Evidence for a novel role of copper-zinc superoxide dismutase in zinc metabolism.

The LYS7 gene in Saccharomyces cerevisiae encodes a protein (yCCS) that delivers copper to the active site of copper-zinc superoxide dismutase (CuZn-SOD, a product of the SOD1 gene). In yeast lacking Lys7 (lys7Delta), the SOD1 polypeptide is present but inactive. Mutants lacking the SOD1 polypeptide (sod1Delta) and lys7Delta yeast show very similar phenotypes, namely poor growth in air and aerobic auxotrophies for lysine and methionine. Here, we demonstrate certain phenotypic differences between these strains: 1) lys7Delta cells are slightly less sensitive to paraquat than sod1Delta cells, 2) EPR-detectable or "free" iron is dramatically elevated in sod1Delta mutants but not in lys7Delta yeast, and 3) although sod1Delta mutants show increased sensitivity to extracellular zinc, the lys7Delta strain is as resistant as wild type. To restore the SOD catalytic activity but not the zinc-binding capability of the SOD1 polypeptide, we overexpressed Mn-SOD from Bacillus stearothermophilus in the cytoplasm of sod1Delta yeast. Paraquat resistance was restored to wild-type levels, but zinc was not. Conversely, expression of a mutant CuZn-SOD that binds zinc but has no SOD activity (H46C) restored zinc resistance but not paraquat resistance. Taken together, these results strongly suggest that CuZn-SOD, in addition to its antioxidant properties, plays a role in zinc homeostasis.

With the appearance of molecular oxygen (O 2 ) in the earth's atmosphere, all aerobic organisms evolved methods to utilize O 2 for energy production. However, reactive byproducts of O 2 metabolism (also known as reactive oxygen species) can be deleterious, and aerobic life forms have developed several systems to combat metabolic and environmental sources of oxygen toxicity. One major constituent in protecting cellular components against reactive oxygen species is superoxide dismutase (SOD). 1 As a part of the primary line of defense against reactive oxygen species, SODs use metal ions to catalyze the dispropor-tionation of superoxide (O 2 . ) to hydrogen peroxide (H 2 O 2 ) and O 2 (1). Eukaryotes, including Saccharomyces cerevisiae, contain two distinct SOD enzymes. Mn-SOD, encoded by the SOD2 gene, resides in the mitochondrial matrix; CuZn-SOD, encoded by the SOD1 gene, is localized mainly in the cytoplasm and nucleus as well as in the mitochondrial intermembrane space. Mutant derivatives of S. cerevisiae that lack either of the SODs exhibit phenotypic deficiencies which are particularly extreme in the sod1⌬ mutants. Even though they are viable, sod1⌬ strains grow poorly in air, are extremely sensitive to redoxcycling drugs, die quickly in the stationary phase, and display aerobic lysine and methionine auxotrophies. It is generally believed that the loss of superoxide-scavenging activity, which is due to the absence of SOD1 gene product, leads to a serious burden of oxidative stress and causes havoc in cellular metabolism and growth processes (2,3). Employing in vivo whole-cell Fe(III) EPR methodology, Keyer and Imlay (4) observed an increase in loosely bound or "free" iron 2 in sod mutants of E. coli. This iron is present in the Fe(II) state, as evidenced by the fact that it is only detectable by EPR at g ϭ 4.3 after treatment of the cells with the iron chelator desferrioxamine, which converts EPR-silent Fe(II) to EPR-detectable Fe(III). A potential source of this increased pool of "free" iron includes a class of enzymes containing exposed [4Fe-4S] clusters, which are superoxide-sensitive (5). The classic example of this type of enzyme is aconitase (6). Attack by O 2 . oxidizes the [4Fe-4S] cluster, resulting in the release of at least one iron ion and inactivation of the enzyme. In E. coli, this process was proposed to be the source of the "free" iron that built up in mutants lacking SOD (4). Recently, adapting the same EPR method, we demonstrated that yeast sod1⌬ mutants also have greatly elevated levels of "free" or "EPR-detectable" iron (iron detectable at g ϭ 4.3 by EPR). In yeast, unlike in E. coli, the iron is present in the Fe(III) state (7). Studies of yeast mutants lacking the SOD1 gene (sod1⌬) have revealed intriguing links connecting CuZn-SOD with transition metal metabolism. Addition of manganese or copper salts to the medium in modest amounts can improve the growth of these strains (reviewed in Ref. 8). More recently, we have shown that sod1⌬ yeast have an increased requirement for iron in aerobic growth, and under certain growth conditions, they display a moderate increase in iron uptake and accumulation (9). Moreover, a variety of second site genetic suppressors of the sod1⌬ phenotype have been isolated (10 -13), and although the gene products involved are located in different cellular compartments and organelles, they all participate in the homeo-stasis of transition metal ions: copper, manganese, or iron.
Out of this scientific endeavor has also come the discovery of the copper chaperone for SOD1 (CCS). This protein is responsible for the delivery of copper to CuZn-SOD and is encoded by the LYS7 gene in S. cerevisiae (14,15). Because the SOD1 polypeptide in lys7⌬ mutants lacks copper in its active site, it is incompetent to catalyze the removal of O 2 . . lys7⌬ and sod1⌬ strains exhibit very similar phenotypic defects (including poor growth in air and aerobic auxotrophies for methionine and lysine), but they are not identical, as we report here. Certain differences exist between these two mutant derivatives, most notably the following two characteristics: 1) sod1⌬ mutants display a 5-fold increase in the levels of EPR-detectable iron relative to lys7⌬ and wild-type yeast and 2) sod1⌬ mutants are sensitive to zinc while lys7⌬ yeast are not. These dissimilarities strongly suggest an in vivo role for SOD1 in metal metabolism, and we have provided further evidence that the SOD1 polypeptide may play a role in cellular resistance to elevated levels of zinc ions.

EXPERIMENTAL PROCEDURES
Reagents, Media, Strains, and Cell Growth-High purity sulfate salts of zinc, copper, and nickel were obtained from Sigma. Methyl viologen, commonly known as paraquat, and the antibiotic G418 were also purchased from Sigma. The yeast strains used in this study are described in Table I. Strains BY4741, JW202, and JW203 were purchased from Research Genetics (Huntsville, AL). The lys7⌬ strain JW101 was obtained by deleting the chromosomal LYS7 locus of EG103 using the lys7::LEU2 construct pSL2527 as described previously (16). Using our sod1⌬::URA3 construct, JW102 (lys7⌬sod1⌬) was prepared by sequential gene deletion starting with the lys7⌬ strain. Southern blot analysis was performed to verify the correct deletion of the LYS7 and SOD1 genes (data not shown). The plasmids pEMBLyex4 and pSALSOD1 (referred to herein as pMn-SOD and carrying the Bacillus stearothermophilus gene for Mn-SOD) were kindly provided by Dr. C. Bowler (17) and introduced into JW201 (the sod1 null yeast in the BY4741 background) by lithium acetate transformation (18) to create strains JW203 and JW204. EG118 (sod1⌬) yeast cells were transformed with YEp351, YEp600, or YEp351-yH46C to yield strains HH101, HH102, or HH103, respectively. YEp351 is a yeast multicopy shuttle vector with a LEU2 selectable marker (19). YEp600 contains a 1.1kilobase genomic fragment containing the wild-type SOD1 gene with its own promoter and carrying an engineered NcoI site at the translation start site (20). YEp351-yH46C was generated by inserting the yH46C fragment (NcoI-SmaI) from pGEM-3Zf(ϩ)-yH46C (21) into YEp600 digested with NcoI-PstI/blunt and introduced into the specified yeast strain by lithium acetate transformation (18). Expression was verified by Western blot analysis.
For experiments, yeast without plasmids were cultured in synthetic complete medium with 2% glucose (SDC) or galactose (SGalC) as mentioned. Yeast with plasmids were cultured in the appropriate dropout medium (SD-Leu for HH101, HH102, and HH103 and SD-Ura or SGal-Ura for JW203 and JW204). SDC medium was composed as described (18) except that the supplements Leu, His, Trp, Met, Ura, and Ade were increased 4-fold. Overnight startup cultures were grown from a single colony in the same type of medium that was used for further growth except when the medium contained galactose. For growth in galactose-containing medium, the startup culture in glucose was washed twice with sterile water prior to inoculation in galactose-containing medium.
EPR Sample Preparation and Measurements-For EPR studies, EG103, EG118, JW101, and JW102 were seeded at an optical density at 600 nm (A 600 ) of 0.1 in 50 ml of SDC medium and grown for 72 h at 30°C. After growth, cells were centrifuged and then briefly treated with desferrioxamine, which binds to labile iron and converts it from Fe(II) (EPR-silent) to Fe(III) prior to washing with 20 mM Tris-Cl, pH 7.4. After washing, the pellet was resuspended in 200 l of 20 mM Tris-Cl, pH 7.4, containing 20% glycerol, and 200 l of the resuspended cells was packed in EPR tubes. The tubes were stored at Ϫ70°C until EPR measurements were performed. EPR spectra were recorded using a Bruker X-band spectrometer using a finger Dewar filled with liquid nitrogen attached to the cavity of the instrument to maintain low temperature. Parameters used for low temperature Fe(III) EPR were as follows: center field, 1500 G; sweep width, 500 G; microwave power, 20 milliwatts; attenuation, 10 dB; modulation amplitude, 20.2 G; modulation frequency, 100 kHz. Further details regarding sample preparation, instrument parameters, and data analysis used for the whole-cell low temperature Fe(III) EPR measurements are delineated in our previous study (7).
Assessment of Metal Ion and Paraquat Sensitivity-Paraquat, zinc, copper, or nickel was added to the indicated medium at the indicated concentration. Experimental cultures in dextrose-containing medium were inoculated at an A 600 of 0.03 to 0.05. Experimental cultures in medium containing galactose were seeded at an A 600 of 0.05 to 0.07. To evaluate the effect of excess zinc ion, the starting pH of the growth medium was at 4 instead of 6 to prevent the precipitation of zinc as zinc hydroxide at pH 6. (This pH change had no obvious effect on growth. Data not shown.) Cultures were grown at 30°C with shaking at 200 rpm for a maximum of 24 h. Total cell growth was determined by measuring turbidity at A 600 . An A 600 of 1 indicates a culture density of 1 ϫ 10 7 cells/ml. Assessment of Total Metal Content-Cells were grown as described above. Typically, duplicate samples containing 15 to 20 ϫ 10 7 cells from 25-ml stationary phase cultures were harvested by centrifugation. The cell pellets were washed once in 5 ml of 10 mM EDTA, pH 8, followed by three washes in 5 ml of metal-free water. The final cell pellets were resuspended in 1 ml of 10% ultrapure nitric acid, and digestion was allowed to take place at 98°C for 18 h. After complete digestion, 950 l was diluted to 7 ml in metal-free water, and these samples were analyzed using a Thermo-Jarrel Ash Iris 1000 inductively coupled plasmaatomic emission spectrometer.

RESULTS
lys7⌬ yeast have normal SOD1 expression, but the SOD1 polypeptide lacks enzymatic activity because the copper is not incorporated into the active site (14,15). Therefore, one would expect lys7⌬ and sod1⌬ strains to possess comparable antioxidant capacity and thus to share similar phenotypes. With the lys7⌬ and sod1⌬ strains constructed from our laboratory wildtype yeast (EG103 background), we have observed numerous similarities between these two strains, such as slow growth, aerobic auxotrophies for lysine and methionine, poor growth in the presence of non-fermentable carbon source, and temperature sensitivity at 37°C. However, further characterization of the lys7 null strain led to the discovery of certain differences between these two yeast mutants. Resistance to Superoxide-First, we examined the effect of paraquat, a known generator of intracellular superoxide, on wild-type, lys7⌬, sod1⌬, and lys7⌬sod1⌬ yeast by culturing cells in the presence of varying concentrations of paraquat for 24 h. As shown in Fig. 1A, at 10 and 25 M of paraquat, concentrations that are toxic for both sod1⌬ and lys7⌬sod1⌬ mutant strains, lys7⌬ cells still grew as well as wild-type cells, but 100 M paraquat was toxic. In contrast, wild-type EG103 can survive in 1 to 10 mM paraquat. Identical experiments were conducted in another genetic background, BY4741, with similar results (Fig. 1B). In the BY4741 background, lys7⌬ cells are resistant to the toxic effects of paraquat at concentrations as high as 100 M (83.2 Ϯ 2.9% control growth), whereas sod1⌬ yeast exhibit significant levels of growth retardation at 25 M (8.0 Ϯ 4.8% control growth).
EPR-detectable Iron-Because it is postulated that an elevated superoxide level is connected to an increase in cellular "free" iron (4,9,22), we wonder whether the decreased sensitivity to paraquat for lys7⌬ yeast correlated with a diminished level of "free" iron as measured by whole-cell Fe(III) EPR spectroscopy. Wild-type, lys7⌬, sod1⌬, and lys7⌬sod1⌬ cells were grown to stationary phase (72 h) in SDC, and EPR samples were prepared and evaluated as described under "Experimental Procedures." As shown in Fig. 2, lys7⌬ cells had a basal level of EPR-detectable iron of 10.5 Ϯ 3.2 M, similar to that observed for wild-type yeast. However, the isogenic sod1⌬ and lys7⌬sod1⌬ yeast had levels that were ϳ5-fold higher (56.8 Ϯ 14.3 M and 47.9 Ϯ 6.0 M, respectively). We also prepared EPR samples of the aforementioned strains without desferrioxamine treatment and observed similar results (data not shown), suggesting that the "free" iron present in all strains was in the ferric state and that the intracellular redox status was similar. Interestingly, in these experiments, elevation of the "free" iron in sod1⌬ mutants was not accompanied by an increase in the total iron content (see below).
Zinc Sensitivity-One of the more puzzling aspects of the phenotypes of sod1⌬ mutants has been their heightened sensitivity to extracellular zinc. We discovered this effect several years ago, and since then, we have sought for an explanation. Wild-type cells tolerate 5 or even 10 mM zinc in the medium (data not shown). On the other hand, cells lacking CuZn-SOD exhibit growth retardation at zinc levels as low as 1 mM. Thus, we decided to assess the ability of excess zinc ions to inhibit growth in lys7⌬ yeast as well. In the experiment shown in Fig.  3A, lys7⌬, sod1⌬, lys7⌬sod1⌬, and the parental wild-type strain (EG103) were cultured in the presence of increasing concentrations of ZnSO 4 . Unlike the sod1⌬ mutant, which displayed marked growth retardation at 0.5 mM ZnSO 4 (ϳ73% control growth), lys7⌬ cells behaved like wild type at all concentrations of ZnSO 4 tested (up to 10 mM, Fig. 3 and data not shown). To ascertain that this phenotypic difference between lys7⌬ and sod1⌬ null strains is not a strain-dependent phenomenon, we tested the effect of excess zinc ions on these two mutants made in the BY4741 background. As shown in Fig. 3B, the resistance of lys7⌬ to elevated levels of ZnSO 4 mirrored that observed for the EG103 background.
In two genetic suppressors of sod1⌬ mutants, pmr1 and bsd2, increased sensitivity toward manganese and copper ions was found to correlate with the accumulation of intracellular manganese and copper, respectively (11,12). Thus, we proceeded to explore whether the sensitivity to excess zinc in the sod1⌬ cells was associated with elevated levels of total cellular zinc. The total cellular level of zinc in the sod1⌬ mutant was compared with those in isogenic lys7⌬ and wild-type strains. In the experiment shown in Fig. 4, zinc levels were measured in cells after 72 h of growth in SDC medium. The sod1⌬, as well as the lys7⌬sod1⌬, mutants accumulate ϳ50% higher levels of zinc than do the wild-type and lys7⌬ strains, further implicating SOD1 in the regulation of heavy metal metabolism.
Specificity of the Extracellular Metal Ion Sensitivity-To evaluate the effects of other transition metal ions on wild-type, lys7⌬, and sod1⌬ yeast, the growth of these strains was tested in the presence of CuSO 4 and NiSO 4 . As shown in Fig. 5A, the extent of growth inhibition is comparable for all three strains. A, parent strain EG103 and isogenic lys7⌬, sod1⌬, and lys7⌬sod1⌬ derivatives. B, parent strain BY4741 and isogenic lys7⌬, sod1⌬, and lys7⌬sod1⌬ derivatives. Liquid cultures inoculated at an A 600 of 0.03 were grown for 24 h at 30°C with shaking at 200 rpm. Total growth relative to that obtained in the absence of paraquat (% control growth) was determined turbidometrically at A 600 . Values are the means of 3 independent experiments except for lys7⌬sod1⌬ strain (n ϭ 2). Error bars indicate 1 standard deviation.

FIG. 2. EPR-detectable iron is dramatically increased in sod1⌬
mutant but not in lys7⌬ yeast. WT, lys7⌬, sod1⌬, and lys7⌬sod1⌬ strains were cultured in SDC medium for 72 h. EPR samples for four independent cultures were prepared and analyzed for each strain except for lys7⌬sod1⌬ yeast (n ϭ 2). EPR-detectable iron levels were calculated by double integration as described under "Experimental Procedures." Results shown are averages, and the error bars indicate standard deviation.
As has been observed previously, low concentrations of CuSO 4 (0.5 mM) enhanced the growth of sod1⌬ and lys7⌬ mutants, but higher concentrations were equally toxic to mutant and wildtype cells. In medium supplemented with [Ni 2ϩ ] ranging from 0 to 1.0 mM, lys7⌬ and sod1⌬ demonstrate similar growth potential (Fig. 5B). Interestingly, low medium [Ni 2ϩ ] seemed to have a more pronounced effect on the growth of wild-type (EG103) yeast. It is possible that the weakened antioxidant capacity of lys7 and sod1 null mutants have caused a constitutive up-regulation of various cellular stress genes, thus render-ing these cells more resistant to other types of environmental stresses. Overall, the experimental evidence depicted in Fig. 5 suggests that the ability of both lys7⌬ and sod1⌬ cells to sequester other transition metal ions probably remains intact on a gross level and that the sensitivity of sod1⌬ mutants to excess metal ions is specific to zinc.
Replacement of CuZn-SOD with Bacterial Mn-SOD-To assess the relative contributions of the enzymatic superoxide dismutase activity and the zinc binding capacity of CuZn-SOD, we restored the superoxide resistance of sod1⌬ yeast by overexpressing a bacterial Mn-SOD in their cytoplasm. Because Mn-SOD is expected to restore SOD activity but not zinc binding capacity in vivo, we were interested to see whether its expression would have any effect on the zinc sensitivity of the sod1⌬ mutant.
The sod1 null mutant in the BY4741 background was transformed with either pSALSOD (pMn-SOD), a yeast expression vector containing the B. stearothermophilus Mn-SOD gene under the control of the GAL-CYC1 promoter, or pEMBLyex4 (pEMBL), the corresponding vector control (17). Growth of the sod1⌬/pMn-SOD strain in galactose resulted in the appearance of Mn-SOD activity and loss of the aerobic lysine auxotrophy (data not shown and Ref. 17). We tested the paraquat sensitivity of the plasmid-containing strains and found that sod1⌬/pMn-SOD cells behave very much like wild-type yeast, resisting paraquat concentrations as high as 1.0 mM (41.1 Ϯ 11.4% control growth for sod1⌬/pMn-SOD compared with 52.8 Ϯ 2.2% for wild type (WT)), whereas the sod1⌬/pEMBL strain was much more sensitive (Fig. 6A). Possibly because of the galactose in the medium, somewhat higher levels (0.2 and 0.5 mM) of paraquat are required to cause toxicity in the sod1⌬/ pEMBL as compared with the non-plasmid-containing sod1⌬ strain. Nevertheless, we can conclude that the prokaryotic Mn-SOD efficiently strengthened the antioxidant capacity of the sod1⌬ yeast as evidenced by greatly increased resistance to paraquat, a superoxide-generating chemical.
We then tested whether the transformed strains were resistant to zinc (Fig. 6B). For medium [Zn 2ϩ ] below 2.0 mM, all four strains behaved similarly (data not shown). Thus, higher concentrations of ZnSO 4 than were previously used were required to evaluate the ability of these transformants to resist zinc toxicity, again possibly because of the use of galactose-based medium. As can be seen in Fig. 6B, expression of cytoplasmic Mn-SOD does not restore zinc resistance to wild-type levels, supporting a role for CuZn-SOD in zinc homeostasis.
Replacement of CuZn-SOD with a Mutant SOD1 Polypeptide, yH46C-In a converse experiment, to provide further mechanistic insight into the zinc sensitivity observed in sod1⌬ cells, we overexpressed yH46C (a mutant SOD1 protein lacking catalytic activity but capable of binding metal ions) in sod1⌬ yeast. Growth of the sod1⌬/yH46C strain was tested in medium supplemented with ZnSO 4 . As shown in Fig. 7, sod1⌬/yH46C grew as well as sod1⌬/ySOD1 in the presence of 1.0 and 2.0 mM ZnSO 4 , but sod1⌬/YEp351 (vector control) exhibited significant growth inhibition at those concentrations. Expression of this mutant SOD had no effect on the resistance to paraquat, confirming that it is catalytically inactive. (At a concentration of 25 M paraquat, growth of sod1⌬/yH46C was just 2.1 Ϯ 0.5% control growth.) DISCUSSION In this study, we demonstrate that certain differences exist between lys7⌬ and sod1⌬ yeast. lys7⌬ yeast are not as sensitive to the redox-cycling drug paraquat as the sod1⌬ mutant, although they are still much less resistant than the isogenic wild-type cells. In addition, lys7⌬ cells exhibit wild-type levels of "free" iron as measured by Fe(III) EPR, whereas sod1⌬ mutants display a 5-fold increase (Fig. 2). Moreover, the lys7⌬ strain grows as well as wild type in elevated zinc, while the sod1⌬ strain fails to thrive (Fig. 3). The increased zinc sensitivity of the sod1⌬ strain may be due to the fact that it accumulates excess zinc relative to the wild-type and lys7⌬ strains (Fig. 4). To explain the phenotypic differences between sod1⌬ and lys7⌬ yeast, two different (albeit not mutually exclusive) hypotheses have been considered. The first hypothesis is that the lys7⌬ strain retains an extremely small amount of SOD activity, below the detection limit of currently available assays. The second is that the copper-free SOD1 protein, which is present in lys7⌬ but not in the sod1⌬ strain, plays a role in zinc metabolism either by acting as a depot for zinc ions or through some other mechanism. To distinguish among these possibilities, we performed experiments designed to separate SOD enzymatic activity from the zinc binding activity. We found that cytoplasmic expression of a bacterial Mn-SOD fully restored the paraquat resistance of sod1⌬ strains but did not fully restore the zinc sensitivity (Fig. 6). Conversely, expression of H46C, a mutant CuZn-SOD with no enzymatic activity, fully restored zinc resistance ( Fig. 7) but had no effect on paraquat resistance.
The first hypothesis (that there may be a residual amount of SOD activity in lys7⌬ yeast) is supported by the facts that the lys7⌬ strains appear slightly more resistant to oxidative stress (e.g. paraquat resistance is a little higher in the lys7⌬ mutants than in the sod1⌬ strains) and that the zinc sensitivity of sod1⌬ strains is only observed in aerobic growth. However, another possible explanation for these phenotypic differences is that paraquat resistance is conferred by superoxide dismutase activity plus some other antioxidant function of the SOD1 polypeptide. For example, inactive apoSOD1 polypeptide might be able to physically protect some vulnerable site(s). (If this is the case, the SOD activity is clearly the more important function because expression of active SOD in the cytoplasm increases paraquat resistance 100-fold or more, while expression of SOD1 made inactive by the absence of Lys7 only increases paraquat resistance a few fold.) The main evidence against the residual activity hypothesis is the fact that SOD activity is undetectable in lys7⌬ strains by any assay we have tried.
The second hypothesis, which we favor, is that the SOD1 protein plays a role in zinc metabolism or homeostasis. CuZn-SOD is an abundant enzyme, and thus it has the capacity to bind substantial amounts of copper and zinc. CuZn-SOD can help protect against copper toxicity in yeast lacking Cup1 (copper metallothionein) (23). That it participates in protection against zinc toxicity is supported by several pieces of evidence: 1) zinc resistance was lost only when the SOD1 polypeptide was missing (as in the sod1⌬ mutant, whether or not it was expressing cytoplasmic Mn-SOD); 2) paraquat resistance was lost only when SOD activity was missing (as in the lys7⌬ mutant and the sod1⌬ mutant, whether or not they were expressing the inactive H46C mutant SOD; and 3) a null mutation in the yeast SOD2 gene, which encodes the mitochondrial manganese SOD, is not associated with an increased sensitivity toward zinc (data not shown). Therefore, the presence of SOD1 polypeptide, either inactive or enzymatically active, is required to confer zinc tolerance. The simplest explanation is that SOD1 acts as a "zinc sink," and its absence causes the cell interior to be exposed to levels of zinc in excess of what can be readily tolerated. However, the data showing elevated zinc levels in the sod1⌬ strain are puzzling in this regard. It may be that the lack of SOD1 somehow interferes with zinc sensing by Zap1 or with zinc transport or storage, leading to a buildup or mislocalization of zinc stores. We are currently exploring these possibilities experimentally.
Another level of complexity is that zinc resistance of the sod1⌬ strain was somewhat improved by cytoplasmic expression of the heterologous Mn-SOD, although it was not fully restored. This may indicate that zinc toxicity has an oxidative stress component that is exacerbated when SOD activity is missing. It is possible that this is related to the differences in "free" iron shown in Fig. 2. Alternatively, it may simply be that a SOD-competent cell is better able to defend itself against other stresses or that cellular antioxidant defenses assist in managing disturbances in metal homeostasis.
Numerous studies have implicated the vacuole in metal ion storage and homeostasis (24 -27), and mutants defective in vacuole function or acidification are sensitive to various metal ions, including copper, nickel, and zinc (28,29). Because sod1⌬ mutants display aberrant vacuolar morphology (22), the possibility remained that the zinc sensitivity was caused by vacuole malfunction in the sod1⌬ yeast. We reasoned that if the zinc sensitivity in our strain was due to defective vacuolar function, the ability of the cells to handle elevated levels of transition metal ions other than zinc would be affected as well. The experimental results depicted in Fig. 5 clearly demonstrate that sod1⌬ cells are as resistant as wild-type cells to increased levels of Cu 2ϩ or Ni 2ϩ . That is, their ability to detoxify metal ions other than zinc remains intact, and the defect responsible for the zinc sensitivity in sod1⌬ strains is not one of general vacuolar function. Taken together, these data strongly suggest that the copper-free SOD present in lys7⌬ cells as well as the normally metallated CuZn-SOD present in WT cells participate in zinc ion homeostasis.
A relationship between zinc metabolism and iron metabolism is also possible. The zinc sensitivity of sod1⌬ mutants may be an indirect consequence of their aberrant iron metabolism. Yeast sod1⌬ mutants exhibit increased levels of "free" iron, which may also come from the destruction of vulnerable [Fe-S] clusters (7), and its accumulation may be exacerbated by the increased cellular transport of iron under conditions of superoxide stress (9). Because iron(II) and Zn(II) are similar chemically with respect to their coordination chemistry, they are transported by some of the same cellular transporters and can be bound by some of the same cellular proteins. Increased iron uptake may result in increased zinc uptake as well (30,31). Thus, attempts by the sod1⌬ yeast to increase iron may backfire in two ways: increased formation of hydroxyl radical due to increased Fenton chemistry (32) and the accumulation of toxic intracellular levels of zinc. Here we report that lys7⌬ yeast, unlike sod1⌬ yeast, do not exhibit altered iron metabolism as evidenced by their wild-type-like level of EPR-detectable "free" iron ( Fig. 2), and they do not accumulate increased levels of zinc inside the cells. Further work will be required to determine whether increased levels of "free" iron and of zinc are a result of the same process or of different ones, but the relationship is intriguing.
In summary, our work has revealed unexpected differences between lys7⌬ and sod1⌬ mutant yeast that imply connections between superoxide toxicity, zinc metabolism, and iron metabolism. Our results indicate that the SOD1 polypeptide participates somehow in zinc homeostasis and that superoxide stress clearly involves iron metabolism. Continued investigations will shed more light on these relationships.