4 4.. 5.7.. '1 u. .‘ "r77, '— Alféfi‘li‘o' . w . . <:#”‘~ .E“:~.—. :14: 3'43“? 41%}: ”L a; 1733‘ :53 I: fig? 51:3: “#:3- . £44 “ w _ . =2 ”b.4427 . - 4 , {:1 A? W 37:74:05.3 ._ 3:": .Egrfl 1‘. 3.! 2;. 3 1.. n 4,414. 4'1“?“ {5“ If”; » 14.4314 .4 £1.53 . war .— ‘ . 4 :i; 4'; I 4.4.444; I? 4 '11 I i ‘1 a . . 4 .44 i 4 " 2': , .4 “‘3‘ ' " 444.113.- 4 .. §. 4'.. \; I;‘:{: A 2.“..‘4 ..‘I“:“4 51‘2- ..:.,. .4. .1 .m. "hunt ~.....-.- 4‘4"” _. ..‘,_ -= fl" “it a‘.‘ W." .7 ,_.._,‘;.‘.,,,L_.. _ 3!” 4.34:2.“ r. .., A Z. 3.. an: . N L 237.. .. “:25 «.7 ‘ U4 ‘1. .‘I 4.44% ‘4; ‘4 wig-Q.r 2"}; n ,... ‘...;..-:.é . - ... .. .4; 4.. ‘1 i141, WW1 I 5531 4:944“ “: ¢xI .‘4 “:IJlis: 3L.” MINI 5‘} ‘IIL: '3‘“ 31.15}:le {11‘ pl "M: “if 32‘}; "4:4 ‘44: ‘41-: :3: “”ng MIN" ' ‘DI‘IIJ'H‘M WI Iéélfi I." 4'95 «5 ~-—.-.Mo . 7:35.? 4 g 4 ‘I‘ .fiifiifl ’44: "4‘: _ ‘J’ 44‘s??? 1—- ‘ 41% ml .4 .: _?“‘ ‘41 . . wig: ‘ A. :_.;<....:?=u? .7... . 2-4 gas“. *— ' 3.... 4%4IIIIIIEQA‘I REEF. 413%!“ A: if“ ‘414; AA :~. M 75:71.1; 4... 1 . “$.33?ng j“ __‘ .' Mu.. 33212-5... '.44I .44 4138:. “I TH E; M H.113? . 3%é‘Qg‘it‘ .ullflhb : 414.: 'M “In?“ W‘M '5‘; h f I . “~72" 0‘ l 1 l I Aw do... : 4“;— « 94413.5:- _ 7 “ ......, g 4% .- 4? 4 3134:: ‘rfi—é 3:" a '(Z.;_‘-, 533? . _“ . 4:332:4-“4- . . . 1 ' 1.‘ ”M K... L“ d..— Er- ”uri-El ~ 4 gram éfisaww -w'”. .. 2:: ‘ a: .12 If.“ V3 M E. i . ,, “A . 3:? _.—. .-... if” H" 1‘15 ‘ 4 2" «35—4- ew «7.1;... m". 43' H' I .- '-_:‘ 9.3”"; 7w 22-? ":2, ,7. gm: .._:3;:€‘Q «ax—1:.— w- “-— ..... n57 Fa: “LL": 4:: .- am ‘7'" gr ”flaw-$5. ”luv 3.” .13“ 7% {m5- ._ - ‘ ..-'" ‘ . in“ .5}. r... .— 33— fiié'ggk .-.._._. ‘97:. -- '73; —..= "’ «4': i.- -. q 1&4 id; {1‘1 13;! III: 04“: IIIIIIIJM Influx“??? 44‘4 4-: I , - aflgtflx ‘ .54": .L‘!‘:;;:I ‘ $93.3“ :. ”,3... ~94; a. 739* .u‘ .-..;:' m...“ 3.5:..- Ax. 1: 1‘ H“ II}! 5:" £4433 ,. "”1. I? I w £244, ‘1 44“" 4:: 4 4““ 444:4 4*4 .4. 4‘68 @4114,“ 4‘? £1 "0 RW‘ 3d, 5"!" $49: 3:341 ”‘i‘Il‘lt $14,; 334% t'gg‘r‘r‘ 3444"" “‘A 4“ 3:? MW: 4‘44: £443.“: "1;: 3:4; 5'“ ”Wm. @4444 4944‘: 3‘44: 44%44‘44: 4‘: W - 4 45:“ ‘4 4’4?! 1:33“ *3 ‘44 I?“ 444:4: “414;; ' 4:11;} J""l;3§;‘ 43 mfig I' '7‘:th at '35 AH‘W ($31 4%: big ,[ a???” fihzi .1 an WIN” it. Who 41%“ i" . i‘r’ wag! ,5; 3‘1““,1 ‘9 fig}: FR 135:. :%_“ V. egg§1f Jé’li‘gg‘mflih ‘; $1 :24 4‘45 ¢ " *4 4.444 fiéfi: £53“? m .. . 4 4: . .. -; =7 fly- A .4 ‘ ',1 :1: '.. , . " _ r42" 4" *2.“ ‘1‘ A 5- ’ I 4.4 LUV . .‘ . M A .- Jgf ""T'r'i? .. - . 5-". r5: .. '13-”:- t.“ . r-'_V, " -‘ ‘ ~ u f . y . . . . 1. 3% 11% fl? . ’ .—- “‘ng l . _ .wfitrufiamw F: .mavfifiwm': ' ‘ A . .;;.» :3: i . - ' '- £317.. ' 55%“ 25' .,_..._ .._. , ._ < .fi ‘ _,;_.; . . ,1. 7'2"?" :2- 1-— .Ja‘. _, - ~ . .— ".344. 2.44 tfifififi.‘ .. -. ,3... .. . 4 ' "53‘35; 45 «~44 . L . a _.. . ,. r ' A: ' A- " ,.... :14‘ "5’ .4- _-. ”W'" ~,~' , W453? W 5:? m i w! .573 w :71; “Eghfi... ~ -314. Q... ,— .._. M M... .f“ 5:: mm ' 1.7.» ’21? .35. ' ‘ gage-2a...ffi '." _ -"’_.' ‘ WAN—.Mflw ‘ .353 .1“ __ _W’" .. ‘32-}:13; W 44* .29”? .u 42-7 r ‘ c731. ' - , 7"...“ x 3;" ._ A ”g: 4 A . . A“ AA 1L ' _f ‘, ._ .«n... f 42% 4 “:4-4 4 ~v~bv "2" - " 3:34.. ;I Jj.’ " am’finp'aR. .3‘ ‘-~ «.m—v... u < FiESiS -o—_...."l‘. . I). V citi- , :9. I”: o . . w,“ .-_i . q“ . ‘ . A _ . ._ ;. . , a -3 " tat?” . i0; .0, ’4. 3. ‘ ‘ ‘ Q ,;7 1:3 ’ - . i k I -. ‘ 'k W :e’ v . ‘ ..‘ . ‘41:. - r hr 13-; This is to certify that the dissertation entitled Cytochrome g_0xidase Purified by Affinity Chromatography in LauryImaItoside: The Effects of Detergent, Lipid DepIetion, and Subunit III Removai on Function presented by Debra A. Thompson has been accepted towards fulfillment of the requirements for Ph.D. Biochemistry degree in Major professor [hue February 1, 1984 MSU is un Affirmative Action/Equal Opportunity Institution 042771 MSU LIBRARIES \. RETURNING MATERIALS: Place in book drop to remove this checkout from your record. FINES wiII be charged if book is returned after the date stamped below. MAR 2 9 2001 @3115 009! I CYTOCHROME £_OXIDASE PURIFIED BY AFFINITY CHROMATOGRAPHY IN LAURYLMALTOSIDE: THE EFFECTS OF DETERGENT, LIPID DEPLETION, AND SUBUNIT III REMOVAL ON FUNCTION By Debra A. Thompson A DISSERTATION Submitted to Michigan State University in partiaI fquiIIment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Biochemistry 1984 ABSTRACT CYTOCHROME £_OXIDASE PURIFIED BY AFFINITY CHROMATOGRAPHY IN LAURYLMALTOSIDE: THE EFFECTS OF DETERGENT, LIPID DEPLETION, AND SUBUNIT III REMOVAL ON FUNCTION By Debra A. Thompson Cytochrome c oxidase is a muItisubunit, intrinsic membrane protein that can be purified using traditionaI methods, but the enzyme composi- tion varies in number of subunits, amount of associated phosphoIipid, and state of aggregation. Therefore, the structuraT requirements for eTectron transfer and proton pumping have been difficuIt to estainsh. The Iimiting factor in the purification of cytochrome §_oxidase has often been the effectiveness of avaiTabIe detergents to squbiIize and disperse the mitochondriaI membrane proteins, and maintain enzyme activity. A new purification method for cytochrome g_oxidase was deveIoped based on the evaIuation of severaI new aIkngchoside detergents synthesized in this Taboratory (Rosevear, P., Van Aken, T., Baxter, J., and Ferguson-MiITer, S. (1980) Biochemistry lg, 4108-4115), which indicates IauryImaItoside is uniquer quaIified for use in purifying cytochrome g oxidase, because it efficientiy disperses the mitochondriaI proteins and sustains the highest IeveIs of oxidase activity reported (turnover number=1100 5'1). Bovine heart oxidase was purified in laurylmaltoside using affinity chromatography on horse cytochrome g:Sepharose 4B. Highly purified rat liver oxidase (~13 nmol heme a/mg protein) was prepared by a more extensive procedure involving hydroxyapatite and horse cytochrome g_affinity chromatography in laurylmaltoside. The rat liver enzyme has high activity (turnover number=270-450 5'1), a low lipid content (1 mol cardiolipin and 1 mol phosphatidyl inositol/mol aa3), and does not contain subunit III. The physical characteristics, respiratory control, proton pumping activity, and kinetic parameters of the rat liver enzyme are reported. Its kinetic behavior is compared with the enzyme in native membranes and reconstituted phospholipid vesicles. The results do not support previous suggestions that subunit III, phospholipid in excess of 1 mol cardiolipin/mol 333, or a dimer consisting of two copies of each subunit are required for normal biphasic binding and electron transfer to cytochrome c. The purified oxidase exhibits respiratory control ratios up to 10 when reconstituted into phospholipid vesicles, providing definitive evidence that proton pumping by subunit III is not solely responsible for the ability of cytochrome g_oxidase to develop and respond to a chemiosmotic proton gradient. v , 'Iriiiu .I u , H’AWMI‘J Ru . u . til"? anvil," .. . ..fl4|1!¥!1iilijifl To My Parents ACKNOWLEDGEMENTS I would like to sincerely thank my mentor, Dr. Shelagh Ferguson-Miller, for guidance and support. Her excellence and enthusiasm as a scientist and teacher are a constant source of encouragement. I would also like to thank Dr. Clarence Suelter for the invaluable introduction to biochemical research that I received in his laboratory. The research presented in this dissertation would not have been possible without the major contributions of Dr. Paul Rosevear and Terrell Van Aken to the development of the alkylglycoside detergents. I owe them, and Steven Castleman, my gratitude for synthesizing the alkylglycosides used in these studies. I would like to express my appreciation to Dr. Angelo Azzi and coworkers for their generous assistance and support in making it possible for me to perform proton pumping measurements on cytochrome £_oxidase in their laboratory. I also thank Dr. John Wilson, Mark Stadt, and Janice Messer for supplying countless preparations of rat liver inner mitochondrial membranes, and Theresa Fillwock for her patience and skill in typing this manuscript. It is a pleasure to acknowledge the support of the many friends that I shared classes, laboratories, and time with at Michigan State. Special among these are Jerome Hochman and Maria Suarez, without whom, had it been possible, it wouldn't have been nearly the fun! Finally, I would like to thank my husband, William Strong, for the love and encouragement he gave me at every step along the way. Parts of Chapters 2 and 3 of this dissertation are reprinted with permission from Biochemistry (1983) gg, 3178-3187; © 1983, the American Chemical Society. TABLE OF CONTENTS LIST OF FIGURES. . . . . . . . LIST OF TABLES O O O O O C O I LIST OF ABBREVIATIONS. . . . . u t O O I o o 0 INTRODUCTION AND LITERATURE REVIEW . . . . . . The Role of Cytochrome c Oxidase in Energy Conservation Subunit Composition . . . . . . Three- dimensional Structure and Active Form of Cytochrome £_0xidase . . . . . . . . . . . . o o o o O O 0 O o o o Subunit Arrangement in Cytochrome £_0xidase . Subunit Functions of Cytochrome £_Oxidase . . Subunit III. . . . . . . . . . . . . . . . Subunit II . . . . The Cytochrome c- Cytochrome c Oxidase Reaction. Dependence of Activity on the Physical Form of g o . o o . o . U o t o ' o 0 O O o O o 0 O Cytochrome £_0xidase . . . . . . . . . . . . . . . . . CHAPTER 1. EVALUATION OF THE SUITABILITY OF ALKYLGLYCOSIDE C DETERGENTS FOR USE Introduction. . . . . . . . Experimental Procedures . . Chemicals. . . . . . . . Spectral Measurements. . Assay Methods. . . . . Alkylglycoside Extraction of Membrane Cytochromes. Results and Discussion. . . Mitochondrial Membranes . . . . . . WITH CYTOCHROME Inner Mitochondrial o t a v o o n o o OXIDASE . Detergent Effects on the Kinetic Characteristics of Rat Liver Cytochrome c Oxidase in Solubilized . D Effect of Detergent and Lipid Environment on the Stability of Purified Cytochrome £_0xidase. . . . Alkylglycoside Fractionation and Solubilization of Inner Mitochondrial Membrane Cytochromes. . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . 36 45 49 58 CHAPTER 2. PURIFICATION OF BOVINE HEART AND RAT LIVER CYTOCHROME E_OXIDASE USING HORSE CYTOCHROME_§ AFFINITY CHROMATOGRAPHY IN LAURYLMALTOSIDE. . Introduction . . . . . . . . Experimental Procedures. Chemicals . . . . . . Spectral Measurements Assay Methods . . . . Protein Determination . . . . . . . . . Affinity Chromatography of Solubilized Bovine Heart Mitochondrial Particles. . . . . . . . . . . . . Rat Liver Cytochrome £_Oxidase Purification . . . . Results. . . . . . . . . . . . . . . . . . . . . . . . Affinity Chromatography of Solubilized Bovine Heart Mitochondrial Particles. . . . . . . . . . . . . Purification of Rat Liver Cytochrome c Oxidase. . . .6. 000 Q 0 o O O ' O 0 o O O O O O O O O O O O I O O O O O O O C Discuss-ion O O O O O O O O O O O O O O O O O O ........ Cytochrome_g Affinity Chromatography. . . . . . . . CHAPTER 3. STRUCTURAL AND FUNCTIONAL CHARACTERIZATION OF PURIFIED RAT LIVER CYTOCHROME £_OXIDASE . . . Intr‘OdUCt-ion O O O O O O O O O O O O 0 O O O O O O 0 Experimental Procedures .................... Chemicals . . . . . . . . ........ . . . . . Spectral Measurements . . . . . . . . . . ........ Assay Methods . . . . . . . . . . . . . . . . . . . Molecular Weight by Gel Filtration ....... . . Phospholipid Analysis . . . . . . . . . . . . . . SDS-Polyacrylamide Gel ElectrOphoresis. . . . . . DCCD Labeling of Cytochrome g_0xidase . . . . . . pH Dependence of Cytochrome_g Redox Potential . . Results. . . . . . . . . . . . . . . . . . . . . . . Physical and Spectral Properties of Rat Liver Cytochrome c Oxidase . . . . . . . . . . . . . . ‘ 0 Kinetic Properties of Rat Liver Cytochrome c Oxidase. Energy Coupling Properties of Rat Liver Cytochrome E- OXidase. O O O . O C C O O C O O C O O O O C . Discussion . . . . . . . . . . . . . . . . . . . . . . Association of Phospholipid with Cytochrome_g Oxidase Kinetic Changes Related to Purification . . Monomer Form of Cytochrome_g Oxidase. . . . SUbunit III 0 O O O O O I O O O O O 0 O I O O I 0 APPENDIX: Publications and Abstracts . . . . . . . . . . BIBLIOGRAPHY. O O O O I O O 0 O O O 0 O O O O O D O O O 0 vi 0 Q 0 g 0 Page 132 141 142 FIGURE 10 11 12 13 LIST OF FIGURES Structural model of cytochrome_g oxidase. . . . . . Alkylglycoside detergents . . . . . . . . . . . . . Molecular size distribution of bovine heart cytochrome £_oxidase equilibrated in octylglucoside, laurylmaltoside, deoxycholate, or Tween 20. . . . . Eadie-Hofstee plots of the kinetics of oxidation of cytochrome_g by rat liver cytochrome_g oxidase in 50 mM Kpi’ pH 6.5. O 0 O O O O O O O O O O O O O Eadie-Hofstee plots of the kinetics of oxidation of cytochrome g_by rat liver cytochrome £_oxidase in 25 mM Tris-cacodylate, pH 7.9 . . . . . . . . . . . Eadie-Hofstee plots of the kinetics of oxidation of cytochrome g by rat liver cytochrome g oxidase in inner mitochondrial membranes solubilized with various detergents. . . . . . . . . . . . . . . . . Time course of oxygen uptake by purified rat liver cytochrome_g oxidase assayed in the presence of various amphipaths. . . . . . . . . . . Hexyllactoside extraction of cytochromes from rat liver inner mitochondrial membranes . . . . . . . . Laurylcellobioside extraction of cytochromes from rat liver inner mitochondrial membranes . . . . . . Laurylmaltoside extraction of cytochromes from rat liver inner mitochondrial membranes in 0.5M KCl . . Laurylmaltoside extraction of cytochromes from rat liver inner mitochondrial membranes in 0.25M KCl. . Time and temperature dependence of laurylmaltoside extraction of cytochromes from rat liver inner mitochondrial membranes . . . . . . . . . . . . . . Purification scheme for rat liver cytochrome c oxidase in laurylmaltoside. . . . . . . . . . . . . vii 3O 37 4O 43 46 51 53 56 59 61 7O FIGURE 14 15 16 17 18 19 20 21 22 23 24 25 26 Chromatography of solubilized bovine heart mitochondria on horse ferrocytochrome g—Sepharose 4B. O O O O O O O O O O O I O O O O O O O O O O O O O Visible absorbance spectra of fractions produced by laurylmaltoside extraction of rat liver inner mitochondrial membranes . . . . . . . . . . . . . . . Hydroxyapatite chromatography of a cytochrome c oxidase enriched fraction from rat liver inner mitochondrial membranes . . . . . . . . . . . . . . . Horse cytochrome_g affinity chromatography of rat liver cytochrome c oxidase. . . . . . . . . . . . . . Visible absorption spectra of rat liver cytochrome g_oxidase purified by affinity chromatography . . . . Gel filtration analysis of the apparent molecular weight of purified rat liver cytochrome g_oxidase- laurylmaltoside complex . . . . . . . . . . . . . . . Two dimensional thin layer chromatography of phospholipids extracted from purified rat liver cytochrome g_oxidase. . . . . . . . . . . . . . . . . Subunit composition of cytochrome c oxidase as revealed by SDS- polyacrylamide gel —electrophoresis using the discontinuous buffer system of Laemmli (1970). O O O O O O O O O O O O O O O O O O ..... DCCD binding to cytochrome c oxidase. . . . . . . . . Subunit composition of cytochromec coxidase analyzed by the highly resolving SDS- polyacrylamide gel electrophoresis system of Merle and Kadenbach (1980). . . . . . . . . . . . . . . . . . . . . . . . Eadie- Hofstee plots of the kinetics of oxidation of cytochrome c by rat liver cytochrome c oxidase in 25 mM Tris- cacodylate, pH 7. 9. . . . . . . . . . . Eadie- Hofstee plots of the kinetics of oxidation of cytochrome c by rat liver cytochrome c oxidase in 50 mM Kpi’ pH 6. 5. O O O O O O O O O O O O O O O O A comparison of the Eadie- Hofstee kinetics of the rat liver cytochrome c oxidase reaction, measured polarographically as rates of oxygen consumption or spectrally as rates of ferrocytochrome_g oxidation . . . . . . . . . . . . . . . . ...... viii 73 78 8O 82 96 98 101 103 105 108 111 113 FIGURE 27 28 29 TABLE Page pH dependence of the redox potential of cytochrome a in cytochrome c oxidase with and without subunit TII I O O O O O O O O O O O O I O O O O O O O D O O O 118 Respiratory control of affinity-purified rat liver cytochrome £_oxidase reconstituted into asolectin vesicles . . . . . . . . . . . . . . . . . . . . . . . 121 Bulk phase pH changes resulting from the oxidation of ferrocytochrome_g by reconstituted cytochrome £_oxidase in phospholipid vesicles . . . . . . . . . . 124 LIST OF TABLES Purification of cytochrome_g oxidase from rat liver mitochondria . . . . . . . . . . . . . . . . . . 77 ix Bicine CCCP DCCD DMSO DPG DTT E'0=Em EDTA Hepes LM MLT Pi PI PMSF PPO RCR SDS SDS-PAGE TMPD TN Tris-HCl LIST OF ABBREVIATIONS N,N—bis(2-hydroxymethyl)glycine Carbonyl cyanide m-chlorophenylhydrazone N,N'—dicyclohexylcarbodiimide Dimethylsulfoxide Diphosphatidylglycerol (cardiolipin) Dithiothreitol Standard redox potential or midpoint potential Ethylenediaminetetraacetic acid 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid Laurylmaltoside Maltose Molecular Weight Inorganic Phosphate Phosphatidyl inositol Phenylmethanesulfonyl fluoride 2,5-diphenyloxazole Respiratory control ratio Sodium dodecyl sulfate Sodium dodecyl sulfate-polyacrylamide gel electrophoresis N,N,N',N'-tetramethyl-p-phenylenediamine Turnover number Tris(hydroxymethyl)aminomethane hydrochloride TS TSH 50 mM Tris-HCl, pH 8, 0.33 M sucrose 50 mM Tris-HCl, pH 8, 0.33 M sucrose, 1 mM histidine xi INTRODUCTION AND LITERATURE REVIEW The Role of Cytochrome c Oxidase in Energy Conservation. Oxidative phosphorylation is the principle means by which eukaryotic cells conserve energy made available from metabolism. It is carried out by a system of protein complexes located in the inner membrane of mitochon- dria, which transfer the electrons derived from substrate molecules to oxygen (Keilin, 1925) in a process that results in a stepwise decrease of their redox potential (Chance and Williams, 1956) and the formation of ATP. The chemiosmotic theory of energy coupling (Mitchell, 1961a,b;1966) proposes that the drop in potential energy experienced by the electrons is used to establish a bulk electrochemical proton gradient across the inner mitochondrial membrane and that the high energy state created by separation of charge and molecular species can be recaptured in chemical form as the high energy phosphate bonds of ATP. According to the original theory, a series of vectorial reactions involving alternating electron and proton carriers is the only mechanism used to produce the electrochemical gradient. However, evidence from a number of laboratories indicates that the respiratory chain proteins may also act as proton pumps, and this apparent violation of a basic tenet of the chemiosmotic mechanism has been the source of intense controversy in the field of bioenergetics (review: Lehninger §t_al,, 1979). In addition, the chemiosmotic mechanism of energy coupling has itself been brought into question by the work of researchers who have evidence for the operation of oxidative phosphorylation under conditions where bulk electrochemical gradients do not exist (Storey and Lee, 1981). An alternative mechanism for energy coupling via localized proton movements within the membrane or protein complexes, or at the water-membrane interface has been proposed (Williams, 1961). Resolution of these fundamentally different views of the respiratory process, and elucidation of the basic mechanism of energy conservation will require a full understanding of all aspects of the electron transfer reactions, from the level of protein-protein and protein-membrane interactions, to the molecular details of the events occurring as electrons move from one redox center to the next. Toward this end, the structure and function of each of the inner mitochondrial membrane protein complexes involved in oxidative phosphorylation have been studied using a variety of interdisciplinary approaches. Cytochrome g_oxidase (ferrocytochrome c: oxygen oxidoreductase, EC 1.3.9.1, or cytochrome gg3), has been studied more intensively than any of the other complexes, and the specific details of its function lie at the heart of the chemiosmotic debate. It performs the final electron transfer reaction in the respiratory process by accepting electrons from the small water—soluble protein, cytochrome g, in the intramembrane space, and transferring them across the membrane to oxygen in a reaction that results in the formation of water. The protons consumed in this process are assumed to come from the mitochondrial matrix, and cytochrome_g oxidase has been classically viewed as contributing to the transmembrane potential via this asymmetric utilization of protons. However, based on measurements of ion movements in the media surrounding whole mitochondria respiring on electrons fed in at the level of cytochrome c, and on the results of similar experiments with purified cytochrome c oxidase reconstituted into liposomes, it has been proposed that the oxidation of ferrocytochrome g_additionally results in the active extrusion of one proton across the inner mitochondrial membrane per electron transferred to oxygen (Wikstrom, I977; Wikstrom and Saari, I977; Sigel and Carafoli, 1978; Azzone §t_§l,, 1979; Casey gt_§l,, 1979). This apparent ability of cytochrome c oxidase to function as a redox-linked proton pump is in direct contradiction to the ligand conducting mechanism of gradient formation proposed by the chemiosmotic theory, and is further disputed on experimental and theoretical grounds (Moyle and Mitchell, 1978; Lorusso §t_al,, 1979; Papa §t_gl,, 1980; Mitchell and Moyle, 1983), including the basic assertion that pH changes cannot be used as evidence for proton pumping due to the dependence of hydrogen ion activity on the concentration of other ions in solution (Stewart, 1982). The natural correlate to the controversy surrounding the function of cytochrome g_oxidase is that the specific structural requirements necessary for all aspects of its native activity are also poorly defined. This problem is complicated since cytochrome g_oxidase is a multisubunit, integral membrane protein containing four metal centers which undergo reversible valence changes during the catalytic cycle. There are two heme A moieties, designated heme a_and heme 33, which differ from one another only in their protein environments, and two copper atoms, CuA and CuB. The hemes are responsible for the characteristic visible spectrum of cytochrome g oxidase (Keilin and Hartree, 1938), and although the complete deconvolution of the spectrum in terms of the contributions of cytochrome §_and cytochrome a3 has been a very complex problem, recent application of the method of principal component analysis (Halaka, 1981) has confirmed the spectral assignments of Vanneste (1966), as also supported by Carter and Palmer (1982). The reduced forms of cytochrome a and cytochrome g3 contribute equally to a single absorbance peak in the Soret region (443 nm) (Y band). The oxidized forms absorb to approximately equal extents at two separate wavelengths in the Soret region (cytochrome a, 426 nm; cytochrome a , 414 nm). The longer wavelength absorbance of the a band (605 nm) is mainly due to the absorbance of cytochrome a, The _a U) (protein associated) absorbance spectrum of one of the copper atoms ' in the near infra-red region (Gibson and Greenwood, 1965). Subunit Composition of Cytochrome c Oxidase. The peptides of cytochrome g_oxidase are coded for by the nuclear and mitochondrial genomes (Schatz and Mason, 1974; Tzagoloff gt al., 1979; Tzagoloff, 1982). The purified enzyme from yeast (Saccharomyces cerevisae) is composed of 7 different polypeptides (Mgsgfl et al., 1973; Poyton and Schatz, I975a; Gutweniger gt_al,, 1981). The three largest subunits (I, II, III) are mitochondrial translation products, and the four smaller subunits (IV, V VI, VII) are translated in the cytoplasm and imported into the mitochondria (Poyton and Schatz I979a,b; Weiss gt al., 1975; Kroon and Saccone, 1974). The enzyme from the mold Neurospora crassa has seven subunits as well (Werner, 1977) present in a one-to-one stoichiometry (Weiss and Sebald, 1978). In the case of mammalian cytochrome g_oxidase, the number of essential subunits is less certain and depending on the source, methods of purification, and system used for analysis, a 6-13 subunit structure has been reported for the purified enzyme (Downer et al., 1976; Merle and Kadenbach, 1980; Buse gt_ l., 1980,1982; Saraste §t_§l,, I981; Kadenbach gt_gl., 1983). This apparent deviation of the mammalian enzyme peptide composition from the seven subunit pattern observed for S. cerevisae and N, Cra§§a_is due to the variability in the amount of subunit III; the presence of three additional bands in the molecular weight region near subunits V and VI (designated a, b, c by Downer et al., 1976, and Vb, VIa, VIb by Merle and Kadenbach, 1980); and the resolution of a group of very small peptides near subunit VII (designated VIIa, VIIb, VIIc, VIII by Merle and Kadenbach, I980). The possibility that a higher degree of structural complexity may be necessary for the function of mammalian cytochrome g_oxidase has led to attempts to achieve a functional definition of the minimum enzyme subunit composition. Several procedures have been designed to reduce the number of peptides in bovine heart cytochrome c oxidase without damaging the electron transfer activity of the enzyme. Limited trypsin digestion removes peptides b and c (Ludwig gt_gl,, 1979), and gel filtration in Triton X-100 removes most of a, b, and c (Downer gt al., 1976). Part or all of these small peptides can also be removed by limited chymotryptic digestion (Carroll and Racker, 1977), preparative electrophoresis in Triton X-100 (Saraste gt_al,, 1980), or ion exchange chromatography at high pH and high concentrations of Triton X-100 (Saraste gt_al,, 1981; Georgevich §t_§l,, 1983) resulting in the removal of subunit III as well. The treated enzyme is active and retains the capacity for high levels of respiratory control in the case of the proteolytic procedure (Carroll and Racker, 1977), but exhibits significantly lower respiratory control ratios after treatment with high concentrations of Triton X-100 (Penttila and Wikstrom, 1981). The general conclusion drawn from these results, plus the variability in the amount of a, b, and c present, and the absence of a, b, and c in enzyme immunoprecipitated from solubilized membranes with antibody to highly purified enzyme (Ludwig gt_al., 1979), is that these peptides are not authentic subunits, but that subunit III is (for review see: Azzi, 1980; Wikstrom §£.fll-, 1981; Capaldi gt_§l,, 1983). However, this view is not supported by the results of Merle gt_gl, (1981) who find all 12 polypeptides present in oxidase immunoprecipitated with antibodies to a single peptide (Subunit IV). Total consensus on this point has not been reached and a regulatory role for these small peptides has been proposed (Merle and Kadenbach, 1982) as well as other functions (Wikstrom gt_al., 1981). As for the apparent controversy surrounding the presence or absence of ”extra” peptides in the mammalian enzyme in the region near subunit VII, this may be more a reflection of the ambiguities in the nomenclature used to refer to these peptides than representative of true differences in experimental results. A review of the literature reveals that many researchers report a heterogeneous polypeptide composition of subunit VII,but that by analogy to the fungal enzymes, a seven subunit nomenclature has persisted in which these peptides are collectively referred to as subunit VII (for review see: Azzi, 1980; Wikstrom gt_al., 1981). Thus, it would appear there is agreement regarding the very low molecular weight subunit composition of the mammalian enzyme, and that these subunits differ from those of S. cerevisae and N; crassa. In addition, the amino acid sequence of these peptides may vary between mammalian species and even in different tissues (Merle and Kadenbach, 1980; Kadenbach gt_al., 1982). Studies on the stoichiometry of the subunits of mammalian cytochrome g_oxidase from different species indicate that the subunits are probably present in one-to-one ratios (Steffens and Buse, 1976; Saraste gt_§l,, 1980). This information, along with molecular weight data for the individual subunits obtained from either amino acid sequence analysis, DNA sequence predictions, or SOS-polyacrylamide electrophoresis can be used to predict the minimum molecular weight for a monomer of cytochrome oxidase containing two hemes and two coppers. For the enzyme from bovine heart this molecular weight is 198,000 (Buse gt_al,, 1982) representing an enzyme with a heme g_to protein ratio of 10 nmol heme g/mg protein. Three-dimensional Structure and Active Form of Cytochrome c Oxidase. The size and shape of the cytochrome g_oxidase protein complex has been investigated using electron microscopy and image reconstruction techniques on two-dimensional crystalline layers of the enzyme. Two methods of protein preparation involving detergent extraction of mitochondrial membranes are used for this technique, and result in either a monomer form in deoxycholate rich, non—membranous sheets (Henderson gt_al,, 1977; Frey gt_§l., 1978), or a dimer form (i.e. two copies of each subunit plus 4 heme A and 4 Cu) oriented in a lipid bilayer when Triton X-100 is used (Fuller gt_al,, 1979; Deatherage gt al, 1982). The results of this type of structural analysis indicate that cytochrome g oxidase is an asymmetric protein about 110 A long, which extends 55 A outside the membrane on the cytoplasmic side and less than 20 A on the matrix side (Figure 1). It traverses the membrane in two distinct domains (or legs) of unequal size separated by about 10 A, that are joined together just outside the bilayer on the cytoplasmic side. The question of whether the monomer or dimer form of cytochrome g oxidase observed in these studies is the physiologically significant species has been difficult to answer, due in part to the uncertainty regarding the actual number and size of the subunits which constitute a monomer. In addition, the many complications arising from interconversions of monomeric, dimeric, and oligomeric forms under various experimental conditions, and difficulties associated with obtaining accurate molecular weight estimates for large asymmetric proteins in detergent solutions (for example: Nozaki et_al,, 1976; le Maire gt_al,, 1980) have been major obstacles in resolving this basic issue of the molecular weight and active form of cytochrome g oxidase. Based on molecular weight estimates by a variety of techniques, the most commonly observed form of purified bovine heart and N, grassa enzymes (apparent Mr of detergent-protein complex m300,000) had been proposed to be an active dimer species composed of two copies of each subunit (Robinson and Capaldi, 1977; Saraste gt__l,, 1981; Ferguson-Miller §t_al,, 1982; Weiss and Kolb, 1979). However, recent clarification of the predicted molecular weight of a monomer of cytochrome g oxidase based on subunit sizes determined from sequence information (Buse gt al., 1980,1982), and the development of procedures favoring more complete dispersion of the protein employing non-ionic FIGURE 1: Structural model of cytochrome g oxidase. The size, shape, and relative positions of monomers in the dimer have been determined for cytochrome g oxidase in two-dimensional crystalline sheets, using electron microscopy with image reconstruction techniques. The arrangement of the subunits in the monomer and the location of cytochrome g_binding are based on the results of chemical labeling and crosslinking studies. 11 detergents, led to the identification of electron transfer competent species with apparent monomer molecular weights (Georgevich gt_§l,, 1983: Suarez gt_gl., 1983,1984; Nalecz gt al., 1983). The possibility that the functional form of the enzyme is a monomer is consistent with the results of molecular weight estimates of the functional unit of cytochrome g_oxidase using target size analysis by radiation inactivation, which indicate that even less than a monomer is required for electron transfer activity (Kagawa, 1967; Thompson gt_al., 1980; Suarez 3: al., 1984). It has also been shown that cytochrome g oxidase from elasmobranch heart (Wilson et_al., 1980; Georgevich §t_al,, 1983) and camel heart (Darley-Usmar gt_al., 1981) exist as a active, seven major subunit species with apparent molecular weights that are much smaller than that of the dimer form of bovine heart cytochrome c oxidase. In addition, active low molecular weight species have been observed in polydisperse preparations of oxidase enzyme from several sources (Thompson et_al,, I980; Darley-Usmar 3: al., 1981). Therefore, the prevailing current view is that the monomer of cytochrome g oxidase is active; and that the dimer form may play a role in regulation, or proton pumping activity (Georgevich §t_gl,, 1983; Penttila, 1983, Nalecz gt_al., 1983), or may simply be an artifact of purification that reflects the tendency of oxidase to associate with other respiratory proteins (Hochman gt_al., 1982, 1983). Subunit Arrangement in Cytochrome c Oxidase. The three dimensional organization of the subunits within the cytochrome g_oxidase membrane has been analyzed using a variety of techniques, including subunit specific antibody binding (Frey gt_al., 1978; Chan and Tracy, 1978; 12 Freedman and Chan, 1983) and covalent chemical labeling. 1) Radio- labeled, hydrophilic (membrane impermeable) protein modifying reagents (Eytan and Schatz, 1975; Eytan gt_al., 1975; Ludwig et_§l,, 1979; Prochaska §t_al,, 1980) were used to determine subunit accessibility to the matrix or cytoplasmic side of the membrane; 2) Hydrophobic (lipid soluble) protein modifying reagents (Bisson §t_gl,, 1979; Cerletti and Schatz, 1979; Prochaska gt al., 1980; Georgevich and Capaldi, 1982) were used to identify subunits in contact with phospholipids; and 3) Crosslinking reagents (Briggs and Capaldi, 1977,1978) were used to determine nearest neighbor relationships of the subunits. The data are not in total accord, but certain spatial assignments can be made with a fair degree of certainty, and the amino acid sequence information indicating regions of hydrophobic and hydrophilic character of the subunits, can be used for further clarification. The current general view of the subunit arrangement is the following (for recent review see: Capaldi gt_al., 1983): The largest peptide, subunit 1, appears to be buried deeply within the bilayer, because it is heavily labeled with hydrophobic reagents and very little with hydrophilic chemicals and antibodies. Subunit II has both hydrophobic and hydrophilic character. It is thought that most of the mass of subunit II is in the hydrophilic domain on the cytoplasmic side of the membrane, and that subunit I plus a portion of subunit II both traverse the membrane in the larger of the two protein domains embedded in the bilayer. The sequence and reactivity of subunit III indicate that it is the most hydrophobic of the subunits. A transmembrane orientation is predicted because of its reactivity on both sides of the membranes, and based on size constraints, it is thought to be the major constituent of the 13 smaller of the two transmembrane domains. Subunit IV is exposed to the mitochondrial matrix, and to some degree to the membrane phase. The position of subunit V in the complex is not clearly established. Studies indicate subunit V is reactive on both sides of the membrane, yet its sequence has very little hydrophobic character suggestive of a transmembrane peptide. Crosslinking studies indicate it is located near subunits I, II, III, and VII. The position of subunit VI has also not been well established. It is fairly unreactive in most labeling procedures, and therefore may be located in the interior of the protein. The peptides collectively referred to as subunit VII, are small transmembrane peptides, exposed to the lipid phase, the matrix, and the intramembrane space, as indicated by their reactivity with hydrophilic and hydrophobic reagents. The results of crosslinking studies suggest that subunits IV, V, and VII are close together in the protein. Subunit Function of Cytochrome c Oxidase. The specific functional roles of the subunits of cytochrome g_oxidase are just beginning to emerge in terms of substrate and prosthetic group binding, electron transfer activity, proton translocating activity, monomer-monomer interactions in the dimer form, and interactions with other respiratory proteins. So far, the functions of subunits II and III are the best defined and will be discussed here. Subunit III will be considered first. Subunit III. Results from two lines of research suggest that subunit III is involved in the proposed proton translocating 14 function of cytochrome g oxidase. Wikstrom and coworkers prepared enzyme depleted of subunit III and several small peptides which although relatively unstable, retains good electron transfer activity (Penttila and Wikstrom, 1981). This oxidase has a reduced capacity for respiratory control when reconstituted into phospholipid vesicles, and no longer exhibits apparent proton pumping activity or the small pH dependence of the redox potential (Em) of cytochrome a observed by these researchers and others (Artzatbanov §t_gl,, 1978). These findings are of interest in light of the proposal that a proton is made available for translocation directly as a result of the reduction of heme a_(Penttila and Wikstrom, 1981; Babcock and Callahan, 1983; Callahan and Babcock, 1983). The results of chemical modification studies employing the cross- linking reagent dicyclohexylcarbodiimide (DCCD) have also been interpreted in terms of a unique functional role for subunit III in proton translocation. DCCD inhibits ATP-linked proton translocation by binding to proton translocating ATPases in mitochondria (Beechey gt al., 1967), chloroplasts (McCarty and Racker, 1967), bacteria (Evans, 1970), Streptococcus (Harold gt_al., 1970) and chromaffin granules (Bashford §t_§l,, 1976). Covalent modification by DCCD also inhibits the apparent proton pumping activity of purified, reconstituted bovine heart cytochrome c_oxidase, while inhibiting electron transfer activity to a lesser extent (Casey gt_§l,, 1979, 1980). This loss of function is attributed to the preferential, though not exclusive, reaction of DCCD with subunit III (Casey gt_al,, 1979) glutamic acid residue 90 (Prochaska gt 31., 1980), located within an amino acid sequence very similar to the DCCD binding site in mitochondrial ATPase (Prochaska gt 15 al., 1980). These results prompted the speculation that subunit III may be either a proton pump or channel used by cytochrome g_oxidase and a strong similarity between the structure of this subunit and that of the proton pump, bacteriorhodopsin, is noted (Azzi, 1980). In contrast, this view of an absolute requirement for subunit III in the proton pumping function of cytochrome g_oxidase is challenged by studies on bacterial cytochrome g_oxidases which contain two heme A and two coppers in a much simpler protein structure of only two subunits. The enzymes from Paracoccus denitrificans (Solioz §t_al,, 1982) and Thermus thermophilus (Sone and Yanagita, 1982; Yoshida gt 91-: 1983) possess apparent proton pumping activity linked to ferrocytochrome_g oxidation when the purified proteins are reconstituted into phospho- lipid vesicles. However, proton pumping has not been demonstrated for the enzymes from Rhodopseudomonas sphaeroides (Gennis gt al., 1982) and Nitrobacter agilis (Sone §t_al., 1983), indicating this function may not be a general property of all bacterial enzymes. It is not yet certain to what extent structural homology exists between the two largest subunits of the eukaryotic and prokaryotic proteins, and the possibility that some bacterial subunits may retain a proton pumping mechanism that has been evolutionarily lost in eukaryotes cannot be entirely excluded. In the case of the enzyme from P. denitrificans, a degree of structural similarity to eukaryotic cytochrome g_oxidase subunit II has been demonstrated by antigenic cross reactivity (Ludwig, 1980). For the enzyme from T. thermophilus, the homology may prove to be much less, since this enzyme has the unusual property of binding heme C in addition to two heme A and two Cu (Yoshida gt_al,, 1983). 16 Subunit III has also been proposed to play a key role in the association of monomers of cytochrome g_oxidase to form dimers. Georgevich gt_3l, (1983) use a procedure employing DEAE-agarose chromatography at high pH and detergent concentrations to decrease the apparent molecular weight of cytochrome_g oxidase to that of a monomer; these conditions also reduce the content of subunit III (Saraste gt 31,, 1981). The removal of subunit III is proposed to be necessary for maintaining the monomeric fonn of bovine heart oxidase and preventing its reassociation in dimers (Georgevich §£.£l-, 1983; Capaldi gt_3l., 1983). However, other researchers (Nalecz et al., 1983; Suarez et al., 1983,1984) observe monomer forms of subunit III containing bovine heart cytochrome g_oxidase in laurylmaltoside, and therefore do not support the view that the presence of subunit III always results in a dimer form of this enzyme. Subunit II. A unique role for subunit II in the electron transfer activity of cytochrome g_oxidase is indicated by increasing evidence that it is directly involved in prosthetic group and substrate binding. Heme 3_is the primary acceptor for the first electron transferred from ferrocytochrome g_to resting cytochrome g_oxidase (Andreasson gt_3l., 1972; Andreasson, 1975). This electron rapidly equilibrates with nearby CuA (Gibson gt_3l., 1965; van Buuren gt__l,, 1974; Wilson gt ‘31., 1975). The location of these prosthetic groups has been investigated by isolating the subunits under conditions designed to maintain their non-covalent associations with the protein. However, the results of this approach have been difficult to interpret due to the difficulty in establishing which of the observed associations are l7 functionally significant, and which result from detachment and random reassociation during the isolation procedures, leading to many conflicting assignments of prosthetic group binding. Recently, Winter gt_3l, (1980) developed an improved electr0phoresis system which minimizes independent migration of the hemes, and permits the separation of the subunits while retaining the heme in a form that has the characteristic visible absorption spectrum of cytochrome 3. Their results indicate that heme A is equally distributed between subunits I and II, and that 90% of the copper is bound to subunit II as well. Specifically, they suggest that heme 3, CuA, and CuB are all liganded to subunit II, and that heme 33 is associated with subunit I. Several indirect lines of evidence provide support for these assignments. 1) EPR analysis of the interactions of heme 3_and CuA with a water soluble paramagnetic compound indicate that both are on the cytoplasmic side of the membrane (Ohnishi gt_3l,, 1979) in the protein domain largely occupied by subunit II. 2) Subunit II shares limited sequence homology to the c0pper binding proteins azurin and plastocyanin in the region of the invariant histidine, cysteine, and methionine residues involved in liganding copper (Steffens and Buse, 1979; Yasunobu gt_3l,, 1979). Evidence from EPR and ENDOR spectro- scopic studies of cytochrome g_oxidase indicates that CuA is liganded by at least one histidine and at least one cysteine residue (Stevens gt 31,, 1982), supporting the proposed analogy with the blue copper proteins. 3) Subunit III and several smaller peptides can be removed from bovine heart cytochrome_g oxidase without altering its electron transfer properties (Penttila, 1983), indicating these subunits are probably not involved in heme and copper binding. 4) The two 18 subunit bacterial enzyme from B, denitrificans contains two heme A and two coppers which appear spectrally identical to those of the mammalian enzyme (Ludwig, 1980). The proposed analogy between the two subunits of the bacterial enzyme and two largest mitochondrially derived peptides of cytochrome g oxidase suggests that only subunits I and II are required for prosthetic group binding. Subunit II has also been implicated as the binding site for cytochrome g, The identities of the subunits directly interacting with cytochrome g_during electron transfer have been investigated by cross— linking cytochrome g_to cytochrome g_oxidase using: yeast cytochrome g modified on its free sulfhydryl with 5,5'-dithiobis (2-nitrobenzoate) (Birchmeier gt_3l,, 1976; Fuller gt_3l., 1981); horse cytochrome g modified on free lysine residues with arylazido photoaffinity labels (Erecinska, 1977; Bisson gt_3l,, 1978a,b; Erecinska gt_3l., 1980; Bisson gt_3l,, 1980); and dithio-bissuccinimidyl propionate (Briggs and Capaldi, 1978) and 1-ethyl-3-[3-(dimethylamino)propyl]carbodiimide (Millett gt_3l., 1982) to covalently crosslink preassociated cytochrome g_- cytochrome g_oxidase complexes. The model of cytochrome g - cytochrome g oxidase interaction that has come from these studies is one in which subunit II comprises the primary binding domain for cytochrome g, and subunit III (Fuller ££.§l-: 1981) and possibly one of the smaller subunits (Erecinska gt_3l,, 1980) also play a role in cytochrome g binding. Logical support for the cytochrome g_binding function of subunit II comes from the structural studies which place most of the mass of this subunit in the intermembrane space where cytochrome g_is sequestered, and from evidence which suggests that heme and copper redox centers are associated with this subunit. 19 The Cytochrome c - Cytochrome c Oxidase Reaction. The binding of cytochrome g_to cytochrome g_oxidase is electrostatic, involving interactions between positively charged lysine residues on the surface of cytochrome g_surrounding the heme crevice (Ferguson-Miller gt 31., 1976; Smith gt_3l,, 1977; Brautigan gt_3l,, 1978; Reider and Bossard, 1978, 1979; Errede and Kamen, 1978; Ferguson-Miller gt_3fl,, 1978; Smith ._t.3l., 1980; Osheroff gt_3l,, 1980) and the negatively charged carboxyl groups of aspartate and glutamate on the surface of cytochrome oxidase (Millett g__3l., 1982, 1983). Electron transfer from the iron of ferrocytochrome g_occurs through its exposed heme edge (Creutz and Sutin, 1974; Ewall and Bennett, 1974; Hodges gt_3l., 1974; Ferguson-Miller gt_3l,, 1976) to the heme 3_of the cytochrome g oxidase located a distance of 25-30 A away when the complex is formed (Vanderkooi gt 31., 1977; Dockter gt_3l,, 1978; Ohnishi gt_3l,, 1979). The formation of the active complex is very efficient due to the proper aligning of the heme edge as a result of the orienting effect of the large dipole moment of cytochrome g_in the electric field of the oxidase, thus making possible rates of electron transfer that approach the diffusion controlled situation (Konig gt_3l,, 1980). The steady state kinetics of this electron transfer reaction have been studied using two types of assay systems: a spectral assay following the decrease in absorption due to conversion of cytochrome g to its oxidized form; a polarographic assay monitoring the decrease in oxygen concentration in the presence of cytochrome g_plus an excess of reducing equivalents. Smith and Conrad (1956) discovered that the reaction of cytochrome g_with cytochrome g_oxidase, which exhibits Michaelis-Menten kinetics when monitored polarographically over a 20 limited concentration range of cytochrome_g (with ascorbate present as the reductant), displays an apparent first order time course in the spectral assay with an observed rate constant that varies with the total amount of reduced and oxidized cytochrome g_present. The widely accepted explanation of these results was proposed as "Mechanism IV“ by Minnaert in 1961. In this scheme, substrate (ferrocytochrome g) and product (ferricytochrome g) bind with equal affinity to cytochrome g oxidase to form active and inhibitory complexes, respectively. Km 02 K1 E2+ + 33+ _; (£2+_E3+)_>(C3+_§-2+) _,(£3+_£3+);>£3+ + 33+ In support of this mechanism, Yonetani and Ray (1965) demonstrated that Km = KI under conditions where pseudo-first order kinetics are observed. A further complication was introduced when it was shown that the reaction of cytochrome g_with cytochrome g oxidase exhibits a biphasic (or multiphasic) kinetic pattern (Nicholls, 1964, 1965) when data are collected over a larger range of substrate concentrations including very low concentrations of cytochrome g_(Ferguson-Miller gt _3l., 1976; Errede gt_3l., 1976). Despite extensive steady state and pre—steady state kinetic analysis of the mechanism of this complex electron transfer reaction, a unified view of the significance of the high and low affinity reactions of cytochrome g_has not emerged. One basic explanation of the data is that there are two separate catalytic reactions of cytochrome g_with cytochrome g_oxidase, occurring at distinct sites on the monomer (Ferguson-Miller, 1976; Errede gt_3l,, 1976). This “two-site" mechanism is supported by binding data indicating two cytochrome g_bind per cytochrome 333 under the conditions of low ionic strength where the biphasic pattern is easily 21 observed (Ferguson-Miller gt_3l., 1976), and the K0 values for the binding reactions are similar to the Km values from the kinetic analysis (Ferguson-Miller gt_3l., 1978,1979). Whether the two sites are identical and negatively cooperative (Veerman gt 31., 1980; Osheroff gt_3l,, 1983) or inherently different, cannot be discerned from the kinetics or binding data. An alternative ”two—site” mechanism has been put forward by Margoliash and Bosshard (1983), suggesting that electron transfer occurs at only one site, but an additional cytochrome g_can bind near the first with lower affinity because of charge repulsion. They propose a mechanism in which the first molecule of cytochrome g_binds tightly to the catalytic site, and its presence reduces the affinity of subsequent interactions of cytochrome g_nearby. At high concentrations of cytochrome 3, when these additional reactions are favored, they cause a reciprocal reduction in affinity at the electron transfer site, resulting in the second kinetic phase. In contrast, a ”one-site” mechanism has been proposed by other researchers who find their pre-steady state kinetic data can be adequately interpreted by the consecutive reactions of two molecules of cytochrome ngith cytochrome g oxidase at a single site. Another "one-site" mechanism suggests that the two different affinity reactions are the result of negative cooperativity within a dimer form of cytochrome g oxidase (Bisson gt 31., 1980; Capaldi gt_3l,, 1982; Nalecz gt_3l., 1983). In one interpretation, the monomer has one high affinity binding site for cytochrome g, and dimerization of the monomers places the binding domains in a configuration which permits only the first cytochrome g to bind with high affinity, the affinity of the second 22 cytochrome g_interaction being lowered due to electrostatic and steric hindrance (Nalecz gt_31., 1983). Dependence of Activity on the Physical Form of Cytochrome c Oxidase. The interpretation of this complex electron transfer reaction is made even more difficult and controversial by the dependence of the apparent kinetic constants on the physical status of the enzyme used for study. Cytochrome g oxidase activity in the native membrane differs from that in the detergent solubilized state (Smith and Camerino, I963; Vanneste _t__1,, 1974), is dependent on the nature of the detergent and the detergent-to-protein ratios used (Mason and Ganapathy, 1970), and undergoes a well documented decline during purification (Vanneste gt 31., 1974). The activity of the purified protein can be increased by thoroughly dispersing the protein with non-ionic detergent and adding back the highly fluid hydrocarbon environment (Vik and Capaldi, 1980) it requires for maximal activity (Vik and Capaldi, 1977). In addition, it has been reported that bovine heart cytochrome g_oxidase activity is specifically dependent on the presence of cardiolipin (Awasthi ££.El-, 1971; Robinson gt_31,, 1980; Fry and Green, 1980; Vik gt_31,, 1981; Robinson, 1982), which has also been proposed to be essential in the low affinity binding of cytochrome g_(Bisson gt_31,, I980; Vik gt 31., I981; Fuller gt 31., 1981). This requirement for cardiolipin has been questioned in the case of the enzymes from yeast (§1_cerevisae) (Watts g_ l., 1978) and dogfish heart (Sgualus acanthias) (Al-Tai gt_31., t— 1983), however. Cytochrome g_oxidase activity has been reported to be further influenced by loss of certain peptides (Penttila and Wikstrom, 1981), specific interaction with an annular layer of lipids (Jost gt 23 31., 1973a,b), and interaction with itself to form a dimer consisting of two complete sets of subunits (Robinson and Capaldi, 1977; Bisson gt 31., 1980; Wikstrom, 1981; Ferguson-Miller gt_31,, 1982). A prerequisite for elucidating the reaction mechanism of this electron transfer complex is the clarification of which properties of the isolated enzyme are characteristic of the native state, and which merely represent purification artifacts or species differences. The approach used in the research presented in this dissertation has been to develop a new purification procedure for rat liver cytochrome g oxidase, since it is well documented that traditional methods of purifying the most commonly studied mammalian form of the enzyme, bovine heart cytochrome g_oxidase, produce non-homogeneous preparations containing varying numbers of subunits, and in varying states of aggre- gation (Wikstrom gt_31., 1981). Furthermore, mitochondria from bovine heart are more difficult to obtain as intact, correctly oriented mitoplasts to use as a well defined system for establishing reliable kinetic constants for native cytochrome g_oxidase with which to evaluate the activity of the purified enzyme. The method developed for purifying rat liver oxidase is based on affinity chromatography in laurylmaltoside, a detergent chosen for its ability to disperse and maintain the enzyme activity in the solubilized membrane and purified form. In addition to a reduced polypeptide composition, cytochrome_g oxidase produced by this procedure has higher activity and purity, and lower lipid content than previously reported. The properties of the purified protein have been investigated to establish the roles of subunit composition, lipid content, and state of self-association in governing the electron transfer activity, 24 respiratory control, and proton translocating activity of cytochrome g oxidase. CHAPTER 1 EVALUATION OF THE SUITABILITY OF ALKYLGLYCOSIDE DETERGENTS FOR USE WITH CYTOCHROME g_OXIDASE The solubilization and purification of cytochrome g_oxidase always requires the use of detergents. Traditionally deoxycholate, cholate, Triton X-100, and Tween have been employed, even though none of these commercially available detergents is entirely suitable for this purpose. Triton X-100 is able to effectively disperse the mitochondrial proteins, but cytochrome_g oxidase has very low activity in this detergent and it is difficult to remove from the enzyme for reconstitution experiments. 0n the other hand, Tween provides a hydrophobic environment that supports fairly high levels of oxidase activity, and the bile salt detergents are readily removable by dialysis, but none of these compounds are particularly good dispersing agents. Ideally, the detergent to use in purifying and analyzing the structure-function relationships of cytochrome g_oxidase should possess a combination of these abilities to disperse the mitochondrial proteins in an active state and be easily removable. In addition, the use of detergents with simple, well-defined structures, high purity, and small uniform micelles greatly facilitates the characterization of the physical and functional properties of membrane proteins (Tanford, 25 26 1972; Tanford and Reynolds, 1976). Detergents meeting all these specifications have not been commercially available in the past. Therefore, in 1980, Rosevear gt__1, began the systematic synthesis of a series of pure, structurally defined alkylglycosides they predicted might have the molecular requirements necessary to activate cytochrome g_oxidase (Figure 2), and in this way they hoped to identify a detergent more appropriate for use with this enzyme. The potential value of the alkylglycoside molecule in detergent applications was of key interest because one of these, octylglucoside, is an effective solubilizing agent for several other membrane proteins (Baron and Thompson, 1975; Stubbs and Litman, 1978; Wittenberger gt 31,, 1978; Felgner gt_31., I979; Petri and Wagner, 1979; Stadt gt_31., 1982), and it is dialyzable due to a high critical micelle concentration (25 mM) (Baron and Thompson, 1975; Ferguson—Miller gt 31., 1982). However, early studies showed that octylglucoside inactivates and denatures cytochrome g_oxidase from gt_gtggg3 and bovine heart (Rosevear gt_31,, 1980). The same studies indicated that laurylmaltoside, a larger analog with a 12 carbon aliphatic tail balanced by a disaccharide head group, might be more suitable. Although the lower critical micelle concentration of laurylmaltoside (0.16 mM; Ferguson-Miller gt_31,, 1982) prevents its easy removal by dialysis, it was found to produce a marked stimulation of cytochrome g oxidase activity and to be a very efficient dispersing agent for the enzyme. Purified bovine heart cytochrome g oxidase eluted as a single peak (Mr m300,000) from a gel permeation column equilibrated in laurylmaltoside under conditions that produced predominantly high molecular weight aggregates in deoxycholate and Tween 20, and multiple FIGURE 2: Alkylglycoside detergents. 27 28 H0 0 0-(CH2)7CH3 0H B-D -OCTYLGLUCOS IDE CH 0H 2 0 H0110 0—(c112)nc113 0H fl-D-LAURYLGLUCOSIDE CHZOH HO 0 H0 H HO O 0 0H CIH 2 OH 0—(CH2)11CH3 /3-D -LAURYLCELLOBIOSIDE Figure CHZOH HO 0 H0 CHZOH OH 0 0 H 0—(CH ) CH 0H 2 11 3 /3-D -LAURYLMALTOS|DE H OH OH HO 0 04C”2)115“3 /3-D -LAURYLLACTOS IDE 0H CH 0H 0H 2 0 HO 0H 1H2 0H /3-D-HEXYL1ACTOS|DE 29 molecular species in octylglucoside (the result of enzyme decomposition and polymerization)(Figure 3). This chapter reports the suitability of laurylmaltoside and several other alkylglycosides as detergents for use in the purification of rat liver cytochrome_g oxidase. The ability of these compounds to solubilize the inner mitochondrial membrane proteins and their effect on the kinetics of cytochrome 3 oxidation were investigated. The results indicate that laurylmaltoside is the most effective detergent for efficiently dispersing the mitochondrial membrane proteins in an uniquely active state, but that 1-oleoylmaltotrioside, a trisaccharide with a longer hydrophobic moiety more similar to the lipid bilayer, provides a better environment for stabilizing the delipidated enzyme activity. These studies also show that detergent solubilization of the inner mitochondrial membrane causes profound changes in the kinetic parameters of the cytochrome g oxidase reaction, but the extent of maximal activation is dependent on detergent structure. FIGURE 3: 30 Molecular size distribution of bovine heart cytochrome g oxidase equilibrated in octylglucoside, laurylmaltoside, deoxycholate, or Tween 20. Cytochrome g oxidase (200 pl, 1 to 22 uM) was gel filtered through LKB Ultrogel 34 (molecular weight range: 20,000 - 350,000) in a 1 x 26 cm column equilibrated in the specified concentrations of detergent plus 100 mM KCl, 10 mM Tris-HCl, pH 7.8, 25°C, at a flow rate of 5 ml/hr. (A) 30 mM (0.9%) octylglucoside. (B) 2 mM (0.1%) laurylmaltoside. (C) 0.5% deoxycholate. (D) 0.5% Tween 20. (from: Ferguson-Miller et al., 1982.) Reprinted by permission © 1982, Elsevier North—Holland, Inc. ABSORBANCE (280 nm) 0.| 31 I I j T T T T T OCTYLGLUCOSIDE A _ DEOXYCHOLATE C T ' T I l I ‘ l 0J5 LAURYL MALTOSIDE _ 0.10 _ 0.05 - o T'l" 5 l2 ‘RE" 20 l l l ‘ l ' l TWEEN 20 0| — 1 1 1 l “% l B 16 V ElUTION VOLUME (ml) Figure 3 EXPERIMENTAL PROCEDURES Chemicals. B-D-octylglucoside, B-D-laurylmaltoside, B-D-laurylglu— coside, B—D-lauryllactoside, B-D—hexyllactoside, B-D-laurylcellobio- side, and B-D-l-oleoylmaltotrioside were synthesized according to Rosevear gt_31, (1980). Deoxycholic acid (Sigma, St. Louis, MO) was recrystallized at least 3 times from hot 80% acetone. Cytochrome 3 (horse heart, Sigma type VI) was purified by chromatography on carboxymethyl-cellulose according to the procedure of Brautigan gt 31. (1978c). Polymerized cytochrome g_was removed immediately before use by gel filtration, in the reduced form, on Sephadex G-75 superfine in 25 mM Tris-cacodylate, pH 7.9. Dimethylsulfoxide was treated with Amberlite MB-3 prior to use (both from Mallinckrodt, Paris, KY) (Fleischer, 1979). Other chemicals were of the best grade available from the sources indicated: cacodylic acid, sodium dithionite (J.T. Baker Chemical, Phillipsburg, NJ or Sigma); N,N,N',N‘-tetramethyl-p- phenylenediamine dihydrochloride (Eastman Chemicals, Rochester, NY); Trizma base (ultrapure) (Calbiochem, LaJolla, CA, or Sigma); ascorbic acid, phosphoric acid, potassium ferricyanide, potassium chloride, sodium chloride, EDTA, sucrose (analytical grade) (Mallinckrodt, Paris, KY); histidine monochloride (Merck, Rahway, NJ); phenylmethanesulfonyl fluoride, l—oleoyl-lysophosphatidyl choline (Sigma); asolectin (Associated Concentrates, Long Island, NY). 32 33 Spectral Measurements. UV and visible spectra were recorded on either a Perkin-Elmer 559 or Aminco-DW2a UV/Vis Spectrophotometer with Midan T microprocessor. The concentrations of cytochromes £1, 3, and 333 in inner mitochondrial membrane fractions were calculated from difference spectra (dithionite reduced minus air-, or ferricyanide- oxidized) using the extinction coefficients of von Jagow and Klingenberg (1972): cytochrome g1, Ac 553-542 nm = 18.7 mM“1 cm‘l; cytochrome 3, As 560-575 nm = 23.4 mM'1 cm'l; cytochrome 333, Ag 605-630 nm = 24.0 mM'1 cm'l. Correction was not made for the overlap of the absorbance spectra of cytochromes g1 and_3; this results in overestimates of their concentrations. Assay Methods. Electron transfer activity of cytochrome g_oxidase was measured polarographically as described by Ferguson-Miller gt 31. (1976, 1978) using a Gilson model K-IC oxygraph equipped with a Yellow Springs Instruments electrode. Steady state kinetic measurements were performed under the conditions detailed in the table and figure legends. Ascorbate (2.8 mM) and TMPD (0.56 mM) were used in the assay system to donate electrons to cytochrome g, Turnover numbers were calculated by multiplying the velocity in nmol 02/5 by 4 (to convert to nmol cytochrome g/s) and dividing by the total nmol cytochrome 333 in the assay. The conditions of detergent treatment used to solubilize inner mitochondrial membranes prior to kinetic analysis were those which resulted in maximal activation of cytochrome g_oxidase activity. In each case, the protein-to-detergent ratio was chosen by determining oxidase activity in a series of samples containing various ratios from 34 0.2—10 mg detergent/mg protein in 0.5 x TSH: laurylmaltoside at 76 mg/ml and membranes at 40 mg protein/ml, diluted in the assay to 0.34 mM detergent and 0.018 nmol cytochrome 333; deoxycholate at 10 mg/ml and membranes at 10 mg protein/ml, diluted in the assay to 0.20 mM detergent and 0.031 nmol cytochrome 333 (Smith and Camerino, 1963); l-oleoyl-lysophosphatidyl choline at 10 mg/ml and membranes at 10 mg protein/ml, diluted in the assay to 0.11 mM detergent and 0.020 nmol cytochrome 333; l-oleoylmaltotrioside at 10 mg/ml and membranes at 10 mg protein/ml, diluted in the assay to 0.08 mM detergent and 0.022 nmol cytochrome 333. Alkylg1ycoside Extraction of Inner Mitochondrial Membrane Cytochromes. Rat liver inner mitochondrial membranes were isolated as intact mitoplasts according to the procedure of Sottocasa gt 31. (1967), as modified by Felgner gt 31. (1979). The membranes were suspended at 30 mg/ml Biuret protein (Jacobs gt 31,, 1956) in 0.33 M sucrose, 50 mM Tris-HCl, pH 8, 1 mM histidine (leSH), then made 20% in DMSO, and frozen at ~65°C. This storage procedure maintains the membrane system intact, and preserves about 95% of the electron transfer activity (Fleischer, 1979). The membranes were thawed undisturbed on ice, diluted 3 fold by addition of 0.66 M sucrose, 100 mM Tris-HCl, pH 8, 2 mM histidine (2xTSH), pelleted by centrifugation at 10,000 xg for 15 min, gently redispersed in 5 fold 2xTSH, and pelleted at 20,000 xg for 20 min. Alternatively, cytochrome g depleted inner mitochrondrial membranes were prepared by twice centrifuging the membranes out of 0.45 M sucrose, 0.15 M KCl at 10,000 xg for 15 min, then once out of 2xTSH at 20,000 xg for 20 min. As indicated in the figure legends, some extractions were performed on fresh (never frozen) mitoplasts. 35 Mitochondrial membranes were usually resuspended at 40 mg protein/ml (Biuret) in 2xTSH, then alkylglycoside detergent plus KCl was added in a volume of water equal to that of the membrane suspension, producing the final concentrations of protein, salt, buffer, and detergent shown in the figure legends. The suspensions were made l mM PMSF by addition of a small volume of a concentrated solution in isopropanol, incubated at 4° or 25°C for the times indicated, then centrifuged at 40,000 xg for 1 h. The concentrations of cytochromes 31, 3, and 333 in the resulting supernatants and pellets were determined spectrally. RESULTS AND DISCUSSION Detergent Effects on the Kinetic Characteristics of Rat Liver Cytochrome c Oxidase in Solubilized Mitochondrial Membranes. To interpret the results of studies on purified enzyme in terms of the native form and function of cytochrome g_oxidase, one needs to have an understanding of the specific ways in which oxidase activity is influenced by the intact membrane and detergent solubilization. Therefore, steady-state kinetic analysis of the oxidation of cytochrome g_by rat liver cytochrome g_oxidase was performed following a variety of detergent treatments of the intact membranes. Figure 4 is an Eadie-Hofstee plot of the data obtained with rat liver inner mitochondrial membranes (cytochrome_g—depleted mitoplasts), and the same membranes solubilized with laurylmaltoside, assayed under conditions that maximize rates of turnover: 50 mM KPi, pH 6.5 (Davies _t__1,, 1964). It is apparent that the kinetic parameters of the enzyme are markedly altered by the dissolution of the membrane, the overall turnover number of the solubilized enzyme is greatly stimulated and the apparent Km values differ from those of the enzyme in the intact membrane. It is usually assumed that the increased activity seen when mitochondrial membrane particles are solubilized is the result of increased availability of the enzyme, due to the existence of a considerable proportion of inverted membrane vesicles in many particle preparations. However, the membranes used in these studies 36 FIGURE 4: 37 Eadie-Hofstee plots of the kinetics of oxidation of cytochrome g_by rat liver cytochrome g_oxidase in 50 mM KPi, pH 6.5. Rat liver inner mitochondrial membranes «3); membranes treated with detergent (:1). The assay conditions and detergent to protein ratios used were those required to produce maximal activities as described in Experimental Procedures. Assay medium (1.75 ml) contained 0.05 nmol cytochrome 333 (intact membranes) or 0.018 nmol cytochrome _333 and 0.34 mM laurylmaltoside (detergent-treated membranes). Cytochrome g_concentrations of 0.07 to 50 uM were used. Turnover numbers (TN) were calculated as described in Experimental Procedures. 38 Illl 500 750 TN (3‘1) 250 30* 0 fan? N.9 x T.,: _ w \ 2... Figure 4 ‘11 39 are intact, right-side-out mitoplasts in which the cytochrome g oxidase should already be totally accessible to cytochrome 3, Moreover, the kinetic changes seen upon addition of detergent are not those expected for a simple increase in enzyme concentration which would result in increased Vma values for both kinetic phases, but little change x in the apparent Km (Ferguson-Miller gt_31,, 1976). Instead, laurylmaltoside solubilization of inner mitochondrial membranes decreases the apparent Km values of the low affinity phase, and results in the development of an additional phase in the kinetic plot, which may reflect a third type of cytochrome g_interaction under these conditions. In spite of the complexity of the kinetic picture, a striking feature of the data is the increase in overall turnover number from 350 5'1 to 1100 5'1 with no large change in the contribution to the activity by the initial high-affinity phase, and no concomitant decrease in apparent affinity for cytochrome thin the second (or third) phase that would account for an increased turnover in terms of an increased rate of dissociation of cytochrome c. Thus, the very high activity of cytochrome g_oxidase in laurylmaltoside dissolved membranes appears to be due to increased rates of turnover of cytochrome g_in its low affinity reactions with oxidase, indicating that some aspect of substrate binding, or the electron transfer pathway, is modified by the altered hydrophobic environment. The steady—state kinetics of the reaction of cytochrome g_with cytochrome g_oxidase in native and laurylmaltoside solubilized inner mitochondrial membranes were also studied under ionic conditions that promote very tight binding of the cytochrome g; 25 nM Tris-cacodylate, pH 7.9 (Ferguson-Miller gt_31,, 1976)(Figure 5). This assay system FIGURE 5: 40 Eadie-Hofstee plots of the kinetics of oxidation of cytochrome g_by rat liver cytochrome 3 oxidase in 25 mM Tris-cacodylate, pH 7.9. Rat liver inner mitochondrial membranes (0); membranes treated with detergent (D). Assay conditions and the detergent to protein ratio (1 mg laurylmaltoside/mg protein) were as described in Experimen- tal Procedures. Assay medium (1.75 ml) contained 0.03 nmol cytochrome 333 (intact membranes) or 0.026 nm of cytochrome 333 and 0.11 mM laurylmaltoside (detergent- treated membranes). Cytochrome g_concentrations from 0.07 to 50 uM were used. 00 ha 1 (M"1 8") TN / [8] x10"8 ED J: 0 41 160 360 500 TN (8“) Figure 5 42 results in much lower apparent Km values and turnover numbers, and the y-axis scale is reduced by a factor of 10 compared to Figure 4. The detergent effect is much less striking under these conditions, however, the data again show that detergent solubilization stimulates the overall turnover number of cytochrome g_oxidase mainly by increasing the contribution of the second phase. The release of‘ oxidase from the membrane also appears to decrease the affinity of cytochrome g_in both kinetic phases, in contrast to the results obtained in 50 mM KPi, pH 6.5 in which the apparent affinity of cytochrome g_in the second kinetic phase increased. The significance of the differences in oxidase activity measured in these two different assay systems will be discussed further in Chapter 3. Other detergents of different chemical character, l-oleoylmalto- trioside, l-oleoyl—lysophosphatidyl choline, and deoxycholate produce very similar kinetic effects to those of laurlymaltoside (Figure 6), but differ in their ability to increase the rate of turnover of the enzyme. For each detergent, a series of detergent-to-protein ratios was tested to determine the conditions required for maximal activation. The maximal turnover rates obtained at 50 mM KPi, pH 6.5, from the data shown in Figure 6 are 600 s'1 in 1-oleoylmaltotrioside (an alkyl— glycoside which stabilizes the activity of purified cytochrome g oxidase: see below), 700 s‘1 in deoxycholate, 850 s'1 in I-oleoyl-lysophosphatidyl choline (an amphipath which markedly stimu— lates purified cytochrome g oxidase activity: Vik and Capaldi, 1980), and 1100 5'1 in laurylmaltoside (the highest turnover number reported for cytochrome g_oxidase). The results indicate that although the chemical nature of the detergent is a factor in determining the FIGURE 6: 43 Eadie- Hofstee plots of the kinetics of oxidation of cyto- chrome g by rat liver cytochrome c oxidase in inner mito- chondrial membranes solubilized with various detergents. Assays were performed in 50 mM KPi, pH 6. 5, with: Lauryl- maltoside at 0.34 mM and 0.018 nmol cytochrome 333 (:1); deoxycholate at 0.20 mM and 0.031 nmol cytochrome aa (A); l-oleoyl~lysophosphatidyl choline at 0.11 mM and 0.0 0 nmol cytochrome_333 (o); 1- oleoylmaltotrioside at 0.08 mM and 0. 022 nmol cytochrome 333 (I ). The protein- -to- detergent ratios used to solubilize the membranes were those which produced maximal activation of cytochrome_g oxidase activity measured polarographically as oxygen consumption under the conditions described in Experimental Procedures. The range of cytochrome g_concentrations used was 0.07 - 40 uM. 44 30 900 600 TN (8") 300 _ m fans: to? x 52 \ 2» Figure 6 45 maximal activity in the low affinity reaction(s) with cytochrome g, the most profound changes in the kinetics (altered Km values and high turnover rates) are produced by simple disruption of the membrane structure and are not dependent on the specific detergent. The effect of detergent solubilization on the kinetics of the second phase suggest that the low affinity interactions of cytochrome_g with the oxidase may involve an electron transfer event or secondary interactions of cytochrome t that are more sensitive to the lipid environment, as well as other influences of the membrane such as protein-protein interactions or membrane surface pressure (Conrad and Singer, 1979). With regard to the latter, increased oxidase activity could be explained as resulting from a change in association with other membrane proteins, or from conformational changes caused by more subtle forces. The fact that the high affinity reaction of cytochrome g_is changed little by detergent treatment of the membrane suggests that whatever is rate limiting in this phase of the reaction is relatively insensitive to the hydrophobic environment. Effect of Detergent and Lipid on the Stability of Purified Cytochrome c Oxidase. The ability of several different hydrophobic environments to substitute for bulk mitochondrial membrane phospholipid in maintaining cytochrome g_oxidase in an active and stable form was evaluated by measuring the activity of highly lipid-depleted rat liver enzyme (prepared as described in Chapter 2), with results shown in Figure 7. In the oxygen consumption tracings, A and B, the conditions used were similar to those reported in the literature to maximally activate purified bovine heart cytochrome g oxidase (Vik and Capaldi, 1980). A, FIGURE 7: 46 Time course of oxygen uptake by purified rat liver cyto- chrome g_oxidase assayed in the presence of various amphipaths. Rat liver cytochrome g oxidase was purified in laurylmaltoside as described in Chapter 2, and activity was measured in 50 mM KPi, pH 6.5, at 0.03 nmol cytochrome 333 with the following concentrations of detergent and lipid: (A) Soybean phospholipids, 0.28 mg/ml and laurylmaltoside, 0.11 mg/ml. (B) 1-0leoylphosphatidyl choline, 0.21 mg/ml. (C) 1-Oleoylmaltotrioside, 0.56 mg/ml and laurylmaltoside, 0.11 mg/ml. (D) Laurylmaltoside, 0.11 mg/ml. Oxygen consumption was measured polargraphically as described in Experimental Procedures. The reactions were started by the addition of 13 uM cytochrome 3 (final concentration). 47 A I 100 nmol 02 c o ‘ ‘ \ J A J 1 1 J minutes Figure 7 48 enzyme in detergent plus asolectin vesicles (soybean phospholipids); B, enzyme in detergent plus l-oleoyl-lysophosphatidyl choline (a compound which provides in micellar form the C18:1 alkyl chain predominat- ing in mitochondrial phospholipids). In both assays, laurylmaltoside replaced Triton X-100 originally used to solubilize the enzyme, and the results were consistent with previous findings; the thoroughly dispersed enzyme was most active when supplied with asolectin or lysophosphatidylcholine (TN=482 5'1 and 404 5'1, respectively), and the activity was constant over the course of the assay. In contrast, when laurylmaltoside alone provided the hydrophobic environment for the delipidated enzyme, the initially high activity (TN=311 5'1) underwent a marked decline during the time required for assay (tracing D). Stabilization of the oxidase activity without additional phospholipid could be achieved by the use of the alkylglyco- side detergent, 1—oleoylmaltotrioside, composed of a C18:1 alkyl tail and trisaccharide head group. This detergent was developed by Van Aken and Ferguson-Miller (unpublished experiments) in an attempt to produce an effective dispersing agent for mitochondrial proteins that would also supply the fluid hydrocarbon environment provided by unsaturated alkyl chains that is required for maximal oxidase activity (Vik gt 31., 1977). Unfortunately, the usefulness of l-oleoylmalto- trioside as a detergent is limited by its low critical micelle concen— tration (<.005 mM), large micelle size (Mr 125,000), and poor dispersing ability (Van Aken and Ferguson-Miller, unpublished experi- ments), and it does not stimulate the activity of oxidase dispersed in laurylmaltoside (TN=320 5'1). Nevertheless, 1—oleoylmaltotrioside 49 is able to prevent the decline in oxidase activity that occurs when laurylmaltoside is used alone (tracing C). This suggests that the C18:1 alkyl chain plays an important role in stabilizing the active conformation of cytochrome g_oxidase in dilute solutions at low pH. Alkyngycoside Fractionation and Solubilization of Inner Mitochondrial Membrane Cytochromes. The ability of the series of alkylglycosides shown in Figure 2 to fractionate, as well as solubilize the cytochromes of rat liver inner mitochondrial membranes was studied with the purpose of identifying a detergent with properties which could be exploited to develop an improved purification procedure for cytochrome g_oxidase. The rat liver enzyme has traditionally been difficult to purify because the ammonium sulfate fractionation procedures used to separate cytochrome 3 reductase and cytochrome g oxidase from yeast and bovine heart are much less effective when applied to rat liver. The finding of Felgner gt_31, (1979), that octylglucoside can be used under carefully controlled conditions to selectively solubilize and reconstitute hexokinase binding protein from rat liver outer mitochondrial membranes, suggested a similar procedure might be useful for cytochrome g_oxidase purification. Studies in this laboratory with octylglucoside (Rosevear 23.91-: 1980) showed that, although it was effective in solubilizing the mitochondrial membrane proteins, it was not able to differentially extract cytochromes 31, 3_or 333 under any conditions of detergent and ionic strength tested. Laurylmaltoside, on the other hand, exhibited a slight tendency to differentially extract cytochromes 3 and 31 at detergent 50 concentrations between 3-5 mM in 1M KCl, and effected complete solubilization of the mitochondrial membranes at considerably lower concentrations than required for octylglucoside. The results of similar studies using two of the other T. (1980), alkylglycosides synthesized by the procedure of Rosevear gt hexyllactoside and laurylcellobioside, are shown in Figures 8 and 9. Hexyllactoside, a dissacharide with a short C5 alkyl chain attached, was not capable of fully extracting any of the cytochromes from the mitochondrial membranes over a range of concentrations from 1-50 mM hexyllactoside in IN KCl; only 13% of the cytochrome 333 was solubilized at the highest detergent concentration tested (Figure 8). Furthermore, hexyllactoside did not exhibit the selectivity required to effect a useful separation of cytochromes 3 and_gl from 333. Its unusual ability to preferentially extract cytochrome 31, a particularly difficult protein to isolate, may be of future value for the purification of this protein. Laurylcellobioside, which differs from laurylmaltoside only in the configuration of the sugar head group, appears to be a slightly better solubilizing agent for the cytochromes than hexyllactoside, when used at low concentrations (1-5 mM laurylcellobioside)(Figure 9). However, the usefulness of this compound as a detergent is limited by its own relatively low solubility. At concentrations above 10 mM, laurylcellobioside precipitated at the low temperature of the incubation and centrifugation steps and apparently pulled proteins in solution down with it. The two other alkylglycosides shown in Figure 2, lauryllactoside and octyllactoside, were also insufficiently soluble for use as detergents (data not shown). FIGURE 8: 51 Hexyllactoside extraction of cytochromes from rat liver inner mitochondrial membranes. Membranes in 1.0 M KCl, 1 x TSH, pH 8, were incubated at the indicated concentrations of hexyllactoside on ice for 10 min, then centrifuged at 40,000 xg for 1 h. The concentrations of cytochrome E3 (A), cytochrome 3 (o), and cytochrome 31 (I) extracted into the supernatant were determined spectrophotometrically, and the soluble cytochrome _c_ oxidase activity (A) was assayed polarographically in 25 mM Tris-acetate, pH 7.9, 250 mM sucrose, with 30 uM cytochrome g as described in Experimen- tal Procedures. The components and activity are plotted as percent of total present in the original membrane suspension (20 mg protein/ml): cytochrome 333, 1.92 uM; cytochrome 3, 2.31 uM; cytochrome 31 1.31 uM; cytochrome_g oxidase activity, TN=333 s-l. 52 Ru 7 _ Ru _ _ :28. nx.~0>._._>_._.0.c. mm._._>_._.o< ww<9xo O 5 5 2 q H nu LU 003 20 _ _ _ 5 O 5 7 5 2 :28. .5 awko._._>_._.o< umo 120 - IOO - Figure 10 58 In an effort to extend the range of detergent concentrations over which selective solubilization of cytochromes 3 and 31 occurs, and thus establish a more controlled and quantitatively reproducible effect, membranes were extracted with laurylmaltoside at further reduced ionic strength, 0.25 M KCl, 1xTSH, pH 8 (Figure 11). The results show that at the lower salt concentration more detergent was required for extraction of all the cytochromes, and that the resolving capability of the system was not increased sufficiently to justify the use of the additional detergent. Experiments in which the protein concentration of the membrane suspensions were varied indicated that the detergent-to-protein ratio was the important factor in determining the selectivity of the solubilization process (not shown). The time and temperature dependence of the selectivity of solubi- lization is shown in Figure 12. Suspensions of inner mitochondrial membranes in 0.5M KCl, 1xTSH, pH 8, and 10 mM laurylmaltoside were incubated at 4° or 25°C for various times from 5-60 min. The results show that the general solubility of the cytochromes is increased by longer incubation times at 25°C, but not at 4°C. It appears that the specificity of the solubilization process is slightly enhanced by short incubation times at both temperatures, and to be optimal for a 10 min incubation at 4°C (the conditions originally used in the above studies of detergent effect). Conclusions. The ability of laurylmaltoside to selectively extract cytochromes 3 and 31 from mitochondrial membranes and fully solubi- lize cytochrome g_oxidase in an uniquely active state, identify it as an important new detergent for use in purifying cytochrome g oxidase. FIGURE 11: 59 Laurylmaltoside extraction of cytochromes from rat liver inner mitochondrial membranes in 0.25 M KCl. Membranes in 1 x TSH, pH 8, and 0.25 M KCl were incubated at the indicated concentration of laurylmaltoside on ice for 10 min, then centrifuged at 40,000xg for 1 h. The concentra- tions of cytochrome 333 (A), cytochrome 3 (g), and cytochrome £1 (I) extracted into the supernatants were determined spectrophotometrically, and the soluble cyto- chrome g oxidase activity (A) was assayed polarographically in 25 mM Tris acetate, pH 7.9, 250 mM sucrose, and 30 uM cytochrome_3 as described in Experimental Procedures. The components and activity were plotted as percent of total present in the original membrane suspension (20 mg protein/ml): cytochrome 3_3, 4.58 mM; cytochrome 3, 3.25 uM; cytochrome_gl, 1.71 pH; cytochrome_3 oxidase activity, TN = 100 s-1. 60 -200 :20... as »._._>_._.o< @3338 o m m n m . “ITS l5 a _ d 20 IO LAURYL MALTOSIDE (mM) _ _ _ 5 O 5 m o 7 w coup—xx; awkoo _ Kg 7: ISO - _ :u a‘ Figure 11 FIGURE 12: 61 Time and temperature dependence of laurylmaltoside extrac- tion of cytochromes from rat liver inner mitochondrial membranes. Membranes in 0.5 M KCl, 1 x TSH, pH 8, and 10 mM laurylmaltoside were incubated at 4°C (closed symbols, solid curve) and 25°C (open symbols, broken curve) for the times indicated. The concentrations of the cytochromes extracted into the supernatants were determined spectro- photometrically as described in Experimental Procedures, and plotted as percent of total present in the original membrane suspension (20 mg protein/ml): cytochrome 333, 2.50 uM (A,A); cytochrome 3, 2.48 uM (0,0); cytochrome £1.1-71uM(D.I)- 62 3 C a a l 40 20 / I M3 b an, 6 III // / / 4» AN 1 II // \..// ,a R . la _ _ _ _ _ a m n w a 229$ o wko0 TIME (min) Figure 12 63 Its suitability is somewhat limited by a low critical micelle concentration resulting in slow removal by dialysis. In addition, the C12 alkyl group is not as effective as the C1821 alkyl chain in maintaining the activity of the purified, delipidated enzyme in dilute solutions. However, the excellent dispersive ability of laurylmaltoside make it well suited for use in the chromatographic purification of this hydrophobic protein. The development of a procedure which uses horse cytochrome ggaffinity chromatography in laurylmaltoside to purify cytochrome g_oxidase from rat liver and bovine heart mitochondria is the subject of Chapter 2. CHAPTER 2 PURIFICATION OF BOVINE HEART AND RAT LIVER CYTOCHROME g_OXIDASE USING CYTOCHROME g_AFFINITY CHROMATOGRAPHY IN LAURYLMALTOSIDE Traditionally, cytochrome g oxidase has been purified from mito- chondria using bile salt detergents and ammonium sulfate fractionation (Fowler gt_31,, 1962; van Buuren, 1972; Kuboyama gt 31., 1972). Due to their charge and lack of dispersing ability, these detergents hindered subsequent purification of the solubilized enzyme by chromatographic procedures. A purification procedure for cytochrome g_oxidase from 31 353333 using horse cytochrome g affinity chromatography was developed when it was found that the non-ionic detergent, Triton X-100, effec- tively solubilizes the mitochondrial proteins from this source. However, published reports indicate this method is not effective in purifying oxidase from bovine heart or yeast sources (Weiss and Kolb, 1979; Bill gt_31,, 1980, 1982; Azzi gt_31,, 1982). It seemed likely that the major obstacle for other tissues might be difficulty in fully dispersing the proteins of the mitochondrial membranes at low ionic strengths. Studies on a number of alkylglycoside detergents have shown that laurylmaltoside is well suited for use in chromatographic purifi- cation (Rosevear gt_31,, 1980) due to its ability both to disperse cytochrome g_oxidase and support high levels of electron transfer activity (Chapter 1). Using laurylmaltoside and an affinity matrix of 64 65 horse cytochrome g_linked via lysine residues to Sepharose 4B, cytochrome g_0xidase from bovine heart mitochondrial particles has been prepared in one chromatographic step with a purity and yield similar to that achieved using an affinity matrix prepared from yeast cytochrome g linked to a matrix via its sulfhydryl residue (Bill gt_31,, 1980, 1982; Azzi gt_31,, 1982). Employing the same matrix and detergent, a more extensive procedure for purifying rat liver cytochrome g oxidase has also been developed, that yields a uniquely active and delipidated enzyme, depleted of subunit III. EXPERIMENTAL PROCEDURES Chemicals. Cholic acid (Sigma, St. Louis, MO) was recrystallized at least 3 times from hot 95% ethanol. Other chemicals were the best grades available from the sources indicated here, or in Chapter 1: hydroxyapatite (Bio-Gel HTP) (Bio-Rad Laboratories, Richmond, CA): Sepharose 4B, cyanogen bromide (Sigma): Triton X-100 (scintillation grade) (Research Products International, Elk Grove Village, IL). Spectral Measurements. UV and visible difference spectra of inner mitochondrial membrane fractions were recorded, and the concentrations of the cytochromes calculated as indicated in Chapter 1. The extinc- tion coefficients for UV and visible absolute spectra of cytochrome 333 (dithionite reduced or air oxidized) were taken from the spectra published by van Buuren gt 31. (1972): cytochrome 333 (reduced) A 605-650 nm = 40 mm-lcm-l and Ae 444-490 nm = 205 mM cm-l; cytochrome 333 (oxidized) A6 420-490 nm = 140 mM'lcm'l. Assay Methods. Electron transfer activity of cytochrome g_oxidase at various stages of purification was measured as in Chapter 1, using the buffer and detergent conditions described in the legend of Table 1. Asolectin vesicles added to the assay mixture, where indicated, were prepared by sonicating to clarity soybean phospholipids (40 mg/ml) in 100 mM HEPES, pH 7.2, 0.1 mM EDTA, 2% cholate for 1 min/ml at 30 watts 66 67 using a Branson sonifier with microtip probe. Titanium particles were removed by centrifugation. Protein Determinations. Protein concentration of mitochondrial particles, inner mitochondrial membranes, and crude fractions of cytochrome g oxidase were estimated by the Biuret procedure according to the modification of Jacobs gt_31. (1956). The fluorescamine method of protein determination was used to follow protein during the chromatography of bovine heart mitochondrial particles in Triton X—100 (Udenfriend gt_31,, 1972). The method of Lowry gt_31, (1951), corrected for any absorbance contributed by the buffer, was used to estimate the protein in purified cytochrome_3 oxidase samples. Affinity Chromatography of Solubilized Bovine Heart Mitochondrial Particles. Keilin-Hartree particles were prepared from bovine heart essentially according to Ferguson—Miller gt 31. (1976), with the excep- tion that a Waring blender was used to disrupt the tissue (2 x 60 second bursts) instead of hand grinding in a mortar. The particles were suspended at N60 mg protein/ml (Biuret) in 100 mM KPi, pH 7.8, and frozen at -20°C. They were thawed and extensively dialyzed against Tris-cacodylate or sodium-Bicine, pH 7.9, at the concentration to be used in the chromatography. The mitochondrial particles were made 20 mg protein/ml (Biuret), and 2 mg of laurylmaltoside 0r 5 mg of Triton X-100 were added per mg of protein. The solution was stirred for 30 min, then centrifuged at 40,000 xg for 30 min. The supernatant, which contained 95-100% of the total oxidase in laurylmaltoside or 70-90% of the total oxidase in Triton X-100, was diluted to 1.5 mg protein/ml 68 (Biuret) in 0.75% (15 mM) laurylmaltoside, or 1% Triton X-100. The protein solution was chromatographed as described in the legend to Figure 14, on an affinity matrix of horse cytochrome g-Sepharose 4B (30 nmol of cytochrome g/ml of Sepharose) prepared according to Weiss and Juchs (1978). Cyanogen bromide activated Sepharose 4B was prepared according to Cuatrecasas gt_31, (1968) with the exception that the reaction was performed at 4°C. Rat Liver Cytochrome c Oxidase Purification. Rat liver inner mitochon— drial membranes were prepared, stored frozen, and washed free of DMSO as indicated under Experimental Procedures in Chapter 1. The pelleted membranes were resuspended at a final concentration of 40 mg protein/ml (Biuret) in 2X TSH. An equal volume of 1.0 M KCl, 16 mM laurylmalto- side was added with stirring to the membranes on ice. PMSF (100 mM in isopropanol) was added to a final concentration of 1 mM. The suspen- sion was incubated 10 min undisturbed on ice, then centrifuged at 40,000 xg for 1 h. The green, liquid "pellet" obtained was resuspended and diluted to a final concentration of 20 mg protein/ml (Biuret) in 0.5 x TSH plus 0.5% sodium cholate, then centrifuged at 40,000 xg for 1 h. The pellet contained 80-100% of the cytochrome 333 and was resuspended in a minimum volume of 50 nM Tris-HCl, pH 8.0, and 0.33 M sucrose (TS) containing 10 mM KPi, then dialyzed overnight against a 100-fold excess of the same buffer. The suspension was made 40 mg protein/ml (Biuret) in the same buffer plus 70-100 mM laurylmaltoside, stirred on ice for at least 30 min, and centrifuged at 40,000 xg for 30 min. A small insoluble pellet was discarded. The supernatant was diluted to 4 mg protein/ml (Biuret) with TS plus 10 mM KPi and 15 mM 69 laurylmaltoside, then applied to a hydroxyapatite column (6 nmol of 333/ml of hydroxyapatite) equilibrated at 4°C with the same buffer plus detergent. After washing with 2.5 column volumes of the same buffer, the cytochrome 333 was eluted with a linear gradient of 3.3 column volumes of KPi, 10-600 mM, pH 8.0, in the above buffer. Fractions containing cytochrome 333 with a 280 nm/420 nm absorbance ratio of 33.0 were pooled and dialyzed overnight vs. a 100-fold excess of 25 mM Tris-cacodylate, pH 7.9, 5% sucrose, with two changes of buffer. The enzyme was applied to an affinity matrix of horse ferrocytochrome g:Sepharose 48 (see above) (l mg protein/ml; 8-9 nmol of 333/ml of Sepharose) that was reduced with 2 column volumes of 5 mM ascorbate in 25 mM Tris-cacodylate, and equilibrated with cold 25 mM Tris-cacodylate, pH 7.9, 5% sucrose, 15 mM laurylmaltoside. After washing with 3 column volumes of the same buffer, the cytochrome 333 was eluted with a linear gradient of 4 column volumes of NaCl, 0-200 mM, in the above buffer. Fractions containing cytochrome 333 with a 280 nm/420 nm ratio of 52.0 were pooled, the enzyme was concentrated in an ultrafiltration cell using an Amicon XM-300 ultrafilter, then diluted with several volumes of 25 nM Tris-HCl, pH 8.0, 5% sucrose and reconcentrated, to exchange the sample buffer and remove excess laurylmaltoside. A summary of this procedure is presented as a flow diagram in Figure 13. 70 FIGURE 13: Purification scheme for rat liver cytochrome g oxidase in laurylmaltoside. 71 FIGURE 13: Rat Liver Cytochrome c Oxidase Purification in Laurylmaltoside Rat liver inner mitochondrial membranes (IMM) (30 mg protein/ml 20% DMSO, frozen -65° ) Thawed, diluted to 10 ml protein/ml with I x TSH Centrifuged, 10,000 xg, 15' 5] P1 once washed IMM) resuspended in 2 x TSH Centrifuged 10,000 xg, 15 min 32 P2 (twice washed IMM) resuspended at 20 mg protein/ml in 5 mM laurylmaltoside,1 x TSH Centrifuged 20,000 x9, 20 min P3 (oxidase enriched membrane fraction) resuspended at 50 mg protein/ml in 0. 5% sodium cholate, 0. 5 x TS SH dialyzed 1:3 vs. 0.5 x TSH, 1h P4 (crude oxidase) resuspended at 40 mg protein/ml in in 1 x TS plus 10 mM KPi, pH 8 and 100 mM laurylmaltoside $5 (soluble oxidase) P5 (insoluble protein) diluted to 4 mg protein/ml and x TS plus 10 mM KPi, pH 8 and 20 mM laurylmaltoside 333/ml hydroxyapatite) equilibrated in 1 Splus 10 mM KPi, pH 8 and 20 mM laurylmaltoside washed with 2. 5 column volumes 1 x TS plus 10 mM KPi, pH 8 and 15 mM laurylmaltoside applied to hydroxyapatite column (6 nmol J’ volumes of KPi, pH 8, 10- 600 mM, in above eluted with linear gradient of 3. 3 column 1’ buffer cytochrome c oxidase fractions with 280/420 nm absorbance i 3.0 dialyzed 1:100 vs 25 mM Tris calcodylate. pH , 5% sucrose overnight, with two changes of buffer applied to horse cytochrome c- Sepharose 48 column (8- 9 nmol 333/ml Sepharose) reduced with 5 «M ascorbate equilibrated with 25 mM Tris cacodylate, pH 8, 51 sucrose washed with 3 column volumes above buffer eluted with linear gradient of 4 column volumes of NaCl, 0-200 mM, in above buffer <———<——-<———— purified cytochrome c oxidase fractions with 280/420 nm absorbance 1 2.0 RESULTS Affinity Chromatography of Solubilized Bovine Heart Mitochondrial Particles. Solubilized bovine heart mitochondria were chromatographed on horse ferrocytochrome g_- Sepharose 4B in laurylmaltoside or Triton X-l00 (Figure l4). Cytochromes 3 and 31 are not significantly retained on the reduced cytochrome 3 matrix, while cytochrome 333 is efficiently bound under the low ionic strength conditions used (u m22.5 mM). After washing and eluting with a salt gradient, the yield of oxidase is 25% when laurylmaltoside is the detergent used (39 nmol heme 3/mg protein)(Lowry), and 20% when Triton X-l00 is used (37 nmol heme 3/mg protein). The results show that both laurylmaltoside and Triton X-l00 are competent in dispersing mitochondrial proteins sufficiently for chromatography, when used at high detergent-to-protein ratios. However, the loss of oxidase during the binding step in Triton X-l00 is consistently at least two fold greater than that which occured in laurylmaltoside under all buffer conditions tested, indicating that Triton X-lOO does not disperse the mitochondrial membrane proteins as well as laurylmaltoside. Tris-cacodylate at 25 mM, pH 7.9, proved to be the best ionic environment for this chromatography in laurylmaltoside (Figure l4A). Bovine heart cytochrome g_oxidase obtained under these conditions has a 280 nm/420 nm ratio of 2.7, and 9.5 nmol heme 3/mg protein. Because of the absorbance of Triton X-l00 at 280 nM, it is necessary to use 72 FIGURE 14: 73 Chromatography of solubilized bovine heart mitochondria on horse ferrocytochrome ggSepharose 4B. (A) Laurylmalto- side-solubilized mitochondria (175 mg protein) were applied to the affinity column (1.5 x 10 cm) that had been reduced with 5 mM ascorbate and equilibrated on the bench top in cold 25 mM Tris—cacodylate, pH 7.9, 5% sucrose, 15 mM laurylmaltoside at a flow rate of 1 column volume/h. The column was washed with 1) 50 ml of the same buffer, 2) 50 ml of 11 mM NaCl in the same buffer, and eluted with 3) 90 ml of a 11-131 mM NaCl gradient in the same buffer. Fractions of 4 ml were collected, and fractions 59-63 were pooled. Protein was estimated by 280-nm absorbance ([3). Concentrations of cytochrome 3 (o) and cytochrome 333 (0) were calculated from difference spectra as described under Experimental Procedures. (B) Triton-solubilized mitochon- dria (145 mg of protein) were applied to the affinity column equilibrated in 45 mM sodium-Bicine, pH 7.9, 5% sucrose, 1% Triton X-100 and were washed and eluted with the NaCl concentrations used in (A). Fractions 54-59 were pooled. Protein was estimated by fluorescamine determina- tion ([3). Cytochrome 3 (o); cytochrome 333 (0); NaCl concentration (---). 74 Se o_ocz Se _ocz m m w m m w _ _ _ _ d .5205: “29.10050 .5205: “29.1095 w 0. 5 o. 5 . n - 0.5 c 5. .I m. E: 0mm 0. wUZ :_pom—omm empmo_com m:_0 A00 cw m< 0 .mmmu_xo uw_cwc:0 xuwcwccc 0:0 upwgmaoxxocvx; c0» :3 mm.o-om.o 0:0 mcowumcmamca ounce cow :5 _._ co m:o_umcu:wocoo pm mmme_xo mo co_uwwuc 0:0 meowmn mczust 00000 any 00 00000 002 mnemoppce—xczwo am asoceca0sc z: 00 000 .0azc :0 00.0 .aoaacocma :5 0.0 .m.0 :0 .apaeamaea 00_mmaoca :5 00 to _E m~.. cw emcsmmms 0cm; Am\mmm.msoc;oow»o _os\m.msoccoouxo _osv mcmaasc c0>occ0u _cawxwz 0 acmwcmoumaocgo 0.03 0000 0mm 0H 0N 00000cca-c asaceca0xc 0.2 0000 0N0 am 000 seaacmapasacec 00_00000x0c0»0 0.0 - 0mm 00 me 00__00 aaocacpxa 000.050 00 umppma umuomcuxm 4.0 - 000 0a 000 000000_ae_0c00_ :0 0 mmcmcDEmE _c_cucozuog_s 0.0 - 000 003 200 c0:e_ 000003 00_30 mE\_oEc Hum x _05: o:_muoc0\m.msmz econezz cm>occ0e u_m_> wmmvwxo quoe mqmum :owp00_c_c:0 __0cm>o .mwcvcocuou_z cm>_0 yam socm wmmuvxo w.msoc200pxu co :o_000_wwc00 .H m0m<3 0mm can _ BDNVHUOSBV FIGURE 16: 8O Hydroxyapatite chromatography of a cytochrome g oxidase enriched fraction from rat liver inner mitochondrial membranes. Cytochrome c oxidase enriched protein (14 nmol 333 at 20 mg protein/mT) (Biuret) obtained by extracting inner mitochondrial membranes with laurylmaltoside as described in Experimental Procedures, was applied to a hydroxyapatite column (0.7 x 27 cm) equilibrated in 1xTS, 10 mM KPi, pH 8, 15 mM laurylmaltoside. The column was washed with 30 ml buffer plus detergent, and eluted by the addition of a 30 ml linear gradient of 10-600 mM KPi, pH 8.0 (---) to the same buffer. Protein was monitored by its absorbance at 280 nm (0). The concentrations of cytochrome 333 (A) and cytochrome 31 (a) were determined from their visible absorption spectra. 81 Figure 16 (W) BlVHdSOHd WOISSVlOd ‘Q "t N (D 1 CD CD l l l l (M) sawoawoouo 0. o M _° - 5.0 T l protein ——-i-—————-r/ 1 1 L 1 l 0. '0. 0. g N .. _ . (“3“ 082) BONVBBOSQV 50 20 ELUTION VOLUME (mls) IO FIGURE 17: 82 Horse cytochrome g affinity chromatography of rat liver cytochrome c oxidase. Cytochrome g_oxidase [164 nmol of 333 at l mg—of protein/ml (Lowry); 8-9 nmol of 333/ml of Sepharose] obtained from hydroxyapatite chromatography of solubilized, fractionated inner mitochondrial membranes was applied to a horse cytochrome g;Sepharose 4B column (1.5 x 21 cm) prepared according to Weiss and Juchs (1978). The column was reduced with 5 mM ascorbate and equilibrated on the bench top in cold 25 mM Tris-cacodylate, pH 7.9, 5% sucrose, 15 mM laurylmaltoside. The column was washed (arrow) with 3 column volumes of buffer, then eluted with a 4 column volume linear gradient of 0-0.2 M NaCl (—--) in the same buffer. Protein (0) and cytochrome oxidase (A) were monitored by their absorbances at 280 and 420 nm, respectively. The 280 nm/420 nm ratio of these absorbances (an indication of the degree of purity) is also shown (:1). 250 0239 20.5: 000 10.00 on. - mo. 83 Figure 17 (W) DON —.0 N6 N V m wuozv/wu OBZ 0. O 01 EDNVQUOSQV 0....“ DISCUSSION Cytochrome c Affinity Chromatography. Horse cytochrome g affinity chromatography has been used to obtain bovine heart cytochrome g oxidase with a purity of 9.5 nmol heme 37mg protein. The most effective chromatographic procedure utilizes the ability of the non—ionic detergent, laurylmaltoside, to disperse the enzyme under conditions of relatively low ionic strength (Rosevear_gt 31., 1980; Ferguson-Miller gt_31., l982). Weiss and Kolb (1979) have previously demonstrated that Triton X-100 provides a good dispersant for cytochrome g_oxidase from 3..gp3§§3, allowing the application of horse cytochrome g affinity chromatography to this enzyme. Our results show that the procedure can also be applied to the bovine heart enzyme when high detergent-to-protein ratios are employed. Other workers (Bill 333_ .31., 1980, 1982) suggest that the previous lack of success of cytochrome_3 affinity chromatography in the purification of the yeast or bovine oxidase, is the result of blockage of the lysine-rich binding site for oxidase on cytochrome g_by cross-linking of the cytochrome to the cyanogen bromide-activated Sepharose via essential lysine residues. They have shown that good purification of oxidase from these sources can be achieved if 3, cerevisiae (yeast) cytochrome g_is linked to thiol-Sepharose 4B via its single sulfhydryl group, which is some distance from the cytochrome 3_oxidase binding domain. However, our data and those of Weiss and coworkers demonstrate that a matrix 84 85 prepared using limited cyanogen bromide activation and horse heart cytochrome g_is equally effective. Weiss and Juchs (l978) show that it is essential to hydrolyze some of the reactive cyanate esters on the resin before adding cytochrome g_in order to avoid multi-point linkages between the protein and the resin that might lead to inaccessibility of cytochrome g_to the oxidase. Thus, the success of the yeast cytochrome 3 method may be accounted for mainly by the low density of cross-links between the thiol resin and cytochrome g, and the high ratio of detergent—to-protein used for solubilization and chromatography. Using the horse cytochrome g_affinity column and laurylmaltoside, a bovine heart cytochrome g_oxidase binding capacity of 0.065 nmol 333 nmol cytochrome gris achieved, greater than that reported by Weiss and Kolb (1979) or Bill ££.El- (1980, 1982). The fact that cytochromes 3_and 3] do not bind to the ascorbate-reduced matrix is convincing evidence that the separation is based on specific interactions, not simple ion exchange or hydrophobic binding (Ozawa gt_31,, T975; Rascati and Parsons, l979), since the 3gl-complex binds reduced cytochrome g_with low affinity (Weiss gt_31,, I978; Ferguson—Miller gt 31., 1979). We conclude that linkage via lysine residues occurs fairly randomly in this procedure, allowing a large proportion of the cytochrome g_to bind correctly with the oxidase. Aside from the economic advantage of using horse rather than yeast cytochrome g, the imino ester linkage of lysine to the resin is much more stable than the disulfide linkage, permitting the long term storage and reuse of the resin. Cytochrome g_affinity chromatography is applied as a final step in the purification of rat liver cytochrome g_oxidase, making maximum use of the capacity of the resin, and resulting in good yield of a highly 86 purified enzyme (routinely 13 nmol heme 37mg protein) from a relatively impure state (t7 nmol heme 3/mg protein). The binding capacity (up to 0.l2 nmol 333/nmol cytochrome g) of the horse cytochrome g—Sepharose affinity column achieved under these circumstances is twice that obtained when the column is used as an initial step in purifying the enzyme from whole mitochondria. Furthermore, a single step purifica- tion as performed by our procedures in laurylmaltoside or Triton X-l00, or as described by Bill gt 31. (l980, l982), gives a less pure enzyme at no higher yield, with much more lipid associated (Azzi gt 31., l982). When affinity chromatography is coupled with hydroxyapatite chromatography for the purification of rat liver oxidase, the enzyme obtained has several unique properties compared to oxidase prepared by conventional techniques (see Chapter 3). Important among these is its lipid-depleted, highly active state. Hydroxyapatite chromatography at low ionic strength appears to be particularly important for removing phospholipid from the enzyme. Evidently the strong interaction of both the protein and the phospholipid with the positively charged calcium ions favors removal of the lipid, but also irreversibly binds some of the enzyme. Addition of l00 mM NaCl to the equilibration and elution buffers gives much higher yield without diminishing the capacity of the hydroxyapatite for oxidase binding (Bernardi, l97l). Under these conditions, the same heme 3_to protein ratios can be obtained in the subsequent affinity chromatography, although considerably more lipid remains associated with the enzyme (5 mol P/mol 333). This is still a low lipid content, and therefore the inclusion of salt in the hydroxyapatite step is recommended for the preparation of oxidase when a highly delipidated enzyme is not required. The purified rat liver 87 enzyme prepared by this procedure is also depleted of subunit 111. Recent results indicate that proteolytic activity present in liver preparations may be important in facilitating the total removal of this peptide, since repeat hydroxyapatite and cytochrome g_affinity chromatography steps were not effective in completely eliminating subunit III from a preparation of enzyme made using fresh mitochondria in which the proteases had been inhibited by several additions by PMSF. The characterization of the rat liver cytochrome g_oxidase produced by this procedure is the subject of Chapter 3. CHAPTER 3 STRUCTURAL AND FUNCTIONAL CHARACTERIZATION OF PURIFIED RAT LIVER CYTOCHROME g_OXIDASE Most generalizations concerning the structure and activity of cytochrome 3 oxidase have been derived from studies on the bovine heart enzyme prepared by classical salting out techniques. It is therefore important to investigate other eukaryotic oxidases, prepared by different purification methods, in order to determine whether the results obtained for the bovine enzyme represent the general case for the isolated protein, and to establish the ways in which these properties relate to enzyme in the native state. Rat liver cytochrome g_oxidase has been prepared by ion exchange and cytochrome g affinity chromatography in laurylmaltoside. The purified enzyme has been analyzed in terms of subunit composition, lipid content, molecular weight, and ability to exhibit respiratory control and proton pumping after reconstitution into phospholipid vesicles. The kinetic profile of the purified enzyme has been compared to that of oxidase in the native membrane, the detergent-solubilized membrane, and soybean phospholipid vesicles. The data clarify the roles of lipid, subunit 111, state of association, and assay conditions in determining the activity of cytochrome g oxidase. 88 EXPERIMENTAL PROCEDURES Chemicals. Sodium dodecyl sulfate (m70% C12; Sigma, St. Louis, MO) used for SDS-PAGE according to Laemmli (1970) was treated with activated charcoal, then recrystallized two times from 95% ethanol. Sodium dodecyl sulfate (398% C12; Pierce Chemical, Rockford, IL) for SDS-PAGE according to Merle and Kadenbach (1980) was used as purchased. Soybean phospholipid (L-a-phosphatidyl choline, Type II-S, Sigma) used to reconstitute cytochrome 3_oxidase for proton pumping experiments was purified according to a modification of the procedure of Kagawa and Racker (1971): phospholipid (25 g) was dissolved in 500 ml dry acetone containing 1 mM DTT and stirred for 2 days at room temperature. The insoluble fraction was collected by filtration and dissolved in 100 ml dry diethylether, then 50 ml of the solution was added with stirring to 450 ml of dry acetone plus 1 mM DTT. The lipids were allowed to precipitate, the solution was cooled on ice, and the precipitate was collected by filtration. The precipitation process was repeated, the lipids were dissolved in diethylether, and the solution was centrifuged for 5 min at ~2000 xg (small pellet discarded). The supernatant was evaporated to dryness using a rotary evaporator, and the lipids were stored under nitrogen at -20°C. Other chemicals were the best grades available from the sources indicated here, or in Chapters 1 and 2: acrylamide and methylenebis(acrylamide) (Bio—Rad 89 9O Laboratories, Richmond, CA or Serva, Garden City Park, NY); protein molecular weight standards (Bethesda Research Laboratories, Rockville, MD, and Sigma); Hepes, choline chloride, Phenol Red, carbonyl cyanide m-chlorophenylhydrazone, valinomycin, lipid standards (Sigma); N,N'-dicyclohexyl [14-C] carbodiimide (CEA, Research Products International, Mount Prospect, IL); fluorescamine (Roche Diagonostics, Nutley, NJ); 2,5—diphenyloxazole (New England Nuclear, Boston, MA); potassium cyanide, potassium sulfate, and reagents for phospholipd extractions and analysis (Mallinckrodt, Paris, KY); urea (ultrapure: Schwarz/Mann, Cambridge, MA); Trizma base and glycine for SDS-PAGE according to Merle and Kadenbach (1980) (Merck, Darmstadt, FRG). Spectral Measurements. UV and visible spectra were recorded and analyzed as indicated in Chapters 1 and 2. Assay Methods. Electron transfer activity of cytochrome g_oxidase was measured polarographically as described in Chapter 1, using the buffer and detergent conditions detailed in the figure legends. Rates of oxidation of ferrocytochrome g_(prepared as in Chapter 1) were measured spectrally by monitoring the 416-410 nm absorbance with an Aminco DW 2a spectrophotometer, under the assay conditions indicated in the legend to Figure 26. Where indicated, affinity purified rat liver cytochrome g_oxidase was reconstituted into preformed phospholipid vesicles essentially according to the dialysis procedure of Casey gt 31. (1979), with the modifications that the cholate concentration was 2% and the enzyme to phospholipid ratios were increased to those given in the figure legends. 91 The proton translocating activity of purified cytochrome_3 oxidase reconstituted into phospholipid vesicles was measured either spectro- photometrically using the pH indicator dye Phenol Red, or using a Radiometer PHM 64 pH meter equipped with Philips CA 14/02 combination pH electrode according to the procedure of Casey gt_31, (1979); the amount of valinomycin in the assay was increased to 4.44 pg. The pH electrode was etched in 50% HF for 8 s, then equilibrated at least 24 h in assay media before use, and was rinsed between measurements with a cycle of: water, ethanol, CHCl3zMe0H (2:1) or diethyl ether, ethanol, water. Purified soybean phospholipids were sonicated into vesicles under nitrogen, and cytochrome_3 oxidase was incorporated using the dialysis procedure described, except that Hepes was omitted from the last dialysis buffer. Bovine heart oxidase vesicles were prepared at 40 mg asolectin/ml, 1% cholate, and 7.5 uM cytochrome 333 (RCR'MS); rat liver oxidase vesicles were 40 mg asolectin/ml, 2% cholate, and 3.8 uM cytochrome 333 (RCR'vIO). Reactions were initiated by the addition of a solution of ferrocytochrome g (carefully adjusted to the pH of the assay media) sufficient for 1 to 3 turnovers of the oxidase. The pH changes due to cytochrome g_oxidase activity were calibrated with respect to additions of 0.5 mM oxalic acid. Molecular Weight Determination by Gel Filtration. Molecular sieve chromatography of purified rat liver oxidase (0.1 ml of 7 uM 333) was performed on a Sephacryl 300 column (0.7 x 43 cm) in 100 mM KCl, 10 mM Tris-HCl, pH 7.8, and 1 mM EDTA, at a flow rate of 5 ml/h. Laurylmal- toside was 0.04% during the chromatography of the oxidase and during 92 the calibration procedure. The column was calibrated with the follow- ing molecular weight standards: Blue Dextran, 2,000,000; thyroglob- ulin, 669,000; ferritin, 440,000; catalase, 250,000; aldolase, 158,000; alcohol dehydrogenase, 140,000; hemoglobin 69,000; bovine serum albumin, 68,000; ovalbumin, 43,000; cytochrome 3, 12,500. Phospholipid Analysis. Phospholipids were extracted from purified rat liver cytochrome g_oxidase (6 mg of protein in 25 mM Tris-HCl, pH 8.0, 5% sucrose, plus laurylmaltoside) with chloroform: methanol (2:1), diethylether ethanol (1:1); and chloroform: methanol: ammonia (100:50:2) according to the procedure of Awasthi gt_31, (1971). In some cases, when laurylmaltoside in the extract made thin-layer chromatography difficult, the lyophilized oxidase was pre-extracted with anhydrous acetone, which removes detergent but no detectable phosphorus from the oxidase (Awasthi gt_31,, 1971). An aliquot of the lipid extract (250 pl; 0.5 ug of P) was chromatographed in two dimensions in the solvent system of Parsons and Patton (1967) on Supelco Redi-Coat 20 precoated TLC plates (0.25 mm). The plates were pre-run in methanol—methylene chloride, 1:1, and activated by heating at 120°C for 1 h. Lipids were visualized by spraying with potassium dichromate (0.6%) in sulfuric acid (55% by weight) and charring at 180°C for 20-30 min, and were identified by comparison to standards. Total extractable phosphorus was quantitated by using the procedure of Bartlett (1959) or Ames (1966). Total phosphorus present in the oxidase was also determined without extraction by directly ashing the protein and measuring phosphorus by the procedure of Ames (1966). 93 SDS-Polyacrylamide Gel Electrophoresis. The subunit composition of purified cytochrome g_oxidase was routinely examined by a modification of the Laemmli (1970) discontinuous SDS-polyacrylamide gel electro- phoresis system using a Bio-Rad vertical slab gel electrophoresis apparatus. The separating gel (1.5 x 100 x 140 mm) contained 0.2% SDS and gradients of 10-22% acrylamide and 5-15% sucrose. Purified cytochrome_3 oxidase was precipitated in 10 volumes of cold 5%-trichloroacetic acid (total time <5 min) to remove it from bulk laurylmaltoside which interferes with SDS binding. The pelleted oxidase was resuspended in 5% SDS, 5% B-mercaptoethanol, 5% glycerol in sample buffer, dissociated by sonication at 10-15°C for at least 30 min in a bath type sonicator, electrOphoresed for 1 h at 15 mA, and then for ~3.5 h at 20-25 mA, or until the dye front migrated off the gel. Peptides were visualized by scanning the absorption of the unstained gel at 280 nm, Coomassie Blue staining, fluorescamine labeling (Ragland gt__1,, 1974), and silver staining (Wray gt_31,, 1981). The high resolution SDS-polyacrylamide gel electrophoresis method of Merle and Kadenbach (1980) was also used in order to more clearly separate and identify the small molecular weight peptides of cytochrome g oxidase; the enzyme was again precipitated with trichloroacetic acid. The electrophoresis apparatus and gel dimensions used for this procedure were those described above. DCCD Labeling of Qytochrome c Oxidase. N,N'-dicyclohexyl [14-CJ-carbodiimide in pentane (54 mCi/mM; 0.5 mCi/mol) was added to a rapidly vortexing solution of purified cytochrome g oxidase (150 ug in 0.1 ml; 140 nmol DCCD/nmol 333) in 0.5 x TS, pH 8, 15 mM 94 laurylmaltoside. The reaction mixture was incubated for 24 h at 4°C, precipitated with 5% trichloroacetic acid, and electrophoresed according to the modification of the Laemmli (1970) procedure described above. Sample lanes were cut into 2 mm slices, incubated overnight in closed vials with 0.5 ml 30% H202 at 60°C, mixed with 5 ml scintillation fluid (Toluene, 2000 ml; Triton X-100, 1000 ml; PPO, 21 gm), and the radioactivity counted with a Beckman model LS7000 scintillation counter. pH Dependence of Cytochrome a Redox Potential. The Em of cytochrome 3_in purified cytochrome g_oxidase was determined essentially according to Penttila and Wikstrom (1981). Cytochrome g oxidase (1.3 uM) in 20 mM K2S04, 20 mM KPi, pH 7.2, and 20 mM laurylmaltoside was incubated overnight with 6 mM KCN at 4°C. The redox state of cytochrome 3 was adjusted to approximately half-reduced by addition of 77 mM potassium ferrocyanide and 0.333 mM potassium ferricyanide. The pH was varied from 6.5-8.0 by additions of NH4OH or CH3C00H, and the redox state monitored by the absorbance difference at 604-630 nm. The values of Em were calculated by determining the redox state of cytochrome 3_spectrally, and using the Nernst equation and the standard redox potential of the ferri/ferrocyanide redox couple under these conditions (E'o = 422 mV). The E'0 of ferri/ferrocyanide was determined using a Pt/Ag-AgCl electrode calibrated with a saturated solution of quinhydrone (E'0=297mV) (Clark, 1972). RESULTS Physical and Spectral Properties of Rat Liver Cytochrome c Oxidase. Figure 18 shows the visible absorption spectra of purified rat liver cytochrome_3 oxidase. The absorption maxima for the reduced form of the enzyme are at 444 nm and 605 nm; the oxidized fonn absorbs maximally at 420 nm. These spectra are characteristic of purified cytochrome g_oxidase, indicating that laurylmaltoside does not interact with the protein in a way that is detrimental to the chromophore. There is no evidence of cytochromes 3 and 31 contamination, which, if present, would absorb strongly in the Soret region, and at 562 nm and 553 nm, respectively. Gel filtration of purified rat liver cytochrome_3 oxidase on Sephacryl 300 revealed the presence of a single 280 nm absorbing species with an apparent molecular weight of 3l0,000 1 30,000, before correction is made for the contribution of the laurylmaltoside micelle to the Stokes radius of the protein (see discussion) (Figure 19). Total lipid phosphorus content of the oxidase is 3.0 mol P/mol cytochrome 333, as determined by analysis of the total lipid extract of the enzyme (Awasthi gt_31,, l97l). Phosphorus analysis performed directly on the unextracted protein (Ames, l966) is in agreement, revealing 3.l mol P/mol cytochrome a . Cardiolipin (DPG) and phosphatidyl inositol (PI) were the only phospholipids detected by two-dimensional thin-layer chromatography of the total lipid extract 95 96 FIGURE 18: Visible absorption spectra of rat liver cytochrome g oxidase purified by affinity chromatography. (A) Air oxidized, (B) dithionite reduced spectra of cytochrome 333 (0.73 nM, 12.5 nmol heme 3/mg protein in 0.15M KPi, pH 8, plus laurylmaltoside) measured as in Experimental Procedures. 97 Figure 18 - 03 4.02 J 01 600 (nm) 500 WAVE LENGT H 400 .15- E! 3 DNVSHOSGV .05 _ FIGURE 19: 98 Gel filtration analysis of the apparent molecular weight of purified rat liver cytochrome g_oxidase—laurylmaltoside complex. Cytochrome g oxidase (0.1 ml, 7 uM cytochrome 333) was applied to Sephacryl—300 (0.7 x 43 cm) equilibrated in 100 mM KCl, 10 mM Tris—HCl, pH 8, 1 mM EDTA, 0.04% laurylmaltoside, and eluted with the same buffer in 0.5 ml fractions at 5 ml/h. The elution profile was monitored by 280 nm absorbance (shown in inset) and the volume corresponding to the peak absorbance is indicated by the bar. Calibration was performed under the same conditions using Blue Dextran and standard proteins (0) as described in Experimental Procedures. 99 1 1 T I I 002 4_ x lOOO- 8 N < 2_ 1‘3 (7,0300- Elui ion volume '— (mls) X l.— I 0 E 2100 - a: < —l D U 3 ‘2’ 30- J 1 l J J l 8 12 16 ELUTION VOLUME (mls) Figure 19 100 (Figure 20), indicating that at most l mole DPG and l mole P1 are associated per mole cytochrome g_oxidase. Figure 21A shows the Coomassie Blue stained gels indicating the subunit composition of purified rat liver oxidase compared to bovine heart enzyme prepared by the method of Suarez gt_31, (1984). The bovine heart enzyme is resolved into seven major subunits in the Laemmli (1970) discontinuous SDS-polyacrylamide gel system. The rat protein contains only six major subunits, in the presence or absence of up to 6 M urea, and is missing subunit III of apparent molecular weight 22,000 (30,000 as determined by DNA sequence estimates (Thalenfeld and Tzagoloff, l980)) that is present in the bovine heart enzyme. There is little or no aggregated material at the top of the gel or high molecular weight species that might suggest that lack of subunit III is the result of poor entry of this hydrophobic subunit into the gel, or aggregation and subsequent anomalous migration. Fluorescence labeling prior to electrophoresis also shows that the rat liver cytochrome g oxidase has been depleted of subunit III (Figure 21B). (The slightly changed migration of some of the small peptides in both the bovine heart and rat liver enzymes results from changes in charge caused by reaction with this reagent, indicating that the binding of SDS does not completely govern the mobility of these peptides). A similar array of subunits is observed by staining the gel for protein using silver or by scanning the absorbance of the unstained gel at 280 nm (not shown). The purified enzymes were also reacted with N,N'-dicycl0hexyl- [14CJ-carbodiimide (DCCD), then electrophoresed (Figure 22). Under conditions where subunit III of bovine heart cytochrome g_0xidase was clearly labeled, the corresponding molecular weight region in the gel FIGURE 20: 101 Two dimensional thin layer chromatography of phospholipids extracted from purified rat liver cytochrome g_oxidase. Cytochrome 3 oxidase (40 nmol 333, 13 nmol heme 3/mg protein) was extracted with the organic solvent systems described in Experimental Procedures, and an aliquot (250 pl, 0.5 pg P) (Ames, 1966) was chromatographed in two dimensions on silica gel (0.25 mm): first dimension (horizontal) chloroform:methanol:waterzaqueous ammonia (130:70:8:0.5); second dimension (vertical) chloro- form:acetone:methanol:acetic acid:water (100:40:20:20:10). The components were visualized by potassium dichromate- sulfuric acid charring. 0, origin; DPG, cardiolipin; PI, phosphatidyl inositol; LM, laurylmaltoside; Mlt, maltose. Mlt near the origin results from the degradation of lauryl- maltoside during the total lipid extraction. Mlt on the second dimension origin line is produced by acidic hydrolysis of laurylmaltoside in the second solvent system. Laurylalcohol migrates very near the solvent fronts (not shown). Mlt 3 .'= I Figure 20 FIGURE 21: 103 Subunit composition of cytochrome g oxidase as revealed by SDS-polyacrylamide gel electrophoresis using the discon- tinuous buffer system of Laemmli (1970). (A) Coomassie blue stained cytochrome g_oxidase. Rat liver oxidase (left lane) (13.5 nmol heme 37mg protein) was purified in lauryl— maltoside according to the method described in Experimental Procedures. Bovine heart oxidase (right lane) (9.0 nmol heme 3/mg protein) was purified according to the procedure of Suarez et al. (1984). Samples were precipitated, dissociated and electrophoresed as indicated in Experi- mental Procedures. The sample wells, stacking gel (6% acrylamide), and separating gel (IO-22% acrylamide gradient) are shown. Numbering of the subunits (I-VII) is according to Downer et al. (1976). Molecular weight designations are those of standard proteins electrophoresed on the same gel: ovalbumin, 43,000; a-chymotrypsinogen, 25,700; B-lactoglobulin, 18,400; lysozyme, 14,300; cytochrome 3, 12,300; bovine trypsin inhibitor, 6,200; insulin (3), 3,000. (B) Fluorescamine-labeled rat liver oxidase (left lane) and bovine heart oxidase (right lane). The precipitated enzyme was labeled with fluorescamine, dissociated, electrophoresed on acrylamide gel, then photographed (unfixed) in UV light. RL 8H Fx’L BH VII Figure 21 FIGURE 22: 105 DCCD binding to cytochrome g oxidase. Affinity purified rat liver oxidase (A) or bovine heart oxidase purified according to Suarez gt_31, (1984) (B) in 0.5 x TSH, pH 8, 15 mM laurylmaltoside was incubated with N,N'-dicyclohexyl [14-C]-carbodiimide (140 nmol DCCD/nmol 333) for 24 h at 4°C, as described under Experimental Procedures. Samples were prepared and electrophoresed as in the legend to Figure 24, and the radioactivity in gel slices determined by scintillation counting. R1 l l l l I l l T soo~ _ A IL 300— - U 100- - I l I _L 4 . . *1 l l l l I I l l soo~ B E 300— ~ U ioo~ - I_‘J¥7- 0L__0 , ,1, . _L 7 7. I ,,, 4;,— Figure 22 107 of the rat liver enzyme contained only background levels of radioactivity. The smaller molecular weight peptides of both enzymes were also labeled by this reagent under the conditions used, as previously observed (Casey gt_31,, 1980; Prochaska gt_31,, 1981). The apparently greater labeling of subunits IV and VI in the rat liver cytochrome g_oxidase is due in part to the fact that radioactivity was added to the samples on a per heme basis, and since the heme-to-protein ratio of the rat liver enzyme was greater, this resulted in the use of a higher ratio of radioactivity-to-protein in the rat liver enzyme. When the reconstituted enzyme was labeled with DCCD, no radioactivity was present in these subunits (R. Casey, personal communication). The consistent lack of visualization of a 22,000 molecular weight peptide in the rat liver enzyme by any of these independent methods of detection indicates that its absence from the resolved peptides is not simply due to its failure to react in a particular staining procedure. The Merle and Kadenbach (1980) system of electrophoresis permits a more accurate comparison of the cytoplasmically synthesized subunits of rat liver and bovine heart cytochrome 3_oxidase (Figure 23). Based on these results, and by comparison to rat liver enzyme purified by ammonium sulfate precipitation and electrophoresed by Merle and Kadenbach (1980; Kadenbach and Merle, 1981), it can be seen that the rat liver enzyme affinity purified in laurylmaltoside has lost one "subunit V" region peptide and two "subunit VI" region peptides. The missing subunits correspond to peptides a, b, and c in the system of Downer gt 31. (1976) and are not considered to be true subunits by Capaldi and coworkers (Capaldi, 1982; Capaldi gt_31,, 1983). However, FIGURE 23. 108 Subunit composition of cytochrome g oxidase analyzed by the highly resolving SDS-polyacrylamide gel electrophoresis system of Merle and Kadenbach (1980). Samples of rat liver oxidase (right lane) (11.5 nmol heme 37mg protein) and bovine heart oxidase (left lane) (9.0 nmol heme 3/mg protein) were prepared and precipitated with trichloro— acetic acid (5%) as in the legend to Figure 21. The numbering of the subunits is according to the nomenclature of Kadenbach and Merle (1983). Note that this sample of rat liver oxidase is of lower purity (11.5 nmol heme 3/mg protein) than that routinely prepared by affinity chroma- tography in laurylmaltoside (12.5-13.5 nmol heme 3/mg protein), and is not entirely depleted of subunit III (see Discussion). Figure 23 110 Merle and Kadenbach (1982) have proposed that the cytoplasmically synthesized peptides of cytochrome 3_oxidase, specifically VIa, VIIa, and VIII, may influence the overall activity of the enzyme in an organ specific fashion, and thus, the loss of peptide b (subunit VIa) from the purified rat liver enzyme is of interest. Subunit VIII was not clearly resolved in either sample shown in Figure 23, and no conclusions can be drawn regarding the effect of the purification on its presence. (Note: the sample used for this analysis was of significantly lower purity than routinely prepared and still retains a small amount of subunit III; this gel system is no better than the Laemmli at resolving this peptide). Kinetic Properties of Rat Liver Cytochrome c Oxidase. The kinetics of purified rat liver cytochrome g oxidase and oxidase reconstituted into phospholipid vesicles, were measured polarographically with horse cytochrome 3_and compared to those of oxidase in the native membrane and in laurylmaltoside-solubilized membranes (Figures 24 and 25). When assayed at relatively low ionic strength and high pH (25 mM Tris-cacodylate, pH 7.9), a clearly biphasic plot is obtained for the enzyme in all cases (Figure 24). Under these buffer conditions, it appears that the major results of purification and delipidation are an overall decrease in catalytic activity, and an increase in the Km for cytochrome 3_in the second kinetic phase. Lipid repletion by reconstitition into vesicles significantly restores the activity of the purified oxidase to a level closer to that of oxidase in solubilized membranes, and decreases the Km value of the second kinetic phase. FIGURE 24: 111 Eadie-Hofstee plots of the kinetics of oxidation of cyto- chrome g,by rat liver cytochrome 3_oxidase in 25 mM Tris-cacodylate, pH 7.9. Affinity purified cytochrome_3 oxidase (0.025 nmol) assayed in the presence of 0.25 mM laurylmaltoside (A). Purified reconstituted cytochrome c oxidase (0.015 nmol) in phospholipid vesicles (0.088 nmol 333/mg phospholipid) assayed in the presence of 5.6 uM CCCP and 1.1 uM valinomycin (0). Oxygen consumption was measured polarographically as indicated under Experimental Procedures. The range of cytochrome g_concentrations used was 0.02-40 nM. The kinetic plots for intact mitochon- drial membranes ( ) and laurylmaltoside solubilized membranes (---) from Figure 5 are shown for comparison. TN /[81 x IO’B (M"s") 112 IOO 200 TN (5") Figure 24 l“ 300 FIGURE 25: 113 Eadie-Hofstee plots of the kinetics of oxidation of cyto- chrome g_by rat liver cytochrome g_oxidase in 50 mM KPi, pH 6.5. Affinity purified cytochrome g oxidase (0.029 nmol) assayed in the presence of 0.25 mM laurylmaltoside (O). Purified, reconstituted cytochrome g oxidase (0.022 nmol) in phospholipid vesicles (0.088 nmol 333/mg phospholipid) assayed in the presence of 5.6 uM CCCP and 1.1 uM valino- mycin (A). Oxygen consumption was measured polarographi- cally as indicated under Experimental Procedures. The range of cytochrome_3 concentrations used was 0.04—50 uM. The kinetic plots for intact mitochondrial membranes (———) and laurylmaltoside solubilized membranes (——-) from Figure 4 are shown for comparison. TN /[SI x 107(M's7') N O 5 30 114 560 TN (s-I) Figure 25 115 At higher ionic strength and lower pH (50 mM KPi, pH 6.5) (Figure 25), non—linear kinetics are also observed, but the affinity of cytochrome g_f0r the enzyme is lower (note the scale is increased by a factor of ten compared to Figure 25). The overall activity of the purified enzyme is much higher under these conditions, and is less affected by reconstitution into phospholipid vesicles. When compared to the activity of the membrane-bound cytochrome g oxidase in this assay system, the enzyme appears not to have changed greatly during purification (membrane bound oxidase: Kml = 3 x 10‘7 M, sz = 1 x 10‘5 M, TN = 350 5'1; purified oxidase: Kml = 2 x 10-7M, sz = 3 x 10-6M, TN = 400 s-1). Figure 26 shows a comparison of the kinetics of purified rat liver cytochrome g_oxidase as determined by spectral and polarographic methods in 50 mM KPi, pH 6.5. When initial rates of reaction with reduced cytochrome g_are measured in the spectral assay, a single kinetic phase results which has an affinity and turnover number that closely corresponds to the low affinity reaction observed in the polarographic assay. Energy Coppling Properties of Rat Liver Cytochrome c Oxidase. Figure 27 shows a pH titration of the cytochrome 3 redox potential in rat liver cytochrome g_oxidase thoroughly depleted of subunit III, and bovine heart enzyme containing normal amounts of subunit III. Although cytochrome 3_in the rat liver enzyme has an Em about 10 mV more positive than that in bovine heart, the magnitude of the redox potential pH dependence is the same for both forms of the enzyme, and varies less than 10 mV over the pH range from 6.5-8.0. A redox FIGURE 26: 116 A comparison of the Eadie-Hofstee kinetics of the rat liver cytochrome g_oxidase reaction, measured polarographically as rates of oxygen consumption (A), or spectrally as rates of ferrocytochrome 3 oxidation (0). Purified rat liver oxidase was assayed in 50 mM KPi, pH 6.5, 0.25 mM lauryl- maltoside. Rates of oxygen consumption were measured as described in Experimental Procedures, with 0.048 nmol cytochrome 333 in the assay, and a range of cytochrome_3 concentrations, 0.06-50 uM. Rates of oxidation of reduced cytochrome 3_were measured by monitqring the absorbance change at 416-410 nm with 2.3 x 10' nmol cytochrome 333 and a range of cytochrome 3 concentrations, 0.04-15 11M. TN/ [3] x 10"(1vl'5') ES (I 117 0 Spectral Kinetics A Polorographlc Kinetics Figure 26 FIGURE 27: 118 pH dependence of the redox potential of cytochrome 3 in cytochrome g_oxidase with and without subunit III. Cyanide inhibited, purified rat liver oxidase (subunit III depleted; 13.0 nmol heme 37mg protein) (A) and bovine heart oxidase (subunit III containing; 9.0 nmol heme 3/mg protein prepared according to Suarez et al. ,1984) (0) were equili- brated with the redox couple, potassium ferro/ferricyanide under the conditions described in Experimental Procedures. The state of reduction of cytochrome 3 was monitored spectrally at 603—630 nm as the pH was varied by additions of NH4 OH or CH COOH, and was used in the Nernst equation to caIculate the Em of cytochrome 3_at each pH. (mV) Em 119 310 - 300 - 290 *' ] l _ 7.0 7.5 pH Figure 27 mM.-4... -.my..-~_v-Ise.: .-.-..'-.‘.1-. ~. .3:- ...-.-. -. . .. 120 potential vs. pH variability of this magnitude is too small to directly implicate cytochrome 3_in the process of coupling an electron transfer event to proton movement. These findings are in partial agreement with those of Penttila and Wikstrom (1981) who observe a similar, very small pH dependence for the Em of cytochrome 3 in subunit III depleted enzyme, but find a 30 mV/pH unit dependence for the subunit III containing enzyme. They regard this as significant evidence for the participation of this redox center and subunit III in the proton pumping function of cytochrome g oxidase. The respiratory control of purified rat liver cytochrome 3_0xidase reconstituted into phospholipid vesicles (Figure 28) was determined by taking the ratio of the rate of oxygen consumption after addition of uncoupler plus ionophore, to the rate obtained in their absence. The very initial activity of the coupled enzyme respiring on cytochrome g is stimulated up to ten-fold by release of the electrochemical gradient across the bilayer. The level of apparent respiratory control achieved with the rat liver enzyme was very dependent on the phospholipid-to- protein ratios used in the reconstitution. A range of 0.05-0.20 nmol cytochrome 333/mg phospholipid (at 2% cholate) resulted in levels of respiratory control from 4-10, with higher ratios of phospholipid yielding better respiratory control. These results clearly show that cytochrome g_oxidase depleted of subunit III is able to develop and respond to a proton gradient resulting from the oxidation of ferrocyto- chrome 3, However, this data does not permit the distinction between control exerted by a pH gradient formed simply by asymmetric utiliza- tion of protons to form water in the interior of the vesicles, and that _.-'-‘~ FIGURE 28: 121 Respiratory control of affinity—purified rat liver cyto- chrome g_oxidase reconstituted into asolectin vesicles. Phospholipid vesicles containing cytochrome g oxidase (0.035 nmol 333; 0.088 nmol aa /mg phospholipid) were prepared by dialysis as descFTged in Experimental Procedures. Oxygen consumption was measured polarograph- ically in 1.75 ml of 40 mM KCl, 10 mM HEPES, pH 7.2, 0.1 mM EDTA, 5.6 mM Ascorbate and 0.28 mM TMPD. Cytochrome g (0.025 umol); carbonyl cyanide m-chlorophenylhydrazone (5.6 AM), and valinomycin (1.1 uM) were added as indicated. 122 MINUTES Figure 28 123 resulting from active proton extrusion from the vesicle interior due to the action of a proton pump. Therefore, it was necessary to try to determine if extruded protons could be detected as an acidification of the external medium. pH changes in the media surrounding cytochrome g oxidase vesicles prepared from rat liver and bovine heart enzymes were measured and the results are shown in Figure 29. Addition of ferrocy- tochrome g_to the intact bovine heart enzyme gives an apparent acidifi- cation that can be calculated to represent 0.4-0.6 Hi/e' under the conditions used (before correction for a short fall of 0-0.2 Hi/e'; see Discussion); rat liver oxidase produces smaller acidifications of 0.1-0.3 H+/e‘ and requires greater correction (a short fall of 0.1-0.4 HT/e'). Less enzyme and cytochrome g were present in the rat liver assay, which would be expected to give less acificiation, and may be a factor in the lower apparent stoichiometries (see Discussion). This lower stoichiometry is not due to an uncoupling effect caused by laurylmaltoside remaining in the vesicles, since this would cause greater leakiness of the vesicles and would be observed as a faster rate of alkalinization after the acidification pulse. In fact, the alkalinization rate is slow and corresponds to that observed with the bovine heart enzyme. The reduced efficiency is also not the result of randomly oriented rat liver enzyme, since almost 100% of this enzyme reincorporates into the phospholipid vesicles with the native mitochon- drial orientation. This was determined by evaluating the ability of membrane impermeable reductants to reduce the reconstituted enzyme (P. O'Shea, personal communication). Unfortunately for the purposes of drawing definite conclusions, the rat liver oxidase (11.5 nmol heme 3/mg protein) used in this experiment still contained a small amount of FIGURE 29: 124 Bulk phase pH changes resulting from the oxidation of ferrrocytochrome 3_by reconstituted cytochrome c oxidase in phospholipid vesicles. The assay media (4.0 ml) contained 75 mM choline chloride, 25 mM KCl, pH 7.4, 4.4 ug valino- mycin, and 0.1 ml bovine heart oxidase vesicles (7.5 uM 333) or rat liver oxidase vesicles (3.8 AM 333). pH vs. time was monitored with a glass combination electrode at 22°C, as described in Experimental Procedures. (———) pH changes initiated by the addition of the indicated amounts of ferrocytochrome g (2.14 mM). (---) pH changes initiated by ferrocytochrome 3_after the addition of 1.0 ug CCCP to the assay media. BOVINE HEART T r’ I : 3 "MOI H... l/ f 1‘ I 6.4 "mop; I “YI- E '1 111 y ’3" \\ l,’ l0 s P—-—-—-4 RAT [IVER ,, T I I I 3nmoIH+ I 11 Figure 29 126 subunit 111, but not enough to be measurable by radioactive labeling with DCCD (R. Casey, personal communication). Furthermore, there was no DCCD labeling in the region of subunits IV and VI that could be construed as indicating the presence of proteolytic fragments of subunit III (Malatesta gt 31., 1983). DISCUSSION Association of Phospholipid with Cytochrome c Oxidase. The rat liver cytochrome 3_oxidase prepared by the method described is consistently highly active (TN = 270-4005’1) and very low in lipid (1 mol DPG and 1 mol PI/mol oxidase (333)). Other methods of purification of the rat liver enzyme yield oxidase with 8-9.5 mol phospholipid/mol enzyme (Hochli and Hackenbrock, l978; Nagasawa gt_31,, l979). Cytochrome g oxidase prepared from bovine heart mitochondria by classical procedures usually contains 25 to 75 mol phospholipid/mol enzyme (Brierley and Merola, T962; Yu gt_31,, T979; Robinson and Capaldi, l977). A number of methods have been developed to remove phospholipid from purified bovine heart oxidase, including repeated ammonium sulfate precipitation in cholate (Tzagoloff and MacLennon, 1965; Awasthi gt 31. 1971; Yu gt_31,, l975), ethanol extraction of the crude enzyme (Hartzell and Beinert, 1974), chromatography in high concentrations of Triton X—100 (Robinson and Capaldi, 1977; Fry and Green, 1980), and glycerol gradient centrifugation in high Triton (Robinson gt 31,, 1980). A minimum level of 0.5-2 mol phospholipid/mol enzyme is obtained by these techniques, but the resulting enzyme is usually relatively inactive (TN 5_40-l00 S'I). Analysis of the bound phospholipid shows that cardiolipin remains tenaciously associated with the oxidase. Because its depletion results in a significant loss of activity, and adding it back gives higher 127 128 activities than obtained with other individual phospholipids, some researchers have proposed that cardiolipin is essential for the function of cytochrome g oxidase (Awasthi gt_31,, 1971; Robinson gt 31,, 1980; Fry and Green, 1980; Vik gt_31,, 1981; Robinson, 1982). In contrast, Watts gt_31, (1978) reported recovery of activity after complete substitution of cardiolipin with phosphatidyl choline in yeast cytochrome g oxidase. However, the low activities obtained in all these lipid repletion experiments make it difficult to draw definite conclusions regarding the precise role of cardiolipin. The rat liver oxidase containing only one mole of cardiolipin, but retaining high activity and responsiveness to added phospholipid may thus prove to be a useful tool for further investigation of lipid requirement. In fact, we have obtained rat liver oxidase with 0.75 mol DPG/mol 333 by subjecting the enzyme to a repeat affinity chromatography step, without loss of activity (TN = 4005'1) (T. Carlson, unpublished data). These findings indicate that a stoichiometric association of cardiolipin with oxidase may not be a specific requirement for electron transfer activity. Kinetic Changes Related to Purification. Rat liver cytochrome g oxidase containing 1 mol of DPG/mol 333 retains the ability to interact with cytochrome g with two kinetically distinct affinities similar to those observed with the membrane-bound enzyme, when assayed under maximal turnover conditions in 50 mM KPi, pH 6.5. When assayed in 25 mM Tris-cacodylate, pH 7.9, conditions designed to favor tight binding of cytochrome 3_to oxidase, the overall activity of the enzyme and the contribution of both kinetic phases is considerably diminished 129 compared to the native enzyme. Vik gt 31. (1981) have previously reported a differential decrease in the magnitude of the low affinity phase upon lipid depletion of bovine heart oxidase and a concomitant loss of cytochrome g_binding. The hypothesis formulated by these authors (Bisson gt_31,, T980; Vik gt_31., l98l; Fuller _t__1., l98l), that a second cardiolipin molecule tightly associated with the oxidase is essential for, and the major component of, the low affinity binding site for cytochrome g_on the enzyme, is not in accord with these observations of a well-defined second phase in the kinetic analysis performed in phosphate buffer, nor with the observation that both kinetic phases in cacodylate are greatly diminished in turnover number in response to lipid depletion. However, the increase in apparent Km of the second kinetic phase could relate to lower affinity binding of a second cytochrome g. This interpretation of the kinetics must be qualified, however, since the Km values cannot be unequivocally determined from these complex plots. The lower turnover of the purified enzyme in cacodylate may reflect decreased accessibility of the bound cytochrome g_to the lipophilic reducing agent, TMPD, in the delipidated enzyme, since in 50 mM phosphate the reduction of cytochrome g_by TMPD occurs mainly in solution (Ferguson-Miller gt 31. l978; Brautigan gt 31, l978c). This postulate is supported by the observation that oxidase reconstituted into phospholipid vesicles has greatly increased turnover numbers compared to the purified enzyme when assayed in cacodylate, but not in phosphate. An effect on TMPD reduction rates alone, however, cannot account for the activity loss during purification observed in these studies or by others, since it has been demonstrated that electron 130 transfer between hemes 3 and 33 is very sensitive to changes in the lipid content of the enzyme (Yu gt_31,, 1975) and removal of cardio— lysin is also correlated with loss of activity measured in the absence of TMPD (Robinson, 1982). When assayed in phosphate buffer, the membrane-bound oxidase and the purified form show similar kinetic constants and overall activity, but during the course of solubilization and purification much higher turnover numbers are observed. The highest catalytic rates, observed in solubilized membranes, were found to be very dependent on the precise conditions and agent used for solubilization. Dramatic depen- dence of turnover rates on the nature of the detergent and detergent- to—protein ratios has been observed previously (Mason and Ganapathy, l970), emphasizing the inherent difficulty in using maximal activities as a measure of the degree to which the enzyme resembles the native state. Although completeness of dispersion is clearly a factor in determining apparent activity (Vik and Capaldi, 1980), other more subtle effects of the membrane on enzyme conformation seem likely to be involved, such as monomer—dimer conversions, interactions with other integral membrane proteins, or possible regulation of activity by cytoplasmically synthesized subunits (Merle and Kadenbach, 1982). With regard to monomer—dimer effects, Nalecz gt_31. (1983) attribute the kinetic differences they see as a result of varying the ionic strength of the spectral assay, to interconversions between a dimer form of bovine heart enzyme having biphasic kinetics and a monomer form with monophasic kinetics. The observation of a single kinetic phase in the spectral assay for the reaction of the monomeric rat liver oxidase with cytochrome this consistent with their results. However, when assayed 131 at the same conditions of ionic strength as used in the polarographic assay, the rat liver enzyme displays clearly biphasic kinetics with a low affinity reaction that closely corresponds to the single phase observed in the spectral assay. Therefore, apparent absence of the high affinity reaction in the spectral assay is most likely a reflection of the low off-constant of cytochrome 3_in this phase, resulting in very low turnover numbers that cannot easily be detected. This is not the case in the polarographic assay system where ascorbate/TMPD can re—reduce cytochrome g_still bound to the oxidase, thus removing the off-constant as the rate limiting step in the high affinity phase of the reaction. Monomer Form of Cytochrome c Oxidase. A number of investigators have considered the idea that the native, active form of cytochrome g oxidase may be a dimer (Robinson and Capaldi, l977; Bisson gt_31,, l980; Wikstrom, 1981; Ferguson-Miller gt_31,, 1982; Nalecz gt 31., 1983). The apparent molecular weight of the rat liver oxidase in laurylmaltoside is found to be 300,000 by gel filtration, in agreement with the value obtained for the bovine heart enzyme under the same conditions (Rosevear gt 31., l980). It is difficult to assess whether this represents a monomer or dimer of oxidase without an accurate estimate of bound detergent, and considering the uncertainty regarding the contribution of the micelle to the Stokes radius of a large asymmetric protein (Nozaki gt 31,, l976; le Maire gt_31., l980). Sedimentation equilibrium studies were therefore performed by Suarez gt 31, (l983,1984) at different solvent densities (Tanford and Reynolds, 1976) to rigorously determine the molecular weight of both the protein 132 and the associated micelle. The laurylmaltoside micelle is found to have a molecular weight of 76,000, slightly larger than previous estimates from gel filtration (Rosevear_gt 31., I980), and slightly less than a single micelle of laurylmaltoside (molecular weight m45,000) is associated with the purified rat liver oxidase. Using this value for bound detergent, the corrected molecular weight for oxidase from the gel-filtration data is W250,000, higher than that predicted for a monomer (198,000: Buse gt,3fl,, 1982), but considerably lower than expected for a dimer. The sedimentation equilibrium results of Suarez gt__1, (l983,1984) also indicate that the size of cytochrome g oxidase is overestimated by the gel filtration procedure, and therefore it can be concluded that the predominant form of cytochrome g_oxidase obtained in laurylmaltoside under these buffer and pH conditions is a highly active monomer. An active monomeric species is consistent with the results of target size analysis of cytochrome g oxidase by radiation inactivation procedures, which show that high turnover and normal binding of cytochrome_3 are associated with a molecular weight unit even smaller than a monomer (Thompson gt_31,, l982). Subunit III. There is a long standing controversy concerning the role of subunit III and even its existence as an integral part of cytochrome g oxidase (Wei and King, 1981). This peptide coded for by the mitochondrial genome, has a molecular weight of 30,000 (Thalenfeld and Tzagoloff, 1980), and behaves anomalously on SOS-polyacrylamide gels (apparent molecular weight, 18,000-24,500; Azzi, 1980). Capaldi and coworkers (Downer_gt 31., I976; Capaldi et al., 1977) have 133 demonstrated that the behavior of this peptide in SDS-gel electro- phoresis systems is highly variable, depending on how the sample is treated before it is applied to the gel, and on the exact conditions under which it is run. Considering the difficulties caused by polymer- ization, insolubility, and inadequate Coomassie Blue staining of this subunit, the peptide composition of rat liver oxidase was examined using a number of different sample treatments, gel systems, and staining techniques. Under conditions where subunit III in bovine heart and crude rat liver oxidase is clearly visualized, no such peptide is seen in the highly purified rat liver enzyme. In addition, reaction of the purified rat liver enzyme with DCCD prior to electro- phoresis (a reagent which reacts primarily with subunit III of cytochrome 3_oxidase when it is present: Casey et al., 1980; Prochaska, 1981) does not label any regions of the gel corresponding to the positions of the bovine heart subunit III or its high molecular weight aggregates. Merle and Kadenbach (1980) have demonstrated unequivocally that subunit III of rat liver oxidase runs in a similar position to that of oxidases from other sources, and can be detected under the buffer and sample preparation conditions used in these electrophoretic analyses. Subunit III has been detected by electro- phoresis of rat liver oxidase prepared by other methods (Hochli and l., l979), indicating that its absence Hackenbrock, l978; Nagasawa gt_ in our preparation is the result of removal during purification, not an inherent difference between the rat and other species. Furthermore, the extensive homology between the nucleotide sequences of subunit 111 from rat liver and bovine heart (Grosskoft and Feldman, 1981; Anderson et al., 1982) assures that these peptides will respond in a 134 in a quantitatively similar manner to various staining techniques. The subunit III depleted state of the purified rat liver enzyme has also been confirmed by the sedimentation equilibrium experiments of Suarez l. (1983), which yield molecular weight values of 161,000 for the g affinity purified rat liver enzyme, and 209,000 for the bovine heart enzyme containing subunit III. This 48,000 difference in molecular weight is remarkably close to that predicted for the difference between a subunit III containing enzyme of 12 subunits (198,000; Buse ££.El-, 1982) and an enzyme depleted of subunit III and the peptides designated by Downer gt 31. (1976) as a, b, and c (139,000). All these indepen- dent methods of confirmation allow the certain conclusion that chromatography in laurylmaltoside effectively removes not only lipid, but also subunit 111 from the liver enzyme. The participation of subunit III in substrate—binding by cyto- chrome g_oxidase has been suggested by the work of several researchers (Fuller gt_31, 1981; Birchmeier gt_31,, 1976; Bisson gt_31., 1980). They observe that the back of cytochrome g oxidase crosslinks to subunit III while the front crosslinks to subunit II, suggesting a close proximity of subunit III to the active site binding domain of the oxidase. The proposal has been made that a dimer form of oxidase may provide a 'cleft' for cytochrome g, with subunit III from one monomer forming the back of the cleft for the other (Capaldi gt 31,, 1983). However, the results of kinetic studies with the purified rat liver enzyme do not support the suggestion that subunit III has a critical role in the high affinity binding of cytochrome g_to cytochrome g oxidase (Wikstrom and Penttila, 1982), since the loss of activity in response to lipid and subunit III removal occurs equally in both 135 kinetic phases, is recoverable by repletion with phospholipid, and is accompanied by relatively small changes in the Km values of the initial kinetic phase. An essential role for subunit III in the process by which cytochrome g_oxidase establishes a potential gradient across the mitochondrial membrane has also been proposed. There is evidence that cytochrome g_oxidase may be able not only to abstract protons from the matrix to form water from 02 (Mitchell, 1961a,b; Moyle and Mitchell, 1978) but also to extrude protons from the matrix by means of a proton pump mechanism (Wikstrom, 1977; Wikstrom and Saari, 1977; Sigel and Carafoli, 1978; Azzone gt_31,, 1979). The involvement of subunit III in this pump mechanism has been suggested by DCCD labeling studies which indicate that this reagent binds preferentially to subunit III and decreases the proton pumping activity of reconstituted bovine heart oxidase (Azzi _t _1,, 1979; Casey gt_31,, 1980; Prochaska gt_31,, 1981), and by the work of Wikstrom and coworkers who deplete the enzyme of subunit III (and several small peptides) and observe decreased respiratory control and loss of proton pumping activity (Penttila and Wikstrom, 1981). The results reported here show that the rat liver enzyme depleted of subunit 111 can be reconstituted into phospholipid vesicles and exhibit respiratory control ratios as high as 10. This level of response to a pH gradient is comparable to the highest reported values for the intact enzyme from bovine heart (Carroll and Racker, 1977). The data obtained with the rat liver oxidase clearly indicate the existence of a mechanism other than a subunit III dependent proton pump for producing a pH gradient, i.e., it confirms the existence of a vectorial electron transfer in this reaction. It is 136 also particularly significant in establishing that subunit III is not involved in the regulation of cytochrome g oxidase activity in response to a pH gradient, whatever its mechanism of production. This conclu- sion is in agreement with the results of Carroll and Racker (1977), who depleted preparations of bovine heart oxidase of subunit III by chymotrypsin digestion and demonstrated undiminished levels of respira- tory control, but it contrasts with the work of Wikstrom and coworkers who depleted bovine heart oxidase of subunit III with high detergent concentrations at alkaline pH, and found that the treated enzyme could respond to a pH gradient with only half its original sensitivity (Penttila and Wikstrom, 1981; Penttila, 1983). Wikstrom and coworkers correlate this reduction in respiratory control with a proposed 50% decrease in the magnitude of the energy conserving ability of the enzyme that occurs coincident with the loss of proton pumping activity in the subunit III depleted enzyme. They also observe the loss of the N30 mV per pH unit sensitivity of the cytochrome 3_redox potential after subunit 111 removal, in support of their view that a shift in the pK of an amino acid near heme 3_is required for proton release by the enzyme. It is not clear whether an observed redox potential pH depen— dence of this magnitude is sufficient to indicate that the reduction of cytochrome 3_is coupled to proton pumping, since a dependence of 60 mV per pH unit is predicted for an oxidation-reduction reaction involving a proton (Clark, 1949), as has been observed for ubiquinone in the cytochrome 331 complex (Dutton and Wilson, 1974). However, it should be noted that changes in the extent of protonation of groups in the vicinity of cytochrome 3 would also result in the dependence of its redox potential on the bulk phase pH, and this added effect could 137 result in either a smaller or larger degree of pH dependence than predicted (Walz, 1979). The results of the experiments reported here on the purified rat liver and intact bovine heart cytochrome g oxidase indicate that there is virtually no pH dependence of the Em of cytochrome 3.in either subunit III depleted or containing enzyme. This discrepancy between the data reported here for the bovine heart enzyme and that reported by other investigators (Penttila and Wikstrom, 1981) is of interest with regard to the proposal that the dimer form of oxidase is required for proton pumping (Wikstrom gt_31,, 1981), since the bovine heart enzyme used in the studies reported here was prepared by the method of Suarez gt 31. (1984) and is a monomer, and the procedure used by Wikstrom and coworkers to deplete their bovine enzyme of subunit III is similar to the procedure of Georgevich gt 21° (1983) used to "monomerize" bovine heart oxidase prepared by methods which produce dimer forms of the enzyme. Taken all together, these results raise the possibility that the interactions between two monomers of oxidase may be important in increasing the pH dependence of the redox potential of cytochrome_3, but that the simple presence of subunit III in the enzyme does not necessarily produce this effect. Although the results with the purified rat liver enzyme establish that subunit III is not required for the high affinity binding of cytochrome g_or for respiratory control, the precise nature of subunit III involvement in the apparent proton pumping function of cytochrome_3 oxidase is less well clarified. The pH measurements on purified rat liver enzyme show that it pumps protons with an efficiency less than half that of the intact bovine heart enzyme when both are measured under the same conditions. A relative activity of this magnitude may 138 be significant, however, since the two subunit enzyme from E; denitrificans has been shown to pump protons with a maximal stoichiometry of only 0.5 H+/e' under conditions where the stoichiometry of the bovine heart enzyme reaction is 1.0 H+/e’ (Solioz gt_31,, 1982). At this time there are two main difficulties in interpreting the proton pumping data presented here. First, the apparent HT/e‘ ratios for the acidification phase resulting from the oxidation of cytochrome g by the coupled enzyme (subunit III depleted or containing) always require correction for any deviation from the predicted stoichiometry of 1 HT/e' consumed in the water forming reaction, observed as an alkalinization in the presence of uncouplers. Due to a consistent short fall in this value for the rat liver enzyme, a larger correction was always subtracted, which in many cases reduced the stoichiometry of apparently authentic and substantial proton pulses to the level of 0-0.1 H+/e'. It will be important in future experiments to address the nature and cause of this short fall for the subunit III depleted enzyme, and to determine the validity of this correction. Second, since a very small amount of subunit III may be present in any given depleted enzyme (as made by this procedure, or others) it is not clear to what extent its presence could contribute to observed acidifications. The exact quantitation of the percentage of this peptide present and assay of a control rat liver enzyme prepared by a technique which does not appear to diminish subunit III, will permit the calculations necessary to determine whether observed H+/e' ratios can reasonably be attributed to the presence of a small amount of subunit III. 139 The data presented here indicate that the difference in proton pumping activity between subunit III depleted and containing enzymes may be quantitative rather than qualitative, in contrast to the results of Penttila and Wikstrom (1981; Penttila, 1983) who find total loss of pumping after subunit III depletion in Triton X-100. These findings suggest the possibility that some procedures of depletion may reduce the efficiency of the pump to zero, while others allow for the reten- tion of partial activity. It should also be noted that reconstitution of some enzyme preparations (namely, bovine heart prepared by the Hartzell-Beinert method (1974)) containing subunit III result in neither good respiratory control, nor any proton pumping, indicating that there are a variety of subtle factors involved. This view is in line with the results of DCCD modification studies which show partial inhibition of the proton pump and retention of respiratory control in the presence of a large molar excess of this reagent (Casey gt_31., 1979, 1980; Prochaska gt_31,, 1981). The present view of subunit III as a possible independent proton pump or channel may be better replaced with a more complex picture in which subunit III interacts with the other peptides of cytochrome g_oxidase to increase the efficiency of its energy conserving mechanism. Purified rat liver cytochrome g oxidase retaining high activity and respiratory control should be useful for further studies of the specific nature of these interac— tions, since subunit III can be isolated in a denatured form by several procedures (Steffens and Buse, 1976; Verheul gt_31,, 1979; Bill and Azzi, 1982), and it may prove possible to renature this peptide and reconstitute the depleted enzyme. Definition of the precise role of subunit III, and each of the other subunits in turn, will help to 140 establish the structural features required to couple electron transfer reactions to the generation of a proton gradient, and will further understanding of the fundamental process by which the energy released from electron transfer is conserved in chemical form. APPENDIX Publications and Abstracts PUBLICATIONS: Suelter, C.H., Thompson, 0., Oakley, G., Pearce, M., Husic, H.D., and Brody, M.S. (1979) Comparative Enzymology of 5'-AMP Aminohydrolase from Normal and Genetically Dystrophic Chicken Muscle. Biochem. Medicine 31, 352-365. Thompson, D.A., Suarez-Villafane, M., and Ferguson-Miller, S. (1982) The Active Form of Cytochrome g Oxidase: Effects of Detergent, the Intact Membrane, and Radiation Inactivation. Biophys. J. 31, 285-292. Thompson, D.A. and Ferguson-Miller, S. (1983) Lipid and Subunit III Depleted Cytochrome g Oxidase Purified by Horse Cytochrome g Affinity Chromatography in Laurylmaltoside. Biochemistry 33, 3178-3187. Suarez, M.D., Revzin, A.R., Kempner, E.S., Thompson, D.A., Ferguson-Miller, S. (1983) Clarification of the Form and Subunit Composition of Mammalian Cytochrome g Oxidase Required for Electron Transfer Activity, in preparation. ABSTRACTS: Suelter, C.H., Brody, M.S., Oakley, G., Thompson, D.A., and Pearce, M. (1978) Enzymology of 5'-AMP Aminohydrolase from Normal and Dystrophic Chicken Breast Muscle. Fed. Proc. 31, 1394. Van Aken, T.B., Thompson, D.A., Rosevear, P., and Ferguson-Miller, S. (1980) Alkylglycoside Detergents: Their Synthesis, Properties, and Use in Purifying Rat Liver Cytochrome Oxidase. Fed. Proc. 32, 2397. Ferguson-Miller, S., Suarez-Villafane, M., and Thompson, D.A. (1982) The Active Form of Cytochrome Oxidase. Biophys. J. 31, 2a. Suarez, M.D., Revzin, A., Swaisgood, M., Thompson, D.A., and Ferguson-Miller, S. (1983) Monomer Forms of Beef Heart and Rat Liver Cytochrome Oxidases in Laurylmaltoside Differing in Subunit Content. Fed. Proc. 13, 1782. Thompson, D.A., and Ferguson-Miller, S. (1983) Affinity Purified Cytochrome Oxidase in Laurylmaltoside: Physical and Kinetic Characteristics of Lipid and Subunit III Depleted Enzyme. Fed. Proc. 42, 1783. 141 BIBLIOGRAPHY BIBLIOGRAPHY Al-Tai, W.F., Jones, M.G., Rashid, K., and Wilson, M.T. (1983) Biochem. J. 392, 901-903. Ames, B. (1966) Methods Enzymol. 3, 115-118. Anderson, S., de Bruijn, M.H.L., Coulson, A.R., Eperon, I.C., Sanger, F., and Young, 1.0. (1982) J. Mol. Biol. 133, 683-717. Andreasson, L.—E. (1975) Eur. J. Biochem. 33, 591-597. Andreasson, L.-E., Malmstrom, B.G., Stromberg, C., and Vanngard, T. (1972) FEBS Lett. 33, 297-301. Antalis, T.M. and Palmer, G. (1982) J. Biol. Chem. 331, 6194-6206. Artzatbanov, V.Y., Konstantinov, A.A., and Skulachev, V.P. (1978) FEBS Lett. 31, 180-185. Awasthi, Y.C., Chuang, T.F., Keenan, T.W., and Crane, F.L. (1971) Biochim. Biophys. Acta 333, 42-52. Azzi, A. (1980) Biochim. Biophys. Acta 331, 231-252. Azzi, A., Bill, K., and Broger, C. (1982) Proc. Natl. Acad. Sci. USA 12, 2447-2450. Azzi, A., Bisson, R., and Casey, R.P. (1979) 13_Membrane Bioenergetics (Lee, C.P., Schatz, G., and Ernster, L., Eds.), pp. 13-20, Addison-Wesley, Reading, MA. Azzone, G.F., Pozzan, T., and Virgilio, F.D. (1979) J. Biol. Chem. .331, 10206-10212. Babcock, G.T. and Callahan, P.M. (1983) Biochemistry 33, 2314-2319. Baron, C. and Thompson, T.E. (1975) Biochim. Biophys. Acta 333, 276. Bartlett, G. (1959) J. Biol. Chem. 225, 466-468. Bashford, C.L., Casey, R.P., Radda, G.K., and Ritchie, G.A. (1976) Neuroscience 1, 399-412. 142 143 Beechey, R.B., Robertson, A.M., Holloway, C.T., and Knight, 1.0. (1967) Biochemistry 3, 3867-3879. Bill, K. and Azzi, A. (1982) Biochem. Biophys. Res. Commun. 133, 1203-1209. Bill, K., Broger, C., and Azzi, A. (1982) Biochim. Biophys. Acta 313, 28-34. Bill, K., Casey, R.P., Broger, C. and Azzi, A. (1980) FEBS Lett. 133, 248-250. Birchmeier, W., Kohler, C.E., and Schatz, G. (1976) Proc. Natl. Acad. Sci. USA 13, 4334-4338. Bisson, R., Azzi, A., Gutweniger, H., Colonna, R., Montecucco, C., and Zanotti, A. (1978a) J. Biol. Chem. 333, 1874-1880. Bisson, R., Gutweniger, H., and Azzi, A. (1978b) FEBS Lett. 33, 219-222. Bisson, R., Jacobs, B., and Capaldi, R.A. (1980) Biochemistry 13, 4173-4178. Bisson, R., Montecucco, C., Gutweniger, H., and Azzi, A. (1979) J. Biol. Chem. 331, 9962-9965. Brautigan, D.L., Ferguson-Miller, S., and Margoliash, E. (1978a) J. Biol. Chem. 333, 130—139. Brautigan, D.L., Ferguson-Miller, S., Tarr, G.E., and Margoliash, E. (1978b) J. Biol. Chem. 333, 140—148. Brautigan, D.L., Ferguson-Miller, S., and Margoliash, E. (1978c) Methods Enzymol. 33, 128-164. Brierley, G.P. and Merola, A.J. (1962) Biochim. Biophys. Acta 33, 205-217. Briggs, M.M. and Capaldi, R.A. (1977) Biochemistry 13, 73—77. Briggs, M.M. and Capaldi, R.A. (1978) Biochem. Biophys. Res. Comm. 33, 553-559. Buse, G., Steffens, G.C.M., Steffens, G.J., Meinecke, L., Biewald, R., and Erdweg, M. (1982) European Bioenergetics Conference Reports, Vol. 2, pp. 163-164, LBTM-CNRS, ed., Lyon. Buse, G., Steffens, G.J., Steffens, G.C.M., Sacher, R., and Erdweg, M. (1980) 13 First European Bioenergetics Conference, pp. 41-42, Patron, Bologna. Callahan, P.M. and Babcock, G.T. (1983) Biochemistry 33, 452-461. 144 Capaldi, R.A. (1982) Biochim. Biophys. Acta 694, 291-306. Capaldi, R.A., Bell, R.L. and Brancheck, T. (1977) Biochem. Biophys. Res. Commun. 14, 425-433. Capaldi, R.A., Malatesta, F., and Darley-Usmar, V.M. (1983) Biochim. Biophys. Acta 126, 135-148. Carroll, R. and Racker, E. (1977) J. Biol. Chem. ggg, 6981-6990. Carter, K. and Palmer, G. (1982) J. Biol. Chem. 2§Z, 13507-13514. Casey, R.P., Chappell, J.B., and Azzi, A. (1979) Biochem. J. 182, 149-156. Casey, R.P., Thelen, M., and Azzi, A. (1980) J. Biol. Chem. ggg, 3994-4000. Cerletti, N. and Schatz, G. (1979) J. Biol. Chem. 264, 7746-7751. Chan, S.H.P. and Tracy, R.P. (1978) Eur. J. Biochem. 89, 595-605. Chance, B. and Williams, G.R. (1956) Adv. Enzymol. 11, 65-134. Clark, W.M. (1972) Oxidation-Reduction Potentials of Organic Systems, p. 264, Krieger Publ. Co., Huntington, NY. Conrad, M.J. and Singer, S.J. (1979) Proc. Natl. Acad. Sci. USA 16, 5202-5206. Creutz, C. and Sutin, N. (1974) J. Biol. Chem. 242, 6788. Cuatrecasas, P., Wilchek, M., and Anfinsen, C. (1968) Proc. Natl. Acad. Sci. USA 61, 636-643. Darley-Usmar, V.M., Alizai, N., Ayashi, A.I., Jones, G.D., Sharpe, A., and Wilson, M.T. (1981) Comp. Biochem. Physiol. 688, 445-456. Davies, H.C., Smith, L., and Wasserman, A.R. (1964) Biochim. Biophys. Acta. §§, 238-246. Deatherage, J.F., Henderson, R., and Capaldi, R.A. (1982) J. Mol. Biol. 1§§, 487-499. Dockter, M.E., Steinemann, A., and Schatz, G. (1978) J. Biol. Chem. 2§§, 311-317. Downer, N.W., Robinson, N.C., and Capaldi, R.A. (1976) Biochemistry 1§, 2930-2936. Dutton, P.L. and Wilson, D.F. (1974) Biochim. Biophys. Acta 346, 165-212. Erecinska, M. (1977) Biochem. Biophys. Res. Commun. 16, 495-501. 145 Erecinska, M., Oshino, R. and Wilson, D.F. (1980) Biochem. Biophys. Res. Commun. 22, 743-748. Errede, B., Haight, G.P., and Kamen, M.D. (1976) Proc. Natl. Acad. Sci. USA 12, 113-117. Errede, B. and Kamen, M.D. (1978) Biochemistry 11, 1015-1027. Evans, D.J. (1970) J. Bacteriol. 125, 1203-1212. Ewall, R.X. and Bennett, L.E. (1974) J. Amer. Chem. Soc. 26, 940. Eytan, G. D. and Schatz, G. (1975) J. Biol.Chem. 262, 767-774. Eytan, G.D., Carroll, R.C., Schatz, G., and Racker, E. (1975) J. Biol. Chem. 262, 8598-8603. Felgner, P.L., Messer, J.L., and Wilson, J.E. (1979) J. Biol. Chem. 262, 4946-4949. Ferguson-Miller, S., Brautigan, D.L., and Margoliash, E. (1976) J. Biol. Chem. 261, 1104-1115. Ferguson-Miller, S., Brautigan, D.L., and Margoliash, E. (1977).ifl The Porphyrins (Dolphin, D., Ed.), pp. 149-240, Academic Press, New York. Ferguson-Miller, S., Brautigan, D.L., and Margoliash, E. (1978) J. Biol. Chem. 252, 149-159. Ferguson-Miller, S., Weiss, H., Speck, S.H., Brautigan, D.L., Osheroff, N., and Margoliash, E. (1979) 16_Cytochrome Oxidase (King, T.E., Orii, Y., Chance, B. and Okunuki, K., Eds.), pp. 219-227, Elsevier/North Holland Biomedical Press, Amsterdam. Ferguson-Miller, S., Van Aken, T., and Rosevear, P. (1982) 12. Electron Transport and Oxygen Utilization (Ho, C., Ed.) pp. 297—303, Elsevier/North Holland Biomedical Press, New York. Fleischer, S. (1979) Methods Enzymol. 66, 28-32. Fowler, L.R., Richardson, S.H., and Hatefi, Y. (1962) Biochim. Biophys. Acta 26, 103. Freedman, J.A. and Chan, S.H.P. (1983) J. Biol. Chem. 266, 5885—5892. Frey, T.G., Chan, S.H.P., and Schatz, G. (1978) J. Biol. Chem. 2.53.9 4389'4395. Fry, M. and Green, D.E. (1980) Biochem. Biophys. Res. Comm. 22, 1238-1246. 146 Fuller, S.D., Capaldi, R.A., and Henderson, R. (1979) J. Mol. Biol. léi- 305-327. Fuller, S.D., Darley-Usmar, V.M., and Capaldi, R.A. (1981) Biochemistry 22, 7046-7053. Gennis, R.B., Casey, R.P., Azzi, A., and Ludwig, B. (1982) Eur. J. Biochem. 126, 189-195. Georgevich, G. and Capaldi, R.A. (1982) Biophys. J. 61, 66-67. Georgevich, G., Darley-Usmar, V.M., Malastesta, F., and Capaldi, R.A. (1983) Biochemistry 22, 1317—1322. Gibson, Q.H., Greenwood, C., Wharton, D.C., and Palmer, G. (1965) J. Biol. Chem. 249, 888-894. Gibson, Q.H. and Greenwood, C. (1965) J. Biol. Chem. 262, 2694-2698. Grosskopf, R. and Feldman, H. (1981) Current Genetics 5, 151-158. Gutweniger, H., Bisson, R., and Montecucco, C. (1981) Biochim. Biophys. Acta 666, 187-193. Halaka, F.G. (1981) Doctoral Dissertation, Michigan State University. Harold, F.M., Pavlasova, F., and Baarda, J.R. (1970) Biochim. Biophys. Acta 126, 235—244. Hartzell, C.H. and Beinert, H. (1974) Biochim. Biophys. Acta 666, 318—338. Henderson, R., Capaldi, R.A., and Leigh, J.S. (1977) J. Mol. Biol. 112, 631-648. Hochli, L. and Hackenbrock, C.R. (1978) Biochem. 11, 3712-3719. Hochman, J.H., Schindler, M., Lee, J.G., and Ferguson-Miller, S. (1982) Proc. Natl. Acad. Sci. USA 12, 6866-6870. Hochman, J.H., Schindler, M., Lee, J.G., and Ferguson-Miller, S. (1983) 16_Biochemistry of Metabolic Processes (Lennon, D., Stratman, F., and Zahlten, R., Eds.), pp. 441—450, Elsevier/North Holland Biomedical Press, New York. Hodges, H.L., Holwerda, R.A., and Gray, H.B. (1974) J. Amer. Chem. Soc. 26, 3132-3137. Jacobs, E.E., Jacob, M., Sanadi, D.R., and Bradley, L.B. (1956) J. Biol. Chem. 226, 147-156. 147 Johnson, M.J. (1949) 16_Respiratory Enzymes (Lardy, H., Ed.), pp. 58—70, Burgess Publ. Co., Minneapolis, MN. Jost, P., Griffith, 0.H., Capaldi, R.A., and Vanderkooi, G. (1973a) Biochim. Biophys. Acta. 611, 141-152. Jost, P., Griffith, O.H., Capaldi, R.A., and Vanderkooi, G. (1973b) Proc. Natl. Acad. Sci. USA 16, 480-484. Kadenbach, B., Hartmann, R., Glanville, H., and Buse, G. (1982) FEBS Lett. 3&- 236-238. Kadenbach, B., Jarausch, J., Hartmann, R., and Merle, P. (1983) Anal. Biochem. 122, 517-521. Kadenbach, B. and Merle, P. (1981) FEBS Lett. 166, 1-11. Kagawa, Y. (1967) Biochim. Biophys. Acta 131, 586-588. Kagawa, Y. and Racker, E. (1971) J. Biol. Chem. 226, 5477-5487. Keilin, D. (1925) Proc. Roy. Soc. (London) B 26, 312-339. Keilin, D. and Hartree, E.F. (1938) Nature 121, 870. Koppenol, W.H. and Margoliash, E. (1982) J. Biol. Chem. 261, 4426-4437. Konig, B.W., Osheroff, N., Wilms, J., Muijsers, A.0., Dekker, H.L., and Margoliash, E. (1980) FEBS Lett. 111, 395-398. Kroon, A.M. and Saccone, C. (1974) The Biogenesis of Mitochondria, Academic Press, London. Kuboyama, M., Yong, F.C., and King, T.E. (1972) J. Biol. Chem. fl, 6375-6383. Laemmli, U.K. (1970) Nature (London) 221, 680-685. Lehninger, A.L., Reynafarje, B., Alexandre, A., and Villalobo, A. (1979) ifl Membrane Bioenergetics (Lee, C.P., Schatz, G., and Ernster, L., Eds.), pp. 393-404, Addison-Wesley, Reading, MA. le Maire, M., Rivas, E. and M¢ller, J.V. (1980) Anal. Biochem. 166, 12-21. Lorusso, M., Capuano, F., Boffoli, D., Stefanelli, R., and Papa, S. (1979) Biochem. J. 162, 133-147. Lowry, 0.H., Rosebrough, N.J., Farr, A.L., and Randall, R.J. (1951) J. Biol. Chem. 126, 265-275. Ludwig, B. (1980) Biochim. Biophys. Acta 626, 177-189. 148 Ludwig, B., Downer, N.W., and Capaldi, R.A. (1979) Biochemistry 16, 1401-1407. Malatesta, F., Georgevich, G., and Capaldi, R.A. (1983) ifl Structure and Function of Membrane Proteins (F. Palmeri et al., Eds.), Elsevier/North Holland Biomedical Press, Amsterdam, in press. Margoliash, E. and Bosshard, H.R. (1983) Trends in Biol. Sci. 6, 316-320. Mason, H.S. and Ganapathy, K. (1970) J. Biol. Chem. 226, 230-237. Mason, T.L., Poyton, R.0., Wharton, D.C., and Schatz, G. (1973) J. Biol. Chem. 266, 1345—1354. McCarty, R.E. and Racker, E. (1967) 262, 3435-3439. Merle, P., Jarausch, J., Trapp, M., Scherka, R., and Kadenbach, B. (1981) Biochim. Biophys. Acta 662. 222-230. Merle, P. and Kadenbach, B. (1980) Eur. J. Biochem. 166, 499-507. Merle, P. and Kadenbach, B. (1982) Eur. J. Biochem. 126, 239-244. Millett, F., Darley-Usmar, V.M., and Capaldi, R.A. (1982) Biochemistry 21, 3857—3862. Millett, F., DeJong, K., Paulson, L., and Capaldi, R.A. (1983) Biochemistry 22, 546-552. Minnaert, K. (1961) Biochim. Biophys. Acta 66, 23-34. Mitchell, P. (1961a) Biological Structure and Function, Proc. First IUB/IUBS Int. Symp., Stockholm, 1960, 2, 581-599. Mitchell, P. (1961b) Nature 121, 144-148. Mitchell, P. (1966) 'Chemiosmotic Coupling in Oxidative and Ezgtosynthetic Phosphorylation,‘ Bodmin, England: Glynn Research Mitchell, P. (1979) Eur. J. Biochem. 26, 1-20. Mitchell, P. and Moyle, J. (1983) FEBS Lett. 161, 167-178. Moyle, J. and Mitchell, P. (1978) FEBS Lett. 66, 268—272. Nagasawa, T., Nagasawa-Fujimori, H., and Heinrich, P.C. (1979) Eur. J. Biochem. 26, 31-39. Nalecz, K.A., Bolli, R., and Azzi, A. (1983) Biochem. Biophys. Res. Commun. 112, 822-828. 149 Nicholls, P. (1964) Arch. Biochem. Biophys. 166, 25-48. Nicholls, P. (1965) 12 Oxidases and Related Redox Systems (King, T.E., Mason, H.S., and Morrison, M., Eds.) Vol. 2, pp. 764-783, Wiley, New York. Nozaki, Y., Schechter, N.M., Reynolds, J.A., and Tanford, C. (1976) Biochemistry 16, 3884-3890. Ohnishi, T., Blum, H., Leigh, J.S., Jr., and Salerno, J.C. (1979) 16_Membrane Bioenergetics (Lee, C.P., Schatz, G., and Ernster, L., Eds.), pp. 21-30, Addison-Wesley, Reading, MA. Osheroff, N., Brautigan, D.L., and Margoliash, E. (1980) J. Biol. Chem. 266, 8245-8251. Osheroff, N., Speck, S.H., Margoliash, E., Veerman, E.C.I., Wilms, J., Konig, B.W., and Muijsers, A.0. (1983) J. Biol. Chem. 266, 5731-5738. Papa, S., Guerrieri, F., Lorusso, M., 1220, G., Boffoli, D., Capuano, F., Capitanio, N., and Altamura, N. (1980) Biochem. J. 122, 203-218. Parsons, J. and Patton, S. (1967) J. Lipid Res. 6, 696-698. Penttila, T. and Wikstrom, M. (1981) 16_Vectorial Reactions in Electron and Ion Transport in Mitochondria and Bacteria (F. Palmieri 21.21:, Eds.), pp. 71-80, Elsevier/North Holland Biomedical Press, Amsterdam. Penttila, T. (1983) Eur. J. Biochem. 166, 355-361. Petri, W.A. and Wagner, R.R. (1979) J. Biol. Chem. 262, 4313-4316. Poyton, R.0. and Schatz, G. (1975a) J. Biol. Chem. 266, 752-761. Poyton, R.0. and Schatz, G. (1975b) J. Biol. Chem. 266, 762-766. Prochaska, L., Bisson, R., and Capaldi, R.A. (1980) Biochem. 12, 3174-3179. Prochaska, L.J., Bisson, R., Capaldi, R.A., Steffens, G.C.M., and Buse, G. (1981) Biochim. Biophys. Acta 661, 360-373. Ragland, W.L., Pace, J.L., and Kempner, D.L. (1974) Anal. Biochem. 62, 24-33. Reider, R. and Bosshard, H.R. (1978) FEBS Lett. 22, 223-226. Reider, R. and Bosshard, H.R. (1979) J. Biol. Chem. 266, 4732-4739. 150 Robinson, N.C. and Capaldi, R.A. (1977) Biochemistry 16, 375-381. Robinson, N.C., Strey, F., and Talbert, L. (1980) Biochemistry 12, 3656-3661. Robinson, N.C. (1982) Biochemistry 21, 184—188. Rosevear, P., Van Aken, T., Baxter, J., and Ferguson-Miller, S. (1980) Biochemistry 12, 4108—4115. Saraste, M., Penttila, T., Coggins, J.R., Wikstrom, M. (1980) FEBS Lett. 111, 35-38. Saraste, M., Penttila, T., and Wikstrom, M. (1981) Eur. J. Biochem. 116, 261-268. Schatz, G. and Mason, T.L. (1974) Ann. Rev. Biochem. 66, 51-87. Sigel, E. and Carafoli, E. (1978) Eur. J. Biochem. 62, 119-123. Smith, H.T., Staudenmayer, N., and Millett, F. (1977) Biochemistry 16, 4971-4974. Smith, L. and Camerino, P. (1963) Biochemistry 2, 1432-1439. Smith, L. and Conrad, H. (1956) Arch. Biochem. Biophys. 66, 403. Smith, M.B., Stonehuerner, J., Ahmed, A.J., Staudenmayer, N., and Millett, F. (1980) Biochim. Biophys. Acta 622, 303-313. Solioz, M., Carafoli, E., and Ludwig, B. (1982) J. Biol. Chem. 261, 1579-1582. Sone, N. and Yanagita, Y. (1982) Biochim. Biophys. Acta 662, 216-226. Sone, N., Yanagita, Y., Hon-Nami, K., Fukumori, Y., and Yamanaka, T. (1983) FEBS Lett. léé, 150-154. Sottocasa, G.L., Kuylenstierna, B., Ernster, L., and Bergstrand, A. (1967) Methods Enz. 16, 448-463. Stadt, M., Banks, P., and Kobes, R. (1982) Arch. Biochem. Biophys. 211. 223-230. Steffens, G.J. and Buse, G. (1976) Hoppe—Seyler's Z. Physiol. Chem. §§Z; 1125-1137. Steffens, G.J. and Buse, G. (1979) Hoppe-Seyler's Z. Physiol. Chem. §§Q. 613-619. Stevens, T.H., Martin, C.T., Wang, H., Brudvig, G.W., Scholes, C.P., and Chan, 5.1. (1982) J. Biol. Chem. 261, 12106-12113. 151 Stewart, P.A. (1982) Biophys. J. 61, 404a. Storey, B.T. and Lee, C.P. (1981) Trends in Biol. Sci. 6, 166-169. Stubbs, G.W. and Litman, B.J. (1978) Biochemistry 11, 215. Suarez, M.D., Revzin, A.R., Swaisgood, M., Thompson, D.A., and Ferguson-Miller, S. (1983) Fed. Proc. 22, 1782. Suarez, M.D., Revzin, A.R., Kempner, E.S., Thompson, D.A., and Ferguson-Miller, S. (1984), in preparation. Tanford, C. (1972) J. Mol. Biol. 61, 59-74. Tanford, C. and Reynolds, J.A. (1976) Biochim. Bi0phys. Acta 661, 133. Thalenfeld, B. and Tzagoloff, A. (1980) J. Bi01. Chem. 266, 6173-6180. Thompson, D.A., Suarez-Villafane, M., and Ferguson-Miller, S. (1982) Biophys. J. 61, 285-292. Tzagoloff, A. (1982) Mitochondria, Plenum Press, New York. Tzagoloff, A., Macino, G., and Sebald, W. (1979) Ann. Rev. Biochem. 26, 419-441. Tzagoloff, A. and MacLennon, D. (1965) Biochim. Biophys. Acta 22, 476-485. Udenfriend, S., Stein, S., Bohlen, P., Dairman, W., Leimgruber, W., Weigle, M. (1972) Science 116, 871-872. van Buuren, K.J.H., Nicholls, P., and Van Gelder, B.F. (1972) Biochim. Biophys. Acta 266, 258-276. van Buuren, K.J.H., Van Gelder, B.F., Wilting, J., and Braams, R. (1974) Biochim. Biophys. Acta 666, 421-429. Vanderkooi, J.M., Landesberg, R., Hayden, G.W., and Owen, C.S. (1977) Eur. J. Biochem. 61, 339-347. Vanneste, W.H. (1966) Biochemistry 6, 838-848. Vanneste, W.H., Ysebaert-Vanneste, M., and Mason, H.S. (1974) J. Biol. Chem. 262, 7390-7401. Veerman, E.C.I., Wilms, J., Casteleijn, G., and Van Gelder, B.F. (1980) Biochim. Biophys. Acta 3229. 117-127. 152 Verheul, F.E.A.M., Boonman, J.C.P., Draijer, J.W., Muijsers, A.0., Borden, D., Tarr, G.E., and Margoliash, E. (1979) Biochim. Biophys. Acta 626, 397-416. Vik, S.B. and Capaldi, R.A. (1977) Biochemistry 16, 5755-5759. Vik, S.B. and Capaldi, R.A. (1980) Biochem. Biophys. Res. Comm. 22, 348-354. Vik, S.B., Georgevich, G., and Capaldi, R.A. (1981) Proc. Natl. Acad. Sci. USA 16, 1456-1460. von Jagow, G. and Klingenberg, M. (1972) FEBS Lett. 22, 278-282. Walz, D. (1979) Biochim. Biophys. Acta EQE; 279-353. Watts, A., Marsh, D., and Knowles, P.F. (1978) Biochem. Biophys. Res. Commun. 61, 403-409. Wei, Y.-H. and King, T.E. (1981) J. Biol. Chem. 266, 10999-11003. Weiss, H. and Juchs, B. (1978) Eur. J. Biochem. 66, 17-28. Weiss, H., Juchs, B., and Ziganke, B. (1978) Methods Enzymol. 66, 98-112. Weiss, H. and Kolb H.J. (1979) Eur. J. Biochem. 22, 139-149. Weiss, H., Schwab, A.J., and Werner, S. (1975).ifl Membrane Biogenesis (Tzagoloff, A., Ed.), pp. 125-153, Plenum Press, New York. Werner, S. (1977) Eur. J. Biochem. 12, 103-110. Wikstrom, M. (1977) Nature 266, 271-273. Wikstrom, M. (1981) 16_Mitochondria and Microsomes (Lee, C.P., Schatz, G., and Dallner, G., Eds.), pp. 249—269, Addison—Wesley, Reading, MA. Wikstrom, M., Krab, K., and Saraste, M. (1981) in Cytochrome Oxidase: A Synthesis, Academic Press, London. Wikstrom, M. and Penttila, T. (1982) FEBS Lett. 142, 183-189. Wikstrom, M. and Saari, H.T. (1977) Biochim. Biophys. Acta 262, 347-361. Williams, R.J.P. (1961) J. Theor. Biol. 1, 1-17. Wilson, M.T., Greenwood, C., Brunori, M., and Antonini, E. (1975) Biochem. J. 121, 145-153. 153 Wilson, M.T., Lalla-Maharajh, W., Darley—Usmar, V., Bonaventura, J., Bonaventura, C., and Brunori, M. (1980) J. Biol. Chem. 266, 2722-2728. Winter, D.B., Bruynincky, W.J., Foulke, F.G., Grinich, N.P., Nicholas, P., and Mason, H.S. (1980) J. Biol. Chem. 266 11408—11414. Wittenberger, B., Raben, D., Lieberman, M.A., and Glazer, L. (1978) Proc. Natl. Acad. Sci. U.S.A. 16, 5457. Wray, W., Boulikas, T., Wray, V., and Hancock, R. (1981) Anal. Biochem. 116, 197-203. Yasunobo, K.T., Tanaka, M., Haniu, M., Samashima, M., Reimer, N., and Eto, T. (1979) 16_Cytochrome Oxidase (King, T.E., Orii, Y., Chance, B., and Okunuki, K., Eds.), Elsevier, North Holland Biomedical Press, Amsterdam. Yonetani, T. and Ray, G.S. (1965) J. Biol. Chem. 226, 3392. Yoshida, T., Lorence, R.M., Choc, M.G., Tarr, G.E., Findling, K.L., and Fee, J.A. (1983) J. Biol. Chem., in press. Yu, C., Yu, L., and King, T. (1975) J. Biol. Chem. 266, 1383-1392. Yu, C., Yu, L., and King, T. (1979) 16_Cytochrome Oxidase (King, T.E., Orii, Y., Chance, B., and Okunuki, K., Eds.), pp. 219-227, Elsevier/North Holland Biomedical Press, Amsterdam. 1117/1111! 3 lZSfl/IT/EIFW/jflr88 1 7 ill/11117111