A Gain of Superoxide Dismutase (SOD) Activity Obtained with CCS, the Copper Metallochaperone for SOD1*

The incorporation of copper ions into the cytosolic superoxide dismutase (SOD1) is accomplished in vivo by the action of the copper metallochaperone CCS (copper chaperone for SOD1). Mammalian CCS is comprised of three distinct protein domains, with a central region exhibiting remarkable homology (approximately 50% identity) to SOD1 itself. Conserved in CCS are all the SOD1 zinc binding ligands and three of four histidine copper binding ligands. In CCS the fourth histidine is replaced by an aspartate (Asp200). Despite this conservation of sequence between SOD1 and CCS, CCS exhibited no detectable SOD activity. Surprisingly, however, a single D200H mutation, targeting the fourth potential copper ligand in CCS, granted significant superoxide scavenging activity to this metallochaperone that was readily detected with CCS expressed in yeast. This mutation did not inhibit the metallochaperone capacity of CCS, and in fact, D200H CCS appears to represent a bifunctional SOD that can self-activate itself with copper. The aspartate at CCS position 200 is well conserved among mammalian CCS molecules, and we propose that this residue has evolved to preclude deleterious reactions involving copper bound to CCS.

The incorporation of copper ions into the cytosolic superoxide dismutase (SOD1) is accomplished in vivo by the action of the copper metallochaperone CCS (copper chaperone for SOD1). Mammalian CCS is comprised of three distinct protein domains, with a central region exhibiting remarkable homology (approximately 50% identity) to SOD1 itself. Conserved in CCS are all the SOD1 zinc binding ligands and three of four histidine copper binding ligands. In CCS the fourth histidine is replaced by an aspartate (Asp 200 ). Despite this conservation of sequence between SOD1 and CCS, CCS exhibited no detectable SOD activity. Surprisingly, however, a single D200H mutation, targeting the fourth potential copper ligand in CCS, granted significant superoxide scavenging activity to this metallochaperone that was readily detected with CCS expressed in yeast. This mutation did not inhibit the metallochaperone capacity of CCS, and in fact, D200H CCS appears to represent a bifunctional SOD that can self-activate itself with copper. The aspartate at CCS position 200 is well conserved among mammalian CCS molecules, and we propose that this residue has evolved to preclude deleterious reactions involving copper bound to CCS.
In eukaryotic cells, copper is delivered to specific protein targets via the action of a family of copper carrier proteins termed "metallochaperones" (1). These molecules are well conserved between yeast and humans and serve to guide the metal to discrete cellular locations and facilitate incorporation of the cofactor into target metalloenzymes (reviewed in Refs. [2][3][4]. One such copper chaperone, COX17, acts in the delivery of copper to mitochondrial cytochrome oxidase (5)(6)(7)(8). A second soluble metallochaperone, ATX1, escorts copper strictly to transport ATPases in the secretory pathway (1, 9 -12). Thirdly, copper delivery and incorporation into cytosolic superoxide dismutase 1 (SOD1) 1 is mediated by the soluble copper carrier, CCS (copper chaperone for SOD), also known in Saccharomyces cerevisiae as LYS7 (13,14). Studies with the yeast metallochaperone have shown that CCS directly incorporates copper into SOD1 despite exquisitely low levels of available free copper (15).
The target of the CCS metallochaperone, SOD1, is a homodimeric copper-and zinc-requiring enzyme that acts to disproportionate superoxide (O 2 . ) to hydrogen peroxide (H 2 O 2 ) and oxygen in a reaction catalyzed by the redox cycling of bound copper (16). However, SOD1 is also capable of catalyzing deleterious reactions involving the redox active copper cofactor. The Cu(I) form of SOD1 can react with H 2 O 2 to generate the highly toxic hydroxyl radical (OH ⅐ ) (17)(18)(19). In fact, it has been suggested that this inherent peroxidase activity of SOD1 may be involved in cases of familial amyotrophic lateral sclerosis in which disease results from dominant mutations in SOD1 (20 -24). It is noteworthy that the human CCS metallochaperone harbors a polypeptide region bearing striking resemblance to SOD1. This region, found in the central 16-kDa portion of CCS, is postulated to serve in target recognition of SOD1 (25), whereas smaller segments at the N and C terminus of CCS are thought to facilitate the binding and release of copper into SOD1 (26). Homology between the central portion of CCS and SOD1 approaches 50% identity and 60% similarity (25). Based on this curious conservation of sequence, CCS was originally postulated to be "SOD4," the fourth mammalian SOD (Gen-Bank TM accession number 1608528).
The concordance of sequence between SOD1 and CCS was the focus of current studies. CCS molecules from diverse mammals contain all of the zinc binding ligands found in SOD1 and three of four histidine copper binding ligands; the fourth histidine is always substituted by an aspartate. We demonstrate here that the presence of this single aspartate prohibits CCS from functioning as a SOD. Substituting this aspartate with a histidine results in a CCS molecule with significant superoxide scavenging activity. Furthermore, this variant of CCS appears to be a bifunctional SOD, capable of self-activation with copper.
The vector for expression of human CCS in yeast cells is pCCS-HIS, containing the human CCS coding sequence under the control of the S. cerevisiae PGK1 (phosphoglycerol kinase) promoter. pCCS-HIS was constructed by mobilizing the PGK1-CCS fusion from pSMCCS (13) by digestion with BamHI and SalI and by insertion at these same sites into pRS423 (HIS3 2; Ref. 30). Human SOD1 was expressed in yeast by the pLC1 vector (URA3 2) where the SOD1 cDNA was placed under the control of the PGK1 promoter as described (31). pSM703 is the PGK1 promoter-containing vector (URA3 2) that served as the parent plasmid for construction of all human SOD1 and CCS expression plasmids. Plasmids pMR002 and pPS030 for the expression of D200H CCS * This work was supported by the Johns Hopkins University National Institutes of Health Center for Environmental Health Sciences and by National Institutes of Health Grant GM 50016 (to V. C. C.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ Supported by NIEHS, National Institutes of Health Training Grant ES07141.
Biochemical Analysis-Cell lysates were prepared essentially as described (13). Strains transformed with the appropriate plasmids were grown approximately 18 h to an A 600 of 1.5 in 50 ml of selecting SD medium (29); cells were harvested and were lysed by glass bead homogenization in 0.2-0.5 ml of a lysis buffer containing 10 mM NaPO 4 (pH 7.8), 0.1 mM EDTA, 0.1% Triton, 50 mM NaCl, 20 g/ml leupeptin, 10 g/ml pepstatin, and 1.0 mM phenylmethylsulfonyl fluoride. Glycerol was added to the final lysates at a concentration of 5%. As needed, lysates were concentrated by Microcon-10 Microconcentrator (Amicon) columns per manufacturer's instructions. For analysis of SOD activity, extracts were applied directly without boiling to a nondenaturing 12% precast polyacrylamide gel (Novex). Following electrophoresis, the gel was subject to nitro blue tetrazolium (NBT) staining for superoxide (32) in a 75-ml solution containing one tablet of NBT (10 mg/tablet), 50 mM KPO 4 (pH 7.8), 0.1 mg/ml riboflavin, and 1 l/ml TEMED. Western blot analyses of lysates were carried out as described (33, 34) using a denaturing 14% polyacrylamide gel. A polyclonal rabbit anti-human SOD1 serum (kind gift of D. Borchelt, Johns Hopkins University) and a polyclonal CCS antibody (obtained from Jeff Rothstein, Johns Hopkins University) were used at 1:5000 and 1:500 dilutions, respectively, and were detected using as secondary antibody, anti-rabbit donkey IgG (Amersham Pharmacia Biotech) conjugated to horseradish peroxidase. Visualization involved the Hybond ECL kit (Amersham Pharmacia Biotech).

RESULTS
The central region of human CCS encompassing amino acids 78 to 232 shares nearly 50% identity with human SOD1. All four of the zinc binding ligands of SOD1 and three of four histidine copper binding ligands are present in CCS; the fourth histidine is replaced by an aspartate residue, which is also a possible ligand for copper (35) (Fig. 1A). Interestingly, this precise pattern of homology to SOD1 is observed with murine CCS (36). Based on sequence analysis alone it would seem plausible that mammalian CCS should possess SOD activity.
To address this question in vivo, human CCS was expressed in a yeast strain lacking SOD1 and was assayed for the ability to overcome the oxidative damage of these cells. Yeast sod1⌬ mutants cannot synthesize lysine when grown in air due to oxidative damage of lysine biosynthetic component(s) (37)(38)(39). This defect is readily corrected by expression of the heterologous human SOD1 (Fig. 1C). However, expression of human CCS failed to complement the lysine auxotrophy of the sod1 mutant (Fig. 1C) even though the protein accumulated to substantial levels (Fig. 1B). By a NBT gel assay for SOD activity (32), CCS exhibited no detectable scavenging of superoxide ( Fig. 2A, lane 5). Therefore, despite the exquisite homology between SOD1 and human CCS, the copper chaperone appears nonfunctional for SOD activity.
The most noteworthy difference between SOD1 and the SOD1 homology domain of human CCS is the substitution of a copper binding histidine with an aspartate (Fig. 1A). To test whether this single amino acid variance precludes SOD activity, the aspartate at CCS position 200 was mutated to a histidine. By Western blot analysis, D200H CCS accumulated to wild type levels when expressed in yeast (Fig. 1B). Surprisingly, expression of this mutant CCS molecule effectively complemented the yeast sod1 mutation (Fig. 1C), suggesting that this metallochaperone had acquired SOD activity. D200H CCS was directly examined for superoxide scavenging activity by the NBT gel assay. These studies were conducted in the back-

FIG. 1. Complementation of a yeast sod1⌬ mutation by a mutant variant of human CCS.
A, an amino acid sequence alignment of the central region of human CCS (residues 78 through 232) and human SOD1. Arrows demarcate hSOD1 copper binding ligands, and ϩ marks zinc binding ligands in hSOD1. Identical and similar residues are denoted by asterisks and dots, respectively, and mutations introduced are indicated at positions above (for CCS) or below (for SOD1) the amino acid sequence. B, Western blot analysis of human CCS expression in yeast. Cell lysates were prepared from the sod1⌬ strain VC107 transformed with either vector pSM703 (Vec), pCCS-His expressing wild type CCS (WT), or with pMR002 expressing D200H CCS. Cell lysates from these strains were subjected to polyacrylamide gel electrophoresis and Western blot analysis using an anti-CCS polyclonal antibody. C, complementation of the aerobic lysine auxotrophy of the sod1⌬ strain. The cells as described in B were spotted onto minimal medium lacking lysine and were allowed to grow aerobically for 4 days. Dilutions represent 2 ϫ 10 5 and 4 ϫ 10 4 cells plated.

FIG. 2. Superoxide dismutase activity associated with D200H CCS.
A, the sod1⌬ sod2⌬ strain KS100 was transformed where indicated with either empty vector, wild type (WT) CCS, or D200H CCS, and cell lysates were examined for superoxide dismutase activity by the NBT gel assay (32). Amount of crude extract protein examined: lane 1, 250 g; lanes 2-4, 50, 150, and 250 g; lane 5, 250 g. B, immunodepletion of D200H CCS. A 100-g protein sample from sod1⌬ sod2⌬ cells expressing D200H CCS was subjected where indicated to immunoprecipitation with preimmune serum or a polyclonal anti-CCS antibody. Supernatants were examined for SOD1 activity by the NBT gel assay. Control, 100-g sample not subjected to immunoprecipitation. ground of a sod1⌬ sod2⌬ strain devoid of any endogenous SOD activity (Fig. 2A, lane 1). As seen in Fig. 2A, a dose response of superoxide scavenging activity was detected in lysates from cells expressing D200H CCS. 2 To establish that the mutant CCS molecule was indeed responsible for this activity, cell lysates were subjected to immunoprecipitation with either an anti-CCS antibody or preimmune control sera prior to analysis of superoxide scavenging activity. As seen in Fig. 2B, treatment with anti-CCS depleted all activity from the cell lysates, demonstrating that D200H CCS is capable of scavenging superoxide.
The gain of activity observed with D200H CCS indicated that aspartate 200 in this molecule is sufficient to prohibit SOD activity. To confirm this, we conducted the reverse mutagenesis experiment in which the fourth histidine copper ligand in human SOD1 was substituted with an aspartate (Fig. 1A). The resultant H120D SOD1 molecule accumulated to significant levels when expressed in the sod1⌬ strain and like other mutant alleles of human SOD1 (31,40,41), exhibited an altered mobility on denaturing gels (Fig. 3A). As seen in Fig. 3B, H120D SOD1 was nonfunctional in complementing the aerobic lysine auxotrophy of the sod1⌬ mutant. Furthermore, when more directly visualized by the NBT gel assay, the H120D SOD1 mutant exhibited no detectable SOD activity (Fig. 3C). The strong inhibitory effect of the H120D mutation on SOD1 activity supports the notion that the aspartate residue at CCS position 200 is sufficient to prevent this molecule from functioning as a SOD.
We next tested whether amino acid Asp 200 in human CCS evolved to facilitate copper transfer to SOD1. To monitor the metallochaperone activity of CCS, wild type and D200H CCS molecules were expressed in a lys7⌬ null strain lacking the yeast CCS molecule (Fig. 4A). Yeast SOD1 is still present in lys7⌬ strains but is normally inactive because the enzyme is apo for copper (13,15,42). However, yeast SOD1 can be fully activated by expression of the human CCS metallochaperone (Fig. 4B). By direct monitoring of SOD1 activity, the wild type and D200H mutant CCS molecules appeared equally capable of charging yeast SOD1 with copper (Fig. 4B). Compared with the experiment of Fig. 2, this study utilized a low level of cell lysate to minimize interference from the superoxide scavenging activity of D200H CCS. To confirm that the activity observed in Fig.  4B indeed reflected SOD1 and not D200H CCS, an immunodepletion experiment was conducted. As seen in Fig. 4C, the activity observed with cells coexpressing SOD1 and D200H CCS was immunodepleted with an antibody directed against yeast SOD1 and not human CCS. Thus, D200H CCS retains full activity as a metallochaperone for SOD1.
Because the D200H CCS molecule retains its function as a copper chaperone and also exhibits superoxide scavenging activity, is it possible that this SOD-like molecule can act as its own metallochaperone? To address this, the mutant CCS molecule was expressed in a strain lacking both SOD1 and the yeast copper chaperone LYS7. This strain cannot grow on medium lacking lysine; however, the D200H CCS mutant rescued oxidative damage and supported lysine independent growth (Fig. 5A). By the NBT gel assay, it is evident that D200H CCS is capable of superoxide scavenging even in the absence of the yeast copper chaperone LYS7 (Fig. 5B, lanes 2-4). Therefore, the D200H mutant of CCS appears to self-activate itself for superoxide scavenging activity. DISCUSSION The mammalian CCS metallochaperone has evolved with a domain exhibiting remarkable homology to its target of copper delivery, SOD1. Based on this high degree of sequence identity, SOD activity would not have been an unreasonable assump- FIG. 3. Inactivation of human SOD1 by a H120D mutation. The sod1⌬ strain VC107 was transformed with either pSM703 (Vec), pLC1 expressing wild type human SOD1 (WT), or with pPS030 expressing the human SOD1 H120D mutant polypeptide (H120D). A, cell lysates were subjected to polyacrylamide gel electrophoresis and Western blot using an anti-human SOD1 polyclonal antibody. B, transformants were tested for complementation of the sod1⌬ lysine auxotrophy by spotting onto minimal medium containing (ϩLYS) or lacking lysine (ϪLYS) as described in Fig. 1. C, 20 g of crude extract protein was subjected to analysis of SOD activity by the NBT gel assay described in Fig. 2B. The positions of human SOD1 and the endogenous yeast SOD2 are indicated.

FIG. 4. Metallochaperone activity of the D200H mutant of CCS.
The lys7⌬ strain SY2950 was transformed with plasmids for expression of wild type and D200H mutant CCS as in Fig. 1. A, Western blot analysis of CCS expression as in Fig. 1B. B, 20 g of crude extract protein was subjected to analysis of SOD activity by the NBT gel assay described in Fig. 2B. C, 20 g of crude extract protein was subjected where indicated to immunoprecipitation with either preimmune serum, the anti-CCS antibody, or the anti-yeast SOD1 antibody, and the resultant supernatants were assayed for SOD activity as in B. Control, extract protein not subjected to immunoprecipitation. Positions of yeast SOD1 and SOD2 are indicated. tion. In fact, a number of copper-containing complexes have been shown to scavenge superoxide (43)(44)(45). However, CCS exhibits no detectable superoxide scavenging activity, and prohibition of SOD activity is in part accomplished by the presence of a single aspartate residue (Asp 200 ) at the putative copper site. Substitution of this aspartate with histidine helps to reconstruct the four-histidine copper site of SOD1 and is sufficient to unlock the superoxide scavenging capacity of CCS.
It is curious that mammalian CCS has evolved to look so much like SOD1 without retaining SOD activity. Presumably, this striking homology is required for CCS recognition of SOD1. SOD1 normally exists as a homodimer (46), and the formation of a transient heterodimer (or multimer) between enzyme and CCS metallochaperone may initiate metal transfer. Physical interaction between SOD1 and the SOD homology domain of CCS has in fact been demonstrated by Gitlin and colleagues (25). Because metal binding can affect the conformation of SOD1 (42), the presence of a SOD1-like copper site in CCS may contribute to the overall domain structure needed for target recognition. It is not known whether copper binds at this site, yet if this were the case, the metal appears incapable of the redox cycling needed for superoxide scavenging. As an added advantage, abrogation of copper redox chemistry should also prohibit the deleterious peroxidase activity typical of SOD1 (18,19).
Based on homology to SOD1, there are four possible copper ligands in the central domain of CCS, and of these, placement of an aspartate at position 200 seems best suited to preclude redox damage to the protein. Uchida and Kawakishi (47) have shown that at the analogous position in human SOD1 (His 120 ) the histidine is highly susceptible to oxidation, whereas the remaining three histidine copper ligands are not. Hence, the presence of Asp 200 in CCS appears perfectly designed to prohibit self-oxidation of the copper chaperone or oxidation of its intimate partner, SOD1.
Although mammalian CCS exhibits great homology to SOD1, far less homology is evident when one compares the central domain of yeast CCS to that of yeast SOD1 (approximately 25% identity at the amino acid level). Furthermore, as revealed through x-ray crystallographic studies by Rosenzweig and co-workers (48), the central domain of yeast CCS is devoid of a metal binding cavity, even though the overall structure is remarkably similar to SOD1. The loss of this metal binding site in yeast CCS may reflect a unique requirement for activation of yeast SOD1. Valentine and co-workers (42) have demonstrated that unlike mammalian SOD1, which forms a symmetrical homodimer, the yeast enzyme adopts an asymmetrical conformation in which there is unequal metallation of the two subunits (42). As fungal SOD1 diverged from the human enzyme, it is likely that their respective metallochaperones evolved concomitantly to conform to specific metallation requirements.
Why do separate molecules exist for SOD1 and its metallochaperone? Our studies here with human D200H CCS indicate that it is possible to manufacture a self-sufficient SOD molecule that can charge itself with copper. Although more energy-consuming, the synthesis of separate molecules for superoxide scavenging and metal incorporation may allow for control of SOD1 activity at the post-translational level. In general, the SOD1 polypeptide is ubiquitously expressed, but the fraction of apo-to holoSOD1 can vary greatly among cell types and tissues (49 -52). As such, CCS represents a potentially effective means for rapidly controlling SOD1 activity in response to cellular needs.