1111

J Biol Inorg Chem (2005) 10: 913–923 DOI 10.1007/s00775-005-0044-y

O R I GI N A L A R T IC L E

Raylene J. Sanchez Æ Chandra Srinivasan William H. Munroe Æ Matthew Alan Wallace Jacob Martins Æ Tina Y. Kao Æ Kate Le Edith Butler Gralla Æ Joan Selverstone Valentine

Exogenous manganous ion at millimolar levels rescues all known dioxygen-sensitive phenotypes of yeast lacking CuZnSOD Received: 2 August 2005 / Accepted: 2 October 2005 / Published online: 8 November 2005
SBIC 2005

Abstract Yeasts lacking copper-zinc superoxide dismutase (sod1D) exhibit a broad range of phenotypes, many of which can be rescued by growth in the presence of high levels of ionic manganese. We undertook a comprehensive survey of the effects of manganese on wildtype and sod1D yeasts and found that 5 mM Mn2+ rescued all known growth-related phenotypes, such as slow growth in air, temperature sensitivity, specific amino acid auxotrophies, no growth in high oxygen, poor growth in nonfermentable carbon sources, and decreased stationary-phase survival. Iron-related phenotypes-elevated electron paramagnetic resonance detectable (‘‘free’’) iron, decreased aconitase activity, and fragmenting vacuoles-as well as zinc sensitivity were also rescued. The activity of manganese superoxide dismutase remained constant or was reduced when the yeasts were grown in the presence of MnCl2, indicating that induction of this alternative superoxide dismutase is not the explanation. In contrast to MnCl2 treatment, addition of two manganese-containing superoxide dismutase mimetic compounds to the growth medium did not provide any rescue of sod1D yeast growth but rather had an sod1D-selective inhibitory effect at micromolar concentrations. Mechanisms by which ionic manganese can effect this rescue, while the mimetic compounds do not, are discussed.

R. J. Sanchez Æ W. H. Munroe Æ M. A. Wallace Æ J. Martins T. Y. Kao Æ E. B. Gralla (&) Æ J. S. Valentine Department of Chemistry and Biochemistry, UCLA, 607 Charles E. Young Drive, East, Los Angeles, CA 90095-1569, USA E-mail: [email protected] Tel.: +1-310-8251946 Fax: +1-310-2069880 C. Srinivasan Æ K. Le Department of Chemistry and Biochemistry, California State University, Fullerton, CA 92834-9480, USA

Keywords Manganese Æ Zinc Æ Copper-zinc superoxide dismutase Æ Oxidative stress Æ Yeast Abbreviations EPR: Electron paramagnetic resonance Æ ICP-AES: Inductively coupled plasma atomic emission spectrometer Æ IMS: Intermembrane space of mitochondria Æ Leu1p: Isopropylmalate dehydratase Æ ROS: Reactive oxygen species Æ SDC: Synthetic complete medium with 2% glucose Æ SD-Lys: Defined medium lacking lysine Æ SD-Met: Defined medium lacking methionine Æ SOD: Superoxide dismutase Æ sod1D: Yeast lacking the CuZnSOD gene Æ sod2D: Yeast lacking the MnSOD gene Æ SOR: Superoxide reductase Æ Tris: Tris(hydroxymethyl)aminomethane

Introduction Non superoxide dismutase (SOD) manganese plays a major role in the antioxidant defense mechanisms of several prokaryotic organisms [1]. The phenomenon was first noted when bacterial species that lack any detectable SOD enzymes were found to contain high (millimolar) levels of dialyzable manganese; the aerobic organism Lactobacillus plantarum, for example, lacks any known form of SOD enzyme and contains 30 mM manganese, which, apparently, functionally replaces SOD enzymes. In other bacterial systems, high levels of non-SOD manganese have been shown to work in conjunction with antioxidant enzymes such as SOD and catalase as a major component of their defense systems against reactive oxygen species [1, 2]. High levels of ionic manganese added to the growth medium also provide a substantial degree of rescue of the dioxygen-sensitive phenotypes of SOD knockout Escherichia coli [3]. The molecular basis for the antioxidant action of non-SOD manganese in prokaryotes is unknown. Early reports suggested that non-SOD ionic manganese

914

possessed SOD activity, on the basis of the xanthine/ xanthine oxidase/cytochrome c assay [4]. However, further studies determined that ionic manganese interferes with this indirect assay by promoting the oxidation of reduced cytochrome c [3]. Furthermore, direct assays using pulse radiolysis or stopped flow techniques have established that ionic manganese at physiological pH has little, if any, SOD activity [5-7]. Despite the fact that ionic manganese lacks significant superoxide dismutase activity at physiological pH, there is considerable evidence that extracts from manganese-treated cells nevertheless have a high degree of superoxide scavenging ability [1, 3, 4, 8]. Manganous ion is well known to react with superoxide at fast rates in vitro to form a relatively long lived transient complex [MnO2]+ [5, 7, 9, 10]. If this transient is rapidly reduced in vivo by an unknown cellular reductant, non-SOD manganese may be acting as a superoxide reductase (SOR), functioning in a manner analogous to the nonheme iron-containing FeSORs [11]. The possibility that non-SOD manganese functions as an antioxidant in vivo has been less explored in eukaryotes than in prokaryotes. Most of the evidence that exists in support of such a function comes from studies of Saccharomyces cerevisiae. The budding yeast S. cerevisiae, like most eukaryotes, contains two SODs: CuZnSOD (SOD1) in the cytosol, the intermembrane space of mitochondria (IMS), and other locations, and MnSOD (SOD2) in the mitochondrial matrix. Yeast cells lacking CuZnSOD (sod1D) have severe aerobic growth defects, die quickly in the stationary phase, and exhibit lysine, leucine, and methionine auxotrophies under normal culture conditions [12, 13]. They also exhibit altered iron and zinc metabolism, when grown aerobically, including elevated electron paramagnetic resonance (EPR) detectable iron levels at g=4.3 and zinc sensitivity [14-16]. The most common genetic suppressors of CuZnSOD knockout (sod1D) yeast, i.e., pmr1-1 and bsd2-1, lead to dramatic increases in intracellular manganese levels [17, 18]. Additionally, the inability of sod1D yeast to survive on medium lacking lysine or methionine as well as its poor growth in 100% dioxygen have previously been shown to be rescued by the addition of 5 mM MnCl2 to the growth medium [8, 17]. Other metal ions, in particular copper and iron, have also been reported to rescue certain phenotypes of sod1D yeast when added to the growth medium, but the rescue in each case was incomplete. Copper, working through a mechanism that requires the presence of metallothionein, restored the ability of sod1D yeast to grow on the nonfermentable carbon source, lactate, but did not restore lysine prototrophy or the ability to grow in an atmosphere of 100% dioxygen [19]. Similarly, iron supplementation restored growth on glycerol, but not methionine prototrophy [16]. In addition to simple metal ions, several classes of manganese coordination complexes have been developed as functional SOD mimics and tested for biological efficacy in different model systems. The two of particular interest to us in this study are

(1) derivatives of Mn(III) salen complexes Euk-8, Euk134, and Euk-189 developed by Eukarion [20] and (2) derivatives of Mn(III) cyclam complexes M40403 and M40404 developed by Metaphore Pharmaceuticals [21]. The Mn(III) salen (Euk) complexes have been reported to increase wild-type Caenorhabditis elegans lifespan [22, 23] and were beneficial in other models of oxidative stress, such as ischemia-reperfusion injury in a rat stroke model [24], reactive oxygen species (ROS)-induced apoptosis [25], and improved survival of sod2 null mice [26]. The Mn(III) cyclam derivative M40403 has been found to be efficacious in reducing ischemia-reperfusion injury [21, 27-29], beneficial in inflammation models in rats [30, 31], and protective of cochlear hair cells [32, 33]. The present study was undertaken to determine the nature and degree of the manganese rescue of sod1D yeast phenotypes and to compare the effect of ionic manganese with that of some manganese-containing SOD mimics that had earlier been shown to have significant degrees of biological activity in other experimental systems. On the basis of earlier studies of L. plantarum and other bacteria entirely lacking in SOD enzymes, we hypothesized that the degree of rescue provided by ionic manganese might be significantly more than that provided by copper and iron if it achieved the required concentration and intracellular location and had the correct ligands in vivo. If so, we expected it to at least partially rescue several more of the well-characterized defects of sod1D yeast than those that had already been discovered [8, 17], and that these data might provide us with valuable clues as to its mechanism of protective action. We expected the manganese-containing SOD mimetic compounds to provide a significant degree of rescue to the sod1D yeast as well. Our results confirmed that ionic manganese at high levels is beneficial to sod1D yeast [8, 17]; however, the extent of the rescue surprised us. We found that addition of exogenous manganous ion to the growth medium at millimolar levels fully rescued all known growth-related phenotypes of yeast lacking CuZnSOD. In other words, we found no significant differences between the phenotypes of sod1D yeast grown in medium containing 5 mM manganous ions and wild-type yeast. In addition, we were surprised to find that the manganese-containing SOD mimetic compounds had no beneficial effect at all and instead retarded growth of the sod1D yeast. These results suggest the possibility that non-SOD manganese may play an antioxidant role in some eukaryotic organisms.

Materials and methods Reagents, media, and cell growth The S. cerevisiae yeast strain EG103 (MATa leu2-3 112 his3D1 trp-289a ura3-52) and the isogenic sod1D strain derived from it, EG118 (MATa leu2-3, 112 his3D1 trp289a ura3-52 sod1D::URA3) [34] were used in this

915

study. MnCl2Æ4H2O and ZnCl2 were purchased from Fisher. Unless otherwise noted, the yeast cells were cultured in synthetic complete medium with 2% glucose (SDC) composed as described in Ref. [35], with the exceptions that the supplements Leu, His, Trp, Met, Ura, and Ade were increased fourfold and the pH was adjusted to 4.0 or 6.0 as indicated. Overnight startup cultures were grown from single colonies for all experiments. Yeast strains were inoculated at a starting OD600 of 0.05 or 0.1 (approximately 0.5-1·106 cells/ mL) in Erlenmeyer flasks with a liquid-to-flask volume ratio of 1:5. Cultures were incubated at 30C, with shaking at 220 rpm in air unless otherwise noted. For indicated experiments, cultures were inoculated and grown as described above except under an atmosphere of 100% dioxygen. Enzyme activity assays For isopropylmalate dehydratase (Leu1p) and aconitase assays, 5·108 cells resuspended in 0.5 mL of standard lysis buffer were subjected to glass bead lysis performed under nitrogen in a septum-sealed glass test tube. Two-hundred microliters of 0.5-mm glass beads with six cycles of 30 s vortexing followed by 30 s on ice was used. Cell debris was removed by centrifugation at 4,000 rpm (3,080g) for 5 min, and the supernatant was assayed for protein concentration by Bradford assay (BioRad). Leu1p activity was determined spectrophotometrically by monitoring the disappearance of citraconate (Aldrich) at 235 nm as described in Ref. [36]. Briefly, a 0.5-mL assay mixture containing 4 mM potassium phosphate, pH 7.0, 0.4 mM citraconate, and 100-300 lg of lysate protein was assayed for 3 min in a 2-mm path length cuvette [37]. Aconitase activity was determined spectrophotometrically as described by Gardner et al. [38]. The assay mixture contained 50 mM tris(hydroxymethyl)aminomethane (Tris)-HCl, pH 7.5, 5 mM sodium citrate, 0.6 mM MnCl2, 0.2 mM NADP+, 2 units of NADP+ isocitrate dehydrogenase, and 50-100 lg of protein. The appearance of NADPH was monitored at 340 nm for 5 min. For SOD activity assays, samples were prepared as just described, but not under nitrogen. SOD activity measurements were made using activity gels, as described in Ref. [39].

Visualization of vacuoles The fluorescent dye FM-4-64 (Molecular Probes), which specifically stains vacuolar membranes, was used. After the cells had reached the log phase (OD600<1.5) in SDC pH 4.0 with or without 5 mM Mn2+ added, approximately 2.5·107 cells were harvested and incubated aerobically with 50 lM FM-4-64 in SDC pH 4.0 (without any added manganese) for 15 min. The cells were then resuspended in SDC pH 4.0 with or without added manganese and grown for at least 1 h prior to visualization using a Zeiss Axioplan 2 microscope. Images were recorded using a Zeiss digital camera. EPR sample preparation EPR samples were prepared as described in Ref. [40]. After 72 h of growth, 10-20 mL of culture was spun down at 4,000 rpm for 10 min at 4C. The cell pellet (approximately 109 cells) was washed once with 10 mL of cold 20 mM Tris-Cl, pH 7.4, resuspended in 200 lL of 20 mM Tris-Cl, pH 7.4, containing 10% glycerol, and transferred into an EPR tube. The sample was frozen in dry ice and stored at -70C until EPR measurements were performed. Whole-cell low-temperature Fe(III) EPR Spectra were recorded with a Bruker X-band spectrometer. Samples were maintained below -178C during the recording of the spectra either using a finger Dewar filled with liquid nitrogen or a variable-temperature gascooled cavity. The parameters for low-temperature Fe(III) EPR using the finger Dewar were as reported in Srinivasan et al. [14]. The parameters for EPR using the variable-temperature cavity were as follows: center field, 1,560 G; sweep width, 500 G; frequency 9.45 GHz; microwave power, 31 mW; attenuation, 10 dB; modulation amplitude, 20 G; modulation frequency, 100 kHz; receiver gain, 2·105; sweep time, 20.97 s; time constant, 81.92 ms; resolution, 2,048 points; number of scans, 16. EPR data processing was carried out using the Bruker WinEPR program. Quantitation and calculation of EPR-detectable iron levels were carried out as described previously [40].

Stationary-phase survival

Measurement of total copper, zinc, iron, and manganese

Yeast cells were inoculated and incubated as described before for 72 h. In order to monitor cell viability in the stationary phase, a serial dilution series was performed and samples were spotted on YPD plates (1% yeast extract, 2% peptone, 2% dextrose, and 2% agar) and incubated at 30C in a low dioxygen environment (roughly 5% dioxygen, Campy Bags, Becton Dickinson Microbiology Systems) for 3 days.

Triplicate samples, each containing approximately 3.5·108 cells (35 OD600 units) per sample, were collected from 50-mL cultures (grown for either 24 or 72 h) by centrifugation (5 min at 4,000 rpm). Samples were washed twice in 4 mL of 10 mM EDTA, pH 8, and twice in 1 mL of high-purity water (VWR Scientific). The cell pellet was resuspended in 1 mL of 20% ultrapure nitric acid (Fisher) and heated at 90-95C for at least 18 h. After

916

complete digestion, 0.9 mL of the sample was diluted to 4.0 mL with high-purity water and analyzed using a Thermo-Jarrel Ash Iris 1000 inductively coupled plasma atomic emission spectrometer (ICP-AES). The intracellular yeast volume of 70 lm3 [41] and the number of cells used for sample preparation were used to quantitate metal levels inside yeast cells for values reported as concentrations. Otherwise, metal levels are reported as parts per billion per OD600 (approximately 107 cells). Manganese-containing SOD-mimetic compounds The manganese salen derivatives, Euk-8, Euk- 134, and Euk-189, were provided by Susan R. Doctrow (Eukarion, Bedford, MA, USA). The SOD-active and SODinactive manganese cyclam derivatives, M40403 and M40404, respectively, were provided by Dennis P. Riley (Metaphore Pharmaceuticals, Fort Lee, NJ, USA). See Structure 1 for structures. Euk-8, Euk-134, and Euk-189 were dissolved in water at 10 mM and M40403 and M40404 at 1.0 mM. Both were filter-sterilized prior to addition to the growth medium at the indicated concentrations.

Results We tested a wide variety of phenotypes that are evidenced in sod1D yeast, and 5 mM ionic manganese fully reversed the phenotype in every case. Five millimolar Mn2+ was chosen because it was the concentration at which the maximum rescue of the sod1D strain was ob-

a N

N Mn

O

Cl

O

R

R Compound EUK-8 EUK-134 EUK-189

R group H OCH3 OCH2CH3

b N

N Cl

N

N

Cl

N

N Mn

Mn N

Cl

N

N

Cl

served (data not shown). Lower concentrations were not as effective, and higher ones were toxic. In fact, 5 mM is borderline toxic; it can be noted in most of the figures that the wild-type yeast was slightly impaired by high manganese treatment. Growth in air or 100% O2 It was previously reported that the addition of millimolar concentrations of Mn2+ to the growth medium rescued the aerobic methionine auxotrophy and growth on rich plates in 100% dioxygen of sod1D yeast [8]. Lapinskas et al. [17] demonstrated that Mn2+ rescued the lysine auxotrophy of the sod1Dsod2D double mutant lacking both CuZnSOD and mitochondrial MnSOD. It was not clear whether this rescue would extend to all other observed phenotypes of the sod1D mutant. Initially, we confirmed that the reported rescue with Mn2+ could be reproduced in the strain most commonly used in our laboratory (EG118). Using defined medium lacking either lysine or methionine (SD-Lys or SD-Met, respectively), we confirmed that the lysine and methionine auxotrophies, characteristic of sod1D yeast grown in liquid medium in air, were rescued similarly regardless of whether the initial medium pH was 4.0 or 6.0 and that both MnCl2 and MnSO4 salts were equally effective (data not shown). To study whether the addition of exogenous Mn2+ would improve the growth of the sod1D strain, EG118, we examined growth under normal atmospheric and 100% dioxygen conditions in glucose and in the nonfermentable carbon source glycerol. Wild-type growth was unaffected by the increase in dioxygen concentration, whereas sod1D strain growth was severely restricted in the presence of high dioxygen in either carbon source. In air, sod1D yeast grows to a lower saturation density than wild-type yeast. Addition of 5 mM MnCl2 effected a modest improvement in growth (20%) for the sod1D strain and a similarly sized decrease in growth for the wild-type strain, such that their final densities were the same (Fig. 1a). The effect of Mn2+ on sod1D yeast grown in 100% O2 was more striking. The addition of 5 mM MnCl2 to SDC medium afforded full protection against the increase in O2 assault that is intolerable to the mutant in unsupplemented medium (Fig. 1b). Again, 5 mM Mn2+ caused a small decrease in growth in the wild-type strain. Both wild-type and sod1D yeast grew to a lower 24-h optical density when glycerol was given as the carbon source. When 5 mM MnCl2 was added to SGC medium, there was no significant effect on wildtype yeast, while sod1D yeast was again rescued to wildtype levels (Fig. 1c).

N

Temperature sensitivity M40403

Structure 1

M40404

Wild-type yeast grew to a 30% lower density when grown at 37C for 24 h than it did when grown at 30C

917

Growth (OD at 600 nm)

a 8 Air 6 4

addition of exogenous Mn2+ provided a dramatic benefit to the sod1D yeast. The growth at 37C with 5 mM Mn2+ was 11 times that of sod1D yeast incubated without added manganese (Fig. 2a), and this was very close to the growth seen in wild-type yeast with added manganese.

2

Stationary-phase survival 0 wild type

Growth (OD at 600 nm)

b

sod1∆

8 100% O2 6 4 2 0 wild type

sod1∆

A previous report showed that another characteristic of sod1D yeast is that they do not survive as well as wildtype yeast in the stationary phase [34]. Wild-type and sod1D yeasts were grown for 72 h (to the stationary phase) with shaking at 220 rpm in air in SDC medium at 30C. Five microliters of a tenfold serial dilution series of these yeast were spotted onto rich medium plates and grown in a low dioxygen environment (Fig. 2b). As was previously observed [42], wild-type yeast survival after this incubation far exceeded the survival of sod1D yeast (Fig. 2b). Again, we saw a significant improvement when the mutant yeast was incubated with 5 mM MnCl2 in liquid culture and a slightly diminished growth of the wild-type yeast, such that wild-type and sod1D yeast

Growth (OD at 600 nm)

c 0.8

SGC

0.6 0.4 0.2 0 wild type

sod1∆

Fig. 1 Mn2+ enhanced growth of yeast lacking the copper-zinc superoxide dismutase gene (sod1D) in air and in a 100% O2 atmosphere. Wild-type EG103 and isogenic sod1D EG118 strains were seeded at OD600 of 0.05 in minimal medium in the absence (white bars) or presence (black bars) of 5 mM MnCl2. a Total growth following 24 h of shaking in synthetic complete medium with 2% glucose (SDC) at 220 rpm in air at 30C was measured turbidimetrically at 600 nm (OD600). b Total growth (OD600) measured after incubation under identical growth conditions in a 100% O2 atmosphere. c Total growth (OD600) measured after incubation in the nonfermentable carbon source, SGC, and in a 100% O2 atmosphere. The values represent the average of two or more independent samples grown on the same day, with error bars indicating the standard deviation between samples. The experiment was repeated at least three times and the data are representative of these trials

(data not shown). By contrast, sod1D yeast exhibited a much more dramatic decrease in density after growth at 37C, reaching an OD600 only 4% of that observed after the normal 30C incubation (data not shown). Ionic manganese (5 mM MnCl2) was added to the cultures to test for rescue of this phenotype. In the case of wild-type yeast, the additional Mn2+ further reduced the growth at 37C (Fig. 2a). However, the

Fig. 2 Added Mn2+ affected growth of wild-type and sod1D yeast incubated at elevated temperatures and extended survival of sod1D in the stationary phase. a The indicated strains were grown in SDC medium in the absence (white bars) or presence (black bars) of 5 mM MnCl2. Total growth following a 24-h growth period, with shaking at 220 rpm in air at 37C, was measured turbidimetrically at 600 nm (OD600). The values represent the average of three independent samples grown on the same day, with error bars indicating the standard deviation between samples. The experiment was repeated at least three times and the data are representative of these trials. b Wild-type EG103 and isogenic sod1D (EG118) strains were incubated in SDC in the absence or presence of 5 mM MnCl2 for 72 h. Cells were viability-tested by spotting 5·104, 5·103, 5·102, and 5·101 cells onto YPD plates (1% yeast extract, 2% peptone, 2% dextrose, and 2% agar) and incubated at 30C in a low dioxygen environment. A characteristic result is shown. The experiment was repeated at least three times using three independent colonies and the same trend was observed

918

Vacuolar fragmentation When yeasts are grown under normal conditions, the majority of wild-type cells have one or two large vacuoles per cell. This type of vacuolar morphology has been termed type A, whereas a pattern of three or more smaller vacuoles has been termed type B vacuolar morphology [44]. Corson et al. [45] reported that yeast lacking CuZnSOD exhibited increased vacuole fragmentation compared with wild-type yeast when grown aerobically. In our wild-type strain grown in SDC at pH 4.0 and treated with FM-4-64 to visualize the vacuolar membranes, less than 25% of the wild-type population contained type B vacuoles. In contrast, sod1D yeast grown in the same conditions showed apparent vacuolar fragmentation-more than two thirds of the cells had three or more smaller vacuoles, i.e., exhibited type B morphology (Fig. 3c). When sod1D yeast was grown in the presence of 5 mM Mn2+, the fraction of sod1D cells exhibiting type B morphology decreased to 36% (Fig. 3c). These values are similar to those observed for the wild-type control and suggest that the addition of exogenous manganese ions also rescues this phenotype. In concert with previous findings suggesting that high levels of manganese are somewhat detrimental to wildtype yeast, 5 mM manganese added to the medium resulted in a slight increase in vacuolar fragmentation in wild-type yeast (data not shown). EPR-detectable (free) iron and total iron levels The amount of ferric iron detectable by EPR at g=4.3 (FeIII g=4.3) is roughly 5 times greater in EG118 yeast (lacking CuZnSOD) than in its wild-type parent strain, EG103 [14]. We measured the amount of FeIII g=4.3 in these

a

Leu1p Activity (∆A/min/mg protein)

The best known targets of in vivo superoxide toxicity are enzymes that contain a [4Fe-4S] cluster with a labile, solvent-exposed iron atom specifically susceptible to superoxide inactivation [43]. Yeasts contain at least three such enzymes-aconitase, Leu1p, and homoaconitase-all of which exhibit decreased activity in sod1D cells [37]. Leu1p is cytoplasmic, like CuZnSOD, and aconitase is in the mitochondrial matrix. Supplementation of the growth medium with 5 mM MnCl2 resulted in a significant increase in the activity of Leu1p in sod1D cells, while the wild-type cells show slightly diminished activity (Fig. 3a). A similar pattern was observed for aconitase (Fig. 3b) though it was not as dramatic.

0.12 0.01 0.08 0.06 0.04 0.02

b

0

wild type

sod1∆

wild type

sod1∆

50

Aco1p Activity (mU/mg protein)

[4Fe-4S] cluster enzyme activity

two strains after 72 h of growth in the presence and absence of 5 mM MnCl2. Growth with Mn2+ resulted in a more than 60% decrease in EPR-detectable iron accumulation in the sod1D mutant relative to growth without added manganese. Conversely, wild-type yeast accumulated approximately 50% more FeIII g=4.3 when grown in the presence of 5 mM Mn2+ (Fig. 4a). Thus, after growth with manganese, we observed only a twofold elevation of FeIII g=4.3 in sod1D yeast over wild-type yeast. We also measured the total iron accumulated by the two strains after growth under identical conditions. In contrast to the EPR-detectable iron, the total iron did not change dramatically with the addition of manganese. Without 5 mM Mn2+, the amounts of total iron amassed by wild-type and sod1D yeasts were statistically indistinguishable. With 5 mM Mn2+, both strains

40 30 20 10 0

c Vacuolar Morphology (percent of cells)

survived to the same degree. Thus, we concluded that the defect in stationary-phase survival characteristic of sod1D yeast at day 3 was also rescued by exogenous manganese.

100 80

type A

60

type B

40 20 0

wild type

sod1∆

sod1∆+Mn

Fig. 3 Growth in the presence of added Mn2+ results in sod1D yeast with Leu1p activity, Aco1p activity, and vacuole morphology more similar to those of wild-type cells. a, b The indicated strains were grown in SDC medium in the absence (white bars) or presence (black bars) of 5 mM MnCl2. a Leu1p activity. b Aco1p activity. The average of data from four or more colonies is reported. c Vacuolar morphology of wild-type and sod1D yeasts was visualized after growth with or without 5 mM MnCl2 followed by growth with 50 lM FM-4-64 as described in ‘‘Materials and methods.’’ Type A vacuolar morphology refers to cells with one to two vacuoles, which is considered normal, whereas a pattern with three or more smaller vacuoles has been termed type B [44]. The average of at least 400 cells from six colonies (three separate fields per colony) is shown

919

100

b

300

Total Iron (µM)

EPR-Detectable Iron (µM)

a 80 60 40 20 0

wild type

sod1∆

in the presence and absence of 5 mM MnCl2. While the growth of wild-type yeast after 24 h was unaffected by the presence of 2 mM ZnCl2 in the growth medium, the sod1D yeast mutant exhibited a dramatic growth defect [15]. The addition of 5 mM MnCl2 to growth medium supplemented with 2 mM ZnCl2 had a small growth inhibitory effect on wild-type yeast and a dramatic stimulatory effect on sod1D yeast (90% growth enhancement). The extra Mn2+ resulted in complete rescue of the zinc sensitivity of sod1D yeast, as the mutant yeast reached the same OD600 as the wild-type yeast (Fig. 5). Accumulation of zinc

200

100

0

wild type

sod1∆

Fig. 4 Growth in the presence of exogenous Mn2+ resulted in decreased levels of electron paramagnetic resonance (EPR) detectable iron in sod1D yeast while total iron levels remained relatively unchanged. The indicated strains were seeded at OD600 of 0.05 in minimal medium in the absence (white bars) or presence (black bars) of 5 mM MnCl2. Iron was measured after a 72-h incubation period. a Low-temperature iron EPR was employed to measure the levels of FeIII g=4.3 accumulation. b Total iron levels were measured using a inductively coupled plasma atomic emission spectrometer (ICP-AES) and are reported as the concentration of iron. The values represent the average of two or more independent samples, with error bars indicating the standard deviation between samples

We considered it possible that Mn2+ rescue results from competition for binding and transport, resulting in a reduction of the amount of zinc entering the cell. To address this question, the zinc content was analyzed in wild-type and sod1D yeasts grown for 24 h in SDC with or without 5 mM MnCl2, in the presence or absence of 1 mM ZnCl2. In these experiments, 1 mM ZnCl2 was used because the sod1D cells, although growth-inhibited, still grew enough at this zinc concentration to provide sufficient sample for metal analysis. Without added zinc, sod1D yeast accumulated almost twice as much zinc as wild-type yeast. However, in the presence of exogenous manganese, the amount of zinc in the two strains was the same (Fig. 6a). In SDC with 1 mM ZnCl2, sod1D yeast amassed approximately 25% more zinc. Supplementation of SDC with 5 mM MnCl2 in addition to 1 mM Zn2+ resulted in a decrease in zinc accumulation in both strains, again restoring wild-type zinc levels in the sod1D mutant (Fig. 6b).

accumulated on average about 10% more total iron, but the difference was not statistically significant (Fig. 4b). Effect of Zn2+ on manganese accumulation

We determined the intracellular concentration of Mn2+ in wild-type and sod1D yeasts in order to determine whether a specific intracellular level of this metal was required for the observed rescue. We used an ICP-AES to measure the manganese content in cells after growth under our standard conditions (SDC pH 4.0, shaking in air at 30C for 24 h) with or without the addition of 5 mM MnCl2. When no extra manganese was added, the sod1D yeast accumulated similar levels of manganese to wild-type cells. When yeast cells were grown on minimal medium supplemented with 5 mM MnCl2, both wildtype and sod1D yeasts acquired high levels of manganese (Fig. 7). Hypersensitivity to zinc Wei et al. [15] demonstrated that 2 mM Zn2+ inhibited the growth of sod1D yeast. This experiment was repeated

To test whether zinc toxicity is related to an effect of zinc on intracellular manganese levels, the total manganese Growth (OD at 600 nm)

Accumulation of Mn2+

10 SDC

8

SDC+5mM Mn

6 4 2 0

+ Zn - Zn wild type

+ Zn - Zn sod1D

Fig. 5 Mn2+ rescued the hypersensitivity of the sod1D yeast to millimolar Zn2+ concentrations. The indicated strains were seeded at OD600 of 0.05 in minimal medium in the absence (white bars) or presence (black bars) of 5 mM MnCl2 and with or without 2 mM ZnCl2 added as indicated. Total growth following a 24-h incubation period was measured turbidimetrically at 600 nm. This is an average of six colonies grown on two different days, with error bars indicating the standard deviation between colonies

920

7 6 5 4 3 2 1 0

SDC SDC +5mM Mn

Growth in the presence of manganese-containing SOD mimics sod1D

b 200

SDC +1mM Zn

160

SDC +1mM Zn +5mM Mn

120 80 40 0

wild type

sod1D

Fig. 6 Effect of manganese on zinc accumulation in wild-type and sod1D yeasts. a Wild-type EG103 and isogenic sod1D yeasts were grown for 24 h in SDC (white bars) or SDC plus 5 mM MnCl2 (black bars). b The same strains were grown for 24 h in SDC with 1 mM ZnCl2 (gray bars) or in SDC with both 1 mM ZnCl2 and 5 mM MnCl2 (white bars). In both experiments, cells were collected, washed, and total zinc was measured with an ICP-AES and is reported as the amount of zinc per optical density (approximately 107 cells). The values are the average of at least 12 samples, harvested and analyzed on two different days. Error bars represent the standard deviation

content was measured in wild-type and sod1D strains after growth in SDC supplemented with 5 mM MnCl2, 1 mM ZnCl2, or both. Growth in SDC plus 1 mM ZnCl2 resulted in a reduction in manganese in wild-type and sod1D yeasts by about 33% when compared with yeast grown in SDC without added metals (Fig. 7a). This influence of zinc on the total manganese content was similar for wild-type and sod1D yeasts grown in SDC plus 5 mM MnCl2; the strains accumulated about 20% less manganese when zinc was present in addition to the excess manganese (Fig. 7b). MnSOD activity One possible explanation for our results could be that excess manganese functions by activating the endogenous MnSOD, which is still present in the mitochondrial matrix of the sod1D strains. Two pieces of evidence argue against this possibility. First, manganese rescue is equally effective in an sod1Dsod2D double knockout strain: MnCl2 fully rescues the lysine auxotrophy [17], hypersensitivity to elevated temperature, and hypersensitivity to ZnCl2 of the sod1Dsod2D double knockout strain (data not shown). Second, we performed gel-based SOD activity assays on wild-type and sod1D strains (Fig. 8). Although the sod1D strain appears to have somewhat more MnSOD activity than the SOD1+ strain, the addition of manganese to the growth medium did not

A number of manganese-containing coordination complexes, studied in cell culture and whole animals, were shown to be effective in preventing damage attributed to oxidative stress [20, 21, 46] (see ‘‘Introduction’’). On this basis, it has been assumed that these compounds are acting as functional SOD mimics. To our knowledge, the effect of manganese-containing SOD mimics has not been studied previously in a yeast model system. We therefore examined several manganese-containing SOD mimics to determine their effectiveness in our yeast model system. The Mn(III) salen derivatives from Eukarion are a unique class of antioxidants that exhibit both SOD activity and catalase activity. We tested Euk-8, Euk-134, and Euk-189, which differ in the nature of the substituents on the salen moiety. They were reported to have the same level of SOD activity but differed in their levels of catalase activity and lipophilicity. The catalase activity of the prototypical Euk-8 (148 lM O2/min) was reported to be lower than that of Euk-134 or Euk-189 (243 and 180 lM O2/min, respectively), and Euk-189 a 0.4

Total Mn (ppb/OD)

wild type

Total Zn (ppb/OD)

result in a further augmentation. Indeed, if anything, the total MnSOD activity decreased in the manganesetreated samples.

SDC 0.3

SDC +1mM Zn

0.2 0.1 0

wild type

b Total Mn (ppb/OD)

Total Zn (ppb/OD)

a

sod1D

20

SDC +5mM Mn

16

SDC +1mM Zn +5mM Mn

12 8 4 0

wild type

sod1D

Fig. 7 Effect of zinc on manganese accumulation in wild-type and sod1D yeasts. a Wild-type EG103 and isogenic sod1D yeasts were grown for 24 h in SDC (white bars) or SDC plus 1 mM ZnCl2 (gray bars). b The same strains were grown for 24 h in SDC with 5 mM MnCl2 (black bars) or in SDC with both 1 mM ZnCl2 and 5 mM MnCl2 (white bars). In both experiments, cells were collected, washed, and the total manganese was measured with an ICP-AES and is reported as the amount of manganese per optical density (approximately 107 cells). The values are the average of at least 12 samples harvested and analyzed on two different days. Error bars represent the standard deviation

921

Fig. 8 SOD activity. Native polyacrylamide gel electrophoresis of yeast extracts, showing endogenous SOD enzyme activity determined by in-gel activity assay a after and b before cyanide

treatment to inhibit the CuZnSOD. The samples are wild-type yeast (1), wild-type yeast treated with 5 mM MnCl2 (2), sod1D yeast (3), and sod1D yeast treated with 5 mM MnCl2 (4)

was found to be more lipophilic than either Euk-8 or Euk-134 [20]. The aerobic growth of wild-type and sod1D S. cerevisiae cultured in SDC pH 6.0 in the presence of each of the Eukarion compounds was recorded after 24 h. Wildtype yeast cells were largely unaffected by treatment with Euk-8, Euk-134, or Euk-189. The poor aerobic growth of sod1D yeast was not improved upon treatment with any concentration of Euk-8, Euk-134, or Euk-189 tested, in stark contrast to the rescue observed with ionic Mn2+. Instead, relatively low concentrations of these compounds prevented growth of these strains. We interpret this effect to mean that the compounds are able to enter the cells. In order to determine whether any beneficial effects of the Eukarion compounds might be observed under more stringent conditions, the 24-h aerobic growth of wildtype and sod1D yeast in SDC pH 6.0 lacking lysine or methionine was monitored in the presence of up to 3 lM Euk-134. Wild-type yeast was not affected by the presence of Euk-134 in either drop-out medium. Very little growth of sod1D S. cerevisiae in either SD-Lys or SDMet was observed, and the growth was not improved with the addition of Euk-134 (data not shown). The Mn(III) cyclam derivative M40403 from Metaphor that we tested has been reported to be highly stable, with a high level of SOD activity and no catalase activity [21, 47]. Aerobic growth of wild-type S. cerevisiae in the presence of M40403 remained unchanged from the untreated control as measured after 24 h (Fig. 9b). sod1D yeast cells cultured with M40403 were not rescued by low levels of this mimic (0.05 and 5 lM, data not shown), and, in fact, their growth was inhibited by about 90% with 25 lM M40403 (Fig. 9b). The structurally related SOD-inactive control compound M40404 had no effect on wild-type yeast or mutant yeast at the same concentrations. The observation of growth inhibition for sod1D yeast cultured with M40403 is strong evidence that this manganese complex enters yeast cells; on the basis of their structural similarities, it seems likely that M40404 enters yeast cells as well.

iron: all of the known sod1D yeast growth phenotypes were rescued fully, including the zinc sensitivity and the iron-related defects such as reduced aconitase and Leu1p activities and elevated levels of EPR-detectable iron. There are some systems in which the effect of elevated levels of manganese may be attributed to an increase in the activity of the mitochondrial MnSOD [48]. We have found that this is not the case in yeast. First, we confirmed the earlier observation that Mn2+ rescues the sod1Dsod2D double mutant, which lacks MnSOD [17]. Second, activity gels showed no increase in activity of MnSOD in cells treated with high manganese (Fig. 8). If anything, the activity of MnSOD was decreased by high manganese in both wild-type yeast and the sod1D mutant. S. cerevisiae normally requires only low levels (less than 3 lM) of manganese in the medium to sustain

Discussion Our studies demonstrate that the beneficial effects of manganese supplementation on sod1D yeast are much more dramatic and complete than those of copper or

Growth (OD at 600 nm)

a 10 WT Euk-134 sod1∆ Euk-134 WT Euk-8 sod1∆ Euk-8 WT Euk-189 sod1∆ Euk-189

8 6 4 2 0

1

10 100 Eukarion Concentration (µM)

Growth (OD at 600 nm)

b 10 8

SDC active (M40403)

6

inactive (M40404)

4 2 0

wild type

sod1∆

Fig. 9 The presence of Eukarion or Metaphore SOD mimics does not improve growth of sod1D Saccharomyces cerevisiae. Total growth following 24 h of shaking in SDC pH 6.0 at 220 rpm in air at 30C was measured turbidimetrically at 600 nm. a Growth of wild-type EG103 yeast (solid symbols) and isogenic sod1D yeast (open symbols) with various concentration of Euk-134 (squares), Euk-8 (triangles), or Euk-189 (circles). b Growth of the same strains in the presence of 25 lM active Metaphore compound M40403 (black bars), the inactive Metaphore control M40404 (gray bars), or the untreated control (white bars). The values represent the average of six or more independent colonies grown on two or more days, with error bars indicating the standard deviation

922

growth. It is interesting to note that in our experiments the concentration of MnCl2 that provided the optimal rescue of the sod1D yeast, i.e., 5 mM, was the same concentration at which a small but significant negative effect is first observed on control wild-type cells. Below that concentration, the addition of MnCl2 to wild-type cells had no effect on growth, and extremely high levels of manganese (8-10 mM) were toxic. One possible explanation is that 5 mM is the lowest concentration of Mn2+ in the medium that provides a rate of influx of Mn2+ that is not totally counteracted by the homeostatic mechanisms in place to handle toxic levels of exogenous metals. The rescue by manganese of the zinc sensitivity of the sod1D yeast was particularly striking to us, since it indicates that the absence of the SOD1 polypeptide does not result in zinc sensitivity so long as the cell can otherwise totally defend itself against superoxide. Mn2+ does not appear to be competing with Zn2+ for import because intracellular zinc levels are not lowered when cells are grown in 5 mM MnCl2 (Fig. 7b). Thus, although the possibility remains that manganese displaces zinc from some important site that mediates zinc toxicity without changing cellular zinc levels, the present data appear to support the hypothesis that Mn2+ rescue of sod1D yeast is due primarily to its antioxidant action in vivo. In our previous study [15], zinc resistance was not restored by cytosolic expression of MnSOD from Bacillus stearothermophilus; this bacterial enzyme did however rescue paraquat sensitivity [15]. This result seems contradictory to our current conclusion that Mn2+ rescue of sod1D yeast is due to its ability to scavenge superoxide. One possible explanation is that Mn2+ may be capable of protecting in a subcellular compartment not accessible to the bacterial MnSOD, a compartment that might be particularly sensitive to the adverse effects of high zinc. A likely candidate is the IMS, for the following reasons. Zinc and manganese can easily access this space via pores in the outer membrane that permit the passage of small water-soluble molecules, but proteins (i.e., the bacterial MnSOD) and other large molecules that are not substrates for specific transport across the membrane are excluded. Superoxide is produced in the IMS, CuZnSOD is ordinarily present, and CuZnSOD in this area was reported to be important for long-term survival of yeasts [42]. High zinc is detrimental to many mitochondrial functions [49]. If this compartment is especially sensitive to high levels of superoxide, then in the absence of active CuZnSOD it could be protected either by correct zinc handling, insured by the zinc-binding function of inactive SOD1p (in yeast lacking the CCS1 (LYS7) gene,1ccs1D, or sod1D yeast expressing H46C-SOD, an SOD-inactive, zinc-binding mutant of SOD1), or by the superoxide 1 This gene was first identified as part of the lysine biosynthetic pathway, hence the name LYS7. Since then, its true functioninsertion of copper into CuZnSOD-has been discovered. We support a recent proposal to rename the yeast gene CCS1 to better reflect its function and to agree with the names of similar genes in other organisms.

scavenging ability of Mn2+ present in the compartment. We find that manganese does accumulate to high levels in mitochondria of cells grown in high manganese (data not shown), lending support to this hypothesis. The deleterious effect of the two classes of manganese-containing SOD mimics, the manganese salen derivatives, Euk-8, Euk-134, and Euk-189, and the manganese cyclam complex, M40403, on the growth of the sod1D yeast was quite surprising to us. In the case of the manganese cyclam derivatives, we have the advantage of being able to compare two very similar coordination complexes of manganese, one of which has high SOD activity and one of which does not. The SODactive compound, M40403, inhibited the growth of the sod1D yeast, while the SOD-inactive compound, M40404, had no effect. The fact that M40403 inhibited growth is evidence that this compound does enter yeast cells; the structural similarity of M40403 and M40404 make it likely that the latter compound enters yeast cells as well. These results suggest that the growth inhibitory effect of the SOD-active manganese-containing compounds on the sod1D yeast may be related to their SOD activity. It is possible that alterations in levels of ROS may lead to the disruption of cell signaling, resulting in poor growth, as the transcription of many genes is induced by ROS [50, 51], but further work is required to evaluate this intriguing hypothesis. What is clear now from this study is that these two classes of manganesecontaining SOD mimetic compounds do not provide any benefit to the sod1D yeast, while the presence of ionic Mn2+ provides full rescue. The antioxidant role of ionic Mn2+ in vivo is thought to be due to its superoxide scavenging ability, and, as discussed in the ‘‘Introduction’’, a number of prokaryotic species are thought to use Mn2+ for exactly this purpose [1]. It now appears that this phenomenon is not restricted to prokaryotes and that the simple eukaryote S. cerevisiae can use manganese in this way as well. Because Mn2+ rescues all the phenotypes tested, we are inclined to believe that it works by chemically removing superoxide. Future studies will aim to get a better understanding of the subcellular localization and ligand environment as well as the inorganic chemistry of manganese that underlies its remarkable antioxidant properties when it is present at high concentrations in cells lacking SOD enzymes. Acknowledgements This work was supported by grant DK46828 to J.S.V. We also gratefully acknowledge the support of the National Science Foundation Predoctoral Fellowship (to R.J.S.), the NIH Chemistry-Biology Interface Predoctoral Training grant (to R.J.S. and M.A.W.), and the University of California Toxic Substances Research and Teaching Program, Lead Campus Program in Toxic Mechanisms (to M.A.W.)

References 1. Horsburgh MJ, Wharton SJ, Karavolos M, Foster SJ (2002) Trends Microbiol 10:496–501

923 2. Daly MJ, Gaidamakova EK, Matrosova VY, Vasilenko A, Zhai M, Venkateswaran A, Hess M, Omelchenko MV, Kostandarithes HM, Makarova KS, Wackett LP, Fredrickson JK, Ghosal D (2004) Science 306:1025–1028 3. Al-Maghrebi M, Fridovich I, Benov L (2002) Arch Biochem Biophys 402:104–109 4. Archibald FS, Fridovich I (1982) Arch Biochem Biophys 214:452–463 5. Cabelli DE, Bielski BHJ (1984) J Phys Chem 88:6291–6294 6. Weiss RH, Flickinger AG, Rivers WJ, Hardy MM, Aston KW, Ryan US, Riley DP (1993) J Biol Chem 268:23049–23054 7. Bielski BH, Cabelli DE (1991) Int J Radiat Biol 59:291–319 8. Chang EC, Kosman DJ (1989) J Biol Chem 264:12172–12178 9. Cabelli DE, Bielski BHJ (1984) J Phys Chem 88:3111–3115 10. Jacobsen F, Holcman J, Sehested K (1997) J Phys Chem 101:1324–1328 11. Niviere V, Fontecave M (2004) J Biol Inorg Chem 9:119–123 12. Gralla EB, Valentine JS (1991) J Bacteriol 173:5918–5920 13. Sturtz LA, Culotta VC (2002) Methods Enzymol 349:167–172 14. Srinivasan C, Liba A, Imlay JA, Valentine JS, Gralla EB (2000) J Biol Chem 275:29187–29192 15. Wei JP, Srinivasan C, Han H, Valentine JS, Gralla EB (2001) J Biol Chem 276:44798–44803 16. De Freitas JM, Liba A, Meneghini R, Valentine JS, Gralla EB (2000) J Biol Chem 275:11645–11649 17. Lapinskas PJ, Cunningham KW, Liu XF, Fink GR, Culotta VC (1995) Mol Cell Biol 15:1382–1388 18. Portnoy ME, Liu XF, Culotta VC (2000) Mol Cell Biol 20:7893–7902 19. Tamai KT, Gralla EB, Ellerby LM, Valentine JS, Thiele DJ (1993) Proc Natl Acad Sci USA 90:8013–8017 20. Doctrow SR, Huffman K, Marcus CB, Tocco G, Malfroy E, Adinolfi CA, Kruk H, Baker K, Lazarowych N, Mascarenhas J, Malfroy B (2002) J Med Chem 45:4549–4558 21. Riley DP (1999) Chem Rev 99:2573–2588 22. Melov S, Ravenscroft J, Malik S, Gill MS, Walker DW, Clayton PE, Wallace DC, Malfroy B, Doctrow SR, Lithgow GJ (2000) Science 289:1567–1569 23. Sampayo JN, Olsen A, Lithgow GJ (2003) Aging Cell 2:319– 326 24. Baker K, Marcus CB, Huffman K, Kruk H, Malfroy B, Doctrow SR (1998) J Pharmacol Exp Ther 284:215–221 25. Pong K, Doctrow SR, Huffman K, Adinolfi CA, Baudry M (2001) Exp Neurol 171:84–97 26. Melov S, Doctrow SR, Schneider JA, Haberson J, Patel M, Coskun PE, Huffman K, Wallace DC, Malfroy B (2001) J Neurosci 21:8348–8353

27. Salvemini D, Wang ZQ, Zweier JL, Samouilov A, Macarthur H, Misko TP, Currie MG, Cuzzocrea S, Sikorski JA, Riley DP (1999) Science 286:304–306 28. Shimizu K, Rajapakse N, Horiguchi T, Payne RM, Busija DW (2003) Neurosci Lett 346:41–44 29. Masini E, Cuzzocrea S, Mazzon E, Marzocca C, Mannaioni PF, Salvemini D (2002) Br J Pharmacol 136:905–917 30. Wang ZQ, Porreca F, Cuzzocrea S, Galen K, Lightfoot R, Masini E, Muscoli C, Mollace V, Ndengele M, Ischiropoulos H, Salvemini D (2004) J Pharmacol Exp Ther 309:869–878 31. Udipi K, Ornberg RL, Thurmond KB 2nd, Settle SL, Forster D, Riley D (2000) J Biomed Mater Res 51:549–560 32. McFadden SL, Ding D, Salvemini D, Salvi RJ (2003) Toxicol Appl Pharmacol 186:46–54 33. 33.Nicotera TM, Ding D, McFadden SL, Salvemini D, Salvi R (2004) Audiol Neurootol 9:353–362 34. Longo VD, Gralla EB, Valentine JS (1996) J Biol Chem 271:12275–12280 35. Kaiser C, Michaelis SAM (1994) Methods in yeast genetics. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York 36. Kohlhaw GB (1988) Methods Enzymol 166:429–435 37. Wallace MA, Liou LL, Martins J, Clement MH, Bailey S, Longo VD, Valentine JS, Gralla EB (2004) J Biol Chem 279:32055–32062 38. Gardner PR, Raineri I, Epstein LB, White CW (1995) J Biol Chem 270:13399–13405 39. Huang TT, Raineri I, Eggerding F, Epstein CJ (2002) Methods Enzymol 349:191–213 40. Srinivasan C, Gralla EB (2002) Methods Enzymol 349:173-180 41. Sherman F (1991) Methods Enzymol 194:3–21 42. Sturtz LA, Diekert K, Jensen LT, Lill R, Culotta VC (2001) J Biol Chem 276:38084–38089 43. Beinert H, Holm RH, Munck E (1997) Science 277:653–659 44. Vida TA, Emr SD (1995) J Cell Biol 128:779–792 45. Corson LB, Folmer J, Strain JJ, Culotta VC, Cleveland DW (1999) J Biol Chem 274:27590–27596 46. Faulkner KM, Liochev SI, Fridovich I (1994) J Biol Chem 269:23471–23476 47. Riley DP, Weiss RH (1994) Abstracts of Papers of the American Chemical Society 208:152-Inor 48. Davis CD, Greger JL (1992) Am J Clin Nutr 55:747–752 49. Cuajungco MP, Lees GJ (1997) Neurobiol Dis 4:137–169 50. Apel K, Hirt H (2004) Annu Rev Plant Biol 55:373–399 51. Azevedo D, Tacnet F, Delaunay A, Rodrigues-Pousada C, Toledano MB (2003) Free Radic Biol Med 35:889–900