Intracellular Mn(I1)-associated Superoxide Scavenging Activity Protects Cu,Zn Superoxide Dismutase-deficient Saccharomyces cerevisiae against Dioxygen Stress*

Three Cu,Zn superoxide dismutase (SOD-1)-defi- cient Saccharomyces cerevisiae mutants do not grow in 100% O2 in rich medium and require Met and Lys when grown in air (Bilinski, T., Krawiec, Z., Liczman-ski, A., and Litwinska, J. (1985) Biochem. Biophys. Res. Commun. 130, 533-539). We show herein that medium manganese(I1) accumulated by the mutants rescues these 02-sensitive phenotypes; 2 mM medium Mn2+ represented the threshold required for cell growth. The accumulation of Mn2+ was not oxygen-inducible since mutants grown aerobically and anaer- obically accumulated the same amount of Mn2+. Mn2+ accumulation is not unique to these mutants since wild type accumulated almost twice as much Mn2+ as did mutant. ESR spectra of the cell extracts and whole cells loaded with Mn2+ were typical of free Mn(I1) ion. These spectra could not account quantitatively for the total cellular Mn2+, however. A screen for soluble antioxidant activities in the Mn2+-supplemented cells detected 0; (superoxide) scavenging activity, with no change in catalase or peroxidase activities. This 0; scavenging activity was CN- and heat-resistant. No achromatic bands were revealed in nondenaturing gels of Mn2+-containing cell extracts stained for 0; scavenging activity.

Three Cu,Zn superoxide dismutase (SOD-1)-deficient Saccharomyces cerevisiae mutants do not grow in 100% O2 in rich medium and require Met and Lys when grown in air (Bilinski, T., Krawiec, Z., Liczmanski, A., and Litwinska, J. (1985) Biochem. Biophys. Res. Commun. 130, 533-539). We show herein that medium manganese(I1) accumulated by the mutants rescues these 02-sensitive phenotypes; 2 mM medium Mn2+ represented the threshold required for cell growth. The accumulation of Mn2+ was not oxygeninducible since mutants grown aerobically and anaerobically accumulated the same amount of Mn2+. Mn2+ accumulation is not unique to these mutants since wild type accumulated almost twice as much Mn2+ as did mutant. ESR spectra of the cell extracts and whole cells loaded with Mn2+ were typical of free Mn(I1) ion. These spectra could not account quantitatively for the total cellular Mn2+, however.
A screen for soluble antioxidant activities in the Mn2+-supplemented cells detected 0; (superoxide) scavenging activity, with no change in catalase or peroxidase activities. This 0; scavenging activity was CNand heat-resistant. No achromatic bands were revealed in nondenaturing gels of Mn2+-containing cell extracts stained for 0; scavenging activity. The Mn2+-dependent 0; scavenging activity in the cell extracts was quenched by EDTA and dialyzable. More than 60% of both the intracellular Mn"+ and the 0; scavenging activity was removed by 2-h dialysis. Dialyzed cells were not viable in air unless resupplemented with either Met or Mn2+. Although Mn2+ supported the aerobic growth of these mutants, excess Mn2+, which correlated with an elevated 0; scavenging activity, was toxic to both mutant and wild type. The results indicate that free or loosely bound Mn2+ ion protects the mutants against oxygen stress by providing an intracellular, presumably cytosolic, 0; scavenging activity which replaces the absent SOD-1.
One-electron reduction of dioxygen leads to the generation of potentially redox active oxygen radicals. Overproduction of these species has been linked to cellular damage, termed oxygen stress (1-3). Antioxidant enzymes include the super-* This research was supported by Grant DK19708 from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ T o whom correspondence and reprint requests should be addressed. oxide dismutases (SOD,* Equation 1) and the catalases and peroxidases, which scavenge the superoxide radical (0;) and HzOZ, respectively. These enzymes appear to play protective roles in aerobes (4-6). 20, + 2H+ + HzO2 + O2 (1) The deletion of both Mnz+-and Fez+-containing SODS in Escherichia coli is lethal to aerobic growth in minimal medium unless specific branched chain amino acids are added (7). The requirement for these amino acids in the double mutant can be correlated to the hypersensitivity of dihydroxy acid dehydratase, an enzyme involved in the biosynthesis of branchedchain amino acids, to hyperbaric oxygen i n vivo (8) and to 0: i n vitro (9). This result indicates that specific cellular proteins may be inherently sensitive to dioxygen-derived free radicals. Such sensitivity may be manifest in the requirement for SOD in aerobic growth.
The yeast, Saccharomyces cereuisiae, has two SOD proteins. SOD-1 is cytosolic and contains 1 mol each of Cu2+ and Znz+ per mol of 16-kDa monomer (10). SOD-2, a 24-kDa polypeptide, contains Mnz+ (11). SOD-2 is synthesized as a preprotein. The 27-amino acid mitochondrial targeting sequence is cleaved following import across the inner mitochondrial membrane (12). Although more than 90% of total cellular yeast 0; scavenging activity is due to SOD-1, a mutant lacking SOD-2 is sensitive to 100% oxygen and is unable to utilize a nonfermentable carbon source in air (13).
Yeast mutants lacking SOD-1 activity also are sensitive t o oxygen. Three such mutants in S. cereuisiae have been isolated and partially characterized by Bilinski et al. (14). These mutants showed a nonparental Met, Lys auxotrophy when grown in air but not when grown anaerobically. All of the mutants failed to grow under 100% Oz even in rich medium. These oxygen-dependent growth defects are due to a mutation of a single nuclear gene which correlates to the lack of active SOD-1 (14, E).' This mutation has been mapped to the SOD1 locus.z The activities of SOD-2 are not different from wild type. The O2 sensitivity of growth was rescued by transformation of these mutants with a plasmid containing wild type yeast SOD-1 gene. Transformation was accompanied by an increase of SOD-1 activity (15,16). Thus, the lack of SOD-1 activity appears to be the causative factor in the mutant growth defects observed under air and oxygen.
Divalent manganese [Mn(II)] is known to dismute 0; i n vitro (17,18). Lactobacillus plantarum, a facultative anaerobe which does not contain any SOD protein species, requires The abbreviations and trival names used are: SOD, superoxide dismutase; SD, synthetic dextrose (minimal medium); YPD, yeast extract-peptone-dextrose (rich medium). E. C. Chang and D. J. Kosman, unpublished observations. Mn2+ when grown aerobically on a nonfermentable carbon source, i.e. when it is obligately aerobic (19). This organism is also resistant to the redox-active quinone, plumbagin, which has been shown to support intracellular 0; generation by this bacterium. Extracts of cells grown aerobically in Mn2+-enriched media exhibited an EDTA-inhibitable 0; scavenging activity which was lost slowly upon dialysis (19,20). This activity is due to the Mn(I1) accumulated by the cell which apparently is retained by a high molecular weight polyphosphate-protein ligand (20). Thus, this and other genera of the family Lactobacillaceae can adapt to aerobiosis, in part, because the species studied accumulate millimolar concentrations of Mn2+. This Mn(I1) provides the 0; scavenging activity provided normally in aerotolerant organisms by micromolar concentrations of SOD (20, 21). Those lactic acid bacteria which contain neither SOD nor high intracellular levels of Mn(I1) are 02-intolerant (21).
Although both SOD-1 (and Mn2+) are important for aerobic growth, excess Mn2+ and SOD are toxic. For example, in E . coli, overexpression of Fez+ SOD is toxic to aerobically grown cells, an effect which correlates with an increase of intracellular H202, the by-product of the dismutation reaction (22). Furthermore, overexpression of SOD-1 increases lipid peroxidation in cultured cells (23) and induces symptoms similar to Down's syndrome in transgenic mice (24). A yeast SOD has not yet been overexpressed in yeast, but Mn2+ has been shown to be mutagenic (25).
We have used Mn2+ as a growth supplement to investigate the relationship between cellular 0; scavenging activity and aerobic growth of yeast using the SOD-1-yeast mutants described above, that is, we wished to test the hypothesis that as in L. plantarum, Mn(I1) would render these mutants aerotolerant. The data suggest that loosely bound Mn(I1) accumulated by these mutants does protect them against oxygen stress via a protein-independent 0; scavenging activity which complements the absence of active SOD-1. On the other hand, excess Mn2+ inhibits growth of both mutant and wild type; the excess Mn2+ in the cell extracts correlates with an elevated 0; scavenging activity.

Materials
Chemicals-Media components were purchased from Difco. The reagents for SOD assay were obtained from Sigma with the exception of hydroxylamine which was purchased from Gallard-Schlesinger (Analar, British Drug House). We have observed that some batches of Grade 111 xanthine oxidase contain more than 1 mM EDTA resulting in chelation of Mn2+ in the cell samples. Grade I xanthine oxidase can be used to avoid this problem. Bovine SOD-1 and catalase were obtained from Boehringer Mannheim and Sigma, respectively. (NH4)2S04, MgS04.7H20, NaC1, and CaC12 used for phosphate-free SD medium were Analar (British Drug House) reagents obtained from Gallard-Schlesinger. The Mn2+ standard for atomic absorption spectrophotometry was purchased from Alfa Products. All other reagents used were reagent grade. The [Mn"] in the distilled water used for growth media and reagents was below the detection limits by flameless atomic absorption spectrophotometry (0.1 pg/pl).
Medium for Cell Growth-The yeast extract-peptone-dextrose (YPD) medium contained 1% (w/v) yeast extract, 2% (w/v) Bactopeptone, and dextrose. The semidefined medium (SD) contained 0.67% (w/v) yeast nitrogen base without amino acids and 2% (w/v) dextrose and was buffered by 10 mM phosphate. This medium contained 9 p M MnS04. For anaerobic growth experiments, 5 mg/liter ergosterol and 1.4 g/liter inositol were added. Phosphate-free SD was prepared as described (26) except that K2HP04 was omitted. This medium was buffered to either pH 4.5 or 6.0 by 10 mM succinate-KOH. Medium pH has to be lower than 5.0 if 4 mM or more Mn2+ is added (to avoid precipitation). Amino acids and/or nucleotides added (mg/liter of medium) for optimal growth of these strains were as follows: Arg and Leu (150) for Dscdl-1C; His (loo), Leu (go), and Ura (25) for Dscdl-4A; Arg and Leu (22.5) and Ura (2.5) for Dscd2-2C; and Trp (280) for AS2-2A.

Methods
Cell Growth-Cell growth was monitored by turbidity a t 660 nm. Prior to experiments, cultures were diluted into fresh medium to give A = 0.1 unless otherwise indicated in the text. Cells were grown a t 30 "C and harvested at the culture density indicated. Growth yield was represented by the culture turbidity after 18-20 h of growth. For anaerobic growth experiments, each 5 ml of culture was flushed with pure N, for 30 min, and the culture was then sealed with a rubber septum.
Preparation of Cell-free Extracts-Cells were collected on an 0.8pm membrane and washed with distilled water (at least 10 X culture volume). The cell pellet was resuspended either in ice-cold distilled water or in 100 mM sodium phosphate buffer (pH 7.8) containing 1 mM EDTA and 0.1% Triton X-100 to one-tenth of the volume of the original cell culture. One-half volume (relative to the latter cell suspension) of glass beads (0.3-0.45 mm) was added. Following vortexing, the homogenization supernatant was collected by centrifugation. Protein concentration in the extract was determined by BCA (bicinchoninic acid) assay (Pierce Chemical Co.) using bovine serum albumin as standard (27).
Enzyme Assays-The EDTA in the nitrite SOD assay used (28) was omitted as noted. The contribution of SOD-2 to the total 0; scavenging activity was determined by adding 1 mM KCN to the assay mixture. The activity of SOD-1 was determined by subtracting the SOD-2 activity from the total 0; scavenging activity. This assay was about 16 times more sensitive than the more widely used cytochrome c assay (29) but yielded identical relative results. Addition of MnS04 did not inhibit the xanthine oxidase-catalyzed production of uric acid from hypoxanthine (assayed spectrophotometrically at 295 nm) suggesting that the Mn2+-dependent 0; scavenging activity observed was not due to an artifact of underproduction of 0; in the SOD assay (data not shown). The 0 2 scavenging activity was also determined in cellular protein samples after separation by gel electrophoresis (30). For samples containing Mn2+, EDTA was omitted from all the reagents used and 3 times more nitroblue tetrazolium and riboflavin were used than are suggested. The catalase activity was determined from the decrease of [H202] observed a t 240 nm (31). Based on a standard catalase preparation (Sigma), 1 unit was equivalent to 861 nmol of H202 min". Determination of Mn2+ and Phosphate-Room temperature ESR spectra were obtained a t 9.5 GHz using a Varian E-9 ESR spectrometer operating at a power of 30 milliwatts. The total amount of Mn2+ in samples was determined by flameless atomic absorption spectrophotometry using a Perkin-Elmer model 360 equipped with model HGA 2100 graphite furnace. Phosphate analysis was as described (32).
Dialysis of Cellular Mn2+-The SOD-1-mutant, Dscdl-lC, was grown to early log phase in SD medium (pH 4.5) supplemented with either Met or 4 mM Mn2+. Following washing, one-half of the cell pellet was resuspended in 250 pl of water, and an extract was prepared as described above. The remainder was resuspended in water and transferred to a dialysis bag (Spectrapor, Spectrum Medical Industries, Inc.) with a molecular mass cut-off of 10 kDa. This cell sample was dialyzed against 100 mM phosphate buffer (pH 4.5) a t room temperature for 2 h prior to washing and preparation of the cell extract.

Mn2+
Supplementation: Effect on Oxygen Stress-Addition of 2 mM Mn2+ to YPD plates rescued growth under pure oxygen of all three SOD-1-mutants, Dscdl-lC, Dscdl-4A, and Dscd2-2C ( I C , 4A, and 2C in Fig. 1). The Met and Lys auxotrophy observed under air could be rescued also if sufficient Mn2+ was added to the growth medium. Fig To test if the protective effect on 02-dependent Met auxotrophy was specific for Mn2+, M P , Cu2+, and Zn2+ were tested also. Magnesium chloride up to 4 mM or ZnC12 up to 1 mM in the growth medium did not provide relief from the aerobic Met/Lys auxotrophy. Addition of CuS04 up to 0.25 mM was minimally effective (data not shown). The 02-dependent Lys auxotrophy in IC and the aerobic Met/Lys auxotrophy observed for 2C and 4A were also rescued by 4 mM Mn2+. More than 8 mM medium Mn2+ inhibited the growth of both wild type and the mutant. This effect is discussed below.
Cellular Accumulation of Mn2+-To investigate if the rescue of Met auxotrophy correlated with accumulation of soluble Mn2+ in the mutants, the amount of Mn2+ in the mutants was determined. All of the mutants accumulated 14 (2C)to 59 (4A)-fold more Mn2+ than controls in which Met was added to support growth (column 3, Table I). In order to show quantitatively if this Mn2+ accumulation was typical for these mutants only or if it was characteristic of wild type as well, a comparison of the accumulation of MnZ+ in a wild-type yeast and in IC grown in air and in nitrogen was made. Wild type cells accumulated 35 pmol of Mn2+/10fi cells under aerobic conditions. The mutant IC accumulated 39 and 28 pmol/lOfi cells aerobically and anaerobically, respectively. Thus, Mn2+ accumulation was essentially independent of strain genotype or p 0 2 . A detailed analysis of the relationship between medium [Mn"] and cell growth with wild type and mutant ( IC) is presented below.
Manganese-associated 0; Scavenging Activity in Cell-free Extracts-The growth defect in these mutants correlates to the lack of SOD-1 (14): In addition, transformation of any of the mutants using a plasmid containing yeast wild type SOD-1 gene rescues the aerobic growth much as Mn2+ supplementation does (15,16). Therefore, the 0; scavenging activity in extracts prepared from Mn2+-grown mutant cell cultures was measured. Generally, there are a t least three components in yeast which can exhibit 0; scavenging activity: SOD-1, which is sensitive to 1 mM CN-; SOD-2, which is resistant to CN-a t 1 mM; and transition metals, whose scavenging activities are typically quenched by EDTA (33). A SOD assay without added CN-or EDTA yielded the total 0; scavenging activity. A CN--resistant 0; scavenging activity was found in all extracts made from cells pregrown in 4 mM Mn2+ (Table   I). This CN-resistance suggests that the elevated 0; scavenging activity was not due to SOD-1. Cyanide-resistant 0; scavenging activities ranged from 101 to 390 units/mg of protein in IC and 4A, respectively. This was approximately a 25-fold increase of 0; scavenging activity (cf. 2C) over the Met-grown control under the same assay conditions. The 0; scavenging activity appeared to correlate to the amount of Mn2+ in the cell extract, but was independent of the protein concentration (data not shown). Bovine SOD-1 (10 units/ml of culture) in the medium did not abolish the Met auxotrophy indicating the protection of cell growth by Mn2+ arose intracellularly.
The CN--resistant 0; scavenging activity in the cell ex-  I Accumulation of Mn2+ and 0; scavenging activity in the cell extracts of SOD-1mutants Yeast mutant cells (IC, 2C, and 4A) from stocks in SD medium were washed twice with water and then diluted into fresh SD medium (pH 4.5) to A = 0.1. The SD medium contained all required nutrients and either Met or 4 mM MnC12. Cell extracts were prepared from cultures in midlog phase ( A = 1.5). The 0; scavenging assay was performed as described under "Methods." EDTA (1 mM) was either added or omitted as indicated. Grade I11 xanthine oxidase (Sigma) was used in this experiment. The average 0; scavenging activity in the absence of EDTA and CN-for the Mn2+-grown samples was 4.2 f 1.1 units/nmol of Mn2+. The specific scavenging activity of aqueous MnS04 was 6.1 (Grade I xanthine oxidase) or 5.3 (Grade I11 xanthine oxidase). Note that addition of CN-in this assay generally enhances dye formation and thus the apparent specific activity of Mn2+ and SOD-2. ND. not determined. tracts was inhibited by EDTA (columns 5 and 7, Table I).
Quantitatively, this inhibition was most significant in the extracts made from cells pregrown in Mn2+-containing medium. This result also suggests that the elevated 0; scavenging activity in these cells was due to the accumulation of Mn2+.
Since there is no SOD-1 activity in any one of the three mutants, the 0; scavenging activity measured in the presence of EDTA was due to SOD-2 only. Generally, the presence of Mn2+ in the medium did not markedly alter the SOD-2 activity (columns 5 and 7, Table I), e.g. a maximum of 2 units/mg of protein increase was observed. This can be compared to the approximately 100-400-unit increase in 0; scavenging activity activity associated with Mn2+ (columns 4 and 6). These results suggest that the Mn2+-dependent relief of Met auxotrophy in these mutants did not result from the induction of SOD-2. Confirming this inference, extracts of wild type yeast grown a t 1, 4,8, and 16 mM medium Mn2+ all showed a SOD-2 activity of 7.9 f 0.2 units/mg of protein ( n = 4). This result suggests that addition of Mn2+ did not induce SOD-2 activity in wild type cells either.
The effect of Mn2+ supplementation on the activity of catalase and cytochrome c peroxidase in extracts of mutant strain IC was examined. Catalase activities were 0.5 and 0.3 pmol of H202 min" mg of protein", in extracts from I C grown in Met-and Mn2+-containing medium, respectively.
(Manganese(I1) added to H2O2 at a concentration equivalent to the [Mn2+] in cell extracts, did not catalyze the disappearance of H2O2.) There was no cytochrome c peroxidase activity detected in any of the cell extracts as determined by both a solution assay (34) and a postelectrophoresis gel-staining assay (35). We conclude that the relief of aerobic Met auxotrophy provided by Mn2+ in the mutants was not due to the induction of SOD-2, cytochrome c peroxidase, and/or catalase or to the removal of endogenous H2O2 by accumulated Mn2+.
To explore the possibility that this Mn2+-dependent 0; scavenging activity might be associated with some other cellular protein(s), cell extracts from I C grown in Mn2+ (100 units of 0; scavenging activity/mg of protein, Table I) were analyzed by 7% nondenaturing polyacrylamide gel electrophoresis, and a SOD assay was done in the gel in the absence of EDTA. As controls, equal amounts of protein from cell extracts of wild type and of mutant 2C transformed with a SOD-1-containing plasmid were also analyzed. The two control samples had specific SOD-1 activities of 144 and 22 units/mg of protein, respectively. Achromatic bands corresponding to SOD-1 were seen in the two controls, but none was observed in all cell samples from 1C (data not shown).
The Mn2+-dependent 0; scavenging activity was also heatresistant. An extract made from mutant cells ( I C ) pregrown in Mn2+-medium to early log phase was boiled for 15 min, then centrifuged. The supernatant was removed, and the precipitate was washed in 10 X the original extract volume. The initial extract, supernatant, and precipitate were analyzed as shown in Table 11. Boiling precipitated 83% of the total cellular protein in the initial extract; 14% of the Mn2+ in the extract co-precipitated with this protein. The 0 2 scavenging activity in the three fractions directly correlated to the amount of Mn2+ present; that is, the specific 0; scavenging activity in each of the three fractions in Table I1 was 4.5 * 0.5 units/nmol of Mn2+, essentially the same specific activity exhibited by MnS04. This result also indicates that the 0; scavenging activity in the cell extract was not functionally associated with specific protein(s). Cellular 0; Scavenging Activity and Cell Growth-The relationship between medium [Mn2+], total 0; scavenging activity, and cell growth is shown in Fig. 2. The 0; scavenging activity was measured in these experiments in the absence of EDTA and CN-. Therefore, the curves in Fig. 2, A and B (open circles) represent the total 0; scavenging activity. Since this activity is dominated by the Mn2+ in the cell extracts, the open circles in Fig. 2 also represent the pattern of the accumulation of Mn2+ as a function of medium [Mn2+]. Wild type cell growth and 0; scavenging activity were not affected by medium [Mn2+] up to 4 mM. At 4 mM Mn2+, 94 units/mg of protein of the total 0; scavenging activity (123 units/mg of protein) was contributed by SOD-1. Above 4 mM medium Mn2+, however, wild type growth was inhibited 16 mM Mn2+ completely inhibited growth. This inhibition of growth appeared to be mediated by 02, since in N P , the percent growth yield of wild type at 8 mM Mn2+ was 100 k 6% of control but that of an aerobically grown culture was only 72 f 4% of the control ( n = 2). Addition of catalase (up to 5.6 units in a 5ml culture containing 16 mM Mn2+ under air) had no effect TABLE I1 Heat treatment and Mn2+-associated 0; scavenging activity in mutant cell extracts Mutant (1C) was grown to AsGOnm = 1.1 in 200 p~ phosphate-SD medium (pH 4.5) containing 4 mM MnC12. Cells were washed with water, and cell extracts were prepared in water. A 5OO-pl aliquot of this extract was incubated at 100 "C for 15 min. The supernatant (Sup) was saved after centrifugation while the precipitate was washed with water (5 X 1 ml), then resuspended in 500 pl of water (Ppt). The last wash was saved, and the Mn2+ content was determined. The amount of Mn2+ in the last wash was 0.31 nmol or 1% of the total. Grade 111 xanthine oxidase (Sigma) was used in this assay. The specific 0; scavenging activity in the three samples normalized by the MnZ+ content was 4.5 k 0.5 units/nmol of Mnz+. on the inhibition of cell growth due to Mn2+ (data not shown). This result suggests the toxic effect caused by Mn2+ is not due to extracellularly produced HzOn. However, Mn2+-dependent inhibition of cell growth in wild type yeast correlated with an increase in total 0, scavenging activity. For example, at 8 mM medium Mn2+, the cell growth was inhibited by 25% and -750 units of 0; scavenging activity/mg of protein were detected in that cell extract. This was a 5-fold increase of scavenging activity compared to cells grown at 4 mM Mn2+. At 16 mM Mn", at which [Mn2+] the cell growth was completely inhibited, the scavenging activity increased by 15-fold (Fig. 2B). The activities of SOD-1 and SOD-2 made little contribution to this increase (legend, Fig. 2). On the other hand, the overnight growth of the mutant ( I C ) was minimal a t medium [Mn2+] up to 2 mM, while 4 mM Mn2+ supported the cell growth at control level (solid line, closed circles, Fig. 2A). At 2 mM Mn2+, the mutant cell accumulated 5.7 * 0.6 pmol of Mnnf/pg of protein ( n = 2, data not shown); the 0: scavenging activity in this cell extract was 55 units/mg of protein (open circles, inset, Fig. 2A). In the absence of added Mn2+, the 0; scavenging activity in mutant cell extract was -2 units/mg (Table I). Therefore, addition of 2 mM Mn2+ to the medium induced a 26-fold increase in 0; scavenging activity. Although this activity level did not support a wild type doubling time by the mutant, it was sufficient for the cells to reach a stationary phase which was 63% of control (dotted line, Fig. 2A). In contrast to wild type, inhibition of mutant growth was not observed at 8 mM Mn2+ (0; scavenging activity, 310 units/mg of protein). At 16 mM Mn2+, however, mutant cell growth was inhibited by more than 75%. This inhibition correlated with a marked increase in scavenging activity from 310 (8 mM) to more than 1000 units/mg of protein.
Dialyzable Mn2+-dependent 0; Scavenging Activity-The physiological significance of the dialyzable Mn2+ in cells was investigated. Mutant strain 1 C was grown to As, nm = 0.8 in a SD medium (pH 4.5) containing 4 mM Mn2+ or Met as control. Both cell samples were washed with distilled water and then dialyzed against 100 mM phosphate buffer. Cell extracts were made at the start of the dialysis ( t = 0) and after 2 h. At that time, the MnZ+ content of the Mn2+pregrown culture decreased from 6.9 to 2.5 pmol of Mn2+/pg of protein or a loss of 63% of the initial cell-associated Mn2+. This loss of Mn2+ was accompanied by a 79% decrease in the total EDTA-inhibitable 0; scavenging activity in the extract (from 42.7 to 8.8 units/mg of protein) suggesting that the dialyzable Mn2+ made the major contribution to the Mn2+dependent cellular 0; scavenging activity. Significantly, after removal of this fraction of intracellular Mn2+ by dialysis, when resuspended in growth medium, the cells needed at least 2 mM medium Mn2+ (or Met) to continue aerobic growth, e.g. the pattern of overnight growth response was the same as that shown in Fig. 2A. The accumulated Mn2+ in wild type cells was lost readily upon dialysis, also, consistent with the result that the accumulation of Mn2+ in yeast is independent of genotype (data not shown). This result suggests that the Mn2+ which provides a diffusible 0; scavenging activity is responsible for protecting against the On-induced growth defects in the mutants.
Investigation of the State of Mn2+ in Cell Extracts-The state of the Mn2+ in the cell extracts and the fractions after heat treatment (Table  11) was analyzed by electron spin resonance spectroscopy (ESR). The original cell extract (Table 11) which contained 56 p~ Mn2+ exhibited an ESR spectrum equivalent to the spectrum of the 9 p~ MnS04 standard made in water shown in Fig. 3a. After heat treatment of the extract, 23 p~ ESR-detectable Mn(II), or 50% of the total initial Mn2+, was found in the supernatant (data not shown). The precipitate contained 8 p~ total Mn2+ (Table  11), but this Mn2+ was ESR-silent. These results suggest that 16% of the total Mn2+ in the original cell extract was free Mn(I1) or bound Mn(I1) within a virtually symmetric ligand field shielded from solvent HzO. That is, most of the Mn2+ in the soluble cell fraction, about 84%, appeared to be associated with some cellular component(s) that eliminated the ESR spectrum typical of free Mn(I1) or was present in another ESR-silent redox state. Following heat treatment, a portion of this Mn2+ was released into the supernatant: essentially all Mn2+ present therein was ESR-detectable (cf. Table 11).
Phosphate is known to abolish the ESR signal typical of hexaaquo-Mn(I1) by shortening the electron spin relaxation time (20), thus the relationship between the amount of ESRdetectable Mn(I1) and phosphate in cell extracts was investigated. Mutant I C was grown in 30, 200, and 400 p~ medium phosphate, and the cell extracts were assayed for Mn2+ and phosphate and examined by ESR. The amount of ESRdetectable Mn(I1) in a cell extract containing 15 p~ total Mn2+ and 132 p~ phosphate (Fig. 36) was equivalent to the 9 p~ Mn(I1) standard (Fig. 3a). The extract examined in Fig.  3c (sample contained 622 pM phosphate) contained 3 times more [Mn2+] (48 p~) than did the sample in Fig. 3b; the ESR spectra had essentially the same amplitudes, however. Similarly, the ESR spectrum shown in Fig. 3d (sample contained 725 p~ phosphate) was one-half of that in Fig. 3a, although the sample contained 4-fold more MnZ+ (35 versus 9 pM). Thus, the amplitudes of the Mn(I1) ESR spectra were inversely proportional to the cellular [phosphate]. DISCUSSION This study demonstrates that addition of MnZ+ protects yeast SOD-1-mutants against 02-induced auxotrophy for Met and Lys and relieves their inability to grow in 100% oxygen in rich medium. This rescue correlates with the intracellular accumulation of Mn2+, and with elevated Mn*+-dependent 0; scavenging activity in cell extracts. This 0; scavenging activity was not due to SOD-1 because: ( a ) the activity was resistant to CN-while SOD-1 is not; ( b ) the activity was EDTA-inhibitable, but that of SOD-1 is not; ( c ) SOD gel assays performed in 7% nondenaturing gels detected no achromatic bands due to SOD-1 or other proteins with 0; scavenging activity. The cellular 0; scavenging activity was not due to SOD-2 because the variation of SOD-2 activity did not correlate to the pattern of the accumulation of intracellular Mn2+; also, no or little SOD-2 activity was detected in the gels, In addition, cell extracts prepared from wild type cells grown at 1-16 mM Mn2+ did not exhibit different SOD-2 activities. Therefore, addition of Mn2+ did not induce the activity of SOD-2 in yeast. Growth in Mn2+ did not affect cellular catalase activity. Peroxidatic activity could not be detected in extracts of cells grown f Mn2+, consistent with the fact that yeast grown on glucose as carbon source expresses little or no peroxidase activity (34) or cytochrome c peroxidase mRNA.' These results rule out the possibility that the rescue of aerobic growth was due to an increase in the activity of any other of the known anti-oxidant enzymes found in S. cereuisiae.
We suggest that the 0; scavenging activity in cell extracts and in the cell is due to the presence of loosely bound Mn(I1). The evidence for this conclusion is summarized as follows. First, the 0; scavenging activity was inhibited by EDTA. Second, the 0; scavenging activity was heat-resistant. This result together with the negative result of the SOD gel assay argue against the involvement of any specific protein(s) in this 0; scavenging activity. Third, the specific 0; scavenging activities of cell extracts from all mutants and wild type preloaded with Mn2+, and of the supernatants and precipitates of cell extracts after heat treatment were essentially the same as that of free Mn2+ (around 5 units/nmol of Mn2+). Fourth, the Mn2+ in 1C and wild type cells was readily lost by dialysis. Finally, Mn2+-depleted mutant cells, which had lost 79% of the accumulated 0; scavenging activity, could not grow in air unless Met or Mn2+ was resupplemented. The reason that less than 20% of the accumulated Mn2+ exhibited an observable Mn(I1) ESR spectrum is probably due to the presence of low molecular weight Mn2+-protein and/or orthophosphate complexes in which the Mn(I1) was readily accessible to HzO. Water in fast exchange at Mn(I1) significantly shortens the electron spin relaxation time resulting in ESR line broadening (36). These complexes and/or the Mn(I1) they contain could be readily dialyzable and are capable of scavenging 0; since, among other characteristics, they do appear to have an accessible coordination site for 0; (17,18,33). Nevertheless, the possibility that the intracellular Mn2+ exists in another, ESRsilent oxidation state cannot be ruled out.
Although Mn2+ may have a cellular function(s) which can ameliorate the toxicity of dioxygen other than supporting an 0; scavenging activity, our results indicate that it is this Mn2+-dependent activity which is essential to the aerobic growth of these mutants. First, quantitatively, the protection of aerobic growth by Mn2+ correlated with an apparent Mn2+dependent 0; scavenging activity. Second, the addition of M$+, which can substitute for Mn2+ in many enzymatic reactions, did not rescue cell growth in these mutants. This result suggests that the rescue due to Mn2+ is associated with its redox chemistry specifically. CuZc, which is also redoxactive and capable of dismutating 0; (37), did provide some supplementation a t 0.25 mM. Since cu2+ was toxic to the strains used at concentrations above 0.5 mM, adding more Cu2+ did not provide more protection. Finally, transformation of all the mutants carrying a scdl allele by a plasmid containing the yeast wild type SOD-1 gene rescued the 02-dependent growth defects. The transformation and the accumulation of intracellular Mn2+ both resulted in an increase of 0; scavenging activity in these mutants while having no effect on the activities of other anti-oxidant enzymes ( 15h2 The rescue of the Met auxotrophy in mutant ( I C ) by Mn2+ exhibited a "delayed-threshold" in that medium [Mn2+] below 2 mM supported no or little growth in 24 h, while cultures grown in 2 mM Mn2+ did approach stationary phase after 48 h ( Fig. 2A, dotted line). As indicated in Fig. 1, 2 mM Mn2+ in rich medium supported mutant growth under 100% 0 2 . The 2 mM threshold could be associated with the Mn2+-dependent 0; scavenging activity which was 55 units/mg of protein under these conditions. However, optimal growth did require 80 units of 0; scavenging activity/mg of protein. This activity is higher than the log phase SOD-1 activity in two of the pJW3-SOD1 transformants of these mutants (15), -15 units of 0; scavenging activity/mg of protein. This latter activity level is the lowest we have observed in an aerotolerant yeast except for the strains which carry physiological suppressors of the scdl allele. These revertants exhibit wild type 0,resistant phenotype despite the absence of SOD-1 activity (38).' An optimal amount of Mn2+ in SOD-1 mutants supported aerobic growth, but excess Mn2+ inhibited cell growth in both wild type and mutant. Excess Mn2+ in humans induces symptoms similar to Parkinson's disease (39) and in yeast is mutagenic (24). Although the mechanism of the toxicity in mammals is not known, the involvement of dioxygen radicals has been shown to be relevant in vitro (40). The Mn2+ toxicity observed here was more apparent in air than in N2 suggesting that oxygen is at least partially responsible for the toxicity of Mn2+ in yeast as well. Nevertheless, 16 mM Mn2+ did inhibit anaerobic cell growth to 60% of an anaerobic control; therefore, above 8 mM, Mn2+ toxicity was clearly less 02-dependent (data not shown).
As shown in Fig. 2, inhibition of growth was correlated with an increase of intracellular 0; scavenging activity. The overproduction of SOD-1 in mice has been shown to induce symptoms similar to Down's syndrome, a process probably mediated by the overproduction of H202 from the SOD reaction (23, 24). Overexpression of Fe2+-SOD in E. coli, which apparently leads to an increase of intracellular H202, is also toxic (22). Since Mn2+ exhibits an 0; scavenging activity in the cell extracts, our results indicate a correlation between elevated 0; scavenging activity and cytotoxicity in yeast. The effect of overexpression of yeast SOD-1 gene in yeast is under investigation.
Although the accumulation of Mn2+ in both SOD-1-yeast mutant and L. plantarum is not 0,-inducible, the latter accumulates Mn2+ more effectively. The Mn2+ in yeast was readily dialyzable while only 40% of the Mn2+ in L. plantarum was lost in 48 h (20). The difference appears related to the co-accumulation of polyphosphate in L. plantarum. The Mn2+-polyphosphate-protein complex in this bacterium is stable and nondialyzable (20). The effect of phosphate on the Mn2+-induced rescue of cell growth in the yeast mutants was investigated (data not shown). The mutant ( I C ) was grown in SD medium containing a defined amount of phosphate in the presence of either Mn2+ (4 mM) or Met. The growth yield of mutant (and wild type) leveled off at 200 PM medium phosphate independent of whether Met or Mn2+ was added as supplement. Although the accumulation of Mn2+ by the mutant in early log phase leveled off at 200 g M medium phosphate, the accumulation of phosphate did not. At 200 p~ medium phosphate, the mutant accumulated 0.3 fmol of phosphate/cell. We estimated from published data (41) that at the end of log phase, yeast accumulates polyphosphate in the range of a pmol/106 cells, or 0.1% of the orthophosphate detected in our experiments. In contrast, high molecular weight polyphosphate accounts for -80% of the total phosphate in L. plantarurn (20). The data suggest that accumulation of phosphate (or polyphosphate) by yeast is not a prerequisite for the O2 resistance provided by Mn2+.
In summary, we infer that intracellular, loosely bound Mn(I1) protects these yeast mutants by providing a cytosolic 0; scavenging activity normally contributed by SOD-1. Since these mutants contain wild type levels of SOD-2, the 0'dependent cellular damage sustained by these mutants is most reasonably cytosolic. We2 have found that mutant 1 C cannot reductively assimilate sulfate for Met biosynthesis, a process which takes place in the cytosol in yeast (42). We postulate that the 02-dependent Met auxotrophy in these mutants may be caused by the attack of 0; or one of its reduction products on one of the cellular elements involved in sulfate assimilation. The notion that the cellular locale of dismutation activity correlates to a specific 0' sensitivity is supported by the observation that while these SOD-1mutants grow on a nonfermentable carbon source (glycerol),' the SOD-2 deletion mutant described does not (13).
Although 0; has been suggested to be the causative agent of oxygen stress, arguments against that exist (43). In aqueous solution, 0; is a weak reductant, and the chemistry of 0; is dominated by its dismutation reaction (3). However, cellular 0; could be an intermediate in the production of another more toxic radical, i.e. the hydroxyl radical (HO.). Since Mn(I1) is capable of reducing HO. to produce Mn(OH)*+ (17), it is possible that part of the effect of Mn'+ is to remove HO . .
A study of the sensitivity of specific cellular element(s) in yeast to different 02-derived free radicals generated in vitro could help to evaluate these possibilities.