sod1 yeast exhibit aberrant vacuoles and genetic suppressors of other oxygen-dependent sod1 null phenotypes rescue the vacuole defect. Surprisingly, iron also is implicated in the fragmentation process as iron addition exacerbates the sod1 vacuole defect while iron starvation ameliorates it. Because the vacuole is reported to be a site of iron storage and iron reacts avidly with reactive oxygen species to generate toxic side products, we propose that vacuole damage in sod1 cells arises from an elevation of iron-mediated oxidation within the vacuole or from elevated pools of "free" iron that may bind nonproductively to vacuolar ligands. Furthermore, additional pleiotropic phenotypes of sod1 cells (including increased sensitivity to pH, nutrient deprivation, and metals) may be secondary to vacuolar compromise. Our findings support the hypothesis that oxidative stress alters cellular iron homeostasis which in turn increases oxidative damage. Thus, our findings may have medical relevance as both oxidative stress and alterations in iron homeostasis have been implicated in diverse human disease processes. Our findings suggest that strategies to decrease intracellular iron may significantly reduce oxidatively induced cellular damage.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Aerobic organisms are chronically exposed to potentially harmful reactive oxygen species generated as by-products of cellular metabolism. One antioxidant enzyme Cu,Zn-superoxide dismutase 1 (Sod1p) plays an important role in detoxifying superoxide radicals (O₂). Superoxide radicals can oxidize iron-sulfur cluster proteins liberating iron (1-4). Thus, Sod1p-deficient yeast (sod1 yeast) may have increased pools of reactive iron (5) as has been shown for SOD¹-deficient Escherichia coli (4). "Free" iron can react with hydrogen peroxide (H₂O₂), and possibly other reactive oxygen species, to generate toxic hydroxyl radicals (OH) by Fenton chemistry (6, 7). Hydroxyl radicals have the potential to damage proteins, nucleic acids, and membranes (4, 8-12).

In recent years, insight into oxidative defense, and SOD1 in particular, has come from studies using the unicellular eukaryote, Saccharomyces cerevisiae. Yeast lacking Sod1p grow slowly in air (13), are sensitive to superoxide generating agents (e.g. paraquat) (13, 14), and exhibit as yet poorly understood metabolic and biosynthetic defects (13-17). To further our understanding of oxidative damage to the cell, we examined the effects of superoxide dismutase deficiency on overall cell structure and organelle morphology and asked whether any such abnormalities may be tied into iron-mediated toxicity. We report here that oxidative stress to sod1 yeast results in fragmentation of the vacuole, an organelle analogous to the lysosome and thought to play a role in iron homeostasis (Refs. 18 and 19). Furthermore, we find lowering iron availability significantly reduces damage to the vacuole, consistent with disruption of iron handling as part of the pathway through which Sod1p deficiency causes vacuole fragmentation.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Yeast Strains, Media, and Growth Conditions.-- The S. cerevisiae strains used in these studies are listed in Table I. sod1 strains were generated by substituting the SOD1 coding region with SOD1 replacement plasmids pKS1 (sod1::LEU2) or pKS3 (sod1::TRP1) as described previously (20). Strains for morphological analysis were propagated in YPD medium (1% yeast extract, 2% peptone, 2% glucose) or in YPG medium (1% yeast extract, 2% peptone, 2% glycerol) in air or in CO₂-enriched anaerobic culture jars shaking at 100-120 rpm at 30 °C. For iron starvation studies, 50 µM of the iron chelator bathophenanthrolinedisulfonic acid was added to YPD, and for iron overload studies 2 mM FeCl₃ was added to YPD.

View this table: [in this window] [in a new window]   Table I Yeast strains used in the morphological analysis of sod1 yeast

Electron Microscopy-- For ultrastructural analysis of yeast by electron microscopy, strains were maintained in log phase for 24 h prior to harvesting. Cells were pelleted, washed in 0.9 M NaCl, washed twice in 0.1 M sodium cacodylate, pH 7.2, fixed in 3% gluteraldehyde, 0.1 M cacodylate at room temperature for 3 h, washed three times in 0.1 M cacodylate, post-fixed in 4% KMnO₄, 0.1 M cacodylate for 1 h at 4 °C, washed several times in water, and incubated in 2% uranyl acetate for 1 h at room temperature. Cells were dehydrated with 10-min serial ethanol washes (35, 50, 70, 95, and four times in 100%) and infiltrated with 50% spurrs, 50% ethanol on a rotator overnight at 4 °C. Cells were pelleted, resuspended in 100% spurrs for 2 h on rotator at 25 °C, repelleted, and resuspended in 100% spurrs an additional 2 h at room temperature. The cell pellet was transferred to capsules, overlaid with fresh spurrs, and polymerized at 65 °C for 24-48 h. Thin sections were cut, stained with lead citrate, and observed on an Hitachi HU12A transmission electron microscope.

Vital Staining of Vacuoles-- For analysis of yeast vacuoles by light microscopy, strains were maintained in log phase for at least 10 h prior to harvesting. Procedures for vital vacuole membrane staining using FM4-64 (Molecular Probes) were as described previously (21). Briefly, cells were harvested and resuspended in YPD medium + 30 µM FM4-64 for 15 min at 30 °C. Cells were pelleted, resuspended in fresh YPD medium, and incubated at 30 °C for an additional 2 h, placed on slides and viewed with standard fluorescence microscopy. Procedures for quinacrine staining were as described previously (22). Logarithmically growing cells were pelleted, resuspended in YPD, 50 mM Na₂HPO₄, pH 7.5, with 200 µM quinacrine and incubated at 30 °C for 5 min. Cells were pelleted, washed, and viewed on slides using standard fluorescence microscopy. All fluorescent images were photographed using a MetaMorph Imaging System (Universal Imaging Corp.) in conjunction with a CCD camera (Princeton Instruments).

Vacuolar Protein Trafficking-- Trafficking of carboxypeptidase Y was assayed as described previously (23). Briefly, isogenic strains 1783 (SOD1) and KS105 (sod1) and isogenic strains SEY6210 (VPS5) and BHY152 (vps5) were pulse labeled with Trans³⁵S-label (ICN) for 10 min. Carboxypeptidase Y was immunoprecipitated from cell lysates, electrophoresed on an SDS-polyacrylamide electrophoresis gel, and visualized by fluorography.

pH, Metal, and Starvation Sensitivity Assays-- To assay the ability of yeast to grow over a broad pH range, 1783 and KS105 yeast strains were diluted to an OD₆₀₀ of 0.05 in minimal medium (synthetic dextrose supplemented with lysine, methionine, cysteine, leucine, tryptophan, histidine, and uracil) (24) which was adjusted to the indicated pH (covering a range from 1.0 to 8.5 in 0.5 increments). Cultures were incubated overnight, and culture growth was measured by optical density.

To assay zinc sensitivity, SOD1 and sod1 yeast were grown overnight in minimal medium, diluted back to an OD₆₀₀ of 0.1 in minimal medium, inoculated with the appropriate concentration of zinc chloride, permitted to grow for 20 h in air or in an anaerobic chamber, and culture growth was measured by optical density.

Starvation sensitivity was assayed as described previously (25). 1783 and KS105 yeast were maintained in log phase for 8 h in synthetic complete medium (24). 10⁸ cells were pelleted by centrifugation, washed in water, resuspended in minimal sporulation media or water, and incubated at 30 °C for 7 days. Cells were counted with a hemocytometer. 500 cells from each culture were plated out onto YPD medium and permitted to grow for 3 days at 30 °C. The number of colonies formed per 500 cells was taken as the measurement of cell survival.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Aberrant Vacuole Morphology in Cu,Zn-Superoxide Dismutase-deficient (sod1) Yeast-- To determine effects of oxidative stress on cell morphology, aerated wild-type (1783) and sod1 (KS105) yeast were examined using electron microscopy (Fig. 1). Because mitochondria are thought to be a major source of superoxide radicals (e.g. Refs. 17 and 26), with leakage of electrons from the respiratory electron transport chain converting up to 5% of consumed oxygen into oxygen radicals (8), we anticipated mitochondrial abnormalities in sod1 yeast. However, the mitochondria of sod1 mutant yeast (arrowheads, Fig. 1, C and D) were indistinguishable from those in wild-type cells (arrowheads, Fig. 1, A and B). What was striking about the ultrastructure of the sod1 yeast was the apparent fragmentation of the yeast vacuole. While the wild-type strain (1783) typically exhibited 1-2 large vacuoles (Fig. 1, A and B), the sod1 mutant yeast (KS105) exhibited 5-15 smaller vacuoles in cross-section (Fig. 1,C and D). Staining such yeast with a vacuole-specific fluorescent dye quinacrine revealed a similar fragmented appearance (see Fig. 2). Imaging through multiple focal planes with a fluorescence microscope revealed that the sod1 yeast contained 25-50 small, individual vacuoles per cell compared with 1-3 vacuoles in wild-type yeast. Electron micrographs demonstrated that the electron densities of many smaller sod1 vacuoles were comparable to full-size wild-type vacuoles indicating that the vacuolar concentration of glycosylated molecules (which contribute significantly to the electron density (27)) was similar. However, some membrane-bound, non-uniform, electron transparent vesicles (arrows, Fig. 1D) were also evident in the vicinity of the vacuoles and may represent vacuole remnants or other aberrant organelles in the sod1 yeast.

View larger version (150K): [in this window] [in a new window]   Fig. 1.   Electron microscopic analysis reveals aberrant vacuole morphology in sod1 yeast. SOD1 (1783) and sod1 (KS105) yeast were grown in rich media containing glycerol (YPG) to induce mitochondrial proliferation and respiration. Labels are N, nucleus; and V, vacuole. Arrowheads point to mitochondria. Arrows point to aberrant membrane-bound vesicles frequently found in sod1 yeast.

View larger version (106K): [in this window] [in a new window]   Fig. 2.   Vacuole fragmentation of sod1 yeast is oxygen dependent. Quinacrine vacuole lumenal staining of SOD1 (1783) and sod1 (KS105) grown for 10 h in rich media (YPD) in air (A and C) or in an anaerobic chamber (B and D).

Vacuole Fragmentation in sod1-deficient Yeast Is Oxygen Dependent-- Because both the vacuole and Sod1p have been implicated in copper homeostasis, an initial hypothesis was that the aberrant vacuole morphology may reflect altered copper handling in sod1 yeast. Sod1p was identified in a genetic screen for copper-binding metallothioneins and was demonstrated to buffer intracellular copper under anaerobic as well as aerobic conditions, suggesting that the role in copper homeostasis is independent of the superoxide scavenging activity (20). Similarly, the vacuole plays a role in copper detoxification (28, 29). To distinguish whether the sod1 vacuolar fragmentation is due to the loss of copper buffering capability or to the absence of superoxide scavenging activity, the morphologies of aerobically and anaerobically grown sod1 and wild-type yeast were compared using electron microscopy (not shown) or quinacrine vital staining to identify the vacuolar lumens (Fig. 2). Vacuoles were fragmented when sod1 yeast were grown under aerobic conditions (Fig. 2C) but not under anaerobic conditions (Fig. 2D). The degree of vacuolar fragmentation in the sod1 strain was exacerbated with increased aeration (data not shown). Similarly, non-aerated sod1 vacuoles were found to fragment within 3 h of transfer to aerated conditions, and the fragmented vacuoles of aerated sod1 yeast recovered within a 4-h period following removal from aeration (data not shown). Since the cell cycle is 2-2.5 h in these yeast, it would appear that vacuolar repair is occurring over this 4-h period rather than synthesis of new vacuoles concomitant with cell division gradually replacing fragmented vacuoles. Altogether, these results indicate that the vacuole fragmentation is an oxygen-dependent, reversible defect presumably mediated by oxygen-free radicals and suggest that accumulation of oxidative damage over time is required to induce such fragmentation.

Vacuolar Fragmentation Can Be Rescued by the Addition of Superoxide Radical Scavenging Mechanisms-- To further elucidate the nature of the vacuole fragmentation, genetic suppressors of various sod1 phenotypes were employed. The first class of suppressors function in redox active metal homeostasis and apparently provide the cell with cytosolic superoxide scavenging activity. Mutations in PMR1, a Golgi P-type ATPase, lead to accumulations of manganese in the cytosol (30) which can scavenge free radicals (31-33). pmr1 mutations have been shown to rescue all known sod1 oxygen-dependent defects including amino acid biosynthetic deficits and paraquat sensitivity (30). As seen by fluorescent FM4-64 vacuole membrane staining in Fig. 3C and quantified in Fig. 4, the pmr1 mutation restores 8 wild-type vacuolar morphology to the sod1 mutant, suggesting that removal of oxygen-free radicals is sufficient to prevent vacuole fragmentation. Treatment of sod1 yeast with 1 mM Mn²⁺ was also sufficient to rescue fragmentation (data not shown) providing additional confirmation that cytosolic superoxide scavenging activity is sufficient to prevent vacuolar abnormalities.

View larger version (131K): [in this window] [in a new window]   Fig. 3.   Vacuole fragmentation of sod1 yeast can be rescued by suppressors that reduce intracellular oxygen radical levels. FM4-64 vacuolar membrane staining of aerated yeast strains (A) wild-type (1783), (B) sod1 (KS104), (C) sod1 pmr1 (KS109), and (D) KS104 transformed with a 2-µM TKL1 plasmid.

View larger version (33K): [in this window] [in a new window]   Fig. 4.   Quantitative effect of sod1, sod2, pmr1, and TKL1 on vacuole fragmentation. Strains (all isogenic to 1783) were maintained in logarithmic phase of growth in rich media (YPD) shaking at 120 rpm for 8 h, then stained with FM4-64, and the number of vacuoles per cell counted. Values represent 300 cells for each genotype scored in three independent experiments (~100 cells scored per genotype per experiment). Standard deviations between experiments were less than 10%.

A second class of suppressors alleviates only a subset of sod1 null deficiencies that apparently arise from a decreased level of NADPH in sod1 yeast (34). sod1 yeast exist in a more oxidized state than wild-type yeast presumably due to utilization of cellular reserves of reductants (such as GSH and NADPH) for the spontaneous reduction of reactive oxygen species (5). Overexpression of transketolase (TKL1) stimulates generation of NADPH by the pentose phosphate pathways and thus rescues the methionine auxotrophy and slow growth defect of the sod1 yeast (34). Despite the ability to rescue these aspects of the sod1 phenotype, overexpression of TKL1 does not rescue the vacuole fragmentation (Fig. 3D). Quantification (Fig. 4) confirms that, for the most part, vacuoles remain fragmented with increased Tkl1p, suggesting that the fragmentation is not related to lowered NADPH levels.

Vacuole Fragmentation Is Specific to the Cytosolic Cu,Zn-Superoxide Dismutase Deficiency-- In addition to Cu,Zn SOD1 which is predominantly cytosolic, eukaryotes also have a nuclear-encoded manganese SOD2 localized within the mitochondrial matrix. To determine whether vacuole fragmentation is a general defect of any superoxide dismutase deficiency, vacuole morphologies of aerated wild-type (1783), sod1 (KS105), sod2 (JS002), and sod1 sod2 (JS001) strains were compared using FM4-64 (Fig. 5), as well as by electron microscopy (data not shown). Yeast deficient for the mitochondrial manganese SOD2 (JS002) had vacuole morphologies indistinguishable from those of wild-type yeast (1783), while sod1 (KS105) and the double mutant sod1sod2 (JS001) yeast exhibited vacuolar fragmentation (Fig. 5; quantified in Fig. 4). This provides evidence that vacuole aberrations are specific to Sod1p deficiency and suggests that the free radicals responsible for vacuole fragmentation are cytosolic or vacuolar in location.

View larger version (131K): [in this window] [in a new window]   Fig. 5.   Vacuolar fragmentation is specific to a deficiency in the cytosolic Cu,Zn-SOD1. The indicated isogenic yeast strains SOD1 SOD2 (1783), sod1 SOD2 (KS105), SOD1 sod2 (JS002), and sod1 sod2 (JS001) were grown aerobically for 10 h in rich media prior to staining with vacuolar membrane specific vital dye FM4-64.

Vacuole Function Is Modestly Compromised in sod1 Yeast-- To determine the potential metabolic consequences of the aberrant vacuoles in sod1 yeast, we examined various vacuole-related functions. The yeast vacuole, like the mammalian lysosome, is an acidic compartment containing abundant hydrolases responsible for macromolecular degradation. In addition, the yeast vacuole plays a role in metabolite storage, pH homeostasis, and sequestration of potentially toxic substances such as metals (for review, see Ref. 35).

Staining with the pH-sensitive vital dye, quinacrine, indicated that the sod1 vacuoles were appropriately acidified (Fig. 2C). However, yeast lacking sod1 were unable to grow at a pH below 3.0 or above 7.0, whereas the wild-type yeast maintained viability over a broader pH range of 2.5 to 7.5 (Fig. 6A). This would suggest that sod1 yeast may not be able to regulate the cytosolic [H+] when extracellular proton levels are particularly high or low. Indeed, other known yeast mutants with a sensitivity to pH are vacuolar mutants (27), suggesting that this sod1 pH sensitivity is vacuole related.

View larger version (29K): [in this window] [in a new window]   Fig. 6.   Additional sod1 phenotypes, potentially vacuole related, include increased sensitivity to pH, starvation, and zinc. A, assay of ability of SOD1 (1783) and sod1 (KS105) yeast to grow over a broad pH range. B, assay of ability to survive 1 week in water (amino acid, sugar, nitrogen, and phosphate starvation) or in minimal sporulation media (sugar, nitrogen, phosphate starvation). C, assay of ability to grow in the presence of increasing levels of zinc.

The vacuole is responsible for the storage of a subset of amino acids and serves as a nitrogen and phosphate reserve (35). When yeast are exposed to adverse starvation conditions, such as that which occur when yeast sporulate or enter stationary phase, they normally up-regulate vacuolar hydrolases to turnover macromolecules to recycle necessary nutrients (36-38). Consistent with previous reports (16, 17), we found that when starved for nitrogen, phosphate, or amino acids, the sod1 yeast exhibited decreased viability (Fig. 6B). This suggests that the fragmented vacuoles may not provide adequate stores of nutrients for starvation conditions or that vacuole-dependent macromolecular turnover processes may be compromised. This is similar to pep vacuolar mutants which also exhibit a sporulation defect and inability to survive starvation conditions (39).

The vacuole (like the mammalian lysosome) additionally plays a role in the sequestration of metals. Mutations in vacuolar components Pep3p and Pep5p result in aberrant vacuole morphology and increased sensitivity to iron, cadmium, and zinc (28). Similarly, sod1 yeast not only have aberrant vacuoles but also an increased sensitivity to iron (40), cadmium (41), and zinc (Fig. 6C), and like the vacuole abnormalities, the sensitivity to these metals is oxygen-dependent. Thus, metal homeostasis and/or sequestration may be altered in these sod1 cells. Altogether, the evidence presented here suggests that oxidatively damaged sod1 vacuoles may be compromised in several aspects of function.

Novel Mechanism of Vacuolar Fragmentation in sod1 Yeast-- Because aberrant vacuole morphology has been described in a number of known yeast mutants, we examined whether the sod1 fragmentation may be occurring through a previously identified pathway. Vacuole fragmentation has been described in yeast mutants, known as Class B vps mutants, which exhibit aberrant trafficking of vacuole proteins (27). To investigate whether loss of Sod1p affects vacuolar protein trafficking, the maturation of the vacuolar protease carboxypeptidase Y was examined. Like many vacuolar proteins, carboxypeptidase Y is synthesized in the endoplasmic reticulum (yielding a polypeptide that migrates with a mobility corresponding to 67 kDa), glycosylated in the Golgi (mobility of 69 kDa), and cleaved upon arrival into the vacuole (mobility of 61 kDa). As seen in Fig. 7A, carboxypeptidase Y maturation was similar in sod1 mutants and in wild-type yeast indicating that the fragmentation is not likely due to a global aberration in vacuolar protein trafficking (as is typical of vps mutants (lane 4, Fig. 7A)) but is occurring via a different mechanism.

View larger version (65K): [in this window] [in a new window]   Fig. 7.   Aberrant sod1 vacuole morphology does not arise from vacuole protein mistrafficking or increased autophagy. A, carboxypeptidase Y (Cpy) forms (69-kDa Golgi precursor and 61-kDa mature vacuolar product) immunoprecipitated from 35S-pulse labeled yeast lysate of strains SOD1 (1783), sod1 (KS105), VPS5 (SEY6210), and vps5 (BHY152) visualized by autoradiography after separation on an SDS-polyacrylamide electrophoresis gel. B, FM4-64 vacuolar membrane staining of SOD1 and sod1 yeast with (PEP4) or without (pep4) the Pep4p protease.

Autophagic yeast mutants also exhibit aberrant vacuole morphology (42, 43). Autophagy is a process in which vesicles of cytosolic material fuse with vesicles containing lysosomal/vacuolar hydrolases in order to degrade cellular components. We hypothesized that aberrant sod1 vacuoles may be the manifestation of increased autophagy, stimulated to turnover oxidatively damaged macromolecules. Pep4p is a vacuolar protease required for the maturation and activation of other hydrolases (44) and without which, autophagy is severely crippled (42). Analysis of sod1 pep4 yeast demonstrated that elimination of Pep4p-dependent autophagy did not affect vacuole fragmentation (Fig. 7B), indicating that elevated autophagy (or at least Pep4p-dependent autophagy) is not a likely cause of or contributor to sod1 fragmentation.

Since drug-induced microtubule disruption has been demonstrated to yield a fragmented appearance of vacuoles (45), we tested whether sod1 vacuole fragmentation arose, at least in part, from microtubule disruption. By immunofluorescence microscopy, we observed a normal staining pattern of microtubules (data not shown). Thus, free radical-induced microtubule disarray is not likely responsible for sod1 vacuole fragmentation either.

Iron Is Implicated in Aberrant Vacuolar Morphology-- Because iron homeostasis is altered in sod1-deficient cells and subcellular fractionation of iron-loaded cells suggest the vacuole is a major site for iron sequestration (18, 19), one additional explanation for vacuole fragmentation in sod1 yeast may be that the vacuole is particularly prone to damage due to higher levels of adverse iron-mediated reactions. Superoxide has been demonstrated to react with labile iron-sulfur clusters of dehydratase enzymes such as aconitase and dehydrogenases (3, 46, 47) to liberate iron, which may then be trafficked to the vacuole/lysosome. Free iron can react readily with the reactive oxygen species to generate toxic hydroxyl radicals which have an estimated lifetime of ~2 ns and radius of diffusion of ~20 Å in aqueous solution (reviewed in Ref. 48). Thus, if hydroxyl radicals were generated by vacuolar iron pools, they would induce damage primarily within the vacuole.

To determine the degree to which iron contributes to vacuole damage, sod1 yeast were either starved for iron or loaded with excess iron and vacuolar morphologies were examined. By deleting Fet3p oxidase (which is thought to convert Fe²⁺ to Fe³⁺ for import by the Ftr1p iron transporter (49-51)), high-affinity iron uptake was impaired in both SOD1 fet3 and sod1 fet3 yeast. As seen in Fig. 8B, absence of the Fet3p transport component ameliorated fragmentation (Fig. 8B). fet3 did not rescue other oxygen-dependent sod1 phenotypes (such as the lysine biosynthetic defect) suggesting that fet3 acts by decreasing intracellular (and vacuolar) iron, thereby diminishing the iron catalyst necessary for adverse chemistry rather than simply acting globally to reduce oxygen radicals in sod1 yeast. Furthermore, growth in iron-rich media increased sod1 fragmentation (Fig. 8F), again consistent with reactive oxygen species reacting with vacuolar iron to provoke local damage.

View larger version (116K): [in this window] [in a new window]   Fig. 8.   Vacuole fragmentation is iron-dependent. FM4-64 vacuolar membrane staining of isogenic SOD1 fet3 (YPH250-fet3) and sod1 fet3 (SL203) yeast grown in low-iron media (A and B) (YPD + 50 mM bathophenanthrolinedisulfonic acid), SOD1 FET3 (YPH250), and sod1 FET3 (SL202) yeast grown in standard media (C and D) or iron-rich media (E and F) (YPD + 2 mM iron).

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Because mitochondria are the primary intracellular sources of superoxide radicals, we anticipated mitochondrial damage in sod1 yeast. However, the most obvious sod1 defects were not in the mitochondria, but rather in the vacuole. Oxidative stress to Sod1p-deficient yeast leads to vacuolar fragmentation through an iron-dependent process. Evidence that the vacuole fragmentation is oxygen (presumably superoxide radical) mediated includes the findings that only aerobically grown sod1 yeast exhibit aberrant vacuoles and genetic supressors (such as pmr1) capable of rescuing all other known aerobic sod1 deficits can also rescue this fragmentation phenotype. Iron deprivation (as in fet3 sod1 strains) also alleviates the fragmentation without providing resistance to O₂ generating reagents or ameliorating other oxygen-dependent phenotypes of sod1 yeast. Thus, reactive oxygen species arising in the absence of Sod1p, in combination with iron, provoke damage to vacuoles.

The notion of iron exacerbating damage under conditions of oxidative stress has been raised previously in other contexts. Fe²⁺ can react with physiological levels of H₂O₂ (and possibly other reactive oxygen species) to produce toxic hydroxyl radicals (reviewed in Ref. 48). In vivo, iron is not found typically in a free state but is sequestered as iron complexes or is bound protectively to enzymes, proteins, or iron carriers (for review, see Refs. 48 and 52). However, superoxide can alter this cellular iron homeostasis. Several groups have found in vivo evidence that superoxide radicals oxidize labile iron-sulfur clusters to liberate iron (1, 2, 3, 46, 47) although the fate and redox state of the "liberated" iron is not known. Additionally, genetic screens have found that mutations in iron-sulfur cluster assembly proteins lessen oxidative damage in sod1 yeast, suggesting that reducing the number of iron-sulfur targets for oxygen radical damage lessens damage in these oxidatively stressed cells (53). Similarly, superoxide radicals have been shown to oxidize iron storage components to liberate iron which is released in a form capable of catalyzing hydroxyl radical production (54). More generally, Wisnicka et al. (40) have suggested that superoxide radicals (in sod1 yeast) may influence the redox status of cellular iron pools by reducing Fe³⁺ to the more reactive Fe²⁺ state. Although total iron is not elevated in sod1 yeast,² the cellular distribution of iron, the ratio of ferrous to ferric iron, and the level of free iron may be altered. In any case, much evidence does suggest that oxygen radicals liberate iron which in turn may exacerbate oxidative damage.

In yeast, the vacuole plays a role in iron homeostasis. Subcellular fractionation of iron-loaded cells suggests that the vacuole is the major site of iron sequestration (18) and yeast lacking a recognizable vacuole structure are unable to accumulate high levels of iron (19). Thus, an iron-enriched yeast vacuole may be particularly susceptible to iron-mediated damage in superoxide-rich sod1 yeast. One possibility is that the vacuole damage may result from a disruption of the iron storage state in the vacuole. Under the influence of increased superoxide levels (which perhaps may be found throughout the Sod1p-deficient cell), vacuolar iron pools may be converted to a more accessible and active Fe²⁺ state that can catalyze hydroxyl radical production and damage macromolecules in the immediate vicinity of the vacuole. A second, not mutually exclusive, possibility is that alterations in iron handling and homeostasis within the cell may result in an increased trafficking of free iron to the vacuole. Increased vacuolar iron may result in increased iron-mediated oxidative reactions within the vacuole or alternatively, the increased iron may simply bind nonproductively to various sites in the vacuole (i.e. sulfur, nitrogen, or oxygen containing ligands). We cannot rule out the possibility that the primary damage actually occurs to cytosolic components that then secondarily afflict vacuolar structure and function. Indeed, we should stress, however, that there is not a global defect in trafficking to the vacuole or vacuole assembly as judged by normal carboxypeptidase Y trafficking results. In any case, both reactive oxygen species and iron contribute to vacuole changes.

Iron-mediated damage occurs not just in the yeast vacuole but also in the mammalian analogue, the lysosome. Similar to the yeast vacuole, the lysosome plays a role in intracellular iron storage, homeostasis, and detoxification (55, 56). Iron overload in the rat liver has been shown to increase lysosomal fragility (56). Although lysosomal fragility has been partially attributed to lipid peroxidation, we should stress that similar lipid membrane damage is not the cause of sod1 vacuole abnormalities because S. cerevisiae do not synthesize the polyunsaturated fatty acids that are susceptible to lipid peroxidation (57). Damage to vacuolar membrane proteins (as well as lysosomal membrane proteins) is a more likely possibility.

By examining structural changes in cellular architecture caused by Sod1p deficiency, we have uncovered a plausible explanation for various puzzling defects in the sod1 yeast: some of the pleiotropic sod1 phenotypes may arise as secondary consequences of a compromised vacuole. For example, sod1 yeast have a sporulation defect (16) and an increased death rate upon entering stationary phase (17). Previously, this sensitivity to adverse conditions was attributed to mitochondrial insufficiency (16, 17). However, although sod1 yeast do not grow as robustly on non-fermentable carbon sources as wild-type yeast, they nevertheless are respiration proficient (e.g. Ref. 15). Instead, we suggest these sod1 abnormalities may be, in part, vacuole related. When yeast are exposed to such adverse conditions, they normally up-regulate vacuolar hydrolases to turnover intravacuolar reserves and cytosolic macromolecules in order to recycle necessary nutrients (36-38). The fact that the sod1 yeast cannot survive through such starvation conditions may be an indication that either vacuolar nutrient storage or the vacuolar-dependent macromolecular turnover processes are compromised. Indeed, many known vacuolar mutants also cannot survive adverse conditions (39). Furthermore, we demonstrate that sod1 yeast have a sensitivity to pH which cannot easily be explained without invoking the role of the vacuole in pH homeostasis. Likewise, an increased sensitivity to transition metals may be directly or indirectly related to vacuole abnormalities. Thus, some of the curious metabolic and biochemical abnormalities in sod1 yeast may in fact be attributable to vacuolar aberrations.

In summary, we demonstrate that oxidative stress to sod1 yeast results in vacuolar fragmentation and either removal of oxygen or iron can ameliorate the damage. This may have medical relevance in that oxidative damage and alterations in iron homeostasis have been implicated in a number of disease states including atherosclerosis, neurodegeneration, arthritis, and aging (for review, see Ref. 48). In some cases oxidative insult is thought to be the primary cause of damage. For example, brain and spinal cord deterioration after ischemic or traumatic injury often appears excessive for the level of trauma (58). One explanation put forth is that iron-binding capacity in the central nervous system and in the cerebral spinal fluid bathing the central nervous system is particularly low. Thus, release of iron by oxidatively or mechanically damaged cells, organelles, and proteins may catalyze toxic oxidative side reactions causing further cell injury. Consistent with this concept, preliminary trials of treatment with iron-chelating agents have had success in diminishing post-traumatic degeneration of brain and spinal cord (59, 60). Thus, iron-mediated oxidative damage is likely a common event in aerobic organisms, and dissection and understanding of the process (and ways of prevention) in yeast may give insight into human disease.

	ACKNOWLEDGEMENTS

We thank Steve Gould for advice on EM analysis of yeast and Scott Emr for advice and reagents to analyze vacuoles. We also thank Dan Kosman (SUNY, Buffalo) for providing yeast strain YPH250-fet3.

	FOOTNOTES

^* This work was supported in part by National Institute of Health Grants NS 27036 (to D. W. C.) and GM 50016 (to V. C. C.) and the Johns Hopkins University National Institute on Environmental Health Sciences Center.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

^** Supported by the Ludwig Institute for Cancer Research. To whom correspondence should be addressed: Ludwig Institute for Cancer Research, 9500 Gilman Dr., La Jolla, CA 92032-0660. Tel.: 619-534-7811; Fax: 619-534-7659; E-mail: dcleveland@ucsd.edu.

² V. C. Culotta, unpublished results.

	ABBREVIATIONS

The abbreviation used is: SOD, superoxide dismutase.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES