The presence of zinc in human cytocuprein and some properties of the apoprotein.

Abstract Purified cytocuprein from human liver, brain, and erythrocytes was found to contain near 2 g atoms of zinc per mole of protein. Both the copper and zinc were strongly bound and only relatively small amounts of them could be removed from the protein by dialysis against EDTA and 1,10-phenanthroline under various conditions. Virtually all of the metal was removed by treatment with potassium cyanide at neutral pH. There appeared to be very little change in the tertiary structure of the protein upon removal of the metal. This was indicated by the observations that the velocity and equilibrium sedimentation properties, the immunological properties, and absorption in the 200 to 240 mµ range of the native and apoprotein are similar. However, such treatment resulted in a diminution in the intensity of absorption of cytocuprein between 240 and 400 mµ and the disappearance of the broad absorption peak with a maximum near 675 mµ.

Purified cytocuprein from human liver, brain, and erythrocytes was found to contain near 2 g atoms of zinc per mole of protein. Both the copper and zinc were strongly bound and only relatively small amounts of them could be removed from the protein by dialysis against EDTA and l,lO-phenanthroline under various conditions. Virtually all of the metal was removed by treatment with potassium cyanide at neutral PH. There appeared to be very little change in the tertiary structure of the protein upon removal of the metal. This was indicated by the observations that the velocity and equilibrium sedimentation properties, the immunological properties, and absorption in the 200 to 240 rnp range of the native and apoprotein are similar. However, such treatment resulted in a diminution in the intensity of absorption of cytocuprein between 240 and 400 rnp and the disappearance of the broad absorption peak with a maximum near 6'75 mp.
The copper-containing proteins known as erythrocuprein, hepatocuprein, and cerebrocuprein have been shown to be identical and the name cytocuprein was proposed for them (1). We have recently found that cytocuprein also contains zinc. The first indication of this was obtained in studies of 'j5Zn exchange by the carbonic anhydrase isozymes in hemolysates (2). In addition to these enzymes, radioactivity was found to be associated with a protein identified as cytocuprein. This result was thought to be caused by a replacement of the copper in the protein by zinc. However, an attempt to remove copper under the conditions used for the zinc exchange reactions was unsuccessful and our attention was directed to the possibility that the native protein actually contained the latter metal. Detailed analyses conducted on cytocuprein from brain, liver, and erythrocytes have shown that the molar levels of zinc and copper in the protein are nearly equal.
An understanding of the roles of zinc and copper in the physical-chemical and biochemical properties of cytocuprein is dependent upon the selective removal of the metals without inducing denaturative changes in the protein. Conditions for the preparation of an apoprotein were determined in order to provide * This work was supported in part by Grant CA-1786 from the United States Public Health Service.
suitable material for such future studies. The properties of the apocytocuprein indicate that no significant changes in tertiary structure occur upon removal of the metal.

METHODS
Cytocuprein was isolated from human erythrocytes, liver, and brain by methods described previously (1,3). All buffers were prepared with water which had been passed through a mixed bed ion exchange resin (Continental Deionized Water Service, Chicago, Illinois). Water used in metal analyses was distilled twice in a quartz apparatus. Protein samples to be analyzed were dialyzed near 4' for at least 48 hours against 0.15 M NaCl which had been passed over a column of Dowex A-l chelating resin (Dow Chemical Company, Midland, Michigan).
Copper analyses were performed by the method of Van de Bogart and Beinert (4). A modification of the dithizone method described by Malmstrom (5) was used to assay zinc. The modifications included wet ashing of the protein (4) and the use of volumes of reagents sufficiently small so that all steps of the procedure through the extraction of the zinc into the organic phase could be carried out in 15-ml Pyrex test tubes. Since the copper of cytocuprein was a potential source of interference in the measurement of zinc, control assays were conducted with copper at levels near those present in the protein samples.
Nitrogen analyses were carried out by a micro Kjeldahl procedure (6), and cytocuprein concentrations are based on a nitrogen content of 16.9% (7). Quantitative precipitin reactions were carried out in a 3.0-ml system as described by Cohn, Wetter, and Deutsch (8). Precipitin reactions in gels utilized the Ouchterlony technique (9). Absorption spectra were measured manually with a Beckman DU-2 spectrophotometer.
Sedimentation analyses employed a Spinco model E ultracentrifuge equipped with both schlieren optics and a photoelectric scanning system (10). A wave length of 265 rnp was generally used with the latter optics.

RESULTS
The levels of zinc and copper in cytocuprein preparations from liver, brain, and erythrocytes are presented in Table I. The experimental variations encountered in the zinc assay are included since the precision of these measurements was not as high as that for the copper assay. The molar levels of the two metals are similar although the values for zinc are generally lower than those for copper. The differences are less than 1 S.D. except for cytocuprein from liver. In this case the levels of both  (5) a Expressed as gram atoms per 33,600 g of protein.
Values in parentheses are the number of analyses conducted.
b Experimental variations given are 1 S.D. The presence of tyrosine in the preparations from liver indicates that the low copper and zinc contents may be caused by contamination by proteins other than cytocuprein (1). The presence of zinc in purified cytocuprein was also confirmed by neutron activation analyses1 and by emission spectroscopy studies. The former of these assitys indicated t'hat metals other than zinc and copper were present in amounts less than 0.1 g atom/33,000 g of protein.
The zinc appears to be an integral part of the native cytocuprein.
This conclusion is supported by the observations that the zinc and copper levels are nearly equal and occur in stoichiometric quantities with respect to the protein.
Failure to remove more than small amounts of zinc by exposing cytocuprein to chelating agents also suggests that the metal is specifically bound. The chelation experiments were carried out by dialyzing cytocuprein under various conditions as outlined in Table II. The zinc and copper contents were unaffected by 1OP ~f l,lO-phenanthroline at pH 7.4 and 4', but the zinc content 'was reduced by 1 These analyses were performed by personnel of t'he Nuclear  (6) through their reaction with rabbit antibody to cytocuprein (U). The precipitin bands were allowed to develop for 40 hours and were stained with Amido black 10B.
13yo at pH 5.5. In one experiment at the latter pH, the dialysis was performed at 57" for 24 hours and then at 4" for an additional 50 hours. This treatment resulted in 12 and 18% decreases in the amounts of zinc and copper, respectively. Some precipitate which had formed while the protein was at room temperature was removed.
Denaturation of the protein may account for the loss of metal experienced under these conditions. About 10% of the zinc was removed by 1O-2 M EDTA at pH 7.4 but the copper content was unaffected.
Ascorbate at an initial level of lo+ M with the EDTA exerted no additional effect.
Markowitz, Cartwright, and Wintrobe (11) v-ere able to remove about half of the copper from cytocuprein by treatment with KCN at 5" for 24 hours. The data in Table II show that virtually all of the zinc and copper can be removed by dialysis against 5 X lo-* M KC'S at pH 8.0. The apocptocuprein prepared by this method \X as studied to determine the effects of the metals on the immunological and the physical properties of the native protein.
The apoprotein appeared to be immunologically identical with the native material when they were examined by the Ouchterlony technique with rabbit antibody to cytocuprein (see Fig. 1). The results of quantitative precipitin reactions shown in Fig. 2, however, indicate that there are differences between the two proteins.
The equivalence point for the apoprotein occurs at a higher antigen level than that for the native protein and less specific precipitate was formed.
This result suggests that some modification of antigenic sites had occurred as a result of removal of the metals.
Velocity sedimentation analysis showed that most of the apoprotein formed one symmetrical boundary but variable amounts of a slower sedimenting component were seen among different preparations (see Fig. 3). The szO,w measured for the major component was 2.86 S at a concentration of 0.8 mg per ml. An apoprotein preparation was passed over a column of Sephadex G-100 to effect removal of the small amount of the slow sedimenting material.  (12) was found to be 32,900. This agrees n-e11 with the molecular weight of the native protein (1). The results of a starch gel elect'rophoretic experiment on cytocuprein and on the apoprotein are shown in Fig. 4. It is apparent that the major component of t'he apoprotein has a much higher anodic mobility than the parent protein.
The cytocuprein used in this experiment had been stored for over 3 months at -20" and had been thawed and refrozen numerous times during this period.
This apparently leads to the generation of several additional minor and more anodic components than noted in the fresh material (I).
The liability of cytocuprein to form more anodic components has been previously noted (3,13). The less anodic of the numerous minor components noted in the apocytocuprein preparation appear to have similar electrophoretic mobilit.ies to the more anodic minor components of the native protein.
It can be seen from Fig. 4 that the major portion of the apoprotein is composed of two electrophoretic components which appear to be analogous to the two cytocuprein components noted in freshly prepared material (3,13).
The absorption spectra of the visible and ultraviolet regions for cytocuprein and the apoprotein are presented in Fig. 5. The apoprotein had a marked decrease in the intensity of absorption in the 250 to 290 rnp range and the broad peak with the maximum at 675 rnp was absent. No change in the absorption of the two proteins over the 200 to 240 rnp range is evident.

DISCUSSION
The zinc in cytocuprein appears to be more firmly bound than that of carbonic anhydrase.
Lindskog and Malmstrom (14) were able to remove all of the zinc from carbonic anhydrase by dialysis at pH 5.0 against 10m2 M 1, lo-phenanthroline for 2 to 7 days. Dialysis of cytocuprein against this chelating agent at pH 5.5 for 3 days removed only a small amount of this metal. These chelating experiments removed zinc more readily than copper, but both metals were strongly bound.
The labeling of cytocuprein -with 65Zn under these conditions as observed earlier (2) appears to have resulted from the exchange of a relatively small port,ion of the zinc.
The observation by illarkowitz et al. (11) that only part of the copper of cytocuprein can be removed by dialysis against cyanide at 3-5" was confirmed by our experiments.
However, at room temperature the zinc and copper were completely removed in 72 hours. The apoprotein has good solubility properties unlike material produced in earlier experiments (15).
The s?~,~ value of cgtocuprein was not affected significantly by the removal of t'he meta'ls, whereas the removal of copper from ccruloplasmin results in a marked decrease in the SK,,,,, value and an increase in the axial ratio (16). The apocytocuprein preparations contained small but variable amounts of protein which sedimented more slowly than the 2.86 S component.
Studies on cytocuprein which had been subject'ed to mild performic acid oxidation or t.o reduction and alkylation have indicated that the protein contains a subunit of molecular weight near 12,000 (13). It is possible that the cyanide has also produced a small amount of a similar subunit through its known action of cleaving disulfide bonds.
However, the production of such material by cyanide appears limited since the major component had the same molecular weight as the native protein.
The intensity of the absorption spectrum in the visible 240 to 350 rnp region for apocytocuprein is considerably less than the native protein and the peak with the maximum at 675 rnp is lost. These changes are caused by the removal of copper and possibly zinc. The reason for the decreased absorption by the apoprotein in the region from 250 to 290 rnp is not clear but may relate to metal ligands, changes in tertiary structure, or both, following removal of the metals.
Changes in peptide structures from a helical to a non-or less helical structure have been shown by Imahori and Tanaka (17) to be accompanied by a large increase in ultraviolet absorption in the region of 190 rnE.1. Rosenheck and Doty (18) have shown that it is feasible to draw inferences from ultraviolet absorption data as to the extent of helical structure in peptides, and Edsall (19) has presented such data for several proteins.
A specific example is the marked decrease in absorption in the 180 to 235 mp region by human carbonic anhydrase B following its conversion to a more helical structure upon acid denaturation (20). A similar effect has been noted for human cytocuprein when its axial ratio is decreased by treatment with ethanol-chloroform (15) under conditions simulating the Tsuchihashi (21) precipitation denaturation of hemoglobin. Removal of metals from cytocuprein did not effect the absorption between 190 and 235 rnp which indicates that no gross changes in tertiary structure and symmetry results from such treatment.
The nondenaturative removal of the zinc from bovine carbonic anhydrase B has been shown to effect no changes in the optical rotary dispersion properties of this enzyme (14).
The observation that cytocuprein contains equimolar levels of copper and zinc suggests that there might be some interaction between these metals. Chelation experiments discussed above show that at least part of the zinc can be removed without loss of copper.
However, studies of reconstitution of the apoprotein appear more attractive at present than attempts to remove one of the metals selectively.
Such studies will permit the evaluation of the role of each of the metals on the light absorption properties, electron paramagnetic resonance spectrum, and immunological and biochemical properties. Although the presence of cytocuprein in liver, brain, and erythrocytes has been recognized for many years, a biochemical role for the protein was not suggested until recently.
McCord and Fridovich (22) have reported that cytocupreins from bovine and human erythrocytes possess superoxide dismutase activity, but the physiological significance of this activity is uncertain. Probably the zinc in cytocuprein is related to this biochemical function.
We have previously suggested the name cytocuprein for the protein which was known as either erythrocuprein, hepatocuprein, or cerebrocuprein, depending on the tissue source (1). In view of the zinc content of this protein, the name cytocuprein is not accurately descriptive.
A more appropriate name should relate to the biochemical function of the protein when it becomes known.