Identification of S100b Protein as Copper-binding Protein and Its Suppression of Copper-induced Cell Damage*

We have isolated from bovine brain a protein with a high capacity to inhibit the copper ion-catalyzed oxidation ofl-ascorbate and identified it as S100b protein, an EF-hand calcium-binding protein, by sequencing its proteolytic peptides. Copper binding studies showed that this protein has four copper-binding sites per dimeric protein molecule with a dissociation constant of 0.46 μm and that in the presence ofl-ascorbate, copper ions bind to a total of six binding sites with a great increase in affinity. Furthermore, we examined whether S100b protein can prevent copper-induced cell damage. Bovine S100b protein was found to suppress dose-dependently the hemolysis of mouse erythrocytes induced by CuCl2. We transformed Escherichia coli cells with pGEX-5X-3 vector containing a cDNA for rat S100b protein, so that this protein could be expressed as a fusion protein with glutathioneS-transferase. The transformed cells were demonstrated to be markedly resistant to a treatment with CuCl2 plus H2O2 as compared with the control cells expressing glutathione S-transferase alone. These results indicate that S100b protein does suppress oxidative cell damage by sequestering copper ions.

The transition metal copper is essential to a variety of cell functions; however, intake of its excess are toxic to living organisms. The toxicity is thought to result, at least in part, from the Fenton or Haber-Weiss reaction, in which copper ions would catalyze the formation of OH ⅐ radical from H 2 O 2 (1). OH ⅐ radical is a powerful oxidant that can damage various cellular components, such as lipid, nucleic acid, and protein. Therefore, copper ions in vivo should be sequestered such that their redox cycling could be hampered, thereby copper ions being unable to participate in the production of OH ⅐ . Considering the occurrence of 0.1-10 mg of copper/100 g of wet tissue or body fluid in human body, the copper sequestration should be recognized as an important mechanism for prevention of oxidative damage.
Under normal conditions copper ions may be tightly bound to proteins, but there are certain oxidant stress conditions where copper ions are released from the tightly binding sites of proteins and become redox-active, e.g. ischemia-reperfusion injury (2), fulminant hepatic inflamation in Wilson's disease (3), Parkinson's disease (4), and rheumatoid arthritis (5). Sequestration of copper ions is also important for prevention of L-ascor-bate oxidation that is extremely accelerated by free copper ions (6,7).
To understand the importance of copper sequestration in prevention of oxidant-induced tissue damage, we have initiated a study to isolate copper-binding proteins from bovine brain and found that S100b protein, an EF-hand calcium-binding protein (8), was one such protein. As a matter of fact, S100b protein was observed to prevent copper-induced hemolysis. By expressing recombinant S100b protein in Escherichia coli cells, we also demonstrated that the S100b protein markedly increased the resistance to H 2 O 2 in the cells that had previously been treated with copper salt.

EXPERIMENTAL PROCEDURES
Materials-Commercial sources of materials are as follows: Sephadex G-75 and pGEX-5X-3 vector from Pharmacia Biotech, Uppsala, Sweden; DE52 cellulose from Whatman International, Ltd., Madison, United Kingdom; pGEM-3Zf(ϩ) vector from Promega Corp.; Pfu DNA polymerase from Stratagene, La Jolla, CA; synthetic oligonucleotides and a SuperScript preamplification system from Biotechnologies, Inc., Gaithersburg, MA; restriction enzymes from New England Biolabs, Beverly, MA; E. coli strain BL21(DE3) pLysS from Novagen, Inc., Madison, WI; Achromobacter protease 1 (lysyl endopeptidase) from Wako Pure Chemical Industries, Ltd., Osaka, Japan; and bovine S100b protein from Sigma. The S100b protein was dissolved in 20 mM Tris-HCl (pH 7.5), and the trace EDTA included was removed by the desalting procedure using a Microcon-3 concentrator (Amicon, Inc., Beverly, MA). Its concentration was estimated in view of the facts that the protein has absorbance of 1.0 at 280 nm at a concentration of 5.4 mg/ml (9) and that its dimeric molecular weight is 21,000 (10). All other chemicals were of analytical grade. Distilled water was purified by passage through a Milli-Q water purification system (Millipore Corp., Bedford, MA). The resistance of the water was 1.8 ϫ 10 7 ⍀ ⅐ cm at 20°C.
Measurement of Copper Binding Capacity-Copper ions in the micromolar range linearly increases the rate of L-ascorbate oxidation, and the increase in the rate is suppressed in proportion to the amount of EDTA, a copper chelator, added. This phenomenon is the basis for the measurement of copper binding capacity that is obtainable from the degree of suppression of the copper ion-catalyzed L-ascorbate oxidation. The standard reaction mixture contained 20 mM Tris-HCl (pH 7.5), 2.0 M CuCl 2 , 50 M L-ascorbate, and sample to be tested for copper binding. Oxidation of L-ascorbate was measured spectrophotometrically at 265 nm at 20°C with a Shimadzu UV-visible recording spectrophotometer UV-160 (Shimadzu, Kyoto, Japan).
Purification of Copper-binding Protein-The purification procedure employed is in principle the same as used to purify copper-containing proteins by Sharoyan et al. (11). The following procedures were carried out at 4°C unless otherwise stated. Acetone powder (45 g) prepared from bovine brain was extracted with 10 mM sodium acetate buffer (pH 6.0). The extract (350 ml) was applied onto a DE52 cellulose column (5 ϫ 12 cm) that had been equilibrated with 10 mM sodium acetate buffer (pH 6.0). The column was washed with 760 ml of 100 mM sodium phosphate buffer (pH 6.0), and eluted with a 800-ml linear gradient, from zero to 1.0 M, of NaCl containing 100 mM sodium acetate buffer (pH 6.0). Fractions of 2.3 ml were collected, and those containing copper binding capacities above 43 nmol of copper bound/mg of protein were combined and concentrated by use of Centriprep-3 concentrators (Amicon, Inc.). The resulting solution was applied onto a Sephadex G-75 column (3 ϫ 100 cm) that had been equilibrated with 100 mM sodium * This work was supported in part by Grant-in-aid 08457055 for Scientific Research from the Ministry of Education, Science, and Culture of Japan. 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.
‡ To whom all correspondence should be addressed. Tel.: 81734-26-8316; Fax: 81734-23-7794; E-mail: nishikim@seishu.wakayamamed.ac.jp. acetate buffer (pH 6.0), and then the column was eluted with the same buffer. Fractions of 2.3 ml were collected, and those containing high specific copper binding capacities were combined and concentrated. Portions (100 l) of the pooled fractions were further subjected to reverse-phase high performance liquid chromatography (HPLC) 1 on a Shodex RSpak D18-613 column (6 ϫ 150 mm, Showadenko, Tokyo) with a linear gradient, from 0 to 70%, of acetonitrile containing 0.1% trifluoroacetic acid at a flow rate of 1.0 ml/min, and the elution was followed spectrophotometrically at 230 nm.
Amino Acid Sequencing Analysis-The purified protein was digested with Achromobacter protease 1 in 10 mM Tris-HCl (pH 9.0) at 37°C for 15 h. The resulting peptides were separated by reverse-phase HPLC as described above. Two peptides were sequenced using an Applied Biosystems 477A gas-liquid phase protein sequencer equipped with an on-line 120A PTH analyzer.
Copper Binding Study by Ultrafiltration Technique-Solutions containing 20 mM Tris-HCl buffer (pH 7.5), 0.87 M bovine S100b protein, and varying concentrations of CuCl 2 (1.25-10 M) with or without 50 M L-ascorbate were dispensed into the cups of Ultrafree-MC filter units (10,000 NMWL, Millipore Corp.). The mixtures (200 l) were prepared by final addition of CuCl 2 . After the filter units were allowed to stand at 20°C for 5 min, they were centrifuged at 4,300 ϫ g at 5°C for 10 min, and the copper concentration of the resulting filtered solutions was measured by the flameless mode with a Shimadzu atomic absorption spectrophotometer AA-646.
Erythrocyte Preparation and Hemolysis-Erythrocytes were collected in heparin from a mouse. Then they were spun down and washed twice with phosphate-buffered saline. The percent hemolysis of the erythrocytes (1.5% hematocrit) in phosphate-buffered saline was determined essentially as described by Caffrey et al. (12). The absorbance at 410 nm for osmotically hemolyzed cells was compared with that for copper-induced hemolysis of cells.
Construction of an Expression Plasmid for Production of a Fusion Protein of Rat S100␤ Subunit with Glutathione S-Transferase-Rat cerebellum RNA was prepared by the method of Chomczynski and Sacchi (13), and a first strand cDNA was synthesized from this RNA with random hexamers as primers. The synthesis was carried out by use of a SuperScript preamplification system according to the manufacturer's manual. A rat S100␤ subunit cDNA was amplified from the cDNA by polymerase chain reaction with a pair of primers that had been synthesized on the basis of the reported nucleotide sequence of rat S100␤ subunit cDNA (14). The primers were a sense primer (5Ј-AGCT-TCTCTGTCTACCCTCCT-3Ј) and an antisense primer (5Ј-CGGGGT-CAGAGTTTCATCCC-3Ј), corresponding to the sequence from nucleotide 38 to 58 and that from nucleotide 548 to 568, respectively. Polymerase chain reaction was carried out in a MiniCycler (MJ Research, Inc., Watertown, MA) with Pfu DNA polymerase under the same conditions as described previously (15) except that the temperature of the annealing step was 55°C. After isopropyl alcohol precipitation, the amplified cDNA was ligated into the HincII site of pGEM-3Zf(ϩ). Further polymerase chain reaction was carried out, with the resulting construct as a template, for preparation of a cDNA fragment to be inserted into the mutiple cloning site of the prokaryotic expression vector pGEX-5X-3. The sense primer was 5Ј-GGGGATCCGGATGTCT-GAGCTGGAGAAG-3Ј, which would generate a BamHI site at the 5Ј end and contains the nucleotide sequence surrounding the initiation codon (nucleotide 119 to 138 of the reported cDNA), and the antisense primer was 5Ј-CGTTGTAAAACGACGGCCAG-3Ј, which corresponds to a part of the LacZ␣ sequence within the vector. The polymerase chain reaction product was digested with BamHI and ligated into the BamHI site of pGEX-5X-3, and the resulting plasmid with the correct orientation (designated pGEX-5X-3/S100␤) was selected by restriction mapping. The final construct can produce a fusion protein of glutathione S-transferase and rat S100␤ subunit.
Copper Resistance by Expression of S100b Protein in E. coli Cells-E. coli cells of the strain BL21(DE3) pLysS were transformed with pGEX-5X-3/S100␤ or pGEX-5X-3, and the resulting respective recombinant clones were cultured at 37°C in LB containing 50 g/ml ampicillin. When the A 600 of the culture became about 0.8, isopropylthio-␤-Dgalactopyranoside was added to a concentration of 1.0 mM, and the cells were cultured at 30°C for 2 h, at which time CuCl 2 was added to a concentration of 1 mM and further cultured at 37°C for 30 min. Then the cells were spun down, washed twice with a solution containing 25 mM Tris-HCl (pH 7.4), 136 mM NaCl, and 2.6 mM KCl and suspended in phosphate-buffered saline. At this point, A 600 of both test and control E. coli cells were adjusted to 1.0 in phosphate-buffered saline, and they were incubated at 37°C for 30 min with or without 1, 2, or 4 mM H 2 O 2 . The number of viable cells was counted after spreading an appropriate portion of the cell suspensions on LB plates containing 50 g/ml ampicillin and culturing overnight.
Analytical Procedures-Protein was assayed by the biuret method (16) in a small scale. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis was carried out by the method of Kadenbach et al. (17), with the exception that the concentration of the separation gel was 12.5%.

Identification of Bovine S100b Protein as a Copper-binding
Protein-Acetone powder of bovine brain was extracted with sodium acetate buffer, and the resulting solution was subjected to DE52 cellulose column chromatography and then to Sephadex G-75 gel chromatography. The purification of copperbinding proteins was followed in terms of copper binding capacity as measured by their inhibitory effect on the coppercatalyzed oxidation of L-ascorbate. The elution profile of the gel chromatography is shown in Fig. 1. A great majority of copper binding capacity was eluted in fractions containing large amounts of protein, and a peak of copper binding capacity, although it was small, appeared in the following fractions whose protein concentrations were very low. The latter fractions with relatively high specific copper binding capacities were combined and further purified by reverse-phase HPLC. Upon elution with a gradient, from 0 to 70%, of acetonitrile containing 0.1% trifluoroacetic acid, there appeared one major and one minor peak at acetonitrile concentrations of 51 and 55%, respectively. Since the major peak showed much higher copper binding capacity than the minor one, we decided to determine the amino acid sequence of the major peak protein.
The protein was digested with lysyl endopeptidase, and the resulting peptides were separated by reverse-phase HPLC by elution with the same acetonitrile gradient as used for the protein purification. Two of the peptides eluted at acetonitrile concentrations of 32 and 38% were isolated, and their amino acid sequences were determined to be AVVALIDVFHQYS-GREGDK and ELINNELSHFLEEIK, respectively. Computer search for these sequences in the available data bases revealed complete coincidence with the partial amino acid sequences (residue 6 -24 and 60 -71, respectively) of bovine S100␤ subunit (10), which is a constituent of S100a protein (␣␤ heterodimer) and S100b protein (␤␤ homodimer). This identification was confirmed by reverse-phase HPLC: the elution time of the authentic S100b protein (commercially available purified pro- 1 The abbreviation used is: HPLC, high performance liquid chromatography. tein) was the same as that of the major peak of the chromatogram of the final step of the purification of the copper-binding protein.
Studies on Copper Binding with S100b Protein-A quantitative experiment on copper binding with bovine S100b protein was carried out by an ultrafiltration technique. Equilibrium binding was observed when solutions containing 0.87 M S100b protein and varying concentrations of CuCl 2 (1.25-10 M) were analyzed (Curve I in Fig. 2A). A Scatchard plot of the data showed that 3.8 atoms of copper bound to a molecule of dimeric protein with a dissociation constant of 0.46 M (Fig. 2B). Interestingly, in the presence of L-ascorbate, 6.5 atoms of copper bound to a molecule of the protein with a higher affinity (Curve II in Fig. 2A). The same strong copper binding was observed when the rate of oxidation of L-ascorbate was measured in the presence of a fixed concentration of S100b protein and varying concentrations of CuCl 2 (Fig. 3). Addition of CuCl 2 exceeding the binding capacity gave a linear increase in L-ascorbate oxidation, the slope of which was essentially the same as that observed by the addition of CuCl 2 to the L-ascorbate solution minus S100b protein. From the result, the number of copper ions bound per molecule of S100b protein was estimated to 5.9. The results of these binding studies indicate that S100b protein has four copper-binding sites per molecule and that the number of copper-binding sites per molecule is increased to six by the presence of L-ascorbate.
Since Ca 2ϩ and Zn 2ϩ are known to bind to S100b protein (18), competition of these metal ions with Cu 2ϩ was examined by observing whether L-ascorbate oxidation would be enhanced by Cu 2ϩ to be released from a Cu 2ϩ -S100b protein (2:1) complex. Addition of 10 mM CaCl 2 did not accelerate the L-ascorbate oxidation, while addition of 1.0 mM ZnSO 4 caused a release of about 60% of the bound Cu 2ϩ as measured by the increase in the rate of L-ascorbate oxidation. These results indicate that the binding constant of S100b protein for copper ions is far larger than that for Ca 2ϩ , if Cu 2ϩ binds to the binding sites for Ca 2ϩ , and that Zn 2ϩ and Cu 2ϩ probably share at least a part of the binding sites.
Inhibition of Copper-induced Hemolysis by S100b Protein-We examined whether S100b protein can inhibit copperinduced hemolysis of mouse erythrocytes (Fig. 4). When erythrocytes were incubated at 37°C in phosphate-buffered saline containing 30 M CuCl 2 , the degree of hemolysis after 2.5-h incubation was 50%. Addition of 1.1 M S100b protein caused about 50% inhibition of the hemolysis, and almost total inhibition was achieved by addition of 2.3 M S100b protein. The degree of hemolysis with 40 M CuCl 2 reached about 90% after 2-h incubation, and S100b protein caused about 50% inhibition of the hemolysis at a concentration of 2.3 M, while it had almost no inhibitory effect at a concentration of 1.1 M. These results clearly indicate that S100b protein inhibits the hemolysis by sequestering copper ions.
Copper Resistance of E. coli Cells by Expression of Recombinant Rat S100b Protein-A cDNA for rat S100␤ subunit was generated by reverse transcriptase-polymerase chain reaction from rat cerebellum mRNA and inserted into the bacterial expression vector pGEX-5X-3 such that the S100␤ subunit could be expressed as a fusion protein with glutathione Stransferase. Cells of E. coli BL21(DE3) pLysS were transformed with the resulting expression plasmid and checked for production of the fusion protein. It was evident that a substantial amount of the fusion protein was produced by induction with isopropylthio-␤-D-galactopyranoside (Fig. 5). The test for copper resistance of E. coli cells expressing S100b protein was carried out as follows. The cells possessing pGEX-5X-3/S100␤ were induced to produce the fusion protein, treated with 1.0 mM CuCl 2 , and then further treated with varying concentrations of H 2 O 2 . E. coli cells transformed with pGEX-5X-3 encoding glutathione S-transferase alone was used as a control for comparison; they produced the enzyme in an amount comparable to that of the fusion protein produced by the cells transformed with pGEX-5X-3/S100␤ (Fig. 5). The pretreatment with copper ions alone gave only a slight difference in the cell viability between the test and control cells. However, the cells expressing S100b protein were markedly resistant to the subsequent treatment with H 2 O 2 at all the concentrations tested (1-4 mM) (Fig. 6). DISCUSSION The results presented in this study clearly demonstrated that S100b protein, an EF-hand calcium-binding protein, can prevent the oxidation of L-ascorbate catalyzed by copper ions. This prevention is due to the sequestration of free copper ions: four copper ions bind per dimeric molecule of bovine S100b protein with a dissociation constant of 0.46 M and in the presence of L-ascorbate, six copper ions bind per molecule with a higher affinity, as demonstrated in this study (Figs. 2 and 3).
The enhancement of copper binding by L-ascorbate may be interpreted to suggest that Cu 1ϩ formed by the reduction of Cu 2ϩ by L-ascorbate has a higher affinity to S100b protein and binds additionally to two sites, but the elucidation of the underlying mechanism of this phonomenon awaits further investigation.
Since addition of copper salt to an erythrocyte suspension is known to cause hemolysis (12), we tested whether S100b protein can prevent this phenomenon and found that it was really the case (Fig. 4). S100b protein at a concentration of 2.3 M was able to almost totally inhibit the hemolysis of mouse erythrocytes (1.5% hematocrit) induced by 30 M CuCl 2 . Assuming the above copper binding ratio of four for S100b protein (without L-ascorbate), the S100b protein present in the reaction system could sequester only 31% of the copper ions, yet it could almost completely inhibit the hemolysis. Under these conditions, the remaining concentration of CuCl 2 (21 M) after sequestration by S100b protein appears to be not enough for induction of hemolysis. In fact, 20 M CuCl 2 was almost ineffective to induce hemolysis (data not shown). In relation to the suppression of copper-induced cell damage by S100b protein, it should be noted that serum albumin, known to tightly bind copper ions (19), likewise exerts inhibitory effects on various oxidative processes involving copper ions. Among such processes are L-ascorbate oxidation catalyzed by copper salt (20) and the hemolysis induced by copper salt alone (12) or by copper salt plus Lascorbate (21).
Culotta et al. (22) have recently reported that the copper toxicity in yeast lacking metallothionein was suppressed by overexpression of copper/zinc superoxide dismutase. In this study, the function of superoxide dismutase was ascribed not to its superoxide scavenging activity but to its role in copper sequestration. We intended to examine whether the copper toxicity can be reduced by expressing S100b protein extraneously in E. coli cells, and actually found that the cells manipulated to express rat S100b protein (Fig. 6) were resistant, to a great degree, to copper-induced oxidative damage in comparison with the control cells not expressing S100b protein. Copper ions would be reduced by reducing substances in cells, and the resulting Cu 1ϩ would be oxidized with H 2 O 2 , producing highly reactive OH ⅐ radical. This redox activity of copper ions may be suppressed by their binding to S100b protein. Even if the FIG. 5. Production of rat S100␤ subunit as a fusion protein with glutathione S-transferase in E. coli cells. Cells of E. coli BL21(DE3) pLysS, transformed with pGEX-5X-3/S100␤, were cultured in the presence (ϩ) or absence (Ϫ) of 1 mM isopropylthio-␤-D-galactopyranoside at 30°C for 2 h to produce the fusion protein. The cells were disrupted by sonication, and the proteins (20 g) were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The major band on each lane marked by ϩ represents the inducibly expressed protein.
FIG. 6. Copper resistance of E. coli cells expressing recombinant rat S100␤ subunit. E. coli cells were allowed to produce a fusion protein of S100␤ subunit with glutathione S-transferase as specified in the legend of Fig. 5 (open bar), and then treated with 1 mM CuCl 2 at 37°C for 30 min. Then the cells were suspended in phosphate-buffered saline (A 600 ϭ 1.0) and incubated at 37°C for 30 min in the presence or absence of the indicated concentrations of H 2 O 2 . The number of viable cells was counted after inoculation of 200 l of each of 10,000-fold diluted cell suspensions. As a control, E. coli cells transformed with pGEX-5X-3 (solid bar) were carried through the same procedure. Data indicate mean Ϯ S.D. of three determinations. *, p Ͻ 0.001 by Student's t test.
copper ions bound to S100b protein were otherwise prone to react with H 2 O 2 and damage the protein itself, loss of the extraneous protein would not be harmful. S100b protein is relatively abundant in the brain, amounting to approximately 2% of the soluble protein of its certain regions (23); therefore, it is possible that S100b protein as a copperbinding protein plays a role in copper hemeostasis as well as it prevents the copper-induced oxidative damage in the brain. In this regard, a recent finding should be noted that ␤A4 amyloid precursor protein reduces Cu 2ϩ to Cu 1ϩ and possibly contributes to neurodegeneration in Alzheimer's disease through generation of reactive oxygen species (24). It is also reported that the copper level of the cerebrospinal fluid is significantly elevated in patients of Parkinson's disease (25). Therefore, it seems likely that S100b protein intervenes the pathogenic processes of these diseases.