- Email: [email protected]
Separation of enzymically active bovine cytochrome c oxidase monomers and dimers by high performance liquid chromatography
Separation of Enzymically Active Bovine Cytochrome c Oxidase Monomers and Dimers by High Performance Liquid Chromatography Theo B. M. Hakvoort, Karin M. C. Sinjorgo, Bob F. Van Gelder, and Anton 0. Muijsers Laboratory of Biochemistry, B. C. P. Jansen Institute, University of Amsterdam, iThe Netherlands
ABSTRACT The aggregation state of two types of bovine heart cytochrome c oxidaae preparations in the presence of laurylmaltoside was investigated by high performance liquid chromatography in two buffers of ionic strengths of 388 mM and 45 mM, respectively. At high ionic strength, it was found that the Fowler cytochrome c oxidase preparation was monomeric (A4, = 2. lOs), while monomers and dimers (2 x uur, M, = 4. 10J) could be isolated from the Yonetani preparation. Under these conditions there was no rapid equilibrium between the two forms. Covalent cytochrome c oxidase-cytochrome c complexes were largely dimeric, and addition of ascorbate and cytochrome c to the oxidase also promoted dimerization. At low ionic strength (I = 45 mM) in the presence of laurylmaltoside the oxidase and the covalent complex with cytochrome c were largely monomeric. In the steady-state oxidation of ferrous horse heart cytochrome c, the monomeric enzyme displayed biphasic kinetics at I = 45 mM. This suggests that the presence of high- and lowaffinity reactions is an intrinsic property of the cytochrome c oxidase monomer.
Abbreviations HPLC, high performance liquid chromatography; acid)
DTNB, 5,5’-dithio-bis-(2-nitrobenxoic
INTRODUCTION Cytochrome
c oxidase is a Y-shaped protein of the inner mitochondrial
membrane.
Address reprint requests to Dr. T. B. M. Hakvoort, Laboratory of Biochemistry, B. C. P. Jansen Institute, University of Amsterdam, P.O. Box 20151, 1000 HD Amsterdam, The Netherlands. Journul of Inorganic Biochenktty 23, 381-388 (1985) 0 1985 Plsevier Science Publishing Co., Inc. 52 Vanderbilt Ave., New York, NY 10017
381 0162-0134/85/$3.30
382
T. B. M. Hakvoort et al.
It possesses a large (Ml) and a smaller (M2) transmembrane domain and a third domain (C) projecting for 60 A out into the intermembrane space, where the reaction with cytochrome c takes place [ 11. Most investigators assume that the bovine heart protein consists of 12-13 subunits, adding up to it4, = 2. lo’, but these numbers are still being disputed [l, 21. It is assumed that the enzyme is dimeric (2 x (2a3) in most nonionic detergents and also possibly in mitochondria, but monomeric forms have been reported for shark and camel cytochrome c oxidase [3, 41. Published procedures for conversion of the dimeric bovine enzyme to monomers involve treatment at elevated pH with high concentrations of Triton and DEAE chromatography. During such treatment the oxidase has a tendency to lose some small subunits and sometimes also the large hydrophobic subunit III [5]. In media of low ionic strength, cytochrome c oxidase shows peculiar biphasic steady-state kinetics, ascribed to a high-affinity cytochrome c reaction site (K, = lOwa M) and a low-affinity one (K,,, = 10m5 M) [6]. The question of whether these kinetics are inherent to an oxidase dimer with cooperativity or are an intrinsic property of the monomer has not yet been settled. In this paper we present separation by a size exclusion HPLC technique of relatively stable monomeric and dimeric bovine heart cytochrome c oxidase in buffers containing laurylmaltoside. Our evidence indicates that the monomeric enzyme can display biphasic kinetics. MATERIALS AND METHODS Bovine heart cytochrome c oxidase was purified according to a modified version of Yonetani [7] and Fowler et al.‘s [8] methods in our laboratory 191. The Fowler procedure involves selective solubilization of cytochrome c oxidase with deoxycholate, whereas, in the Yonetani procedure, all cytochromes are solubilized by excess cholate followed by selective denaturation of complex III. Horse heart cytochrome c was purified according to Margoliash and Walasek [lo]. The covalent 1 : 1 complex of bovine heart cytochrome c oxidase-yeast iso-lcytochrome c was prepared as described by Fuller et al. [I 11, using DTNB. Sizeexclusion chromatography of bovine heart cytochrome c oxidase was performed on a TSK SW 3000 (600 x 7.5 mm) column (Toyo Soda, Japan) at 20°C at a flow rate of 0.5 ml/min. The HPLC system consisted of LKB 2150 pumps, Rheodyne 7125 injector and an LKB 2158 Uvicord SD detector operated at 275 nm or 405 nm, light path 2.5 mm. Equilibration of the column and elution of the proteins was done with either a buffer containing 100 mM Trisjacetate, 1 mM EDTA, 300 mM NaCl (pH 7.5, I = 388 mM; the “high ionic strength buffer”) or a buffer containing 50 mM Trislacetate, 1 mM EDTA (pH 7.5, I = 45 mM; the “low ionic strength buffer”). To both buffers laurylmaltoside (0.2% w/v) was added as a detergent. Calibration of the TSK SW 3000 column was done by using the following markers: NADH-Q-oxidoreductase M, = 1.6 x lo6 (a generous gift from P. Bakker); thyroglobulin M, = 669 x 103, ferritin M, = 440 x 103, catalase M, = 235 x 103, aldolase M, = 158 x lo3 (all purchased from Pharmacia, Sweden); and ceruloplasmin M, = 130 x lo3 (a generous gift from R. A. Cuperus). Cytochrome c oxidase activity was measured spectrophotometrically as previously described 1121. Polyacrylamide gel electrophoresis in the presence of
Separation of Bovine Cyt c Oxidase
383
sodium dodecyl sulphate was performed according to Kadenbach [13]. Sedimentation velocity analysis was conducted according to Schachman [ 141, using an MSE analytical ultracentrifuge model Mk2. Yeast cytochrome c and DTNB were purchased from Sigma Chemical Company (St. Louis, MO, USA),andlauQlmaltoside from Calbiochem-Behring Corp. (La Jolla, CA, USA).
RESULTS Aggregation State of Cytochrome c Oxidase at High Ionic Strength A sample of purified cytochrome c oxidase, prepared according to our modified Fowler procedure [9], was subjected to gel permeation HPLC in a medium of high ionic strength (I = 388 mM) with 0.2% (w/v) laurylmaltoside as a detergent. As shown in Figure 1, the enzyme eluted as a single sharp 275 nm absorbance peak with an apparent molecular mass of 2. lo5 as calculated from the retention times of a series of calibration proteins (see Methods). A sample of cytochrome c oxidase purified according to the method of Yonetani [7], however, showed two sharp 275 nm absorbance peaks of about equal intensity (Fig. lB, solid line), the positions of which correspond to proteins with a mass of 2. lo5 and 4. 105, respectively. A sedimentation run in the analytical ultracentrifuge in the same medium showed the presence of two species differing by a factor 2 in apparent molecular mass. We conclude that in the preparation both monomers (1 x a~~) and dimers (2 x uu3) were present. A duplicate HPLC run of the Yonetani cytochrome c oxidase preparation with 405 nm detection (Fig. lB, broken line) again showed two peaks of nearly equal absorbance. This indicates that the heme a/protein ratio is the same in both peaks, which is in line with the idea that the two peaks represent monomeric and dimeric forms of the enzyme. Furthermore, both peaks of the Yonetani-type preparation, as well as the single peak of the Fowler-type preparation, displayed the same subunit pattern within the limits of resolution when assayed in the SDS-urea polyacrylamide gel electrophoretic system according to Kadenbach [ 131 (not shown in the figures). The trace of high molecular weight contaminant visible in the elution pattern of Figure 1B is low in 405 nm absorbance and probably represents NADH-ubiquinol oxidoreductase [ 151. The observation that cytochrome c oxidase dimers and monomers elute as distinct peaks suggests that no rapid equilibrium exists under these high ionic strength conditions; otherwise, one fused peak at intermediate position would result. Indeed, rechromatography of samples taken from the peak fractions showed that dimer and monomer remained dimeric and monomeric, respectively (not shown in the figures). The possibility that the occurrence of two peaks was a detergent effect caused by residual Tween-80, originally present in our Yonetanitype enzyme preparation, therefore, is less likely.
Effect of Cytochrome c on Aggregation State In order to investigate the effect of cytochrome c on the aggregation state of the oxidase, we prepared covalent complexes from DTNB-activated yeast iso-lcytochrome c and cytochrome c oxidase. Noncovalent cytochrome c-cytochrome c
384
T. B. M. Hakvoort et al.
0
ELUTION
30
60 TI ME (MINI
FIGURE 1. Elution profile of gel permeation HPLC of cytochrome c oxidase in a buffer of 388 mM ionic strength in the presence of 0.2% laurylmaltoside. Absorbance was measured at 275 nm (---) and 405 nm (---). (A) Fowler-type oxidase preparation; (B) Yonetani-type oxidase preparation; (C) covalent 1 : 1 complex of Fowler-type oxidase and yeast iso-1-cytochrome c. For details see Methods section.
oxidase complexes will fully dissociate at the high ionic strength used during chromatography. The activated thiol group of Cys 103 in yeast iso-1-cytochrome c slowly forms a disulphide bridge with a thiol group in subunit III of the bovine oxidase [ 111. Starting with either the Fowler-type or the Yonetani-type preparation, we prepared covalent cytochrome c-cytochrome c oxidase complexes with an overall stoicheiometry of 1 c per aa3. The Fowler 1 : 1 complex showed 60% dimers (Fig. 1C) and the 1 : 1 complex made from the Yonetani-type oxidase was 80 % dimeric when assayed at high ionic strength by HPLC. The 405 nm/275 nm absorbance ratio was the same for both peaks, suggestive of a monomeric and a dimeric 1 : 1 complex, respectively. So, in the covalent complex the equilibrium shifted to the dimer side. When Fowler-type cytochrome c oxidase (175 PM) was incubated with horse cytochrome c (122 PM) at high ionic strength in the presence of 10 mM ascorbate, the mixture rapidly reached anaerobiosis. A sample taken soon afterwards showed 10% dimers, a second sample taken 1 hr after anaerobiosis was 50% dimeric.
Separation
of Bovine Cyt c Oxidase
385
A 0.16-
0.12-E E N ti
0.08-
: CL z z
0.04-
0 0
-1. 10
L
20
40
50
ELUT ION T I ME (rnrd
60
70 TN (~‘1
FIGURE 2. (A) HPLC elution profile of Fowler-type cytochrome c oxidase in a buffer of 45 mM ionic strength containing 0.2% laurylmaltoside. (B) Eadie-Hofstee plot of spectrophotometric cytochrome c oxidase activity assay of the monomeric peak fraction (indicated by bar in Fig. 2A) in the same 45 mM ionic strength medium. Reduced cytochrome c 0.1-60 pM. The curved line is a computer fit through the points, the mathematical sum of two straight lines (indicated in the figure) corresponding to the high- and low affinity reactions, respectively. For details see Ref. 12.
Whether this conversion is due to the anaerobiosis, to the slow turnover, to the presence of cytochrome c (at high ionic strength not significantly bound to the oxidase), or to formation of pulsed oxidase [ 161, remains to be investigated. The cytochrome c oxidase of course reoxidizes upon injection into the HPLC buffer flow. Aggregation State of Cytochrome c Oxidase at Low Ionic Strength Cytochrome c oxidase is most stable and soluble at high ionic strength. Decreasing the ionic strength induces the risk of protein precipitation in a column that is difficult to regenerate once clotted. We discovered it was possible to run the oxidase in 50 mM Tris/acetate (pH 7.5, I = 45 mM) plus 0.2 % laurylmaltoside. The Fowler-type preparation, again, displayed a sharp monomeric peak (Fig. 2A) and even the Yonetani-type preparation eluted at the monomer position, though the peak showed broadening at the front (high-molecular) side. At low ionic strength, the Yonetani-type preparation was not very stable. Chromatography frequently led to the loss of some small subunits. The covalent complexes were still partially dimeric at low ionic strength. At low ionic strength incubation of the Fowler-type preparation under turnover conditions only led to slight broadening of the
386
T. B. M. Hakvoort et al.
monomer peak. We may tentatively conclude, therefore, that in the presence of laurylmaltoside low ionic strength promotes formation of monomeric cytochrome c oxidase. In the same medium of ionic strength 45 mM with 0.2% laurylmaltoside, in which the Fowler-type cytochrome c oxidase behaved as a pure monomer (Fig. 2A), we determined the cytochrome c oxidase activity of an HPLC peak sample. The spectrophotometric assay was used with reduced cytochrome c as substrate. As shown in Figure 2B, a curved Eadie-Hofstee plot was found that could be analyzed as the sum of a high-affinity and a low-affinity phase (Fig. 2B, straight lines). It is concluded that monomeric cytochrome c oxidase displays the characteristic biphasic kinetics often described for the enzyme at low ionic strength. The oxidation of ferrocytochrome c remains strictly first-order, so there is no change in activity during the assay. Furthermore, it is unlikely that the strong dilution of the monomeric HPLC peak sample in the activity assay would induce dimerisation.
DISCUSSION A significant result of the present study is the possibility to isolate pure monomeric and dimeric forms of bovine cytochrome c oxidase in the same buffer and detergent. This implies that the activation energy for interconversion and equilibration of the two forms must be rather high, at least in our high ionic strength medium. Our results concerning the stability of monomers and dimers are in agreement with the observations of Azzi et al. [17], who found two separate peaks under certain conditions in gel filtration experiments of cytochrome c oxidase in the presence of laurylmaltoside. At low ionic strength, our Yonetanitype preparation essentially was monomeric with a broadened leading edge of the peak. This suggests that, under these conditions, a more rapid equilibrium is established between the two aggregation states. Thompson and Ferguson-Miller [ 181 studied rat liver cytochrome c oxidase by gel filtration in a medium with an ionic strength of approximately 100 mM in the presence of laurylmaltoside. They concluded that the enzyme was predominantly present in the monomeric form. Their cytochrome c oxidase preparation, however, was depleted of subunit III. Other studies [5] have shown that conditions inducing removal of subunit III, for instance high pH in the presence of 5% Triton X-100, also promote monomerisation of the oxidase. This may be explained by the assumption that subunit III is situated in the M2 membrane traversing domains of the enzyme molecules that are thought to be in close contact in the dimer [ 191. The monomeric as well as the dimeric cytochrome c oxidase isolated in our study does contain subunit III. The presence of subunit III might be a prerequisite for formation of dimers. We have found that, under conditions where the oxidase is monomeric, the covalent cytochrome c oxidase-cytochrome c complex is largely dimeric. This can be understood in light of Capaldi et al.‘s [20] suggestion that cytochrome c binds in a cleft between the monomers in a cytochrome c oxidase dimer. At pH 7.5 the cytochrome c oxidase protein has a negative net charge, and the positively charged cytochrome c may help overcome electrostatic repulsion between the oxidase monomers in the dimer.
Separation of Bovine Cyt c Ox&se
387
Because laurylmaltoside only recently has been introduced as an excellent detergent for cytochrome c oxidase [21] most studies on the aggregation state of the enzyme have used other nonionic detergents such as Tween-80 and Triton X100, or the anionic (deoxy)cholate. Bovine heart cytochrome c oxidase tends to be 90% dimeric in nonionic detergents [3], but the chicken and dogfish heart enzymes were present as a mixed population of monomers and dimers. Cytochrome c oxidase isolated from camel [4] and hammerhead shark [22], however, were reported to be more than 90 % monomeric. Most investigators agree that monomeric, as well as dimeric oxidase, are capable of rapid electron transfer from reduced cytochrome c to oxygen [4, 5, 18, 223. Initially, the dimer was suggested to be the active species [3], but Wilson et al. [22] reported that monomeric dogfish oxidase was the most active form, and subsequently Capaldi et al. [5] stated that any role for the dimer would be limited to a control function and/or an as yet undefined role in coupling electron transfer to proton pumping. Explicit statements about the absence or presence of the so-called high- and low affinity cytochrome c binding sites in monomers (only to be observed at low ionic strength [6]) are rare. Thompson et al. [18] imply that the monomer displays the biphasic kinetics. On the other hand, Nalecz et al. [17] conclude that biphasic kinetics are caused by negative cooperativity in the dimer, and monomeric oxidase only shows the high affinity, low turnover site. As yet we have no explanation for this apparent discrepancy between our results and those of Azzi et al. One should be cautious in extrapolating the aggregation state of the oxidase sample to the actual situation prevailing during the activity measurement. Nevertheless, we feel our results suggest that the presence of high- and low affinity cytochrome c reactions are an intrinsic property of the monomeric enzyme. The authors thank H. Plot for performing the sedimentation velocity experiments. This work was supported by grants from the Netherlands Organization for the Advancement of Pure Research (2. W.O.) under the auspices of the Netherlands Foundation for Chemical Research (S.O.N.).
REFERENCES 1.
2.
R. A. Capaldi, F. Malatesta, and V. M. Darley-Usmar,
Biochim. Biophys. Acta 726, 135-148 (1983). G. Buse, G. C. M. Steffens, and L. Meinecke, in Structure and Function of Membrane Proteins, E. Quagliariello and F. Palmieri, Eds., pp. 131-138, Elsevier Science Publishers,
5.
Amsterdam (1983). N. C. Robinson and R. A. Capaldi, Biochemistry 16, 375-381 (1977). V. M. Darley-&mar, N. Alizai, A. I. Al-Ayash, G. D. Jones, A. Sharpe, and M. T. Wilson, Comp. Biochem. Physiol. 68b, 445-456 (1981). G. Georgevich, V. M. Darley-Usmar, F. Malatesta, and R. A. Capaldi, Biochemistry 22, 1317-
6. I. 8. 9.
1322 (1983). S. Ferguson-Miller, D. L. Brautigan, and E. Margoliash, J. Biol. Chem. 251, 1104-1115 (1976). T. Yonetani, J. Biol. Chem. 235, 845-852 (1960). L. R. Fowler, S. H. Richardson, and Y. Hatefi, Biochim. Biophys. Acta 64, 170-173 (1962). C. R. Hartze.11, H. Beinert, B. F. van Gelder, and T. E. King, Methods Enzymol. 53, 54-66
3. 4.
10. 11.
12.
(1978). E. Margoliash and 0. F. Walasek, Methods Erqvmol. 10, 339-348 (1967). S. D. Fuller, V. M. Darley-Usmar, and R. A. Capaldi, Biochemistry 20, 7046-7053 (1981). K. M. C. Sinjorgo, J. H. Meijling, and A. 0. Muijsers, Biochim. Biophys. Acta 767, 47-56
(1984).
388
T. B. M. Hakvoort
et al.
13. 14. 15.
B. Kadenbach, J. Jaraush, R. Hartmann, and P. Merle, Anal. Biochem. 129, 517-521 (1983). H. K. Schachman, Ultrucentrifugation in Biochemistry, Academic Press, New York (1959). R. Penniall, W. B. Elliot, K. J. H. van Buuren, and B. F. van Gelder, J. Bioenerg. Biomembr. 8,
16.
271-292 (1976). M. Brunori, A. Colosimo,
G. Rainoni, M. T. Wilson, and E. Antonini, J. Biol. Chem. 254, (1979). K. A. Nalecz, R. Bolli, and A. Azzi, Biochem. Biophys. Res. Commun. 114, 822-828 (1983). 10769-10775
17. 18. 19. 20. 21. 22.
D. A. Thompson and S. Ferguson-Miller, Biochemistry 22, 3178-3187 (1983). T. G. Frey, M. J. Costello, B. Karlson, J. C. Haselgrove, and 1. S. Leigh, J. Mol. Biol. 162, 113130 (1982). R. A. Capaldi, V. M. Darley-Usmar, S. Fuller, and F. Millett, FEBS Lett. 138, l-7 (1982). P. Rosevaer, T. van Aken, J. Baxter, and S. Ferguson-Miller, Biochemistry 19, 4108-4115 (1980). M. T. Wilson, W. Lalla-Maharajh, V. M. Darley-Usmar, J. Bonaventura, C. Bonaventura, and M. Brunori, J. Biol. Chem. 255, 2722-2728 (1980).
Received and accepted November 1984
ABSTRACT The aggregation state of two types of bovine heart cytochrome c oxidaae preparations in the presence of laurylmaltoside was investigated by high performance liquid chromatography in two buffers of ionic strengths of 388 mM and 45 mM, respectively. At high ionic strength, it was found that the Fowler cytochrome c oxidase preparation was monomeric (A4, = 2. lOs), while monomers and dimers (2 x uur, M, = 4. 10J) could be isolated from the Yonetani preparation. Under these conditions there was no rapid equilibrium between the two forms. Covalent cytochrome c oxidase-cytochrome c complexes were largely dimeric, and addition of ascorbate and cytochrome c to the oxidase also promoted dimerization. At low ionic strength (I = 45 mM) in the presence of laurylmaltoside the oxidase and the covalent complex with cytochrome c were largely monomeric. In the steady-state oxidation of ferrous horse heart cytochrome c, the monomeric enzyme displayed biphasic kinetics at I = 45 mM. This suggests that the presence of high- and lowaffinity reactions is an intrinsic property of the cytochrome c oxidase monomer.
Abbreviations HPLC, high performance liquid chromatography; acid)
DTNB, 5,5’-dithio-bis-(2-nitrobenxoic
INTRODUCTION Cytochrome
c oxidase is a Y-shaped protein of the inner mitochondrial
membrane.
Address reprint requests to Dr. T. B. M. Hakvoort, Laboratory of Biochemistry, B. C. P. Jansen Institute, University of Amsterdam, P.O. Box 20151, 1000 HD Amsterdam, The Netherlands. Journul of Inorganic Biochenktty 23, 381-388 (1985) 0 1985 Plsevier Science Publishing Co., Inc. 52 Vanderbilt Ave., New York, NY 10017
381 0162-0134/85/$3.30
382
T. B. M. Hakvoort et al.
It possesses a large (Ml) and a smaller (M2) transmembrane domain and a third domain (C) projecting for 60 A out into the intermembrane space, where the reaction with cytochrome c takes place [ 11. Most investigators assume that the bovine heart protein consists of 12-13 subunits, adding up to it4, = 2. lo’, but these numbers are still being disputed [l, 21. It is assumed that the enzyme is dimeric (2 x (2a3) in most nonionic detergents and also possibly in mitochondria, but monomeric forms have been reported for shark and camel cytochrome c oxidase [3, 41. Published procedures for conversion of the dimeric bovine enzyme to monomers involve treatment at elevated pH with high concentrations of Triton and DEAE chromatography. During such treatment the oxidase has a tendency to lose some small subunits and sometimes also the large hydrophobic subunit III [5]. In media of low ionic strength, cytochrome c oxidase shows peculiar biphasic steady-state kinetics, ascribed to a high-affinity cytochrome c reaction site (K, = lOwa M) and a low-affinity one (K,,, = 10m5 M) [6]. The question of whether these kinetics are inherent to an oxidase dimer with cooperativity or are an intrinsic property of the monomer has not yet been settled. In this paper we present separation by a size exclusion HPLC technique of relatively stable monomeric and dimeric bovine heart cytochrome c oxidase in buffers containing laurylmaltoside. Our evidence indicates that the monomeric enzyme can display biphasic kinetics. MATERIALS AND METHODS Bovine heart cytochrome c oxidase was purified according to a modified version of Yonetani [7] and Fowler et al.‘s [8] methods in our laboratory 191. The Fowler procedure involves selective solubilization of cytochrome c oxidase with deoxycholate, whereas, in the Yonetani procedure, all cytochromes are solubilized by excess cholate followed by selective denaturation of complex III. Horse heart cytochrome c was purified according to Margoliash and Walasek [lo]. The covalent 1 : 1 complex of bovine heart cytochrome c oxidase-yeast iso-lcytochrome c was prepared as described by Fuller et al. [I 11, using DTNB. Sizeexclusion chromatography of bovine heart cytochrome c oxidase was performed on a TSK SW 3000 (600 x 7.5 mm) column (Toyo Soda, Japan) at 20°C at a flow rate of 0.5 ml/min. The HPLC system consisted of LKB 2150 pumps, Rheodyne 7125 injector and an LKB 2158 Uvicord SD detector operated at 275 nm or 405 nm, light path 2.5 mm. Equilibration of the column and elution of the proteins was done with either a buffer containing 100 mM Trisjacetate, 1 mM EDTA, 300 mM NaCl (pH 7.5, I = 388 mM; the “high ionic strength buffer”) or a buffer containing 50 mM Trislacetate, 1 mM EDTA (pH 7.5, I = 45 mM; the “low ionic strength buffer”). To both buffers laurylmaltoside (0.2% w/v) was added as a detergent. Calibration of the TSK SW 3000 column was done by using the following markers: NADH-Q-oxidoreductase M, = 1.6 x lo6 (a generous gift from P. Bakker); thyroglobulin M, = 669 x 103, ferritin M, = 440 x 103, catalase M, = 235 x 103, aldolase M, = 158 x lo3 (all purchased from Pharmacia, Sweden); and ceruloplasmin M, = 130 x lo3 (a generous gift from R. A. Cuperus). Cytochrome c oxidase activity was measured spectrophotometrically as previously described 1121. Polyacrylamide gel electrophoresis in the presence of
Separation of Bovine Cyt c Oxidase
383
sodium dodecyl sulphate was performed according to Kadenbach [13]. Sedimentation velocity analysis was conducted according to Schachman [ 141, using an MSE analytical ultracentrifuge model Mk2. Yeast cytochrome c and DTNB were purchased from Sigma Chemical Company (St. Louis, MO, USA),andlauQlmaltoside from Calbiochem-Behring Corp. (La Jolla, CA, USA).
RESULTS Aggregation State of Cytochrome c Oxidase at High Ionic Strength A sample of purified cytochrome c oxidase, prepared according to our modified Fowler procedure [9], was subjected to gel permeation HPLC in a medium of high ionic strength (I = 388 mM) with 0.2% (w/v) laurylmaltoside as a detergent. As shown in Figure 1, the enzyme eluted as a single sharp 275 nm absorbance peak with an apparent molecular mass of 2. lo5 as calculated from the retention times of a series of calibration proteins (see Methods). A sample of cytochrome c oxidase purified according to the method of Yonetani [7], however, showed two sharp 275 nm absorbance peaks of about equal intensity (Fig. lB, solid line), the positions of which correspond to proteins with a mass of 2. lo5 and 4. 105, respectively. A sedimentation run in the analytical ultracentrifuge in the same medium showed the presence of two species differing by a factor 2 in apparent molecular mass. We conclude that in the preparation both monomers (1 x a~~) and dimers (2 x uu3) were present. A duplicate HPLC run of the Yonetani cytochrome c oxidase preparation with 405 nm detection (Fig. lB, broken line) again showed two peaks of nearly equal absorbance. This indicates that the heme a/protein ratio is the same in both peaks, which is in line with the idea that the two peaks represent monomeric and dimeric forms of the enzyme. Furthermore, both peaks of the Yonetani-type preparation, as well as the single peak of the Fowler-type preparation, displayed the same subunit pattern within the limits of resolution when assayed in the SDS-urea polyacrylamide gel electrophoretic system according to Kadenbach [ 131 (not shown in the figures). The trace of high molecular weight contaminant visible in the elution pattern of Figure 1B is low in 405 nm absorbance and probably represents NADH-ubiquinol oxidoreductase [ 151. The observation that cytochrome c oxidase dimers and monomers elute as distinct peaks suggests that no rapid equilibrium exists under these high ionic strength conditions; otherwise, one fused peak at intermediate position would result. Indeed, rechromatography of samples taken from the peak fractions showed that dimer and monomer remained dimeric and monomeric, respectively (not shown in the figures). The possibility that the occurrence of two peaks was a detergent effect caused by residual Tween-80, originally present in our Yonetanitype enzyme preparation, therefore, is less likely.
Effect of Cytochrome c on Aggregation State In order to investigate the effect of cytochrome c on the aggregation state of the oxidase, we prepared covalent complexes from DTNB-activated yeast iso-lcytochrome c and cytochrome c oxidase. Noncovalent cytochrome c-cytochrome c
384
T. B. M. Hakvoort et al.
0
ELUTION
30
60 TI ME (MINI
FIGURE 1. Elution profile of gel permeation HPLC of cytochrome c oxidase in a buffer of 388 mM ionic strength in the presence of 0.2% laurylmaltoside. Absorbance was measured at 275 nm (---) and 405 nm (---). (A) Fowler-type oxidase preparation; (B) Yonetani-type oxidase preparation; (C) covalent 1 : 1 complex of Fowler-type oxidase and yeast iso-1-cytochrome c. For details see Methods section.
oxidase complexes will fully dissociate at the high ionic strength used during chromatography. The activated thiol group of Cys 103 in yeast iso-1-cytochrome c slowly forms a disulphide bridge with a thiol group in subunit III of the bovine oxidase [ 111. Starting with either the Fowler-type or the Yonetani-type preparation, we prepared covalent cytochrome c-cytochrome c oxidase complexes with an overall stoicheiometry of 1 c per aa3. The Fowler 1 : 1 complex showed 60% dimers (Fig. 1C) and the 1 : 1 complex made from the Yonetani-type oxidase was 80 % dimeric when assayed at high ionic strength by HPLC. The 405 nm/275 nm absorbance ratio was the same for both peaks, suggestive of a monomeric and a dimeric 1 : 1 complex, respectively. So, in the covalent complex the equilibrium shifted to the dimer side. When Fowler-type cytochrome c oxidase (175 PM) was incubated with horse cytochrome c (122 PM) at high ionic strength in the presence of 10 mM ascorbate, the mixture rapidly reached anaerobiosis. A sample taken soon afterwards showed 10% dimers, a second sample taken 1 hr after anaerobiosis was 50% dimeric.
Separation
of Bovine Cyt c Oxidase
385
A 0.16-
0.12-E E N ti
0.08-
: CL z z
0.04-
0 0
-1. 10
L
20
40
50
ELUT ION T I ME (rnrd
60
70 TN (~‘1
FIGURE 2. (A) HPLC elution profile of Fowler-type cytochrome c oxidase in a buffer of 45 mM ionic strength containing 0.2% laurylmaltoside. (B) Eadie-Hofstee plot of spectrophotometric cytochrome c oxidase activity assay of the monomeric peak fraction (indicated by bar in Fig. 2A) in the same 45 mM ionic strength medium. Reduced cytochrome c 0.1-60 pM. The curved line is a computer fit through the points, the mathematical sum of two straight lines (indicated in the figure) corresponding to the high- and low affinity reactions, respectively. For details see Ref. 12.
Whether this conversion is due to the anaerobiosis, to the slow turnover, to the presence of cytochrome c (at high ionic strength not significantly bound to the oxidase), or to formation of pulsed oxidase [ 161, remains to be investigated. The cytochrome c oxidase of course reoxidizes upon injection into the HPLC buffer flow. Aggregation State of Cytochrome c Oxidase at Low Ionic Strength Cytochrome c oxidase is most stable and soluble at high ionic strength. Decreasing the ionic strength induces the risk of protein precipitation in a column that is difficult to regenerate once clotted. We discovered it was possible to run the oxidase in 50 mM Tris/acetate (pH 7.5, I = 45 mM) plus 0.2 % laurylmaltoside. The Fowler-type preparation, again, displayed a sharp monomeric peak (Fig. 2A) and even the Yonetani-type preparation eluted at the monomer position, though the peak showed broadening at the front (high-molecular) side. At low ionic strength, the Yonetani-type preparation was not very stable. Chromatography frequently led to the loss of some small subunits. The covalent complexes were still partially dimeric at low ionic strength. At low ionic strength incubation of the Fowler-type preparation under turnover conditions only led to slight broadening of the
386
T. B. M. Hakvoort et al.
monomer peak. We may tentatively conclude, therefore, that in the presence of laurylmaltoside low ionic strength promotes formation of monomeric cytochrome c oxidase. In the same medium of ionic strength 45 mM with 0.2% laurylmaltoside, in which the Fowler-type cytochrome c oxidase behaved as a pure monomer (Fig. 2A), we determined the cytochrome c oxidase activity of an HPLC peak sample. The spectrophotometric assay was used with reduced cytochrome c as substrate. As shown in Figure 2B, a curved Eadie-Hofstee plot was found that could be analyzed as the sum of a high-affinity and a low-affinity phase (Fig. 2B, straight lines). It is concluded that monomeric cytochrome c oxidase displays the characteristic biphasic kinetics often described for the enzyme at low ionic strength. The oxidation of ferrocytochrome c remains strictly first-order, so there is no change in activity during the assay. Furthermore, it is unlikely that the strong dilution of the monomeric HPLC peak sample in the activity assay would induce dimerisation.
DISCUSSION A significant result of the present study is the possibility to isolate pure monomeric and dimeric forms of bovine cytochrome c oxidase in the same buffer and detergent. This implies that the activation energy for interconversion and equilibration of the two forms must be rather high, at least in our high ionic strength medium. Our results concerning the stability of monomers and dimers are in agreement with the observations of Azzi et al. [17], who found two separate peaks under certain conditions in gel filtration experiments of cytochrome c oxidase in the presence of laurylmaltoside. At low ionic strength, our Yonetanitype preparation essentially was monomeric with a broadened leading edge of the peak. This suggests that, under these conditions, a more rapid equilibrium is established between the two aggregation states. Thompson and Ferguson-Miller [ 181 studied rat liver cytochrome c oxidase by gel filtration in a medium with an ionic strength of approximately 100 mM in the presence of laurylmaltoside. They concluded that the enzyme was predominantly present in the monomeric form. Their cytochrome c oxidase preparation, however, was depleted of subunit III. Other studies [5] have shown that conditions inducing removal of subunit III, for instance high pH in the presence of 5% Triton X-100, also promote monomerisation of the oxidase. This may be explained by the assumption that subunit III is situated in the M2 membrane traversing domains of the enzyme molecules that are thought to be in close contact in the dimer [ 191. The monomeric as well as the dimeric cytochrome c oxidase isolated in our study does contain subunit III. The presence of subunit III might be a prerequisite for formation of dimers. We have found that, under conditions where the oxidase is monomeric, the covalent cytochrome c oxidase-cytochrome c complex is largely dimeric. This can be understood in light of Capaldi et al.‘s [20] suggestion that cytochrome c binds in a cleft between the monomers in a cytochrome c oxidase dimer. At pH 7.5 the cytochrome c oxidase protein has a negative net charge, and the positively charged cytochrome c may help overcome electrostatic repulsion between the oxidase monomers in the dimer.
Separation of Bovine Cyt c Ox&se
387
Because laurylmaltoside only recently has been introduced as an excellent detergent for cytochrome c oxidase [21] most studies on the aggregation state of the enzyme have used other nonionic detergents such as Tween-80 and Triton X100, or the anionic (deoxy)cholate. Bovine heart cytochrome c oxidase tends to be 90% dimeric in nonionic detergents [3], but the chicken and dogfish heart enzymes were present as a mixed population of monomers and dimers. Cytochrome c oxidase isolated from camel [4] and hammerhead shark [22], however, were reported to be more than 90 % monomeric. Most investigators agree that monomeric, as well as dimeric oxidase, are capable of rapid electron transfer from reduced cytochrome c to oxygen [4, 5, 18, 223. Initially, the dimer was suggested to be the active species [3], but Wilson et al. [22] reported that monomeric dogfish oxidase was the most active form, and subsequently Capaldi et al. [5] stated that any role for the dimer would be limited to a control function and/or an as yet undefined role in coupling electron transfer to proton pumping. Explicit statements about the absence or presence of the so-called high- and low affinity cytochrome c binding sites in monomers (only to be observed at low ionic strength [6]) are rare. Thompson et al. [18] imply that the monomer displays the biphasic kinetics. On the other hand, Nalecz et al. [17] conclude that biphasic kinetics are caused by negative cooperativity in the dimer, and monomeric oxidase only shows the high affinity, low turnover site. As yet we have no explanation for this apparent discrepancy between our results and those of Azzi et al. One should be cautious in extrapolating the aggregation state of the oxidase sample to the actual situation prevailing during the activity measurement. Nevertheless, we feel our results suggest that the presence of high- and low affinity cytochrome c reactions are an intrinsic property of the monomeric enzyme. The authors thank H. Plot for performing the sedimentation velocity experiments. This work was supported by grants from the Netherlands Organization for the Advancement of Pure Research (2. W.O.) under the auspices of the Netherlands Foundation for Chemical Research (S.O.N.).
REFERENCES 1.
2.
R. A. Capaldi, F. Malatesta, and V. M. Darley-Usmar,
Biochim. Biophys. Acta 726, 135-148 (1983). G. Buse, G. C. M. Steffens, and L. Meinecke, in Structure and Function of Membrane Proteins, E. Quagliariello and F. Palmieri, Eds., pp. 131-138, Elsevier Science Publishers,
5.
Amsterdam (1983). N. C. Robinson and R. A. Capaldi, Biochemistry 16, 375-381 (1977). V. M. Darley-&mar, N. Alizai, A. I. Al-Ayash, G. D. Jones, A. Sharpe, and M. T. Wilson, Comp. Biochem. Physiol. 68b, 445-456 (1981). G. Georgevich, V. M. Darley-Usmar, F. Malatesta, and R. A. Capaldi, Biochemistry 22, 1317-
6. I. 8. 9.
1322 (1983). S. Ferguson-Miller, D. L. Brautigan, and E. Margoliash, J. Biol. Chem. 251, 1104-1115 (1976). T. Yonetani, J. Biol. Chem. 235, 845-852 (1960). L. R. Fowler, S. H. Richardson, and Y. Hatefi, Biochim. Biophys. Acta 64, 170-173 (1962). C. R. Hartze.11, H. Beinert, B. F. van Gelder, and T. E. King, Methods Enzymol. 53, 54-66
3. 4.
10. 11.
12.
(1978). E. Margoliash and 0. F. Walasek, Methods Erqvmol. 10, 339-348 (1967). S. D. Fuller, V. M. Darley-Usmar, and R. A. Capaldi, Biochemistry 20, 7046-7053 (1981). K. M. C. Sinjorgo, J. H. Meijling, and A. 0. Muijsers, Biochim. Biophys. Acta 767, 47-56
(1984).
388
T. B. M. Hakvoort
et al.
13. 14. 15.
B. Kadenbach, J. Jaraush, R. Hartmann, and P. Merle, Anal. Biochem. 129, 517-521 (1983). H. K. Schachman, Ultrucentrifugation in Biochemistry, Academic Press, New York (1959). R. Penniall, W. B. Elliot, K. J. H. van Buuren, and B. F. van Gelder, J. Bioenerg. Biomembr. 8,
16.
271-292 (1976). M. Brunori, A. Colosimo,
G. Rainoni, M. T. Wilson, and E. Antonini, J. Biol. Chem. 254, (1979). K. A. Nalecz, R. Bolli, and A. Azzi, Biochem. Biophys. Res. Commun. 114, 822-828 (1983). 10769-10775
17. 18. 19. 20. 21. 22.
D. A. Thompson and S. Ferguson-Miller, Biochemistry 22, 3178-3187 (1983). T. G. Frey, M. J. Costello, B. Karlson, J. C. Haselgrove, and 1. S. Leigh, J. Mol. Biol. 162, 113130 (1982). R. A. Capaldi, V. M. Darley-Usmar, S. Fuller, and F. Millett, FEBS Lett. 138, l-7 (1982). P. Rosevaer, T. van Aken, J. Baxter, and S. Ferguson-Miller, Biochemistry 19, 4108-4115 (1980). M. T. Wilson, W. Lalla-Maharajh, V. M. Darley-Usmar, J. Bonaventura, C. Bonaventura, and M. Brunori, J. Biol. Chem. 255, 2722-2728 (1980).
Received and accepted November 1984