Purification and Characterization of Ag , Zn-Superoxide Dismutase from Saccharomyces cerevisiae Exposed to Silver *

Cu,Zn-superoxide dismutase plays an important role in protecting cells from oxygen toxicity by catalyzing the dismutation of superoxide anion into hydrogen peroxide and oxygen. In Saccharomyces cerevisiae Cu,Znsuperoxide dismutase is coregulated with copper-thionein by copper via the transcription factor ACE 1. We demonstrate here that presence of &NO, in the culture medium leads to a five times increase of Cu,Zn-superoxide dismutase mRNA, with a concomitant six times decrease of the enzyme activity. Susceptibility of yeast to silver was apparently inversely related to Cu,Zn-superoxide dismutase activity. From silver-treated yeast a Cu,Zn-superoxide dismutase with impaired dismutase function was purified and was shown to contain silver, which was located to the copper sit .These data suggest that Cu,Zn-superoxide dismutase may play an additional direct role in the defense of S. cerevisiae against metal stress by functioning as metal chelator.

The yeast Saccharomyces cerevisiae contains two species of superoxide dismutase (EC 1.15.11), i.e. the copper, zinc-, (Cu,Zn-SOD)l a n d the manganese-containing forms. The former one is localized in the cytosol while the manganese enzyme is restricted to the mitochondrial matrix. The three-dimensional structure and the mechanism of action of these enzymes are well characterized (1,2), but their biological properties are still a matter of debate. It has been demonstrated that these enzymes play an important role in protection against damage related to 0, toxicity. In fact, mutants that are defective in either Cu,Zn-or manganese-SOD are not able to grow in the presence of increased oxygen concentration or of compounds that generate oxygen radical intermediates (3,4). Both in yeast and in Escherichia coli manganese-SOD is able to change its level in the cell, depending on oxygen concentration of the medium (5,6) or on the exposure to redox active molecules in the presence of oxygen (7,8). In yeast several results indicate that Cu,Zn-SOD responds in a similar fashion to such stimuli (9, lo), in line with its proposed role as a major antioxidative enzyme of the cell.
In S. cerevisiae Cu,Zn-SOD and manganese-SOD behave differently in response to other stress factors such as exposure to copper (11). Cu,Zn-SOD has been reported to increase its ac-* This work was supported in part by the Consiglio Nazionale delle Ricerche Special Projects Chimica fine and CT 04. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ To whom correspondence should be addressed: , Dept. of Biology, University of Rome Tor Vergata, via della Ricerca Scientifica, 00133 Rome, Italy.
The abbreviations used are: SOD, superoxide dismutase; CUP, gene coding for copper-thionein; ACE 1, activator of CUP 1 expression; PAGE, polyacrylamide gel electrophoresis, tivity upon changes of the copper concentration in the medium, whereas manganese-SOD does not. The effects on Cu,Zn-SOD cannot be explained only on the basis of oxygen-dependent redox cycling of the metal ion because a copper-dependent increase of Cu,Zn-SOD is observed also under anaerobic conditions (12).
The presence of copper at the catalytic site of Cu,Zn-SOD suggests that copper availability limits the enzymatic activity, thus exerting a regulation at the post-transcriptional level. This is the case of anaerobic cultures of S. cerevisiae where the presence of a copper-free proenzyme requiring copper for activation has been demonstrated (12). Furthermore, copper can modulate yeast Cu,Zn-SOD at the transcriptional level via ACE 1 (131, a transcription factor which, upon binding of Cu(I), activates the promoter of CUP 1, the gene coding for copperthionein (14, 15). An ACE 1-binding site has been localized in the Cu,Zn-SOD promoter (131, and its ability to bind ACE 1 in vitro has been demonstrated (16).
These data point to a possible role of Cu,Zn-SOD in sharing a copper-sequestering function with metallothionein in yeast.
To further investigate this aspect, we have studied the effects of exposure to silver on yeast Cu,Zn-SOD. Ag(1) has physicochemical and electronic properties similar to copper (171, but it is not capable of generating reactive oxygen species by redox cycling. Moreover, silver has been shown to be able to induce metallothionein via the ACE 1 factor to a comparable extent as copper (13,18).

MATERIALS AND METHODS
Chemicals-Silver nitrate was purchased by Aldrich. Yeast extract was obtained from Difco. DE32 was from Whatman. Diphenylthiocarbazone was from Sigma.
Chelex 100, low molecular weight standard, and goat anti-rabbit horseradish peroxidase conjugate immunoblotting kit were obtained from Bio-Rad.
All other materials were of reagent grade and were obtained from the best commercial sources available.
Growth Conditions-Aculture grown overnight in basal medium was used as the inoculum. Cultures were grown aerobically in a rotatory shaker (Orbit Environ Shaker, LabLine Instruments) at 30 "C and 180 revolutions/min, with a flask volume/medium volume ratio of 4:l.
M I ) was supplied as silver nitrate to reach final concentrations of 25, 50, 100, and 200 PM from filter sterilized solution (Millex GV 0.22-pm filter unit, Millipore Corp., Bedford, MA) freshly made each time. As control yeast were also grown in the presence of 50 VM NaNO,. Cells were harvested from yeast cultures grown in the late exponential phase by centrifugation at 4 "C for 20 min at 2,500 x g. Cells were washed twice with ice-cold distilled water and stored at -20 "C until needed. Cellular extracts from each culture were prepared from wet cells by vortexing (30 s for five times with chilling between passes) in the presence of glass beads. The debris and the intact cells were removed by centrifugation at 23,000 x g for 30 min. Purification of Cu,Zn-Superoxide Dismutase from Silver-treated Yeast-Cu,Zn-SOD was purified from 500 g of yeast cells by the method of Goscin and Fridovich (19) with some modifications. The chromatographic step was modified utilizing a linear gradient of potassium phosphate (2.5-150 mM) at pH 7.8. Under these conditions the enzyme was eluted at a ionic strength between 40 and 60 m M buffer as four distinct peaks, instead of 20 m~ as a single peak (19). They were all active and were pooled, concentrated, and subsequently dialyzed overnight against 20 m M Tris-HC1, pH 7.8. The sample was then applied onto a Hiload Sepharose Q 16/10 column from Pharmacia LKB Biotechnology Inc. equilibrated with 20 m~ Tris-HC1, pH 7.8, and eluted with a linear NaCl gradient (0-200 I " ) . Fractions containing SOD activity were pooled and concentrated. Fast protein liquid chromatography was performed with a Pharmacia liquid chromatography unit. SOD activity was measured by a polarographic method (20) with an AMEL model 466 polarographic analyzer. The determinations were conducted with sodium tetraborate buffer at pH 9.8 which allows measurement only of the Cu,Zn isoform (21). Data were expressed in micrograms/milligrams with reference to purified yeast Cu,Zn-SOD. Manganese-SOD was assayed spectrophotometrically by the xanthinexanthine oxidase method in the presence of 2 m M KCN (22).
Proteins were determined by the method of Lowry et al. (23). X-band low temperature EPR spectra were recorded with a Varian E-9 spectrometer, equipped with a Stelar temperature controller, and interfaced to a Stelar Prometheus Data System. The detection limits of EPR spectroscopy under our experimental conditions are 10 -6 M Cu(I1).
The Ag,Zn-superoxide dismutase derivative was also prepared in vitro by addition of AgNO, at 1:l ratio with respect to the available copper-binding sites of the copper-free enzyme which was obtained as described previously (24). This derivative and the native protein give the same color yield with the Lowry reagent.
Northern Blot Analysis-Total RNA was prepared and subject to standard Northern blot analysis on nitrocellulose filters, using a synthetic oligonucleotide (+82 to +111) of the coding sequence of Cu,Zn-SOD as described elsewhere (13). Quantitation of the autoradiograms was obtained by densitometric scanning with a LKB Ultrascan XL Laser Densitometer coupled with an LKB 2400 Gel Scan software package. Data are given in arbitrary densitometric units, expressing normalized areas of integrated peaks from densitometric scanning.
Electrophoresis and Immunoblotting-Polyacrylamide gel electrophoresis was performed in a slab gel system from Bio-Rad. 7.5% acrylamide non-denaturing gels (25) were stained for total protein with Coomassie Blue R-250 or for SOD activity by the method of Beauchamp and Fridovich (26). SDS-PAGE (27) was performed at 12.5% acrylamide using, as molecular mass standards, proteins ranging from 97400 to 14400 Da. These gels were then stained with Coomassie Blue R-250 or subject to Western blot on nitrocellulose membrane with a Bio-Rad transblot apparatus by the method of Towbin et al. (28).
Antisera against yeast Cu,Zn-SOD were raised in rabbits and were a gift from Dr. Franco Marmocchi, University of Camerino, Italy. They proved to react efficiently with S. cerevisiae Cu,Zn-superoxide dismutase on nitrocellulose filter, giving a linear response between 10 and 150 ng of purified yeast Cu,Zn-superoxide dismutase.
Yeast Cu,Zn-SOD was localized on Western blots aRer incubation with the specific antisera by the horseradish peroxidase method. Quan- Metal content was determined with a Perkin Elmer 3030 atomic absorption spectrometer equipped with a graphite furnace. All solutions were prepared with Chelex 100-treated water, and acid-washed vials were used.
For determination of copper, silver, and zinc, the purified SODS from silver-treated and silver-untreated cells were diluted with 6 M HNO, and hydrolyzed for 16 h at room temperature.
For zinc determination, the samples were previously treated with 0.3% diphenylthiocarbazone in carbon tetrachloride to remove spurius zinc (30).

RESULTS
Effect of Silver on Cell Growth--Treatment of S. cerevisiae with AgNO, produced no significant change in the growth rate or cell morphology when silver concentration into the growth medium was kept below 25 1". When cells were grown in the presence of 25 or 50 p~ silver the time required to reach the stationary phase increased from 10 h, observed for S. cerevisiae grown in standard conditions, to 22 and 30 h, respectively (Fig.  1). Moreover, cell growth was completely inhibited at concentrations of AgNO, higher than 50 1". Resistance of yeast to silver was apparently related to Cu,Zn-superoxide dismutase activity. In fact, we found a high positive correlation between growth yield and enzyme activity (Fig. 2). Ag(1) toxicity was found t o be related to the reducing power of the medium. For this reason, to obtain reproducible growth curves, we added the metal a t a fxed time after medium sterilization. The effects observed in the presence of AgNO, were due to the Ag' ions. In fact the growth curve and the Cu,Zn-SOD activity of yeast grown in the presence of 50 p~ NaNO, were identical to those of control cells (not shown).
Copper content of the control cells was 0.27 & 0.07 nmoYlO' cells. Silver content of acid-digested cells treated with 50 1 " AgNO, was 0.14 2 0.06 nmol/lO' cells (the value is a mean of five determinations, each one done in triplicate). This metal fraction was not removed by washing the cells either with metal chelators or with acid and was assumed to be intracel-
Effect of Silver on Cu,Zn-Superoxide Dismutase-Silver has already been demonstrated to be capable of inducing Cu,Znsuperoxide dismutase mRNA transcription in yeast (13). In the present study treatment with 50 p~ AgNO, was shown to induce a significant increase of Cu,Zn-SOD mRNAwith a marked decrease in the Cu,Zn-superoxide dismutase activity which dropped by nearly one order of magnitude (Table I). Furthermore, the decrease in activity was not paralleled by significant changes in the protein level (Table I), as judged by Western blot analysis, also taking into account the observed decrease in antigenicity of the Cu,Zn-superoxide dismutase isolated from silver-treated yeast (see below).
It has previously been demonstrated that Cu,Zn-superoxide dismutase activity and the immunoreactive protein are roughly proportional to the copper content of cells, and in anaerobically growing cells a large fraction of the enzyme is in the form of an inactive protein, which is reactivated by the addition of copper (12). Therefore homogenates from cells grown in 50 p~ AgNO, were incubated with CuSO,. As shown in Table I, Cu,Zn-SOD activity increased significantly in 23,000 x g supernatants of silver-treated cells after incubations with 1 mM CuSO, and subsequent extensive dialysis, never reaching, however, the value obtained with control cells. Furthermore, native electrophoresis gels of extracts prepared from cells growing in the presence of silver showed multiple isophorms of Cu,Zn-superoxide dismutase, with increased mobilities (Fig. 3).
Treatment with 50 p~ AgNO, does not affect the activity of manganese-SOD which resulted in 15.8 2 6.1 units/mg protein with respect to 16.3 2 8.2 units/mg protein for untreated cells.
Characterization of Cu,Zn-SOD Purified from Silver-treated S. cerevisiae-The enzyme purified from silver-treated cells, as described under "Materials and Methods," gave an electrophoretic pattern of several bands in both protein and activity stained non-denaturing gels that was similar to that observed with silver-treated extracts (Fig. 4,A and B ) . In SDS-PAGE the sample gave only a single band with an apparent relative molecular mass (18,000 Da) identical to that of purified S. cerevisiae Cu,Zn-superoxide dismutase (not shown).
Isoelectrofocusing of the purified protein shows three main active bands focused a t pH 4.5, 4.3, and 4.2, respectively, identical to that observed with silver-treated extracts (Fig. 5).
The purified protein cross-reacted less efficiently with a polyclonal antibody raised against purified yeast Cu,Zn-SOD, the loss in antigenicity varying between 30 and 50% among the various purifications. A similar result was also obtained with genuine Ag,Zn-SOD (data not shown). Table I1 shows that the metal content of the protein purified from the silver-treated cells was strongly altered as compared with that purified from untreated cells. The data indicate that silver content complements the large decrease of copper, suggesting binding of Ag(1) at the copper site. In fact, the amount of zinc determined in the protein sample purified from silver-  treated cells was similar to that of the enzyme purified from untreated cells. A lower k , was observed for the silver-containing protein, which was, however, identical to that of the native protein if expressed on the basis of the copper content (Table  11). This suggests that activity is due to the presence of residual native enzyme molecules. Accordingly, the EPR spectrum of the silver-copper, zinc protein was identical to that of Cu,Zn-superoxide dismutase purified from untreated yeast (Fig. 6), although the copper signal appeared much lower than what expected on a copper/protein molar ratio 2:1, namely close to 0.4:l.

DISCUSSION
Ag (1) is a potent inhibitor of growth and fermentation in S. cerevisiae, with massive release of intracellular K+ occurring because of irreversible damage to the structure and integrity of the plasma membrane (31,32). Under our conditions, however, S. eerevisiae, displayed silver tolerance even though a lag phase in growth curve was observed. Interestingly, growth yield in the presence of silver was strictly related to superoxide dismutase activity, confirming that the enzyme is necessary to yeast for proper growing in the presence of oxygen.
However, direct effects of silver on Cu,Zn-SOD (in the absence of concomitant changes of manganese-SOD activity) were observed, and they were shown to operate at different levels. In fact, we were able to detect a specific increase in the Cu,Zn-SOD mRNA in response to silver ions, confirming that silver, like copper, can regulate Cu,Zn-SOD at the transcriptional level (13). On the other hand, a marked decrease in the enzyme activity with a concomitant slight decrease in the immunoreactive protein was observed. However, the synthesized protein is not able to perform a fully efficient catalytic function. This result was explained by the purification and characterization of a silver-containing SOD with impaired dismutase function due to the presence of silver in place of copper.
The results of analytical, spectroscopic, and catalytic determinations indicate that the majority of the protein molecules are Ag(1)-Zn-superoxide dismutase with Ag(1) most likely to be located to the copper site. This preferential binding site is in line with in vitro binding experiments showing that silver binds to bovine Cu,Zn-SOD with a higher affinity for the copper site than for the zinc site, and that the occupancy by zinc of the native zinc site enhances the binding of silver to the copper site (17).
As reported above, several isoforms were observed in SOD purified from silver-treated yeast. Those isoforms showed different behaviors in isoelectrofocusing and in non-denaturing acrylamide gel electrophoresis; such heterogeneity most probably reflects small differences in both net charge and overall folding, indicating a variable metal content of the various isoenzymes. The latter conclusion is supported also by the observation of a substoichiometric copper content determined in the protein purified from silver-treated yeast.
It is also interesting to note the higher antigenicity displayed by the Cu,Zn-SOD as compared either to the silver derivative obtained from silver-treated yeast or to in vitro prepared Ag,Zn- SOD. This fact may be linked either to alteration of protein folding or to conformational masking of immunogenic regions due to the presence of silver. As a matter of fact, an altered immunoprecipitating activity respect to the metal content of the Cu,Zn-SOD was previously observed (33).
On the basis of these data, it is appealing to speculate that Cu,Zn-superoxide dismutase in S. cerevisiae, being an abundant, stable, high-affinity metal-binding protein, may play a role in the homeostasis of copper and other metal ions which are harmful to yeast. This role would be complementary to that of metallothionein, which is coregulated with Cu,Zn-superoxide dismutase in S. cerevisiae (13,16).
Acomparable situation is found in E. coli where the genes for iron-and manganese-superoxide dismutases are regulated by the iron responsive fur repressor protein. Furthermore, iron modulates the activities of both iron-SOD and manganese-SOD at the transcriptional and at the post-translational levels (34-36).
The physiological benefit of this type of regulation may reside in an ancestral function of the superoxide dismutase as a metal-binding protein, which develops a dismutase function when copper is bound at the solvent-accessible metal-binding site and oxygen is available.