Cytochrome c Oxidase Components II. A STUDY OF THE COPPER IN CYTOCHROME c OXIDASE*

The possible role of copper in the terminal respiratory enzyme, cytochrome c oxidase, has been investigated in a number of studies. Early nutritional research (l), which has been con-firmed by later work showed that copper deficiency pro-duced low cytochrome c osidase activity. Isolation studies also implicated in highly purified The in these iron. will remove all the copper without a loss of enzyme An&sic-Copper and iron analyses were performed according to published procedures (13, 14). Heme was determined as the pyridine hemochromogen (14). Activity was measured by fol-lowing the rate of oxidation of reduced cytochrome c spectro-photometrically (24).

The possible role of copper in the terminal respiratory enzyme, cytochrome c oxidase, has been investigated in a number of studies.
Early nutritional research (l), which has been confirmed by later work (Z-5), showed that copper deficiency produced low cytochrome c osidase activity.
Isolation studies also implicated copper in this enzyme system. Reilin and Hartree found relatively large amounts of copper in their heart muscle preparation (6). The Rutgers group (7,8), and later others (g- 14), also found a high copper content in their highly purified preparations.
The amount of copper in these preparations was reported to be within the range of 1 to 4 atoms of copper to each atom of heme iron.
No method that will remove all the copper without causing a loss of enzyme activity has yet been found (13,(15)(16)(17)(18).
Numerous chelators have been used in different ways in order to effect this removal, but none has proved successful.
Both electron spin resonance spectrophotometry and chelating agents have been used in attempts to determine the oxidation state of the copper under a variety of experimental conditions. Results of such experiments have, however, been conflicting.
On the one hand, Ehrenberg and Yonetnni (21) obtained from their cytochrome c oxidase preparation an electron slain resonance signal with characteristics typical of a noncatalytic cupric acid-amino acid complesl (21,23), the hyperfine spacing being 184 gauss, which corresponds to a normal splitting constant of 0.019 cm-l.
They also found that reduced cytochrome c did not affect the magnitude of the cupric signal under anaerobic conditions. These observations led to the conclusion that the copper did not take part in the enzymic reaction by changing valence. This opinion was reinforced by chemical studies employing a cuprous chelator (19,20).
On the other hand, the S17isconsin workers obtained signals from their cytochrome c osidase preparation (22) indicating that two types of copper were present: (a) "native" copper, which had an electron spin resonance signal similar to that of the catalytic copper proteins, lactase and ceruloplasmin, was reducible by substrate, and had no resolvable hyperfine absorption, and (b) "secondarily bound" copper, which was similar to the copper ob-  (21). Moreover, they (13) were not able to duplicate Yonetani's results (19,20) with the cuprous chelator on their preparation.
In the present study, we have re-examined the copper contents and the spectral properties of purified cytochrome c oxidase preparations.

EXPERIWSNTAL PROCEDURE
In preparing all reagents, special precautions were taken to minimize extraneous copper.
Distilled water was passed through a commercial deionizer before use. After the phosphate buffer was prepared in the usual manner, it was passed through a Chelex 100 column.
In preparing the Chelex 100 columns, the resin was first purified. Resin, 100 g, was suspended in 800 ml of deionized distilled water, stirred for 5 minutes, and allowed to settle for 30 minutes, after which the fine particles were removed by decantation. This procedure was repeated four times.
The resin, free of fine particles, was then suspended in 800 ml of 1 K hydrochloric acid, allowed to settle for 1 hour, and then decanted off. This process was repeated.
The resin was then suspended in 800 ml of 1 N sodium hydroxide and allowed to settle for 1 hour before the supernatant fluid was decanted.
This procedure was repeated.
The resin was thoroughly washed with deionized distilled water and suspended in 800 ml of 0.1 M phosphate buffer, pH 7.4. The pH of the suspension was adjusted with 2 N hydrochloric acid to 7.4, and the resin was allowed to settle for 30 minutes.
The supernatant fluid was decanted and replaced with 800 ml of 0.1 M phosphate buffer, pH 7.4. The pH of the suspension was determined and, if necessary, again adjusted to 7.4.
The resin adjusted in this manner was poured into a chromatographic tube to make a column 3 cm in diameter and 20 cm in height.
Two hundred milliliters of 0.1 bf phosphate buffer, pH 7.4, were passed through the column and discarded. The next 2 liters of the phosphate buffer were collected for use.
The ammonium sulfate and cholic acid used were recrystallized compounds.

Enzyme
Preparation-The cytochrome c oxidase employed in these studies was isolated as previously described (14), except that in the last ammonium sulfate fractionation steps, all the reagents employed contained less than 1 PM copper.
The horse heart cytochrome c used was Sigma type II, purified as previously described (25). Reduced cytochrome c was obtained by a previously published method (14).

An&sic-Copper
and iron analyses were performed according to published procedures (13,14). Heme was determined as the pyridine hemochromogen (14). Activity was measured by following the rate of oxidation of reduced cytochrome c spectrophotometrically (24). Sephadex Column-The bathocuproine-chelated copper was separated from cytochrome c oxidase by means of a column of Sephadex G-75 that had been suspended in cholate-ammonium sulfate-phosphate buffer. The buffer was prepared by dissolving 113.4 g of ammonium sulfate in a liter of 0.1 M phosphate buffer, pH 7.4, containing 0.6% cholate, and then readjusting the pH to 7.4. This solution contained less copper than 1 /AM. Just before hydrated Sephadex was poured into a column, the solution was made 0.1 RIM with respect to bathocuproine sulfate and 0.58 IRM with respect to sodium dithionite. The Sephadex was then poured into a chromatographic tube to make a column 1.8 cm in diameter and 40 cm in height. The column was washed with 10 column volumes of the cholate-ammonium sulfate-phosphate buffer (14) containing bathocuproine and dithionite. Six milliliters of cytochrome c oxidase were allowed to enter the column and were then eluted with solvent prepared as described above. Fractions were collected automatically and the optical density of the material in each tube was determined at two different wave lengths.
In most experiments, it was desirable to remove the dithionite and bathocuproine. This was accomplished by passing the enzyme through a second column of Sephadex, which was prepared as described above except that the dithionite and bathocuproine were omitted.
ESR Spectroscopy-ESR spectroscopy was performed in the usual manner with a Varian V-4500 spectrometer having a NOkc field modulation. The spectra ("signals") are presented as derivatives of the absorption with respect to field strength, the field increasing from left to right.
Spectra were measured at about -165" with the use of field modulation amplitudes of 3. With other copper proteins, this technique gave results within +~4% of the cupric copper concentration measured chemically (26).

RESULTS
Shortly after cytochrome c oxidase was added to the Sephadex G-75 column, two bands appeared. The most rapidly moving of the bands was the green cytochrome c oxidase, whereas the slower moving band was the orange-yellow chelate of copper. By following the absorbancy changes at 598 mp, a chromatograph ( Fig. 1) was obtained. The absorption spectrum of the orangeyellow peak was identical with that obtained under the same conditions with a solution of copper and bathocuproine. The absorbancy at 598 rnp was attributable to the cytochrome c oxidase, whereas the absorbancy at 475 rnp reflected the concentration of the copper bathocuproine.
The spectrum of a reduced cytochrome c oxidase preparation is shown both before and after the addition of bathocuproine in Fig. 2. Fig. 2A represents the spectrum obtained before the preparation has passed through the Sephadex column, and Fig.   2B is the same preparation after it has passed through the Sephadex column. It is apparent that most of the copper chelate has been removed by the Sephadex treatment. From the changes in absorbancy at 475 rnp, it was determined that about 83.4% of the copper was removed. It was never possible to remove the chelatable copper completely by this method. Even after the preparation had been been passed through the Sephadex column a second time, between 10 and 20% of the copper remaining could be chelated with bathocuproine. Table I  nents of the cytochrome c oxidase preparation, as well as activity measurements. It is apparent from these results that only 40 % of the copper originally present in the preparation is chelated by the bathocuproine; the remaining copper appears to be tightly bound. Even after the preparation has passed through the Sephadex column and the chelatable copper has been removed,  There is, in fact, an increased specific activity that can be attributed to the removal of extraneous protein from the preparation by this treatment.
The ratio of copper to heme a is given in Table I. In order to make a comparison with other preparations, however, it must be pointed out that the iron to copper ratio would be 10 to 20% less (14). These results indicate the presence of a nonheme a iron in the preparation, as was previously noted (14). The values show that there is more copper than iron in these preparations.
ESR Results-A typical electron spin resonance signal obtained from cytochrome c oxidase before it was treated with Sephadex is shown in Fig. 3. This spectrum is almost identical with the signal previously observed by Ehrenberg and Yonetani; it had gmax = 2.04, g-parallel = 2.2, and a hyperfine splitting constant, A = 0.02 cm-l, which is equivalent to a splitting of about 188 gauss. The corresponding values reported by Ehrenberg and Yonetani were gmax = 2.04, g-parallel = 2.185, and 184 gauss. The sample was 0.55 mM in total copper by chemical analysis, and 0.20 mM in cupric copper by ESR comparison with a standard. It was similar to other preparations studied by ESR in that dithionite abolished the cupric copper signal, which was partially attenuated by reduced cytochrome c under anaerobic conditions. We found it experimentally difficult to obtain 109~o reduction by reduced cytochrome c, and in our system the cuprous copper slowly reoxidized even in a closed cuvette; however, values between 30 y0 and 66% reduction were consistently obtained. The copper was rapidly and completely reoxidized when oxygen was admitted to the sample.
Upon adding KBHl at 0" to cytochrome c oxidase preparations that had not been passed through the Sephadex column, it was found that the ESR spectrum was altered. Much of the hyperfine structure was lost, and the new signal resembled strongly those reported by Beinert et al. (22) and considered by them to contain only "native" copper. Since some change of sample geometry took place upon KBHd addition, the reduction of copper by this reagent was not estimated quantitatively, but it was roughly equivalent to 30% of the ESR-detectable copper.
The ESR spectrum of a cytochrome c oxidase preparation after Sephadex treatment is depicted in Fig. 4. This signal, although it still showed some hyperfine splitting in the region of g = 2.2, almost entirely lost the small peak near g = 2.0, on the side of the major absorption, and strongly resembled the signal of cytochrome c oxidase observed by Beinert and his co-workers. The sample was 0.19 mm in total copper chemically determined; by the ESR comparison method, it contained 0.10 IDM cupric copper. A reduction of about 2Ooj, could be obtained by addition of KBH( under anaerobic conditions, and the signal then resembled almost exactly the signals reported by the Wisconsin group for native cytochrome c oxidase.

DISCUSSION
The present study does not establish that copper is directly involved in cytochrome c oxidase activity.
It does, however, bring into question the validity of Yonetani's interpretations which are based on the use of bathocuproine sulfate to chelate the copper of cytochrome c oxidase (19,20). Although bathocuproine chelates a portion of the copper present in a purified cytochrome c oxidase, enzyme activity is not affected, and when this copper is removed, the preparation still contains slightly more copper than heme a. For the most part, this remaining copper is firmly bound and cannot be chelated in either the cuprous or cupric form with bathocuproine.
Attempts were made to employ the outlined procedure with Sephadex in order to remove cupric copper chelated with EDTA.
Combinations of EDTA and bathocuproine were also used. In no instance was the ratio of copper to heme a ever reduced to less than 1.1: 1. That Griffiths and Wharton (13) have obtained similar results has been interpreted as presumptive evidence of a role for copper in the cytochrome c oxidase activity.
In agreement with other workers (13,18,22), we have found that the tightly bound cytochrome c oxidase copper can be directly reduced by ferrocytochrome c. This reduction takes place equally well whether carbon monoxide is present or absent.
Spectral data have established that carbon monoxide reacts directly with the heme groups of cytochrome us. The classic action spectrum of Warburg (27) has demonstrated that carbon monoxide inhibits respiration and that this inhibition can be reversed by light.
This reversal is brought about by absorbed light, which causes dissociation of the carbon monoxide complex of what is now known to be cytochrome us. The fact that it occurs only with hemoproteins and not with copper proteins establishes that, if the copper in cytochrome c oxidase is functional, it does not react with carbon monoxide.
The Wisconsin workers (13,22) have suggested that their experimental results are compatible only with a reaction of copper on the oxygen side of the cytochrome or by a pathway independent of the cytochrome a and u3 components. Thus, they propose a scheme of CN Cytochrome c -+ "cytochrome u" -+ copper -I+ oxygen.
Since cyanide reacts with both copper and the heme group of cytochrome ~3, their experimental results are compatible with this interpretation.
When carbon monoxide is employed, however, it is clear that the inhibition is due to the heme at the cytochrome level. Therefore, according to the scheme outlined above, an inhibition of the reduction of the copper would be expected in the presence of carbon monoxide.
The data of the present study do not support this proposal.
Since cyanide also combined with the heme group, a decreased rate of reduction of copper should also result if cytochromes a and aa are involved in its reduction.
This expectation appears to be equally unfulfilled (13,18,22). A number of possible mechanisms for copper function still have not been excluded by the available experimental data. The ESR evidence we have obtained shows that a cytochrome c oxidase preparation may have a signal typical of the Ehrenberg-Yonetani (21) enzyme, but that it is converted into a preparation with the characteristics of the enzyme described by Beinert et al. either by KBHl reduction or by Sephadex treatment in the presence of bathocuproine disulfonate. We can also confirm the observation of the latter investigators (22) that only a part of the chemically determinable copper can be detected by ESR. The reason for this discrepancy is unknown. It is probably best to withhold judgment concerning exchange interaction between cuprous and cupric copper as an explanation for the phenomenon.
The similarity of the cytochrome oxidase signal to those of lactase and ceruloplasmin has also been explained on the basis of this hypothesis (28), but it has been shown that copper proteins containing only 1 copper atom per molecule have similar ESR spectra (26). SUMMARY The copper present in a purified cytochrome c oxidase preparation has been investigated by chemical methods and electron spin resonance (ESR) spectrophotometry.
The preparation has been shown to contain two types of copper: (a) loosely bound copper, which can be chela.ted directly with bathocuproine disulfonate, and (b) tightly bound copper, which cannot be chelated unless the enzyme is denatured.
Removal of the loosely bound copper does not affect enzyme activity. After its removal, the ratio of copper to heme a of the preparation approaches 1.
The tightly bound copper can be reduced by ferrocytochrome c in a nitrogen or carbon monoxide atmosphere and is rapidly oxidized in air. In agreement with a previous report (22), the ESR signal of this bound copper contains no hyperfine structure, and only a portion of the chemically determined copper can be detected by ESR.