A physiological role for Saccharomyces cerevisiae copper/zinc superoxide dismutase in copper buffering.

The copper toxicity of yeast lacking the CUP1 metallothionein is suppressed by overexpression of the CRS4 gene. We now demonstrate that CRS4 is equivalent to SOD1, encoding copper/zinc superoxide dismutase (SOD). While overexpression of SOD1 enhanced copper resistance, a deletion of SOD1, but not SOD2 (encoding manganese SOD), conferred an increased sensitivity toward copper. This role of SOD1 in copper buffering appears unrelated to its superoxide scavenging activity, since the enzyme protected against copper toxicity in anaerobic as well as aerobic conditions. The distinct roles of SOD1 in copper and oxygen radical homeostasis could also be separated genetically: the pmr1, bsd2, and ATX1 genes that suppress oxygen toxicity in sod1 mutants failed to suppress the copper sensitivity of these cells. The Saccharomyces cerevisiae SOD1 gene is transcriptionally induced by copper and the ACE1 transactivator, and we demonstrate here that this induction of SOD1 promotes protection against copper toxicity but is not needed for the SOD1-protection against oxygen free radicals. Collectively, these findings indicate that copper/zinc SOD functions in the homeostasis of copper via mechanisms distinct from superoxide scavenging.

The copper toxicity of yeast lacking the CUP1 metallothionein is suppressed by overexpression of the CRS4 gene. We now demonstrate that CRS4 is equivalent to SOD1, encoding copper/zinc superoxide dismutase (SOD). While overexpression of SOD1 enhanced copper resistance, a deletion of SOD1, but not SOD2 (encoding manganese SOD), conferred an increased sensitivity toward copper. This role of SOD1 in copper buffering appears unrelated to its superoxide scavenging activity, since the enzyme protected against copper toxicity in anaerobic as well as aerobic conditions. The distinct roles of SOD1 in copper and oxygen radical homeostasis could also be separated genetically: the pmr1, bsd2, and ATX1 genes that suppress oxygen toxicity in sod1 mutants failed to suppress the copper sensitivity of these cells. The Saccharomyces cerevisiae SOD1 gene is transcriptionally induced by copper and the ACE1 transactivator, and we demonstrate here that this induction of SOD1 promotes protection against copper toxicity but is not needed for the SOD1-protection against oxygen free radicals. Collectively, these findings indicate that copper/zinc SOD functions in the homeostasis of copper via mechanisms distinct from superoxide scavenging.
The transition metal copper represents a paradox to living organisms because trace amounts of this ion are essential to promote growth, yet concentrated levels of this same metal can drastically impair cell growth and function. Although many hypotheses have been proposed to explain copper toxicity in biological systems, a popular model exploits the ability of the metal to participate in so-called Fenton or Haber-Weiss chemistry (1)(2)(3). In this model, copper ions would catalyze the conversion of hydrogen peroxide (H 2 O 2 ) to the powerful oxidant hydroxyl radical (OH ⅐ ), which has the capacity to damage cellular components. Alternatively, copper toxicity may also be mediated through the inappropriate binding of the metal to nitrogen, oxygen, and sulfur ligands in biomolecules, thereby inactivating enzymes and disrupting cellular function (4).
In order to balance the growth inhibitory and stimulatory effect of copper ions, all organisms have evolved with various metal homeostasis factors that properly control the cellular accumulation, distribution, and detoxification of the metal. A number of copper transporters have been identified that act in the passage of copper across external membranes or in the intracellular delivery of the metal to cellular stores. For example, copper-transporting ATPases have been identified in bacteria (5), yeast (6,7) and man (8 -11). All of these ATPases share a remarkable degree of homology, exemplifying the need to tightly conserve the homeostasis of essential copper ions. Non-ATPase copper transporters have also been identified in bacteria (12) and yeast (13), and these share a common metal binding motif.
As another means of maintaining copper ion homeostasis, many eukaryotic organisms contain one or more forms of a metal binding metallothionein that acts to chelate, sequester, and thereby detoxify copper ions. In the bakers' yeast, the CUP1 copper-containing metallothionein (14 -16) has been demonstrated to not only protect against copper but also to guard against the oxidative damage associated with superoxide anion (17). This metallothionein of Saccharomyces cerevisiae, and another encoded by the CRS5 gene (18), are known to be transcriptionally regulated by copper ions through the action of a copper and DNA binding trans-activator, ACE1 (18 -22). Interestingly, the ACE1 copper sensor factor has been shown to also induce the S. cerevisiae SOD1 gene in response to elevated copper (23,24). SOD1 encodes a copper-and zinc-containing superoxide dismutase (SOD), 1 and the rationale for the coinduction of this free radical detoxifying enzyme with the metallothioneins was not completely understood (23,24). Nevertheless, these studies on S. cerevisiae CUP1 metallothionein and SOD have substantiated the notion that considerable overlap exists between systems controlling copper ion homeostasis and oxygen radical metabolism.
To identify additional factors involved in copper ion homeostasis, we developed a selection strategy to isolate S. cerevisiae genes that, when overexpressed, have the ability to suppress the copper toxicity associated with loss of the CUP1 metallothionein. This strategy led to the isolation of two genes, CRS4 and CRS5, that, when present in multi-copy, conferred a relatively high degree of copper resistance to yeast lacking CUP1. In our earlier studies, we discovered that CRS5 encodes an additional yeast metallothionein, only the second to be identified in S. cerevisiae (18). Expecting to uncover a third S. cerevisiae metallothionein, we have here sequenced and characterized the CRS4 gene. Surprisingly, CRS4 was found to be identical to the S. cerevisiae SOD1 gene, encoding copper/zinc superoxide dismutase. The biological role of this enzyme in copper buffering in yeast is discussed herein.
Stocks of all strains were maintained on YPD medium, and tests for copper ion and paraquat sensitivity were carried out using a synthetic dextrose medium as described previously (30). Anaerobic cultures were maintained by growth in an O 2 -depleted culture jar (BBL GasPak) as described previously (26).
Plasmids and Gene Analyses Methods-The original CRS4 (SOD1)containing constructs p12 and p28 were isolated from a S. cerevisiae genomic library as described previously (18). These isolates represent identical 6.0-kb Sau3AI fragments of S. cerevisiae genomic DNA inserted at the BamHI site of the 2 m LEU2 yeast shuttle vector, YEp13 (31). The p28⌬sp construct was obtained by removing a 1.2-kb SphI fragment of p28 by digesting at vector and insert SphI sites and by religating. The pRS28 construct contains the 6.0-kb insert of p28 mobilized by digestion with EagI and NheI and inserted into the EagI and SpeI sites of the 2 m HIS3 vector, pRS423 (32). The p28⌬sac and p28⌬r1 constructs were obtained by digesting pRS28 with SacI or EcoRI, respectively, at sites in the polylinker and insert and by recircularizing through ligation. The p28⌬hp construct was similarly obtained by digesting at insert HpaI and polylinker EcoRV sites and by religating.
Construction of the pVC731-736 series of SOD1 containing plasmids was as follows: sequences corresponding to SOD1 residues Ϫ257 to ϩ586 with respect to the translational start site were amplified by the polymerase chain reaction (PCR). The upstream SOD1 primer (shown in Fig. 7A) was either wild type (for construction of pVC734 and pVC736) or contained a CGC to TAT substitution at position Ϫ227 (for pVC731 and pVC733). The PCR products were ligated to the pCRII vector (Invitrogen) to generate pCR16 (wild-type promoter) and pCR15 (mutant ACE1 site); the inserts were then mobilized by digestion with BamHI and XhoI and inserted into the BamHI and SalI sites of the CEN vector, pRS315, to generate pVC731 and pVC734 and also into the 2-m vector, pRS415, to generate pVC733 and pVC736.
Construction of the sod1⌬::LEU2 and sod1⌬::TRP1 gene replacement plasmids and deletion of the SOD1 chromosomal locus were performed as follows. A fragment containing SOD1 sequences Ϫ703 to ϩ18, with respect to the translational start, was amplified by PCR and inserted into vector pCRII (Invitrogen). This construct was digested with SacI and PstI to liberate the SOD1 sequences, which were then subcloned into the SacI and PstI sites of pRS305. Downstream SOD1 sequences ϩ314 to ϩ586 (with respect to translational start) were introduced at the HindIII and PstI sites of this construct by inserting a HindIII-PstI fragment from pVC16, generating the sod1⌬::LEU2 deletion construct, pKS1. The sod1⌬::TRP1 deletion construct pKS3 was constructed by subcloning the ApaI-SacI fragment from pKS1 into the ApaI and SacI sites of pRS404. To delete the chromosomal SOD1 gene, the pKS1 or pKS2 plasmid was linearized with PstI and used to transform haploid leu2 or trp1 yeast, respectively, by electroporation. Correct deletion of SOD1 sequences ϩ18 to ϩ314 was confirmed using PCR.
Identification of CRS4 as SOD1 was accomplished by sequencing both strands of p28⌬hp and p28⌬sac using the Sequenase method (U. S. Biochemical Corp.) modified for double-stranded DNA. Sequence information was stored by the DNA Strider software package and upon a search of the N.C.B.I data base, CRS4 was found to encompass the S. cerevisiae SOD1 and URA8 open reading frames. Northern and Southern blot analyses were conducted as described previously (33).

Suppression of Copper Toxicity by the S. cerevisiae CRS4
Gene-Copper toxicity in yeast cells lacking a functional CUP1 metallothionein can be suppressed by overexpression of the S. cerevisiae CRS (Copper-resistant suppressors) gene (18). As shown in Fig. 1, yeast lacking CUP1 metallothionein (cup1⌬) were unable to grow on plates containing 25 M copper whether cultured aerobically or anaerobically. When the cup1⌬ yeast strain was transformed with either the multi-copy CRS4 or CRS5 gene, cells exhibited a strong resistance to 25 M copper under both aerobic and anaerobic conditions ( Fig. 1). Hence, overexpression of CRS4 or CRS5 can alleviate copper toxicity in cup1⌬ yeast in a manner that is not dependent upon atmospheric oxygen.
To examine the extent to which CRS4 protects cells from copper toxicity, a liquid culture test was utilized to obtain measurements of the mean inhibitory concentration for the metal (concentration required to inhibit total cell growth by 50%). On the YEp13 yeast episomal plasmid, CRS4 is amplified to 5-10 copies/cell (Southern blot analyses, not shown), and in cup1⌬ yeast, this elevation in CRS4 gene dosage increased the mean inhibitory concentration for copper by approximately 400% (Fig. 2A). To test whether this increased resistance to copper is specific to cup1⌬ strains, we also assayed for CRS4 function in a cup1 s strain harboring a single copy of the metallothionein gene. As shown in Fig. 2B, these strains are normally more resistant to copper toxicity than are cup1⌬ yeast, and this resistance was increased further by overexpression of CRS4; however, the CRS4-containing plasmid only increased the mean inhibitory concentration by approximately 35% in cup1 s cells (Fig. 2B). The apparent nonadditive effects of CRS4 and CUP1 suggested that these two genes may play functionally redundant roles in copper detoxification.
Identification of CRS4 as the S. cerevisiae SOD1 Gene-Two identical isolates of CRS4 (p12 and p28) were obtained from the genomic library, and these comprise a 6.0-kb fragment of genomic DNA (Fig. 3). To identify the CRS4 gene in this fragment, segments of p28 were sub-cloned and tested for the ability to confer copper resistance. A 1.5-kb SphI-HpaI segment of this DNA was found to suppress copper toxicity in yeast lacking CUP1 (Fig. 3). Sequence analyses of this region revealed the presence of two previously cloned genes, the S. cerevisiae URA8 (34) and SOD1 (35) loci. Of these, isolated sequences of SOD1 were found to be necessary and sufficient to suppress the copper toxicity of cup1⌬ yeast (Fig. 3), demonstrating that CRS4 is equivalent to SOD1. The S. cerevisiae SOD1 gene (35) encodes a copper and zinc containing SOD that is known to scavenge superoxide anion (O 2 . ) radicals and is thought to play a critical role in protecting cells against oxidative damage. Yeast strains containing mutations in the SOD1 gene exhibit a number of aerobic defects including a deficiency in lysine biosynthesis and in sulfur metabolism and a sensitivity toward atmospheric oxygen and paraquat, a O 2 . -generating agent (25,26,(35)(36)(37)(38)(39). All of these defects are dependent upon the presence of oxygen, indicating that SOD plays a pivotal role in protecting cellular constituents against oxygen-related toxicity.
To explore the role of SOD1 in copper buffering, we compared the growth of a sod1⌬ mutant lacking the endogenous SOD1 gene to that of a SOD1 wild-type strain. Resistance to copper was also tested in a sod2⌬ mutant lacking the mitochondrial manganese containing SOD and a sod1⌬ sod2⌬ double mutant strain lacking both forms of SOD. The strains utilized for these studies also contain a wild-type tandemly amplified CUP1 locus and are considerably more resistant to copper toxicity than are cup1⌬ and cup1 s yeast strains. As shown in Fig. 4, the strains containing a sod1⌬ single or sod1⌬ sod2⌬ double mutation exhibited an increased inhibition of growth relative to wild-type yeast in 60 M copper medium, and this increased sensitivity toward copper was not observed with sod2⌬ single mutants. In comparison, strains containing a mutation in SOD2 exhibited an elevated sensitivity toward 30 M cadmium that was not observed with sod1 mutants (Fig. 4). In other studies, the sod1⌬ and sod2⌬ mutants consistently exhibited impaired growth relative to wild-type strains at 30 -90 M CuSO 4 and 5-50 M CdCl 2 , respectively (not shown). The increased copper sensitivity of sod1⌬ mutants was also observed with anaerobically grown cells (not shown), consistent with our observation that overexpression of SOD1 suppressed copper toxicity in both aerobic and anaerobic conditions (Fig. 1). Thus the SOD1 protein appears to play a role in copper buffering in addition to oxygen radical scavenging.
Copper Toxicity and the Genetic Suppressors of SOD1 Deficiency-We have previously shown that the oxygen-dependent defects of sod1⌬ mutants (lysine and methionine auxotrophies, sensitivity toward atmospheric oxygen and paraquat) can be suppressed by alterations in the expression of three nuclear genes, ATX1 (40), PMR1 (27), and BSD2 (28). Overexpression of the ATX1 gene, or mutational inactivation of either PMR1 or BSD2, will suppress to some degree all of the aerobic defects of yeast containing a sod1⌬ mutation. In the present studies, we have tested whether these suppressors of oxygen toxicity can also reverse the copper sensitivity associated with loss of the copper/zinc SOD. These experiments would additionally address the potential overlap between the apparent roles of SOD1 in copper buffering and in oxygen radical detoxification.
As shown in Fig. 5A, sod1⌬ mutants still exhibited a striking sensitivity toward copper toxicity when the anti-oxidant gene ATX1 was overexpressed. Furthermore, a pmr1 mutation was also incapable of reversing the copper toxicity of sod1⌬ cells, since the sod1⌬ pmr1⌬ mutant still exhibited impaired growth in 30 -70 M concentrations of copper compared with SOD1 wild-type yeast containing a pmr1⌬ mutation (Fig. 5B). A mutation in BSD2 was likewise ineffective in suppressing the FIG. 2. Resistance to copper toxicity in yeast transformed with the CRS4 gene. Growth of the indicated yeast strains was tested in the presence of increasing concentrations of CuSO 4 . Liquid cultures seeded to an optical density at 600 nm of 0.1 were grown for 16 h at 30°C, and total growth was determined turbidimetrically at an optical density at 600 nm. Where indicated, strains were transformed either with YEp13 vector or with YEpCRS4, the p28 CRS4 containing plasmid. The relevant genotypes of the yeast strains tested were as follows: A, 51.2c ϭ cup1⌬::URA3; B, VC-sp6 ϭ cup1 s . Results represent the averages of 2-3 independent trials with range Յ15%. copper sensitivity associated with loss of SOD1 (Fig. 5C). BSD2 gene mutations are themselves associated with an increase in sensitivity toward copper (28), and an additive effect on copper toxicity was observed with the combination of bsd2 and sod1 mutations (Fig. 5C). Hence, perturbations in the function of ATX1, PMR1, and BSD2 can suppress only the oxidative defects and not the copper toxicity of sod1⌬ mutants. These results provide genetic evidence that SOD1 plays separate physiological roles in copper ion buffering and oxygen radical detoxification.
The Role of ACE1 in SOD1-mediated Protection against Copper-It has previously been demonstrated by Tamai et al. (17) and by Ciriolo et al. (42) that the S. cerevisiae SOD1 gene is positively regulated at the level of transcription by copper ions and ACE1, the same copper sensor trans-activator that regulates the yeast CRS5 and CUP1 metallothionein genes (23,24). The rationale for this curious co-regulation of SOD1 and the metallothioneins by copper was not well established (23,24). In the present studies, we utilized two approaches to test whether the induction of SOD1 by copper ions is essential to its role in protecting yeast against the toxicity of oxygen radicals and/or copper. In the first line of study, the ACE1 gene was deleted in the background of both sod1⌬ and SOD1 wild-type yeast, and cells were tested for sensitivity toward paraquat (a generator of O 2 . ) and copper. As shown in Fig. 6A, SOD1 wild-type cells exhibited a striking resistance toward paraquat over sod1⌬ yeast, and this SOD1-mediated protection against free radical damage was not altered in strains containing an ACE1 gene deletion. In contrast, the functional SOD1 gene did not efficiently protect against copper toxicity in strains lacking ACE1 (Fig. 6B). This finding indicates that the ACE1 factor is needed for the SOD1 protection against copper but not oxygen free radicals.
We also tested the significance of ACE1 in SOD1 function by inactivating the ACE1 binding site in the SOD1 promoter. A pair of 2-m-containing constructs were made that contain FIG. 3. Identification of the CRS4 gene. Plasmid p28 was isolated from a yeast genomic YEp13 library as a clone that confers copper resistance to cup1⌬ yeast. The remaining constructs represent fragments of the p28 insert subcloned into the 2-m yeast vectors YEp13, pRS423, or pRS425 as described under "Experimental Procedures." The constructs were used to transform strain 51.2c, and all were found to support growth of this cup1⌬ strain on minimal medium plates containing 25 M CuSO 4 (as in Fig. 1 4. Effects of sod1⌬ and sod2⌬ gene deletions on tolerance to heavy metals. The indicated strains of yeast were cultured as described in Fig. 2 in minimal medium that was either untreated (control) or supplemented with 60 M CuSO 4 (copper) or 30 M CdCl 2 (cadmium). Total growth was measured as in Fig. 2, and results represent the averages of two to four independent trials (the complete range of values obtained are represented by error bars). Strains tested are as follows: SODϩ, EG103; sod1⌬, KS101; sod2⌬, EG110; sod1⌬ sod2⌬, KS100. All strains contain an amplified CUP1 locus and normally exhibit a high level of copper resistance.  Fig. 4. % Control Growth was obtained by dividing the A 600 value obtained with copper-treated yeast over that of control untreated cells. Strains of yeast utilized are as follows: A, sod1⌬-pATX1, KS107 transformed with the p18 construct harboring ATX1 (40); SODϩ-pATX1, EG103 transformed with p18; B, sod1⌬ pmr1⌬, VC119; SODϩ pmr1⌬, PJKP1-2; C, sod1⌬ bsd2⌬, VC120; SODϩ bsd2⌬, XL103⌬b2. All strains contain an amplified CUP1 locus. Results represent averages of two to three independent trials where range Յ10%. In the various experiments, control untreated cells grew to an A 600 of 2.5-4.0.
SOD1 sequences Ϫ257 to ϩ586, with respect to the translational start site. These constructs, pVC733 and pVC736, were identical, except pVC733 contained a CGC to TAT triple mutation in the ACE1 binding site of the SOD1 promoter ( Fig. 7A; Refs. 20,21,23,24). Both constructs were used to transform a sod1⌬ strain and were monitored for copper induction by Northern blot analysis. As shown in Fig. 7A, the two SOD1 constructs were expressed to equal degrees in the absence of copper; however, only the pVC736 construct containing the wild-type ACE1 binding site responded to copper through induction of SOD1 mRNA. To test whether this increase in SOD1 mRNA is essential to its role in buffering oxygen radicals versus metals, sod1⌬ strains harboring the pVC733 mutated, and the pVC736 wild-type constructs were tested for sensitivity to both oxygen and copper. Lysine auxotrophy is an excellent marker for oxygen toxicity in yeast lacking copper/zinc SOD (26,35,36,38,39), and both pVC733 and pVC736 were fully capable of complementing the lysine auxotrophy of sod1⌬ cells (not shown). The wild-type and mutant constructs were also equally effective in complementing the paraquat sensitivity of sod1⌬ yeast (Fig. 7B). Hence, the ACE1 binding site does not appear to impact on the ability of SOD1 to protect against oxidative damage. However, in contrast to results obtained with paraquat (Fig. 7B), the pVC736 wild-type construct was substantially more effective in complementing the copper sensitivity of sod1⌬ yeast than was the pVC733 mutant construct, and this result was observed in the background of both sod1⌬ and sod1⌬ sod2⌬ yeast (Fig. 7C). Identical experiments were conducted with the same SOD1 wild-type and mutant constructs expressed on a yeast CEN vector, and although the CEN constructs were less efficient at complementing the sod1⌬ mutation, the same qualitative results were obtained; the wildtype SOD1 construct was considerably more effective in suppressing copper toxicity (data not shown). Together with the studies on ace1⌬ yeast (Fig. 6), these findings strongly indicate that the role of SOD1 in copper buffering, but not in oxygen radical detoxification, is dependent upon the ACE1 trans-activator and a functional ACE1 binding site in the SOD1 promoter. A, the Ϫ257 to Ϫ225 primer utilized to construct the pVC736 and pVC733 SOD1 plasmids is shown. Asterisks mark the CGC to TAT substitution made in pVC733. Both plasmids were used to transform sod1⌬ yeast, and where indicated, cultures of transformed cells were treated for 1 h with 60 M CuSO 4 prior to isolation of total mRNA and analysis by Northern blot. The blot was sequentially hybridized to SOD1 and ACT1-specific probes. When normalized against ACT1, the pVC736 and pVC733 mRNA exhibited a 3.9-and 1.2-fold induction upon copper treatment. B, paraquat sensitivity in the indicated yeast strains was measured as in Fig. 6A. % Control Growth was obtained by dividing the A 600 value obtained with paraquat treated yeast over that of control untreated cells. C, sensitivity toward copper toxicity was measured in the indicated strains as in Fig. 4. Strains utilized were as follows: SODϩ, EG103; sod1⌬, KS107; sod1⌬ sod2⌬, EG133. pVC733 and pVC736 indicate strains transformed with these SOD1 containing episomal plasmids. All strains contain an amplified CUP1 locus. DISCUSSION The studies presented here strongly suggest that copper/zinc SOD serves a physiological role in copper ion buffering. We demonstrate that the S. cerevisiae SOD1 gene can act as a multicopy suppressor of copper toxicity. SOD1 was only one of two loci found to have the capacity to suppress the strong copper sensitivity associated with loss of the S. cerevisiae CUP1 metallothionein. The other, CRS5, encodes an additional metallothionein (18). SOD1 not only protects against copper toxicity as a multicopy locus, but the single copy chromosomal gene also participates in copper buffering since a sod1⌬ gene deletion was associated with an increased sensitivity toward copper. Of the two forms of SOD, copper buffering appears specific to the copper/zinc containing enzyme; a null mutation in the SOD2 gene encoding the mitochondrial manganese SOD was not associated with an increased sensitivity toward copper.
We provide several lines of evidence that the apparent role of copper/zinc SOD in copper buffering is distinct and perhaps unrelated to the well established role of this enzyme in free radical detoxification. First, the previously reported defects associated with loss of S. cerevisiae SOD1 (aerobic growth inhibition, methionine and lysine auxotrophy, increased mutation rate) were absolutely dependent on the presence of oxygen (25,28,(35)(36)(37)(38)(39). In contrast, the copper sensitivity associated with sod1⌬ mutants was observed independent of atmospheric oxygen. The roles of SOD1 in copper and oxygen radical homeostasis can also be separated at the genetic level. A mutation in the PMR1 or BSD2 genes or overexpression of ATX1 will suppress the oxygen-related defects associated with loss of SOD1 function (26 -28, 40); however, these same genetic perturbations failed to overcome the increased copper sensitivity of sod1⌬ mutants. We initially found this result surprising, since all three of these suppressors are themselves involved in copper homeostasis. PMR1 encodes a P-type ATPase homologue that is believed to function in the delivery of manganese, copper, and calcium ions into the Golgi (27); BSD2 is involved in the homeostasis of copper and cadmium (28), and ATX1 appears to encode a small copper carrier (40). Evidently these suppressors increase the local concentration and availability of redox active transition metals but do not offer any metal buffering function.
It has been proposed that SOD can circumvent the Haber-Weiss reaction (the two-step copper or iron catalyzed production of OH ⅐ ) by scavenging the O 2 . that would normally provide the reducing equivalents for the transition metal (1-3, 41). However, O 2 . scavenging cannot account for the protective effect of SOD1 with respect to copper ions since SOD1 protects against copper toxicity under anaerobic conditions. Instead, we propose that the copper resistance associated with SOD1 gene dosage reflects the ability of the encoded enzyme to directly bind and sequester the metal and thereby help prevent the dysfunctional binding of copper to biological metal ligands. In this model, SOD1 has the potential to function as a metallothionein, as previously proposed by Ciriolo et al. (42). Fee and co-workers (43) have proposed that bacterial forms of SOD can similarly function in iron metabolism.
In the budding yeast, several lines of evidence suggest functional relatedness between the metallothioneins and copper/ zinc SOD. Not only does SOD1 suppress metallothionein-related defects (this study), but the converse holds true; overexpression of CUP1 metallothionein will suppress defects associated with loss of SOD1 (17). Furthermore, SOD1 and the CUP1 and CRS5 metallothioneins are coordinantly induced at the transcriptional level by copper ions and the copper transactivator ACE1 (18 -20, 23, 24). Originally, the rationale for copper induction of SOD1 was not clear, as the gene is constitutively expressed at a relatively high level and ACE1 only serves to increase this expression by a few times. We demonstrate here that the copper induction of SOD1 by ACE1 is not essential to the anti-oxidant role of the enzyme. SOD1 still protects against oxidative damage in yeast containing an ace1⌬ deletion and with SOD1 genes lacking a functional ACE1 binding site. In contrast, the functional ACE1 factor and ACE1 binding sequences were found to be necessary for the protection against copper toxicity afforded by SOD1. We therefore propose that the constitutive level of copper/zinc SOD is sufficient to protect cells against oxidative damage by efficiently scavenging O 2 . . However, when exposed to toxic copper levels, the cell responds by further increasing SOD1 levels. This elevation in SOD1, together with the induction of the metallothioneins, provides an effective method for metal buffering. It is noteworthy that mutations in the human SOD1 gene have been associated with a familial form of amyotrophic lateral sclerosis (FALS), a fatal neurodegenerative disorder affecting the motor neurons (44). All FALS mutations are dominant, and in certain instances, the disease cannot be explained simply by a loss of SOD1 O 2 . scavenging activity (45)(46)(47)(48)(49)(50). The apparent role of copper/zinc SOD1 in copper buffering may in fact have important implications with regard to the etiology of FALS (47).