Proc. Natl. Acad. Sci. USA
Vol. 76, No. 6, pp. 2664-2668, June 1979 Biochemistrv
Ion-channel component of cytochrome oxidase (ion transfer complex/energy coupling)
MITCHELL FRY AND DAVID E. GREEN Institute for Enzyme Research, University of Wisconsin-Madison, Madison, Wisconsin 53706
Contributed by David E. Green, March 19, 1979
ABSTRACT Cytochrome oxidase is resolvable into an electron transfer complex and an ion transfer complex. The ion transfer complex has been shown to have the capability for inducing nonspecific ion transport into liposomes. Subunit I of cytochrome oxidase has been identified as an ion-channelforming protein. The paired moving charge (PMC) model of mitochondrial energy coupling (1) has laid the theoretical foundation for the practical dissection of energy coupling systems. Theory predicted that energy coupling systems be resolvable into exergonic (energy driving) and endergonic (energy driven) components. The endergonic unit(s) must necessarily be in close juxtaposition to the exergonic unit for effective charge coupling between electron/cation transfer (2). The most obvious examples of energy coupling systems are the electron transport complexes of mitochondria. It has been shown in this laboratory that such complexes are resolvable into an electron transfer complex (ETC) and an ion transfer complex (ITC) (3-5). We found that the ITC of cytochrome oxidase was composed essentially of subunits I, II, and III (4); it is these three subunits that are coded for by the mitochondrial genome (6, 7). An essential requirement for this endergonic center (and others) is that it should possess the capability for ion transport. The ITC of cytochrome oxidase may now be described in terms of only one subunit-i.e., that subunit responsible for ion transport. Treatment of the inner membrane of mitochondria with sodium dodecyl sulfate leads to the dissociation of more than 30 subunits (8). The functions of these subunits are recognized only when they contain prosthetic groups such as heme, or when they are recognized as an essential component of some active enzyme preparation. The large molecular weight subunits of cytochrome oxidase belong to this class of nondescript entities that are often conveniently classified as "structural proteins." The present communication indicates that one of these large cytochrome oxidase subunits holds the key to the problem of mitochondrial energy coupling.
incorporated into phospholipid liposomes. The superficial resemblance of liposomes to biological membranes is obvious and has led to many studies of their properties, particularly their permeability characteristics. Their permeability characteristics for inorganic cations, anions, and water are qualitatively similar to those of biological membranes (9). The ITC-mediated influx of cations into the aqueous spaces of liposomes w as measured by two techniques. In one, a salt solution of the test cation containing the radioactively labeled species was added to a liposome suspension to produce a concentration gradient of that salt across the liposomal membrane. The amount of salt thus entering the liposome was then determined by scintillation counting after removal of untrapped salt by gel filtration or centrifugation. Alternatively, changes in the light-scattering properties of a liposome suspension were measured in response to the addition of salt solutions of increased osmolarity. Equilibrium and kinetic data have shown that liposomes prepared with KCI in the osmotically active spaces obey the Boyle-van't Hoff law and behave as ideal osmometers (10, 11). Salts or other compounds that normally do not penetrate the liposomal membrane will cause an osmotic imbalance resulting in liposome shrinkage (increase in light scattering) or swelling (decrease in light scattering), representing liposome water permeability (12). ITC Incorporation into Liposomes. The ITC subunits have a high degree of hydrophobicity (13) and when resolved from cytochrome oxidase are in a very aggregated and water-insoluble state. Before effective incorporation of ITC into liposomes could be achieved, the ITC had to be first converted to a more dispersed form. After dialysis of ITC protein against 100 vol of distilled water (1 mM in EDTA) for 12 hr at 4°C, an aqueous suspension of lipid (either asolectin or egg lecithin) was added to one half the weight of ITC protein. This mixture was sonicated with a Branson Sonifier at 50 W for 6 min while maintained in an iced water bath. Lipidated ITC protein was then removed by centrifugation at 30,000 rpm in a no. 30 Spinco rotor for 30 min and finally resuspended in 0.25 M sucrose/25 mM KCI/20 mM Tris-HCI, pH 7.4, to approximately 20 mg of protein per ml. Two methods of ITC incorporation into liposomes were routinely used. The lipidated ITC could be added to an aqueous suspension of phospholipid and the whole sonicated for 6 min at 50 W while maintained in an iced water bath. The clear liposome suspension was then left for at least 2 hr at room temperature under nitrogen before use. Alternatively, lipidated ITC could be added directly to a similarly sonicated phospholipid suspension and incorporation achieved by a 15-min incubation at 37°C. Either method of incorporation produced similar results with regard to ITC function.
METHODS The ITC of Cytochrome Oxidase. The ITC of cytochrome oxidase as isolated by an acid/alcohol method (4) was composed essentially of subunits I, II, and III. By slight modifications in this method of resolution of ITC, the proportions of these three subunits may be varied (see Fig. 1). Thus it becomes possible to test a range of ITC fractions for the capability of ion transport and thereby deduce the active element(s). Demonstration of ITC Function. The essential role of the ITC-i.e., mediation of ion transport-was studied with ITC The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U. S. C. §1734 solely to indicate this fact.
Abbreviations: ETC, electron transfer complex; ITC, ion transfer complex. 2664
Biochemistry: Fry and Green RESULTS Variations on our original theme for the resolution of cytochrome oxidase into ITC and ETC (4) produced ITC fractions that varied in their relative proportions of subunits I, II, and III (Fig. 1). From an ITC fraction enriched in subunits I and III (similar to that shown in Fig. 1A), sodium dodecyl sulfate gel filtration on Sephadex G-150 following the experimental procedure of Yu and Yu (14) yielded electrophoretically pure fractions of subunits I and III. Sodium dodecyl sulfate/urea gel electrophoresis was performed according to Swank and Munkres (15). The major modification to our original method of resolution of cytochrome oxidase was to suspend samples of washed cytochrome oxidase in 0.1 M phosphate buffer prior to resolution. This procedure markedly improved the resolution of subunits I and III in the ITC fraction and resulted in almost complete recovery of protein in the ETC fraction. Previously we had used cytochrome oxidase isolated in phosphate buffer but resuspended in Tris-HC1 buffer prior to resolution (4). Fig. ID shows the ITC fraction obtained from cytochrome oxidase suspended throughout in Tris-HCl buffer. Incorporation of ITC, ETC, and purified subunits into asolectin or egg lecithin liposomes induced a range of permeabilities towards mono- and divalent cations (Table 1). Examination of the subunit profiles of the various fractions in Fig. 1 indicated the importance of subunit I in inducing cation permeability, and the use of purified subunit I confirmed this indication. ETC fractions containing only small amounts of subunit I were relatively devoid of ion transport capability. The ITC-induced permeability towards metal cations was nonspecific with regard to valency, both monovalent and divalent cations being transported with equal effectiveness. Liposomes with incorporated ITC were also permeable to uncharged, hydrophilic molecules such as glucose and sucrose. The ITC was equally effective with regard to ion transport in liposomes
Proc. Natl. Acad. Sci. USA 76 (1979)
2665
constructed of either egg lecithin or asolectin phospholipids. WhhI'Ab6ed cation is present during sonication of liposomes, with or without incorporation of ITC, identical amounts were found to be trapped in both cases. This observation precluded the possibility that ITC protein might be binding large amounts of cation and assured that the presence of ITC in liposomes had little effect on the state of such vesicles. Increasing the proportion of liposomes with incorporated subunit I led to a parallel increase in the number of liposomes able to trap cations through ITC mediation (Table 2). Calculation indicated that approximately two molecules of subunit I per liposome was necessary for maximal cation influx. Further increase in the amounts of incorporated subunit I per liposome had little effect on the amount of trapped cation, a further indication of the stability of liposomes with incorporated subunit I. Because liposome suspensions were diluted some five times by separation through gel filtration, we have also substantiated these results by the sedimentation of liposomes, thus avoiding any unforeseen dilution effects. The results shown in Table 3 confirmed those experiments using the column technique. Changes in the light absorbance of a liposome suspension in response to osmotic fluctuations is a quick and convenient method by which to follow the permeability of such liposomes towards various salts. The data in Table 4 fully supported those measurements on the distribution of radioactive cations. Control liposomes (no ITC) have a limited permeability towards cations; production of a high external salt concentration therefore resulted in liposome shrinkage by virtue of the efflux of water from such liposomes. However, the results given in Table 4 clearly -indicated that subunit I and ITC fractions were able to mediate the influx of salts into liposomes with a concomitant net swelling of the liposome. Necessarily high concentrations of salts in these experiments precluded the use of divalent cations because these interfered with the assay and liposome stability.
Migration FIG. 1. Densitometric traces of sodium dodecyl sulfate/urea/polyacrylamide gels of various ITC and ETC fractions of cytochrome oxidase. Cytochrome oxidase was resolved by an acid/alcohol procedure as described (4) with the following modifications: (A) ITC cytochrome oxidase suspended in 0.1 M phosphate buffer, pH 7.4, prior to resolution without reduction by sodium dithionite. Exposure of cytochrome oxidase to the acid/alcohol mixture was initially at 380C with rapid cooling to 00C followed by immediate separation of the ITC precipitate by centrifugation. (B) The ETC of A was harvested after neutralization of the mixture. (C) An ITC fraction obtained from an additional exposure of the ETC of B to the acid/alcohol mixture at 380C for 1 hr. (D) An ITC of cytochrome oxidase reduced with sodium dithionite prior to resolution and suspended in 50 mM Tris-HCl, pH 8.0, with incubation at 380C for 10 min in acid/alcohol. (E) Gel profile of subunit I purified by sodium dodecyl sulfate filtration on Sephadex G-150 from an ITC fraction similar to A.
2666
Biochemistry: Fry and Green
ITC fraction incorporated*
Proc. Natl. Acad. Sci. USA 76 (1979)
Table 1. The ITC-induced influx of cations into liposomes Amount of label incorporated into liposomes expressed Type of as a percentage of label trapped by sonicationt 4CJGlucose f14C Dextrant liposome 45Cat 85Srt 22Na 86Rb I
A 44 Asolectin 46 37 :36 32 8 B Asolectin 11 13 7 8 7 8 C Asolectin 48 49 42 38 38 8 D Asolectin 33 38 26 21 25 9 Subunit I 93 95 Asolectin 88 82 81 9 5 5 Subunit III Asolectin 4 6 8 9 NoITC Asolectin 6 7 4 7 A Egg lecithin 50 49 38 37 41 2 B 6 Egg lecithin 12 13 8 10 8 C 53 Egg lecithin 52 44 40 44 5 D Egg lecithin 36 39 29 26 28 7 Subunit I Egg lecithin 98 98 91 89 84 4 No ITC Egg lecithin 5 5 3 4 An aqueous phospholipid suspension (50 mg/ml), 20 mM in Tris-HC1 (pH 7.4), 25 mM in KCI, 0.25 M in sucrose, and 1 mM in the chloride salt of the test cation, was sonicated for 6 mih at 50 W while maintained in an iced water bath. Where ITC was incorporated it was present at a concentration of 0.4 mg of protein per 50 mg of phospholipid. After sonication the liposome suspension was allowed to stand 2 hr at room temperature under nitrogen before use. To I -ml aliquots of the liposome suspension was added 100 ,ul of the sucrose/KCl/Tris-HCl medium containing the labeled compound to give a final concentration of the test compound of 5 mM and a final concentration of KCI of 100 mM. After 1 min of incubation at room temperature the untrapped label was removed from the liposomes by passage through a Sephadex G-25 (coarse) column (1.5 X 20 cm), eluting with sucrose/KCI/Tris.HCI, 5 mM in the unlabeled test compound and 100 mM in KCl. Elution time was approximately 5 min. * See Fig. I for ITC fractions. t Liposomes (with and without ITC) were sonicated in the presence of labeled test compound (5 mM). After a 2-hr equilibration, untrapped label was removed by gel filtration. The amount of trapped label was taken to represent the maximum amount of sequesterable compound. Results are therefore expressed as a percentage of this upper limit. I Results with these cations were calculated after subtracting control influx values (no ITC). This was done because appreciable amounts of these cations are bound to the surface of the liposomes (particularly those formed from asolectin). § I'4C]Dextran (carboxyl labeled), Mr 16,000.
The results shown in Tables 1-4 were obtained by applying a salt gradient across the liposomal membrane. Therefore, salt influx into liposomes, mediated by ITC, was driven by a salt concentration gradient. However, when labeled cation was added to the external liposome mileu avoiding any significant salt gradients, then very little cation was taken up by liposomes with incorporated ITC. Under these conditions, such cation uptake was by a slow exchange process, only some 10% of the external test cation being sequestered by liposomes with in-
corporated ITC after 1 hr of incubation. In order to show that the ITC was capable of mediating the efflux of cations from liposomes, EDTA was used to "pull" pretrapped calcium from liposomes with incorporated ITC. External calcium, as well as that bound to the external liposome surface, was effectively removed from normal liposomes by EDTA treatment without release of trapped calcium. However, liposomes with incorporated subunit I effectively lost all their trapped calcium by EDTA chelation.
Table 2. Effect of subunit I concentration on amount of cation influx into liposomes* Amount of label incorporated into liposomes expressed as a percentage Subunit I, Liposome system of label trapped (asolectin), approximate moles mg subunit I per by sonication per liposomet 45Ca 22Na 50 mg phospholipid
Table 3. ITC-mediated influx of cations into liposomes studied by sedimentation of liposomes ITC fraction Amount of label incorporated into liposomes expressed incorporated as a percentage into egg of label trapped by sonication* lecithin 45Ca 22Na liposomes* lI4ClGlucose
0 0.1 0.2 0.3 0.4 0.5 1.0
0 0.4
0.8 1.2 1.6 2.0 4.0
20 42 64 95 96 92
5 17 38 61 89 92 86
* See Table 1 for details. The particle weight of a unilamellar, single-compartment liposome is about 2.1 X 106 (16). The electron microscopic appearance of the sonicated liposomes used in these studies (formed from asolectin) indicate them to be essentially multilamellar from which we have assumed an approximate particle weight of 6 X 106. The M, of subunit I is assumed to be approximately 35,000 (17).
32 34 A 48 6 8 15 B 43 52 50 C 81 92 96 Subunit I A 3-ml liposome suspension (50 mg of egg lecithin per ml) was diluted to 4 ml with the sucrose/KCI/Tris-HCl medium containing the labeled cation to give a final concentration of the test compound of 5 mM and a KCl concentration of 100 mM. The liposome suspension was then sedimented in a Spinco 50 rotor by centrifugation at 50,000 rpm for 1 hr. The liposome pellet was gently rinsed twice with the sucrose/KCl/Tris-HCl medium (containing 100 mM KCI) and finally resuspended to a total volume of 1 ml. Samples (0.5 ml) of this were taken for isotope counting. Results are expressed after subtraction of control (no ITC) values. * See Table I for details.
Proc. Natl. Acad. Sci. USA 76 (1979)
Biochemistry: Fry and Green Table 4. The ITC-induced influx of cations into liposomes followed by light-scattering change-&: Magnitude of light-absorbance change induced by addition of salt (A)tA ITC fraction RbCl NaCl KCl incorporated* +0.04 +0.04 +0.04 No ITC -0.05 -0.06 -0.05 A +0.03 +0.03 +0.03 B -0.07 -0.08 -0.07 C -0.04 -0.04 -0.03 D -0.16 -0.16 -0.16 Subunit I To a 3-ml (1-cm light path) silica cuvette was added 2.5 ml of the sucrose/KCI/Tris-HCl medium and 0.5 ml of liposome suspension [final lipid (asolectin) concentration = 8.3 mg/ml]. After a 1-min equilibration, 50 ul of sucrose/KCl/Tris-HCl medium was added containing sufficient amount of the test chloride salt to give a final concentration of 100 mM. Mixing was achieved by stirring and the absorbance of the suspension was read 1 min later at 520 nm. Measurements were made at room temperature in a Cary model 118 spectrophotometer. +, Absorbance increase (liposome shrinkage); -, absorbance decrease (liposome swelling). * See Fig. 1. t Liposomes (with and without ITC) were prepared as described in Table 1.
Liposomal influx of salts can produce osmotic changes within the liposome and such changes might induce "osmotic strain" in the liposome structure. This could result in the rapid iysis and resealing (with the concomitant trapping of salts) of the phospholipid structures. We have therefore checked the effect of such changes on liposome permeability to cations. Liposome suspensions were diluted by addition of water containing labeled test cation in sufficient amounts to cause a degree of liposome swelling approximately equal to that caused by salt additions. These results indicated that very little cation influx occurred during liposome swelling, with or without incorporated ITC. Few effective inhibitors of the ITC-mediated influx of cations into liposomes have been found to date. Fluorescein mercuric acetate and lanthanum are potent inhibitors, as was phosphatidylethanolamine. The effects of this phospholipid are particularly interesting and are the subject of further investigation referred to in the discussion. DISCUSSION The experiments reported in this communication support the idea that subunit I of cytochrome oxidase is an active transmembrane ion-channel protein. Incorporation of subunit I or ITC fractions into liposomes mediates the rapid influx of both monovalent and divalent cations under pressure of a salt gradient imposed across the liposomal membrane. Under conditions in which no salt gradient exists there occurs only a minimal exchange of cations mediated by the ITC. Control experiments have assured that such ion transport is indeed ITC mediated and not simply the result of "leaky" liposomes. The assay of channel-forming proteins is a new and challenging area of biochemistry. Such proteins must be incorporated into a lipid membrane phase in order to study their functions. [Kagawa (18) has given a recent review on this subject. I In the present studies we have used liposomes to study channel properties; such a technique gives a general awareness of the characteristics of channel activity but cannot measure the refined aspects of channel function. In unresolved cytochrome oxidase it is proposed that electron transport is coupled to movement of cations; cation movement is an endergonic process driven by the exergonic electron transport (2). Similarly, ion transport mediated by the resolved
2667
ITC of cytochrome oxidase is only demonstratable when coupleod Iaan !"exergoniccenter"-i.e., a salt concentration gradient. The nonspecificity for ions transported and the large molecular weight of subunit I argue for an ion-channel structure rather than an ion carrier (19). Subunit I possesses the general
properties required of a transmembrane channel protein. It is a highly hydrophobic protein with about 35% polar amino acids (20), buried deep in both the isolated enzyme and the mitochondrial membrane (21). In the membrane-bound enzyme, subunit I is almost completely inaccessible to all surface probes
(22, 23). Wikstrom et al. (24) have also proposed a channel structure for subunit I (and subunit II). These authors propose an ionic channel structure with specificity for protons being achieved by protein-bound "ordered" water. We believe the present results are more consistent with a relatively nonspecific ion channel. Is the ion transport capability of resolved ITC demonstrable in unresolved cytochrome oxidase? The answer to this question is affirmative. Enzymic activity can be shown to proceed in membrane-incorporated cytochrome oxidase with concomitant ion movements (3, 25). However, cytochrome oxidase incorporated into liposomes can also mediate the influx of ions under nonenergized conditions, the driving force here being a salt concentration gradient. Therefore, cytochrome oxidase can behave in a similar fashion to the resolved ITC. Respiratory control in membrane-incorporated cytochrome oxidase is an accepted but not understood phenomenon and is strongly dependent on the blend of phospholipids used for incorporation (26-28). Noted in the present article was the inhibition exerted on ion transport by phosphatidylethanolamine. We have been able to show that a variation in the blend of liposomal phospholipids can "stop or start" the ITC-mediated transport of ions; a similar correlation can be shown for nonenergized, unresolved cytochrome oxidase (provided the cytochrome oxidase is first depleted of its bound lipids). Thus, respiratory control can be equated with ion-channel capability in cytochrome oxidase and as such is not directly connected to electron transport. Hunter and Capaldi (29) have presented evidence that respiratory control in membranous vesicles of cytochrome oxidase is a localized molecular phenomenon and not a membrane phenomenon as implicit in the chemosmotic model of mitochondrial energy conservation. Note Added in Proof. Recent studies in this laboratory have established that subunits I and III of cytochrome oxidase are indistinguishable on the basis of their amino acid compositions. This raises the question as to which subunits of cytochrome oxidase actually constitute the ITC "'complex." Because subunit III is clearly an inactive species in ion translocation, the implication is that the dimeric polypeptide (i.e., subunit I) is the active form of the subunit ion channel.
This investigation was supported in part by Program Project Grant GM 12847 of the National Institute of General Medical Sciences. 1. Green, D. E. & Reible, S. (1974) Proc. Natl. Acad. Sci. USA 71, 4850-4854. Blondin, G. A. & Green, D. E. (1975) BioScience 28, 18-24. Kessler, R. J., Blondin, G. A., Vande Zande, H., Haworth, R. A. & Green, D. E. (1977) Proc. Natl. Acad. Sci. USA 74, 36623666. 4. Fry, M., Vande Zande, H. & Green, D. E. (1978) Proc. Natl. Acad. Sci. USA 75,5908-5911. 5. Fry, M. & Green, D. E. (1978) Proc. Natl. Acad. Sci. USA 75, 2. 3.
6.
5377-5380. Rubin, H. S. & Tzagoloff, A. (1973) J. Biol. Chem. 247, 594603.
2668
Biochemistry: Fry and Green
7. Mason, T. L. & Schatz, G. (1973) J. Biol. Chem. 248, 13551360. 8. Harmon, H. J., Hall, J. D. & Crane, F. L. (1974) Biochim. Blophys. Acta 344, 119-155. 9. Bangham, A. D., Standish, M. M. & Watkins, J. C. (1965) J. Mol. Biol. 13, 253-264. 10. de Gier, J., Mandersloot, J. G. & van Deenen, L. L. M. (1968) Biochim. Biophys. Acta 150,666-675. 11. Bittman, R. & Blau, L. (1972) Biochemistry 11, 4831-4839. 12. Jain, M. K., Toussaint, D. G. & Cordes, E. H. (1973) J. Membr. Biol. 14, 1-16. 13. Phan, S. H' & Mahler, H. R. (1976) J. Biol. Chem. 251, 270276. 14. Yu, C. A. & Yu, L. (1977) Biochim. Biophys. Acta 495, 248259. 15. Swank, R. J. & Munkres, K. D. (1971) Anal. Biochem. 39, 462-477. 16. Huang, C. (1969) Biochemistry 8,344-352. 17. Downer, N. W., Robinson, N. C. & Capaldi, R. A. (1976) Biochemistry 15, 2930-2936. 18. Kagawa, Y. (1978) Biochim. Biophys. Acta 505,45-93.
Proc. Natl. Acad. Sci. USA 76 (1979) 19. Urry, D. W., Long, M. M., Jacobs, M. & Harris, R. D. (1975) Ann. N. Y. Acad. Sci. 264, 203-220. 20. Poyton, R. 0. & Schatz, G. (1975) J. Biol. Chem. 250, 752761. 21. Eytan, G. D., Carroll, R. C., Schatz, G. & Racker, E. (1975) J. Biol. Chem. 250, 8598-8603. 22. Eytan, G. D. & Schatz, G. (1975) J. Biol. Chem. 250,767-774. 23. Eytan, G. D. & Broza, R. (1978) J. Biol. CIem. 253, 31963202. 24. Wikstrom, M., Saari, H., Penttila, T. & Saraste, M. (1978) FEBS Symp. 45, 85-94. 25. Hinkle, P. C., Kim, J. J. & Racker, E. (1972) J. Biol. Chem. 247, 1338-1339. 26. Eytan, G. D., Matheson, M. J. & Racker, E. (1976) J. Biol. Chem. 261,6831-6837. 28. Kagawa, Y., Johnson, L. W. & Racker, E. J. (1973) Biochem. Biophys. Res. Commun. 50,245-251. 27. Racker, E.'(1972) in Molecular Basis of Electron Transport, eds. Schultz, J. & Cameron, B. F. (Academic, New York), p. 45. 29. Hunter, D. R. & Capaldi, R. A. (1974) Biochem. Biophys. Res. Commun. 56, 623-628.
Vol. 76, No. 6, pp. 2664-2668, June 1979 Biochemistrv
Ion-channel component of cytochrome oxidase (ion transfer complex/energy coupling)
MITCHELL FRY AND DAVID E. GREEN Institute for Enzyme Research, University of Wisconsin-Madison, Madison, Wisconsin 53706
Contributed by David E. Green, March 19, 1979
ABSTRACT Cytochrome oxidase is resolvable into an electron transfer complex and an ion transfer complex. The ion transfer complex has been shown to have the capability for inducing nonspecific ion transport into liposomes. Subunit I of cytochrome oxidase has been identified as an ion-channelforming protein. The paired moving charge (PMC) model of mitochondrial energy coupling (1) has laid the theoretical foundation for the practical dissection of energy coupling systems. Theory predicted that energy coupling systems be resolvable into exergonic (energy driving) and endergonic (energy driven) components. The endergonic unit(s) must necessarily be in close juxtaposition to the exergonic unit for effective charge coupling between electron/cation transfer (2). The most obvious examples of energy coupling systems are the electron transport complexes of mitochondria. It has been shown in this laboratory that such complexes are resolvable into an electron transfer complex (ETC) and an ion transfer complex (ITC) (3-5). We found that the ITC of cytochrome oxidase was composed essentially of subunits I, II, and III (4); it is these three subunits that are coded for by the mitochondrial genome (6, 7). An essential requirement for this endergonic center (and others) is that it should possess the capability for ion transport. The ITC of cytochrome oxidase may now be described in terms of only one subunit-i.e., that subunit responsible for ion transport. Treatment of the inner membrane of mitochondria with sodium dodecyl sulfate leads to the dissociation of more than 30 subunits (8). The functions of these subunits are recognized only when they contain prosthetic groups such as heme, or when they are recognized as an essential component of some active enzyme preparation. The large molecular weight subunits of cytochrome oxidase belong to this class of nondescript entities that are often conveniently classified as "structural proteins." The present communication indicates that one of these large cytochrome oxidase subunits holds the key to the problem of mitochondrial energy coupling.
incorporated into phospholipid liposomes. The superficial resemblance of liposomes to biological membranes is obvious and has led to many studies of their properties, particularly their permeability characteristics. Their permeability characteristics for inorganic cations, anions, and water are qualitatively similar to those of biological membranes (9). The ITC-mediated influx of cations into the aqueous spaces of liposomes w as measured by two techniques. In one, a salt solution of the test cation containing the radioactively labeled species was added to a liposome suspension to produce a concentration gradient of that salt across the liposomal membrane. The amount of salt thus entering the liposome was then determined by scintillation counting after removal of untrapped salt by gel filtration or centrifugation. Alternatively, changes in the light-scattering properties of a liposome suspension were measured in response to the addition of salt solutions of increased osmolarity. Equilibrium and kinetic data have shown that liposomes prepared with KCI in the osmotically active spaces obey the Boyle-van't Hoff law and behave as ideal osmometers (10, 11). Salts or other compounds that normally do not penetrate the liposomal membrane will cause an osmotic imbalance resulting in liposome shrinkage (increase in light scattering) or swelling (decrease in light scattering), representing liposome water permeability (12). ITC Incorporation into Liposomes. The ITC subunits have a high degree of hydrophobicity (13) and when resolved from cytochrome oxidase are in a very aggregated and water-insoluble state. Before effective incorporation of ITC into liposomes could be achieved, the ITC had to be first converted to a more dispersed form. After dialysis of ITC protein against 100 vol of distilled water (1 mM in EDTA) for 12 hr at 4°C, an aqueous suspension of lipid (either asolectin or egg lecithin) was added to one half the weight of ITC protein. This mixture was sonicated with a Branson Sonifier at 50 W for 6 min while maintained in an iced water bath. Lipidated ITC protein was then removed by centrifugation at 30,000 rpm in a no. 30 Spinco rotor for 30 min and finally resuspended in 0.25 M sucrose/25 mM KCI/20 mM Tris-HCI, pH 7.4, to approximately 20 mg of protein per ml. Two methods of ITC incorporation into liposomes were routinely used. The lipidated ITC could be added to an aqueous suspension of phospholipid and the whole sonicated for 6 min at 50 W while maintained in an iced water bath. The clear liposome suspension was then left for at least 2 hr at room temperature under nitrogen before use. Alternatively, lipidated ITC could be added directly to a similarly sonicated phospholipid suspension and incorporation achieved by a 15-min incubation at 37°C. Either method of incorporation produced similar results with regard to ITC function.
METHODS The ITC of Cytochrome Oxidase. The ITC of cytochrome oxidase as isolated by an acid/alcohol method (4) was composed essentially of subunits I, II, and III. By slight modifications in this method of resolution of ITC, the proportions of these three subunits may be varied (see Fig. 1). Thus it becomes possible to test a range of ITC fractions for the capability of ion transport and thereby deduce the active element(s). Demonstration of ITC Function. The essential role of the ITC-i.e., mediation of ion transport-was studied with ITC The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U. S. C. §1734 solely to indicate this fact.
Abbreviations: ETC, electron transfer complex; ITC, ion transfer complex. 2664
Biochemistry: Fry and Green RESULTS Variations on our original theme for the resolution of cytochrome oxidase into ITC and ETC (4) produced ITC fractions that varied in their relative proportions of subunits I, II, and III (Fig. 1). From an ITC fraction enriched in subunits I and III (similar to that shown in Fig. 1A), sodium dodecyl sulfate gel filtration on Sephadex G-150 following the experimental procedure of Yu and Yu (14) yielded electrophoretically pure fractions of subunits I and III. Sodium dodecyl sulfate/urea gel electrophoresis was performed according to Swank and Munkres (15). The major modification to our original method of resolution of cytochrome oxidase was to suspend samples of washed cytochrome oxidase in 0.1 M phosphate buffer prior to resolution. This procedure markedly improved the resolution of subunits I and III in the ITC fraction and resulted in almost complete recovery of protein in the ETC fraction. Previously we had used cytochrome oxidase isolated in phosphate buffer but resuspended in Tris-HC1 buffer prior to resolution (4). Fig. ID shows the ITC fraction obtained from cytochrome oxidase suspended throughout in Tris-HCl buffer. Incorporation of ITC, ETC, and purified subunits into asolectin or egg lecithin liposomes induced a range of permeabilities towards mono- and divalent cations (Table 1). Examination of the subunit profiles of the various fractions in Fig. 1 indicated the importance of subunit I in inducing cation permeability, and the use of purified subunit I confirmed this indication. ETC fractions containing only small amounts of subunit I were relatively devoid of ion transport capability. The ITC-induced permeability towards metal cations was nonspecific with regard to valency, both monovalent and divalent cations being transported with equal effectiveness. Liposomes with incorporated ITC were also permeable to uncharged, hydrophilic molecules such as glucose and sucrose. The ITC was equally effective with regard to ion transport in liposomes
Proc. Natl. Acad. Sci. USA 76 (1979)
2665
constructed of either egg lecithin or asolectin phospholipids. WhhI'Ab6ed cation is present during sonication of liposomes, with or without incorporation of ITC, identical amounts were found to be trapped in both cases. This observation precluded the possibility that ITC protein might be binding large amounts of cation and assured that the presence of ITC in liposomes had little effect on the state of such vesicles. Increasing the proportion of liposomes with incorporated subunit I led to a parallel increase in the number of liposomes able to trap cations through ITC mediation (Table 2). Calculation indicated that approximately two molecules of subunit I per liposome was necessary for maximal cation influx. Further increase in the amounts of incorporated subunit I per liposome had little effect on the amount of trapped cation, a further indication of the stability of liposomes with incorporated subunit I. Because liposome suspensions were diluted some five times by separation through gel filtration, we have also substantiated these results by the sedimentation of liposomes, thus avoiding any unforeseen dilution effects. The results shown in Table 3 confirmed those experiments using the column technique. Changes in the light absorbance of a liposome suspension in response to osmotic fluctuations is a quick and convenient method by which to follow the permeability of such liposomes towards various salts. The data in Table 4 fully supported those measurements on the distribution of radioactive cations. Control liposomes (no ITC) have a limited permeability towards cations; production of a high external salt concentration therefore resulted in liposome shrinkage by virtue of the efflux of water from such liposomes. However, the results given in Table 4 clearly -indicated that subunit I and ITC fractions were able to mediate the influx of salts into liposomes with a concomitant net swelling of the liposome. Necessarily high concentrations of salts in these experiments precluded the use of divalent cations because these interfered with the assay and liposome stability.
Migration FIG. 1. Densitometric traces of sodium dodecyl sulfate/urea/polyacrylamide gels of various ITC and ETC fractions of cytochrome oxidase. Cytochrome oxidase was resolved by an acid/alcohol procedure as described (4) with the following modifications: (A) ITC cytochrome oxidase suspended in 0.1 M phosphate buffer, pH 7.4, prior to resolution without reduction by sodium dithionite. Exposure of cytochrome oxidase to the acid/alcohol mixture was initially at 380C with rapid cooling to 00C followed by immediate separation of the ITC precipitate by centrifugation. (B) The ETC of A was harvested after neutralization of the mixture. (C) An ITC fraction obtained from an additional exposure of the ETC of B to the acid/alcohol mixture at 380C for 1 hr. (D) An ITC of cytochrome oxidase reduced with sodium dithionite prior to resolution and suspended in 50 mM Tris-HCl, pH 8.0, with incubation at 380C for 10 min in acid/alcohol. (E) Gel profile of subunit I purified by sodium dodecyl sulfate filtration on Sephadex G-150 from an ITC fraction similar to A.
2666
Biochemistry: Fry and Green
ITC fraction incorporated*
Proc. Natl. Acad. Sci. USA 76 (1979)
Table 1. The ITC-induced influx of cations into liposomes Amount of label incorporated into liposomes expressed Type of as a percentage of label trapped by sonicationt 4CJGlucose f14C Dextrant liposome 45Cat 85Srt 22Na 86Rb I
A 44 Asolectin 46 37 :36 32 8 B Asolectin 11 13 7 8 7 8 C Asolectin 48 49 42 38 38 8 D Asolectin 33 38 26 21 25 9 Subunit I 93 95 Asolectin 88 82 81 9 5 5 Subunit III Asolectin 4 6 8 9 NoITC Asolectin 6 7 4 7 A Egg lecithin 50 49 38 37 41 2 B 6 Egg lecithin 12 13 8 10 8 C 53 Egg lecithin 52 44 40 44 5 D Egg lecithin 36 39 29 26 28 7 Subunit I Egg lecithin 98 98 91 89 84 4 No ITC Egg lecithin 5 5 3 4 An aqueous phospholipid suspension (50 mg/ml), 20 mM in Tris-HC1 (pH 7.4), 25 mM in KCI, 0.25 M in sucrose, and 1 mM in the chloride salt of the test cation, was sonicated for 6 mih at 50 W while maintained in an iced water bath. Where ITC was incorporated it was present at a concentration of 0.4 mg of protein per 50 mg of phospholipid. After sonication the liposome suspension was allowed to stand 2 hr at room temperature under nitrogen before use. To I -ml aliquots of the liposome suspension was added 100 ,ul of the sucrose/KCl/Tris-HCl medium containing the labeled compound to give a final concentration of the test compound of 5 mM and a final concentration of KCI of 100 mM. After 1 min of incubation at room temperature the untrapped label was removed from the liposomes by passage through a Sephadex G-25 (coarse) column (1.5 X 20 cm), eluting with sucrose/KCI/Tris.HCI, 5 mM in the unlabeled test compound and 100 mM in KCl. Elution time was approximately 5 min. * See Fig. I for ITC fractions. t Liposomes (with and without ITC) were sonicated in the presence of labeled test compound (5 mM). After a 2-hr equilibration, untrapped label was removed by gel filtration. The amount of trapped label was taken to represent the maximum amount of sequesterable compound. Results are therefore expressed as a percentage of this upper limit. I Results with these cations were calculated after subtracting control influx values (no ITC). This was done because appreciable amounts of these cations are bound to the surface of the liposomes (particularly those formed from asolectin). § I'4C]Dextran (carboxyl labeled), Mr 16,000.
The results shown in Tables 1-4 were obtained by applying a salt gradient across the liposomal membrane. Therefore, salt influx into liposomes, mediated by ITC, was driven by a salt concentration gradient. However, when labeled cation was added to the external liposome mileu avoiding any significant salt gradients, then very little cation was taken up by liposomes with incorporated ITC. Under these conditions, such cation uptake was by a slow exchange process, only some 10% of the external test cation being sequestered by liposomes with in-
corporated ITC after 1 hr of incubation. In order to show that the ITC was capable of mediating the efflux of cations from liposomes, EDTA was used to "pull" pretrapped calcium from liposomes with incorporated ITC. External calcium, as well as that bound to the external liposome surface, was effectively removed from normal liposomes by EDTA treatment without release of trapped calcium. However, liposomes with incorporated subunit I effectively lost all their trapped calcium by EDTA chelation.
Table 2. Effect of subunit I concentration on amount of cation influx into liposomes* Amount of label incorporated into liposomes expressed as a percentage Subunit I, Liposome system of label trapped (asolectin), approximate moles mg subunit I per by sonication per liposomet 45Ca 22Na 50 mg phospholipid
Table 3. ITC-mediated influx of cations into liposomes studied by sedimentation of liposomes ITC fraction Amount of label incorporated into liposomes expressed incorporated as a percentage into egg of label trapped by sonication* lecithin 45Ca 22Na liposomes* lI4ClGlucose
0 0.1 0.2 0.3 0.4 0.5 1.0
0 0.4
0.8 1.2 1.6 2.0 4.0
20 42 64 95 96 92
5 17 38 61 89 92 86
* See Table 1 for details. The particle weight of a unilamellar, single-compartment liposome is about 2.1 X 106 (16). The electron microscopic appearance of the sonicated liposomes used in these studies (formed from asolectin) indicate them to be essentially multilamellar from which we have assumed an approximate particle weight of 6 X 106. The M, of subunit I is assumed to be approximately 35,000 (17).
32 34 A 48 6 8 15 B 43 52 50 C 81 92 96 Subunit I A 3-ml liposome suspension (50 mg of egg lecithin per ml) was diluted to 4 ml with the sucrose/KCI/Tris-HCl medium containing the labeled cation to give a final concentration of the test compound of 5 mM and a KCl concentration of 100 mM. The liposome suspension was then sedimented in a Spinco 50 rotor by centrifugation at 50,000 rpm for 1 hr. The liposome pellet was gently rinsed twice with the sucrose/KCl/Tris-HCl medium (containing 100 mM KCI) and finally resuspended to a total volume of 1 ml. Samples (0.5 ml) of this were taken for isotope counting. Results are expressed after subtraction of control (no ITC) values. * See Table I for details.
Proc. Natl. Acad. Sci. USA 76 (1979)
Biochemistry: Fry and Green Table 4. The ITC-induced influx of cations into liposomes followed by light-scattering change-&: Magnitude of light-absorbance change induced by addition of salt (A)tA ITC fraction RbCl NaCl KCl incorporated* +0.04 +0.04 +0.04 No ITC -0.05 -0.06 -0.05 A +0.03 +0.03 +0.03 B -0.07 -0.08 -0.07 C -0.04 -0.04 -0.03 D -0.16 -0.16 -0.16 Subunit I To a 3-ml (1-cm light path) silica cuvette was added 2.5 ml of the sucrose/KCI/Tris-HCl medium and 0.5 ml of liposome suspension [final lipid (asolectin) concentration = 8.3 mg/ml]. After a 1-min equilibration, 50 ul of sucrose/KCl/Tris-HCl medium was added containing sufficient amount of the test chloride salt to give a final concentration of 100 mM. Mixing was achieved by stirring and the absorbance of the suspension was read 1 min later at 520 nm. Measurements were made at room temperature in a Cary model 118 spectrophotometer. +, Absorbance increase (liposome shrinkage); -, absorbance decrease (liposome swelling). * See Fig. 1. t Liposomes (with and without ITC) were prepared as described in Table 1.
Liposomal influx of salts can produce osmotic changes within the liposome and such changes might induce "osmotic strain" in the liposome structure. This could result in the rapid iysis and resealing (with the concomitant trapping of salts) of the phospholipid structures. We have therefore checked the effect of such changes on liposome permeability to cations. Liposome suspensions were diluted by addition of water containing labeled test cation in sufficient amounts to cause a degree of liposome swelling approximately equal to that caused by salt additions. These results indicated that very little cation influx occurred during liposome swelling, with or without incorporated ITC. Few effective inhibitors of the ITC-mediated influx of cations into liposomes have been found to date. Fluorescein mercuric acetate and lanthanum are potent inhibitors, as was phosphatidylethanolamine. The effects of this phospholipid are particularly interesting and are the subject of further investigation referred to in the discussion. DISCUSSION The experiments reported in this communication support the idea that subunit I of cytochrome oxidase is an active transmembrane ion-channel protein. Incorporation of subunit I or ITC fractions into liposomes mediates the rapid influx of both monovalent and divalent cations under pressure of a salt gradient imposed across the liposomal membrane. Under conditions in which no salt gradient exists there occurs only a minimal exchange of cations mediated by the ITC. Control experiments have assured that such ion transport is indeed ITC mediated and not simply the result of "leaky" liposomes. The assay of channel-forming proteins is a new and challenging area of biochemistry. Such proteins must be incorporated into a lipid membrane phase in order to study their functions. [Kagawa (18) has given a recent review on this subject. I In the present studies we have used liposomes to study channel properties; such a technique gives a general awareness of the characteristics of channel activity but cannot measure the refined aspects of channel function. In unresolved cytochrome oxidase it is proposed that electron transport is coupled to movement of cations; cation movement is an endergonic process driven by the exergonic electron transport (2). Similarly, ion transport mediated by the resolved
2667
ITC of cytochrome oxidase is only demonstratable when coupleod Iaan !"exergoniccenter"-i.e., a salt concentration gradient. The nonspecificity for ions transported and the large molecular weight of subunit I argue for an ion-channel structure rather than an ion carrier (19). Subunit I possesses the general
properties required of a transmembrane channel protein. It is a highly hydrophobic protein with about 35% polar amino acids (20), buried deep in both the isolated enzyme and the mitochondrial membrane (21). In the membrane-bound enzyme, subunit I is almost completely inaccessible to all surface probes
(22, 23). Wikstrom et al. (24) have also proposed a channel structure for subunit I (and subunit II). These authors propose an ionic channel structure with specificity for protons being achieved by protein-bound "ordered" water. We believe the present results are more consistent with a relatively nonspecific ion channel. Is the ion transport capability of resolved ITC demonstrable in unresolved cytochrome oxidase? The answer to this question is affirmative. Enzymic activity can be shown to proceed in membrane-incorporated cytochrome oxidase with concomitant ion movements (3, 25). However, cytochrome oxidase incorporated into liposomes can also mediate the influx of ions under nonenergized conditions, the driving force here being a salt concentration gradient. Therefore, cytochrome oxidase can behave in a similar fashion to the resolved ITC. Respiratory control in membrane-incorporated cytochrome oxidase is an accepted but not understood phenomenon and is strongly dependent on the blend of phospholipids used for incorporation (26-28). Noted in the present article was the inhibition exerted on ion transport by phosphatidylethanolamine. We have been able to show that a variation in the blend of liposomal phospholipids can "stop or start" the ITC-mediated transport of ions; a similar correlation can be shown for nonenergized, unresolved cytochrome oxidase (provided the cytochrome oxidase is first depleted of its bound lipids). Thus, respiratory control can be equated with ion-channel capability in cytochrome oxidase and as such is not directly connected to electron transport. Hunter and Capaldi (29) have presented evidence that respiratory control in membranous vesicles of cytochrome oxidase is a localized molecular phenomenon and not a membrane phenomenon as implicit in the chemosmotic model of mitochondrial energy conservation. Note Added in Proof. Recent studies in this laboratory have established that subunits I and III of cytochrome oxidase are indistinguishable on the basis of their amino acid compositions. This raises the question as to which subunits of cytochrome oxidase actually constitute the ITC "'complex." Because subunit III is clearly an inactive species in ion translocation, the implication is that the dimeric polypeptide (i.e., subunit I) is the active form of the subunit ion channel.
This investigation was supported in part by Program Project Grant GM 12847 of the National Institute of General Medical Sciences. 1. Green, D. E. & Reible, S. (1974) Proc. Natl. Acad. Sci. USA 71, 4850-4854. Blondin, G. A. & Green, D. E. (1975) BioScience 28, 18-24. Kessler, R. J., Blondin, G. A., Vande Zande, H., Haworth, R. A. & Green, D. E. (1977) Proc. Natl. Acad. Sci. USA 74, 36623666. 4. Fry, M., Vande Zande, H. & Green, D. E. (1978) Proc. Natl. Acad. Sci. USA 75,5908-5911. 5. Fry, M. & Green, D. E. (1978) Proc. Natl. Acad. Sci. USA 75, 2. 3.
6.
5377-5380. Rubin, H. S. & Tzagoloff, A. (1973) J. Biol. Chem. 247, 594603.
2668
Biochemistry: Fry and Green
7. Mason, T. L. & Schatz, G. (1973) J. Biol. Chem. 248, 13551360. 8. Harmon, H. J., Hall, J. D. & Crane, F. L. (1974) Biochim. Blophys. Acta 344, 119-155. 9. Bangham, A. D., Standish, M. M. & Watkins, J. C. (1965) J. Mol. Biol. 13, 253-264. 10. de Gier, J., Mandersloot, J. G. & van Deenen, L. L. M. (1968) Biochim. Biophys. Acta 150,666-675. 11. Bittman, R. & Blau, L. (1972) Biochemistry 11, 4831-4839. 12. Jain, M. K., Toussaint, D. G. & Cordes, E. H. (1973) J. Membr. Biol. 14, 1-16. 13. Phan, S. H' & Mahler, H. R. (1976) J. Biol. Chem. 251, 270276. 14. Yu, C. A. & Yu, L. (1977) Biochim. Biophys. Acta 495, 248259. 15. Swank, R. J. & Munkres, K. D. (1971) Anal. Biochem. 39, 462-477. 16. Huang, C. (1969) Biochemistry 8,344-352. 17. Downer, N. W., Robinson, N. C. & Capaldi, R. A. (1976) Biochemistry 15, 2930-2936. 18. Kagawa, Y. (1978) Biochim. Biophys. Acta 505,45-93.
Proc. Natl. Acad. Sci. USA 76 (1979) 19. Urry, D. W., Long, M. M., Jacobs, M. & Harris, R. D. (1975) Ann. N. Y. Acad. Sci. 264, 203-220. 20. Poyton, R. 0. & Schatz, G. (1975) J. Biol. Chem. 250, 752761. 21. Eytan, G. D., Carroll, R. C., Schatz, G. & Racker, E. (1975) J. Biol. Chem. 250, 8598-8603. 22. Eytan, G. D. & Schatz, G. (1975) J. Biol. Chem. 250,767-774. 23. Eytan, G. D. & Broza, R. (1978) J. Biol. CIem. 253, 31963202. 24. Wikstrom, M., Saari, H., Penttila, T. & Saraste, M. (1978) FEBS Symp. 45, 85-94. 25. Hinkle, P. C., Kim, J. J. & Racker, E. (1972) J. Biol. Chem. 247, 1338-1339. 26. Eytan, G. D., Matheson, M. J. & Racker, E. (1976) J. Biol. Chem. 261,6831-6837. 28. Kagawa, Y., Johnson, L. W. & Racker, E. J. (1973) Biochem. Biophys. Res. Commun. 50,245-251. 27. Racker, E.'(1972) in Molecular Basis of Electron Transport, eds. Schultz, J. & Cameron, B. F. (Academic, New York), p. 45. 29. Hunter, D. R. & Capaldi, R. A. (1974) Biochem. Biophys. Res. Commun. 56, 623-628.
