a I . . a. I '2‘? xvv 9: x» 3 f «F JEEULuh .«mor... mu? w... .E fix.“ .91 .HL... azmnhtumfl. L . 3%“. 1 THCSTS wlsllllllllll 3 1293 01701 4758 This is to certify that the dissertation entitled STUDY OF THE ELECTRON TRANSFER AND PROTON PUMPING MECHANISMS IN CYTOCHROME _c_ OXIDASE presented by Jie Qian has been accepted towards fulfillment of the requirements for Ph . D . degree in Jimhfimislry . ' /W Major professor Date 7],? /93 MS U is an Affirmative Action/Equal Opportunity Institution 0-12771 v v-—~‘, \ v "v' '\ V‘— PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINE return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE ma alumnus-m4 STUDY OF ELECTRON TRANSFER AND PROTON PUMPING MECHANISMS IN CYTOCHROME c OXIDASE By Jie Qian A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Biochemistry 1998 ABSTRACT STUDY OF ELECTRON TRANSFER AND PROTON PUMPING MECHANISMS IN CYTOCHROME c OXIDASE By Jie Qian The mechanism of proton translocation and electron transfer in cytochrome c oxidase is a critical question in the field. For detailed study of this process, a Rhodobacter sphaeroides cytochrome c oxidase overexpression strain was constructed, increasing the oxidase expression level up to four-fold over wild type. An improved rapid purification system was developed by using histidine as eluant to purify the histidine—tagged Rhodobacter sphaeroides cytochrome oxidase with nickel affinity chromatography. Site-directed mutagenesis was used to test proposed important residues in proton pumping and electron transfer. Mutants were generated in three regions in subunit I of Rhodobacter sphaeroides cytochrome c oxidase: D132A and D132E, on an interior loop between helix II and helix III; D407A, D407N and D407C, on an exterior loop between helix IX and helix X; and R482K, R482Q, R482A, and R482P, on an exterior loop between helix XI and helix XII. It has been proposed that Asp132 and Asp407 might be involved in proton pumping and Asp407 could also be a manganese ligand. Arg482 has been suggested to play a role in the direct electron transfer between Cu A and heme a. The D132E mutant exhibited wild type characteristics in all aspects, while D132A showed loss of proton pumping activity with inhibition of electron transfer activity; however, no major structural changes were observed at the heme sites, as evidenced by visible absorbance. It was concluded that the carboxyl group at this position is essential for proton pumping. Reconstituted D132A demonstrated inhibition instead of stimulation of electron transfer activity when the membrane potential or proton gradient was released, opposite to wild type enzyme. The long chain fatty acid, arachidonic acid, stimulated electron transfer activity of the purified D132A up to seven-fold and restored normal respiratory control, suggesting that arachidonic acid can repair the mutant enzyme by compensating for loss of the carboxyl group, and improving proton uptake from the inside of the membrane. All the mutants at the Asp407 position had electron transfer activity, proton pumping activity, and Mn/Mg binding properties similar to those of wild type, and the structure at the redox active metal centers of these mutants appeared to be native. These results ruled out a requirement for Asp407 in the proton pumping or manganese binding. All the nonconservative mutants at Arg482 showed decreased electron transfer activity but complete loss of proton pumping ability, implying that it could play a role in proton pumping. The internal electron transfer rate between CuA and heme a for R482Q and R482P was shown to be much slower than that for wild type, suggesting that the through-bond pathway in which Arg482 is involved is important for electron transfer from CuA to heme a. All mutants lost heme a 3 to different degrees, and the mutants R482Q, R482A and R482P are incapable of binding Mn/Mg, indicating a stabilization function for the hydrogen bonding network at the interface of subunit I and subunit II. Copyright by Jie Qian 1998 To my husband, Michael and my family in China cncoux helped tommi Williar ACKNOWLEDGMENTS I am grateful to my advisor Dr. Shelagh Ferguson-Miller for her guidance and encouragement during my graduate studies at Michigan State University. She not only helped me to learn more about science, but also about people. I also thank my committee members: Drs. Gerald T. Babcock, Zachary Burton, Honggao Yan and William Wells for advice and support throughout my Ph.D. program. I am grateful to - Dr. Gerald Babcock for collaboration in spectroscopy analyses and for discussing results with me. I am also grateful to Dr. Frank Millett at the University of Arkansas for collaboration in the rapid electron transfer rate analyses. I would like to thank my past and present coworkers in the lab: Drs. Jon Hosler, Denise Mills, and Carrie Hiser, who gave me excellent suggestions in scientific writing; Dr. John Fetter, who was directly involved in my project; Dr. Laurence Florens, who helped me tremendously in computer operation as well as in scientific thinking; Dr. Uan Gen Kang for being a very nice and supportive counselor for my scientific career; Yasmin Hilmi, for her friendship; Yuejun Zhen, for being a fellow Chinese in the lab. Special thanks are due to Patty Voss of Dr. John Wang’s lab, who taught me molecular biology techniques during my first lab rotation and helped me adjust to the new environment. I would like to thank two undergraduate students, Amanda Looney and Dan Dorgan, for their contributions to my projects. I wish to thank Drs. Michelle Pressler, Wenjun Shi and Curt Hoganson of Dr. Gerald T. Babcock’s group for doing resonance Raman and EPR assays on my vi sat kin Pie and mi broe grit: samples; and Dr. Lois Geren of Dr. Frank Millett’s group for doing ruthenium fast kinetics assays on my mutants. Ienjoy sharing a lot of thoughts with Dr. Michelle Pressler and appreciate her friendship. Iam very fortunate to have met and married Mike here in MSU. My last two and half years with Mike have been the happiest time in my life. His love, patience and support helped me through many frustrating periods. Our discussion of science broadened my knowledge and stimulated me to become a better scientist. I am also grateful to Mike’s family who are very loving and caring. I am especially grateful to my kind grandmother, my mother and my sister. Without their love, encouragement and sacrifice, it would not have been possible for me to get my education and pursue my dreams. I am very grateful to have met my host family in America, Joan and Bob Wirtz. Their kindness and hospitality make me feel at home. Finally, I would like to thank the many people I have met in Michigan for their friendship: Jane Boles, Wei Guo, Tong Hao, Shannon Langdon, Yue Li, Wenjing Liu, Bing Ren, Jun Sheng, Jie Song, Ying Tang, Lu-Lu and CY. Wang, Dian-Peng Xu, Ying Wang, Hong Zhang, and Wenge Zhang. They make my life here much more enjoyable. vii TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES xii xiii ABBREVIATIONS xvi CHAPTER I LITERATURE REVIEW H 1. Background on Cytochrome c Oxidase l. 1 . Introduction 1.2. Structure and Function of Cytochrome c Oxidase 1.2.1. Metal Centers 1.2.2. Subunits of Cytochrome c Oxidase 2. Proton Transfer Pathways 2.1. Proton Wire Theory 2.2. Bacteriorhodopsin - A Model Proton Pumping System for Cytochrome c Oxidase 2.3. Models of Proton Pumping Coupled to Electron Transfer ......................... 2.3.1. Direct Coupling Models 2.3.2. Indirect Coupling Models 2.4. Bound Water 2.5. Site-Directed Mutagenesis Studies on Asp132 and Asp407 to Study Their Involvement in Proton Pumping 3. Electron Transfer Pathways 3.1. Electron Entry Site 3.2. Cytochrome c Docking Algorithms 3.3. Cytochrome c - Cytochrome c Oxidase Interaction 3.4. Electron Transfer Pathway Models 3.4.1. Marcus Theory 3.4.2. “Through-Space” Model 3.4.3. “Through-Bond” Model 3.4.4. Testing “Through-Space” and “Through-Bond” Theory in Cytochrome c Oxidase References viii 22 24 24 36 38 39 4O 40 4 1 4 l 42 43 43 45 47 raw; CHAPTER H ASPARTATE 132 IS ESSENTIAL FOR PROTON PUMPING PROCESS IN RHODOBACTER SPHAEROIDES CYTOCHROME c OXIDASE Abstract Introduction Materials and Methods Site-Directed Mutagenesis Growth of Rhodobacter sphaeroides Membrane Preparation Hydroxyapatite Chromatography Purification by FPLC Reconstitution of Cytochrome c Oxidase Visible Spectroscopy Oxygen Consumption Assay Proton Pumping Assay Respiratory Control Ratio (RCR) Assay Results Visible Spectra, Electron Transfer and CO Binding Assay ................................ Reconstitution of Enzyme and Respiratory Control Assay ................................ Proton Pumping Assay Fatty Acid Titration on D132A Mutant Discussion References CHAPTER III CONSTRUCTION OF OVEREXPRESSION STRAIN AND IMPROVEMENT OF PURIFICATION OF RHODOBACTER SPHAEROIDES CYTOCHROME c OXIDASE 62 62 62 65 66 67 67 68 68 68 69 7O 70 7O 74 77 82 87 89 Abstract Introduction Materials and Methods Construction of Overexpression Plasmid Small Scale Cytoplasmic Membrane Preparation Visible Spectroscopy Purification of His-Tagged Cytochrome c Oxidase Construction of Overexpression Plasmid for D132A Mutant ................... Construction of His-Tagged Plasmid for D132A Mutant Results Overexpression System Development in Rhodobacter sphaeroidesmm ix 90 91 93 93 93 96 96 97 '97 99 99 CH: SPH Histidine-Tagged Cytochrome c Oxidase Purification Construction of Overexpression and His-Tagged Plasmid for D132A Discussion References CHAPTER IV ASPARTATE 407 IN RHODOBACTER SPHAEROIDES CYTOCHROME c OXIDASE IS NOT REQUIRED FOR PROTON PUMPING OR MANGANESE BINDING Abstract Introduction Materials and Methods Site-Directed Mutagenesis Enzyme Purification Reconstitution of Cytochrome c Oxidase EPR Spectroscopy Resonance Raman Spectroscopy Mn Content Determination Results Visible Spectra, Electron Transfer, and CO Binding Assays Respiratory Control and Proton Pumping Mn and CuA Binding Sites Resonance Raman Spectroscopy Discussion References CHAPTER V ARGININE 482 IS IMPORTANT FOR STABILIZING THE INTERFACE OF SUBUNIT I AND SUBUNIT H OF RHODOBACTER SPHAEROIDES CYTOCHROME c OXIDASE Abstract Introduction Materials and Methods Site-Directed Mutagenesis Enzyme Purification Reconstitution of Cytochrome c Oxidase Stopped-Flow Proton Pumping Assay X 102 102 108 112 114 115 116 120 120 120 123 123 123 123 124 124 '130 130 137 140 145 147 148 149 153 153 153 153 156 SL'MM Protein Assay Pyridine Hemochromogen Assay Ruthenium Kinetics Assay Optical Spectroscopy Other Assays Results Visible Spectra, Electron Transfer, and CO Binding Assays ........................... Respiratory Control and Proton Pumping Purity and Heme Content Mn/Mg Site and Cu A Environment Analysis of Intrinsic Electron Transfer Rates Discussion References SUMMARY AND PERSPECTIVES 1. Asp132 Is Essential for Proton Pumping in Rhodobacter sphaeroides Cytochrome c Oxidase 2. Improvement of Expression and Purification of Cytochrome c Oxidase Is Important for Site-Directed Mutagenesis Studies 3. Asp407 Is Not Required for Proton Pumping or Mn Binding in Rhodobacter sphaeroides Cytochrome c Oxidase 4. Arg482 Is Important for Stabilizing the Interface between Subunit I and Subunit II, and Plays an Indirect Role in Proton Pumping .................. 5. Future Research References xi 157 157 157 158 158 159 159 165 168 169 175 183 188 191 191 192 193 193 195 197 Ch LIST OF TABLES Chapter II 1. Comparison of Wild Type and Mutant Rhodobacter sphaeroides Cytochrome c Oxidase: Electron Transfer Activity of Purified and Reconstituted Enzyme, Respiratory Control Ratio (RCR), Proton Pumping Efficiency, and CO Binding Efficiencym. Chapter IV 1. Comparison of Wild Type and Mutant Rhodobacter sphaeroides Cytochrome c Oxidase: Electron Transfer Activity of Purified and Reconstituted Enzyme, Respiratory Control Ratio (RCR), Proton Pumping Efficiency, and CO Binding Efficiencyw Chapter V 1. Comparison of Wild Type and Mutant Rhodobacter sphaeroides Cytochrome c Oxidase: Electron Transfer Activity of Purified and Reconstituted Enzyme, Respiratory Control Ratio (RCR), Proton Pumping Efficiency, and CO Binding Efficiencyw 2. Comparison of Wild Type and Mutant Rhodobacter sphaeroides Cytochrome c Oxidase: Ana/Am, [Heme A], [Protein], [Heme A]/[Protein], Purity, [aa3], [Heme A]/[aa,], Am/Aws, and Heme a, Content 3. Fast Kinetics Studies of Wild Type and Mutant Cytochrome c Oxidase .......... xii 127 160 169 180 Ch Cham LIST OF FIGURES Chapter I 1. Three-dimensional crystal structure of bovine cytochrome c oxidase at 2.8 A resolution 2. Crystal structures of the three common subunits of Paracoccus denitrificans and bovine cytochrome c oxidase 3. Structure of metal centers in subunit I and subunit II in cytochrome c oxidase 4. Important residues in subunit I of cytochrome c oxidase 5. Amino acid sequence alignment of subunit I in Rhodobacter sphaeroides, Paracoccus denitrificans and bovine cytochrome c oxidase ........ 6. Proposed proton input channels in the direct coupling model 7. The structure of the putative HZO/proton exit channel in cytochrome c oxidase Chapter II 1. Wild type amino acid sequence and oligonucleotide sequences of wild type and mutants at position 132 in subunit I of Rhodobacter sphaeroides cytochrome c oxidase 2. Comparison of visible spectra of purified wild type and mutant Rhodobacter sphaeroides cytochrome c oxidase 3. Respiratory control assay of reconstituted wild type and mutant Rhodobacter sphaeroides cytochrome c oxidase 4. Proton pumping activity assay of reconstituted wild type and mutant cytochrome c oxidase 5. Fatty acid titration on D132A activity xiii 10 12 29 35 73 76 79 81 Chapter III 1. Construction of the overexpression plasmid pJQ200 by subcloning of the COXI and COXII/III operons of Rhodobacter sphaeroides cytochrome c oxidase into expression vector pRK415-l ......................... 95 2. Comparison of the expression of Rhodobacter sphaeroides cytochrome c oxidase from the overexpressed construct with regular wild type strains, 2.4.1, Ga and CY91 101 3. Purification of histidine-tagged wild type Rhodobacter sphaeroides cytochrome c oxidase with a Ni-NTA affinity column ................................ 105 4. Comparison of the visible spectra of the regular construct of D132A with that in the overexpressed construct pJQ2OO and in the short COXI operon 107 Chapter IV 1. Structure of the interface between subunit I and subunit II of beef heart cytochrome c oxidase 118 2. Wild type amino acid sequence and oligonucleotide sequences of wild type and mutants at position 407 in subunit I of Rhodobacter sphaeroides cytochrome c oxidase 122 3. Comparison of visible spectra of purified wild type and mutant cytochrome c oxidase of Rhodobacter sphaeroides 126 4. pH dependence of activity of purified wild type and D407A cytochrome c oxidase 129 5. Proton pumping activity assay of reconstituted wild type and mutant cytochrome c oxidase 132 6. EPR spectra of purified wild type and mutant Rhodobacter sphaeroides cytochrome c oxidase measured at 110 K ........................... 134 7. EPR spectra of purified wild type and D407A mutant Rhodobacter sphaeroides cytochrome c oxidase measured at 10 K .............................. 136 8. Resonance Raman spectra of purified wild type and mutant cytochrome c oxidase 139 9. Structure of the interface between subunit 1 and subunit II of beef heart cytochrome c oxidase at 2.3 A resolution .................................. 144 xiv wu‘ Chapter V 1. Structure of the interface between subunit I and subunit II of cytochrome c oxidase 151 2. Wild type amino acid sequence and oligonucleotide sequences of wild type and mutants at position 482 in subunit I of Rhodobacter sphaeroides cytochrome c oxidase 155 3. Comparison of visible spectra of purified wild type and mutant Rhodobacter sphaeroides cytochrome c oxidase 162 4. pH dependence of activity and relative activity of purified wild type and R482? cytochrome c oxidase 164 5. Stopped-flow proton pumping assay on wild type, R482K and R482Q cytochrome c oxidase 167 6. EPR spectra of purified wild type and mutant Rhodobacter sphaeroides cytochrome c oxidase grown in high [Mn] and low [Mg] ....................... 172 7. EPR spectra of purified wild type and mutant Rhodobacter sphaeroides cytochrome c oxidase in high [Mg] and low [Mn] 174 8. Comparison of optical spectra] of wild type and mutant cytochrome c oxidase 177 9. Dependence on ionic strength of second-order rate constant of the reaction between Ru55Cc, and cytochrome c oxidase ......................... 182 XV ABBREVIATIONS A.U., absorbance units bR, bacteriorhodopsin CCCP, carbonyl cyanide m-chlorophenylhydrazone CO, carbon monoxide DCCD, N, N’-dicyclohexylcarbodiimide DEAE, diethylaminoethyl EDTA, ethylenediamine tetraacetic acid EPR, electron paramagnetic resonance ESEEM, electron spin echo envelope modulation FPLC, fast protein liquid chromatography FI‘ IR, Fourier transform infrared spectroscopy H‘le', ratio of protons pumped per electron transferred HEPES, 4-(2-hydroxyethyl)-1-piperazine-ethanesu1fonic acid Irn, imidazole kb, kilo-base pair(s) kDa, kilo—dalton PAGE, polyacrylamide gel electrophoresis PCR, polymerase chain reaction RCR, respiratory control ratio SDS, sodium dodecyl sulfate TRIS, Tris(hydroxymethyl)aminomethane TMPD, N, N, N’, N’-tetramethyl-p-phenylenediamine val, valinomycin xvi CHAPTER] LITERATURE REVIEW LITERATURE REVIEW 1. Background on Cytochrome c Oxidase 1.1. Introduction Cytochrome c oxidase (EC1.9.3.1) is the terminal enzyme in many bacterial and all mammalian respiratory chains. It utilizes the electrons that are ultimately obtained from food to reduce oxygen to water. Cytochrome c is the direct electron donor for cytochrome c oxidase. The chemical reaction is as follow: 8 Hf + 4 cyt-c2+ + 02 —> 2 H20 + 4 H; + 4 cyt-c3+, where H: represents protons taken up in the inside of the respiratory membrane and H; represents protons released to the outside of the membrane. The oxygen chemistry generates redox energy which is utilized to drive proton translocation from the cytoplasm to the periplasm for bacteria, and from the matrix to the intennembrane space of mammalian mitochondria (Wikstrbm et al., 1981). The electrochemical gradient derived from both proton consumption and translocation drives ATP synthesis via the FIFo adenosine triphosphate (ATP) synthase (Mitchell & Moyle, 1965). Cytochrome oxidase has a critical role in energy metabolism; therefore, deciphering the puzzle of this exciting enzyme is essential for understanding the molecular mechanism of energy transduction. Cytochrome c oxidase is a large metalloprotein. Bovine, and more recently Rhodobacter sphaeroides cytochrome c oxidases are two of the most extensively studied aa3 type cytochrome oxidases. Bovine cytochrome c oxidase has thirteen subunits (Figure 1) (Merle & Kadenbach, 1980; Kadenbach et al., 1983). The three largest subunits are encoded on mitochondrial DNA, and synthesized inside the mitochondrion. The remaining ten subunits are encoded in the nucleus and made in the cytoplasm. In contrast, Rhodobacter sphaeroides cytochrome oxidase only has three subunits (Figure 2), which are similar to those encoded by mitochondrial DNA, the three largest subunits in the bovine enzyme (Figure 2). Rhodobacter sphaeroides cytochrome oxidase also 2 Figure l. Threedimensional crystal structure of bovine cytochrome c oxidase at 2.8 A resolution. The figure was created with the computer program Rasmol using the coordinates from bovine oxidase (Tsukihara et al., 1996). All thirteen subunits are represented in different colors, with the subunit number in the same color as the subunit structure. Hm I Figure 2. Crystal structures of the three common subunits of Paracoccus denitrificans (A) and bovine (B) cytochrome c oxidase. Each subunit has a different color. The structures of subunits I, II and III of oxidase from these two species appear almost identical. The figures were created with Rasmol using the coordinates from Paracoccus denitrificans (Iwata et al., 1995) and bovine (Tsukihara et al., 1996) cytochrome c oxidase, both at 2.8 A resolution. B. bovine A. Paracoccus denim'ficans Figure 2 shots than subui asag been plasrr 12. S M11 1 them 7 shows similar electron transfer activity, proton pumping ability, and spectral characteristics, as the bovine oxidase; therefore, these three subunits are the major subunits for catalytic functions, and Rhodobacter sphaeroides cytochrome oxidase serves as a good model to study the mammalian enzyme. The genes encoding cytochrome oxidase subunits and accessory proteins have been deleted from the Rhodobacter sphaeroides genome and can be reintroduced on plasmids, making it possible to generate mutants by site-directed mutagenesis. 1.2. Structure and Function of Cytochrome c Oxidase A great achievement in the bioenergetics field is the successful crystallization of both the bovine and Paracoccus denitrificans cytochrome c oxidases and the solution of the structures to atomic resolution (Figure 1, Figure 2). Since Paracoccus denitrificans is very clOsely related to Rhodobacter sphaeroides, the latter’s three dimensional structure can be excellently predicted from the former by homology modeling. The large size of this protein and the membrane-bound nature has made crystallization an extremely difficult task. Earlier efforts resulted in a 2.5 A structure of only the soluble domain of the subunit II in a quinol oxidase with an engineered dinuclear copper center (Wilmanns et al., 1995). However, after many years of effort, in 1995, Yoshikawa and coworkers (T sukihara et al., 1995) and Michel and coworkers (Iwata et al., 1995) independently solved the structures of the mammalian and bacterial versions of this enzyme almost simultaneously. The fully oxidized Paracoccus denitrificans structure has been solved in the presence of azide and a specific antibody fragment, while the bovine oxidase has been solved in the oxidized native state. Remarkably, the two enzymes are amazingly similar in the structure of the common subunits, even though these two species are not closely related to each other (Figure 2). The structures are also in reasonably good agreement to the topology predicted by hydropathy analysis. In 1996, the complete structure of bovine oxidase was solved to the resolution of 2.8 A (Tsukihara et al., 1996), which shows the detailed structure of the thirteen subunits. In 1997, Michel and co-workers q...“ 501w 16501 mole hour ache in th ml the far {151310, elem! Mpin 8 solved a two-subunit Paracoccus denitrificans cytochrome c oxidase structure at 2.7 A resolution (Osterrneier et al., 1997), which includes 52 tentatively defined water molecules. In 1998, Yoshikawa and coworkers solved fully oxidized and fully reduced bovine heart cytochrome c oxidase to 2.3 A resolution (Yoshikawa et al., 1998). These achievements play a very important role in furthering our understanding of the key issue in the field, which is, “How is the electron transfer process coupled to proton translocation across the membrane?” 1.2.1. Metal Centers Cytochrome c oxidase is a metalloprotein, containing four redox active metal centers, heme a, heme a 3, CuA and CuB, confirmed by both bacterial and mammalian oxidase crystal structures (Figure 3) (Iwata et al., 1995; Tsukihara et al., 1995). The bovine oxidase also contains one zinc, one magnesium, (Einarsdottir, 1995) and one calcium, resolved by the crystal structures from both groups (Tsukihara et al., 1995; Ostermeier et al., 1997; Yoshikawa et al., 1998). Heme a, heme a3 and CuB are located in subunit 1, with the latter two forming a binuclear active center where 02 is reduced to H20. The heme A porphyrin ring has formyl and famesyl substituents. The function of the famesyl tail is unknown, but has been variously suggested to play a role in proton translocation (Woodruff et al., 1991), to orient the heme in the protein, and to facilitate electron transfer between heme a3 and CuB (Caughey et al., 1975). A function in proton pumping is not likely, since another oxidase in Paracoccus denitrificans has proton pumping ability but does not contain a famesyl group on the heme (Wikstrbm et al., 1981; Woodrufi‘, 1993). Heme a is ligated by Hile2 and His421, heme a3 by His419 and Cu]; by Hi5333, His334, and H15284 (Figure 4) (the numberings in this thesis are Rhodobacter sphaeroides numberings, unless otherwise noted); all of the ligands are totally conserved in subunit I of heme-copper oxidases (Figure 5) (Calhoun et al., 1993; Hosler et al., 1993). The liganding structure was demonstrated by mutagenesis and Figure 3. Structure of metal centers in subunit I and subuzit H in Qtochrome c oxzdaae. Subunit I (dark lue and heart 11 (light flue); are represented as a Ca- backborie trace. Cu, (yellow baIIsi. heme a {red}. heme a, Ire-ii. and Cue (yellow ball; are the four redox active metal centers. Mg (magenta ball: is a redox inactive metal center. Cu, is a dinuclear center locate in subunit [1; the rest of the metal centers are located in subunit 1. The figure was created using the coordinates from 2.8 A resolution bovine oxidase crystal structure (Tsu -ihara et al.. 1996). Mg CUB heme a3 Figuire 3 11 Figure 4. Important residues in subunit I of cytochrome c oxidase. The CuB (blue ball) ligands H284, H333 and H334 are in yellow; the heme a 3 (red) ligand H419 and heme a (red) ligands H102 and H421 are in green; the Mg (yellow ball) ligands H411, D412 and E254 (subunit II) are in magenta, magenta and cyan, respectively; the Mg also has three liganding water molecules, which are not shown in this figure. D132 (yellow) and E286 (cyan) have been hypothesized to be involved in a pumped proton channel, and K362 (blue) has been suggested to play a role in the proposed substrate proton channel. R481 (gray) and R482 (blue) have been predicted to be important for electron transfer between Cu A and heme a. The three positions at which I introduced mutations are D132 (yellow), D407 (blue) and R482 (blue), with underlined names. This figure was created using the coordinates from 2.8 A resolution bovine cytochrome c oxidase (Tsukihara et al., 1996). 12 3’ we. K362 Figure4 13 Figure 5. Amino acid sequence alignment of subunit 1 in Rhodobacter sphaeroides, Paracoccus denitriflcans, and bovine cytochrome c ox1dase. All the important amino acids, H102, H284, E286, H333, H334, K362, H411, D412, H419, H421, R481 and R482 (see Figure 4), are colored green except that D132, D407 and R482 are in red. All of these residues are highly conserved among these three species. :MVt :’I‘)\I.S 1.1.1 RAI'ZI A SLJP [JVC‘IIJ I .‘-'VAl-"l‘ VYMI'IMI‘II .MA!’ I V63] . I .‘IV( 'I‘"P VYM](MI'ZL.QIIIV’ WA! :A 25 .i'I‘NIIKI) I(Jll;Yl.F"I‘/\(i ‘1 . “.‘Z.)"1'NIIKIV) If I'I‘I.YI .IJ-‘KI ICWF'MII'W‘NIIKI) 1(IVI.YI.F"IY_I(J J, Ii WI-‘M: 3" W! I 1’ MAI’AA VH( 311(‘1 I") ll-l1)’l'l(( u‘F‘l'"I‘ 21 . MF I N 5 . . ll!)l€l?l JI"I"7' MA I ’AA 1 I It‘ll-1": 1 1 1. 1‘? l J"! L! h .w I l..- nu v 1" no I 14 m 2&3 ....MAZ>>B nan m.mmmfim>mm mUGZAZNAZB m diddQBQmmx Ova mdfimmfimwmm mmmABBMABD m Cadmmwnmmx nun maomhfimwmm mmBABZMABD BB...AQ>BQ >mmxm2A2> Dm>mmmZUmq émmz3>2m>2 >mx0¢md9>m Dan >HOHmmdmm¢ mmH>m> DHMmmmzwom dmm23>22¢9 >MW Dun H>0thamm< mmdmdwamm> Dav EZZBflmflwmw QHMmmmzwom Gammommmfiz Z>U>WZHh nan OQZH¢E>EQZ md>hxm3¢>> GammommmHQ ZmOHRZZZLm 00v AOO<3mm>Om UmZ¥OH3>>> ONQ OCLHOE>EUJ m2>>xmm¢>> Gammommmfiq Zh223hm but AMUfiZflmwom OmzwaOmfim bM‘ UflmHUh>¢UJ m2>>xmm<§> >>BDIA>HQA mmZHOBA nmn 00>BhamHm0 AfiZZZ¢mm3M nan HZGGEJEdAZ mh>¥>UEmH4 >>EDI>>¢DA m40mfl>>09> OQM 00>Bhdmdm0 m<342m5xmm GUM HmUUZZB

¥HUEm>€ >>BO$>>¢D> m40mq>HUF> Dan 00>Bmdmdm0 AflZAZmfixam Dhn HmOOSZE

E HHZB¢ mBQ>QZO>Bm MON 23:43>Hm04 mOHWZS<3>S MEN 02>Ommmxxw m>>B>H$mHZ HBZBdASth OOEAWSU QNM me<3>>b04 HOH¢¢EZ COM mdhmemxxd .mBMH>:mHH H>25422m>m OOBAmAO>mwa >UH4>2<>>E Dan mA>OhHm3M¢ .bE¢H>mm>H UmUmAHAH>> waxwmhzmqm MMN ONAHQDUOOd QOhBBZAZM MHN QBAASEHUdd fibmqudqd> OmquHHH>> ummwmmZJHm GUN ONA>QQOOO< mammOEUmzm 6‘" QZAAZBHmAmQAH43 OmémA>HHW> mmethAHm hhfl O>A>DDDOOW mammfifiwmzm hm" DBQAZBHQAM3>LA m90>0m24mm HRH MZZHHBBHmZ HdUJHmm>U¢ and ASAWEHBAQ> m<0<34420¢ 49Hh>m3¢md m>¥mABZUm< 0°" MZZAEEEHHZ HflUJHmmm>U¢MB..Bm ¢B>hHm3¢mA m>MSZBZUm< bdfl KZZAmBBHZZ HdeHmmdwm had qm>>93090 £0 ..... ¢m> Odd mefl320m AQQWQ>30>O m0200200m< d'd Adm<>0qd>0 U>>23>mdzz «NH AdmtdZdew HIAQZE>ZOE Ammwn>3OHU WUJOOZOOmd and bam¢>ddm90 ¢>>A3>quz and Examdzdmdw HEAQ2N>ZOm 00HZHQ2>ZE mH2>h¢m¢B> on >>Z>HOQQO. .......... av ................ AABU UOhACQH>>h mZZA>Um>BH Ha S>234x02m6 UNdmdfldem db 4 ......... .ONAUZ>O>U OUbA4mH>>h EZZAHOIGEH ma 2>23A302m9 UZM>40mm amN402m0>U mowdmdeqd mA¢BO>ZU<3 an <0hAA>ABUH meZBmmABm n ZHNZ ................ mmoqm2m2>> BmU>MHAO>H H. UAHUH ammZBWZNSm flu maOMBDme OmOm>E> Bm<>mHAU>Q an OUEEANA>UH 0M225m2h3m OH Emmwmmam.. NSOZHfldddz cue mefl>om can atom nan nemnm an. mcfl>om one n-ee on. nomnm nun mcw>om coo n-ea one samnm nun mcfi>om a.» n-ea can nemnm nna mcw>om «on n-ea pan nemnm mad mcfi>om «nu n-ea pnu sauna and mcfl>om «pa n-ea apu gamma up mcfl>om an” p-09 add annex n. mcw>om no n-ee an enmnm H wcfi>om « n-ee a sauna 5P be: the 15 spectroscopic studies (Calhoun et al., 1993; Hosler et al., 1993) and confirmed by three dimensional X-ray crystal structure (Iwata et al., 1995; Tsukihara et al., 1995). Both hemes are situated approximately 13 A below the cytosolic side of the membrane, but there is a 108° angle in between them. They are perpendicular to the plane of the membrane. Since the shortest distance between heme a and heme a3 is 4.7 A, one possible electron transfer pathway is a direct pathway from heme a to heme a 3 via the heme edges. For many years it was a controversial issue as to whether the Cu site contained one or two coppers, but recent EPR and metal analysis studies suggested that CuA was a 2 Cu center, which was confirmed by the crystal structures (Figure 3, Figure 4). The Cu- Cu distance is 2.6 A. The Cu site is a dinuclear center that acts as a one-electron acceptor (Steffens et al., 1993). The dinuclear copper center is similar to that of nitrous oxide reductase (NZOR), and is predicted to have evolved from the latter enzyme (van der Oost et al., 1994). It is generally considered that electrons are transferred from soluble cytochrome c, which binds to subunit II, to CuA, then from Cu to heme a, and finally to the heme arCuB center (Hill, 1993). The Cu-Cu dinuclear center is ligated by six ligands, HisZ60, Met263, Hi5217, Glu254, Cy5252 and Cy3256 (see Figure 1 in Chapter V), all of which are in subunit II, as originally predicted by sequence homology and mutation on the Paracoccus denitrlficans enzyme CuA domain. Interestingly, both of the crystal structures reveal a bridging amino acid, Glu254, between the Cu site and Mg site. The Cu, and Mg sites are situated at the interface of subunit I and subunit 11 (Figure 3), and the shared ligand of these two metals might suggest that they are important for the structural stability at the interface. As predicted by mutagenesis (Hosler et al., 1995) and confirmed by the crystal structures (Tsukihara et al., 1995; Ostermeier et al., 1997), Mg has two more ligands in subunit 1, His411 and Asp412, and at least one water molecule (Espe et al., 1995). Three water molecules are observed to bind Mg in the later high resolution structure of beef '- ‘fl‘fl' hea dun EPR al., 1 P101) et al. al., I Whicl Mata the SI 1998} 16 heart oxidase (Yoshikawa et al., 1998). Mn competes for this site with Mg if present during growth. The Mg/Mn site is redox inactive in oxidase, but serves as an excellent EPR sensitive probe of the Mn/Mg region, as is also the case for the CuA site (Hosler et al., 1995). The significance of the Mn/Mg center is still controversial. Recently, it has been proposed that magnesium might play a structural role in stabilizing the enzyme (F lorens et al., 1998). A water channel is proposed, based on the crystal structure (Tsukihara et al., 1996), and the Mn/Mg site is located at the beginning of the proposed water channel, which could also be a proton exit channel, above the heme a 3-Cu3 binuclear center. Mutagenesis studies of the Mg ligands are consistent with a role for the metal site in the structure and function of the proposed water / proton exit channel (Florens et al., 1998) A calcium site has been identified in both the mammalian and bacterial oxidases by Yoshikawa and Michel's group. It is located at the loop between helix 1 and helix [1 in subunit 1. However, the assignment for the ligands for the calcium is somewhat different. From the bovine oxidase crystal structure, Ca binds to carbonyls of G1u54, Gly59 and Ile484 (Ser441 in bovine), a water and the carboxyl of G1u54. But in Paracoccus denitrificans, Ca is coordinated by the backbone carbonyl oxygen of Glu54, Ala57 (HisS9 in Paracoccus denitrificans) and Gly59 as well as the side chain oxygen of Glu54 and Gln61 (Ostermeier et al., 1997). Therefore, only Glu54 and Gly59 are the common ligands. This is a new metal discovered in cytochrome c oxidase whose function is unknown. In the mammalian oxidase, zinc is coordinated by four cysteines (Cys60, Cys62, Cys82 and Cys85, bovine numbering) in subunit Vb forming a tetrahedral zinc site. There is a zinc finger motif from Cys60 to Cy582, but the overall structure is not a typical zinc finger. The function of zinc is not clear (Tsukihara et al., 1995). tonal Para “iii a all? 31 Ifrrnir. Sfmici most i mil t We: Ct’ems and “it its Ch; mar-01’]: 3%: it A .hlSllc 17 1.2.2. Subunits of Cytochrome c Oxidase Rhodobacter sphaeroides cytochrome c oxidase has only three subunits, with a total molecular weight of 125 kDa (Hosler et al., 1992). Since it is highly homologous to Paracoccus denitrificans oxidase, we make the assumption that the three dimensional structures of these two are essentially identical in most respects. Although the beef heart oxidase has thirteen subunits, with a total weight of 204 kDa (Buse et al., 1978; Suarez et al., 1984), the crystal structures of the catalytic core of these two species are remarkably similar (Figure 2). Therefore, I will first describe the bacterial oxidase and indicate the additional features of the more complex bovine oxidase later. 1.2.2.1. Bacterial Cytochrome c Oxidase 1.2.2.1.]. Subunit I Subunit I is the largest submit in Rhodobacter sphaeroides cytochrome c oxidase, with a molecular weight of 62.1 kDa. It consists of twelve transmembrane helices, which are at an angle of 20°-30° to the vertical line from the membrane plane. Both the N- terminus and C-terminus are on the cytoplasmic side. The twelve helices form three semicircles with a quasi-threefold axis of symmetry (Iwata et al., 1995). Subunit I is the most important subunit functionally, since it contains three out of the four redox active metal centers (heme a, heme a3 and CuB) (see section 1.2.1 of this chapter). A variety of studies have indicated that the proton pumping mechanism is driven by oxygen reduction events at the heme a 3-CuB binuclear center. Several models including “direct coupling” and “indirect coupling” models have been proposed and are discussed in section 2.3 of this chapter. One of the CuB ligands, His333, has been proposed to have at least two conformations, since this residue was not resolved in the X-ray structure of the oxidized, azide bound form (Iwata et al., 1995). This histidine residue is suggested to be the key to a "histidine shuttle model" - a model of direct coupling of electron transfer and proton sec in 613 18 translocation (Morgan et al., 1994; Iwata et al., 1995). More details are discussed in section 2.3.1 of this chapter. A covalent linkage between another CuB ligand, Hi3284, and the nearby residue Tyr288 is seen in the bovine and Paracoccus denitrzficans crystal structures (Ostermeier et al., 1997; Yoshikawa et al., 1998). This tyrosine is also hydrogen bonded to the OH group of the famesyl moiety of heme a3, and possibly to the bound 0; at heme a3, suggesting it may be involved in the oxygen reduction mechanism (Yoshikawa et al., 1998). Based on this unique structural characteristic and time-resolved resonance Raman studies, Proshlyakov and coworkers proposed a mechanism which involves a tyrosine radical in the oxygen reaction (Proshlyakov et al., 1998). A recent crystal structure from Yoshikawa's group also suggests that there is a peroxide bridging heme a3 and CuB in the fully oxidized “resting” form (Yoshikawa et al., 1998), a controversial interpretation (Gennis, 1998). 1.2.2.1.2. Subunit II Subunit 11 consists of an N-terminal external loop, two transmembrane helices and a C-terminal soluble domain located on the outside of the membrane (periplasmic side) (Figure 2). The molecular weight is 32.9 kDa. The extrarnembrane domain contains ten [3 sheets forming a barrel. Both the N- and C- terminals interact tightly on the periplasmic side. The Cu site is located in the external C-terminal domain (Iwata et al., 1995). Cytochrome c binds to subunit II and donates electrons to Cu. Previous studies suggest that there are high affinity and low affinity cytochrome c binding sites (Smith et al., 1977; Ferguson-Miller et al., 1978). Mutational studies confirm the importance of carboxylic residues in cytochrome c binding (Zhen et al., 1997) (more detail are discussed in section 3.3). The role of Cu in electron transfer reactions has been an important issue. Mutagenesis and spectroscopy studies have made it clear that Cu is the initial electron entry site and plays an important role in electron transfer reactions (see more discussion in section 3.1 of this chapter). 11.2 l9 1.2.2.1.3. Subunit III Subunit III consists of seven transmembrane helices, with a molecular weight of 30.1 kDa (Figure 2). The N-terminus is on the cytoplasmic side and C-terminus is on the cytosolic side. It does not contain a redox active center. The seven transmembrane helices form two bundles. Helix I and II are separated from helix III-VII by a V-shape cleft. A lipid molecule is resolved in the clefi in the crystal structure of the oxidase, which is proposed to be a docking site of other membrane proteins, such as membrane- bound cytochrome c552. Subunit III has been considered to be involved in proton pumping activity, since when Glu98 located in the middle of transmembrane helix III is modified by DCCD, the enzyme loses proton pumping activity (Casey et al., 1979; Casey et al., 1980). However, recent mutagenesis studies have demonstrated that mutations introduced to Glu98 and another conserved carboxylic acid residue Asp259 have no effect on proton ptunping efficiency or electron transfer activity (Haltia et al., 1991). In fact, removal of subunit 111 does not eliminate proton pumping (Thompson et al., 1985), nor change the response of oxidase to proton or charge gradients (Gregory & Ferguson-Miller, 1988). Although subunit III is not essential for the normal function of the enzyme, it may be important to stabilize or assemble the mature oxidase complex. 1.2.2.2. Beef Heart Cytochrome c Oxidase The bacterial oxidase is a heterotrimer that does not further self associate into dimers in the crystal structure, while the mammalian oxidase is made up of one each of 13 different subunits that organizes further into a dimer in the crystal form. Whether this dimer has functional significance is not clear (Suarez et al., 1984; Ferguson-Miller & Babcock, 1996; Tsukihara et al., 1996). 1.2.2.2.1. The Core Subunits (I, II, and III) The structures of the three largest subunits in the beef heart oxidase enzyme are amazingly similar to those of Paracoccus denitrificans. However, there are several differences. First, Hi3333, the ligand of CuB that was not resolved in the original bacterial 3"? nt Essen Me: an} 20 crystals, is observed in the electron density map of all forms of the bovine oxidase, indicating that it assumes only one conformation. Second, a crystal structure of a reduced form of the mammalian cytochrome oxidase shows a conformational change in a loop including residue Asp51 (bovine numbering), compared to the oxidized form. Based on this result, an indirect coupling model (see section 2.3.2 of this chapter) for proton pumping has been proposed, involving movement of Asp51 (Yoshikawa et al., 1998). Although Asp51 is conserved in animals, it is not conserved in plant and bacterial oxidases. An evolution in the proton pumping mechanism has been suggested (Yoshikawa et al., 1998). Third, two possible nucleotide binding sites have been discovered, which are occupied by cholate. This lends support to theories that ATP binding affects electron transfer activity and proton pumping efficiency (T aanman et al., 1994; Frank & Kadenbach, 1996). 1.2.2.2.2. The Remaining 10 Subunits The nuclear encoded subunits cover most of the surface of the core subunits but are not in direct contact with the binuclear center; therefore, they do not appear to be essential for the catalytic function of the oxidase. A role in stabilizing the protein and protecting against release of radical intermediates has been suggested. It has been shown that smaller subunits are involved in regulation of the oxidase (Kadenbach et al., 1991). Subunits VIa, VIb, and We have domains above the binuclear center and near the cytochrome c binding site, as well as a proposed water channel. Therefore, they might have some effect on cytochrome 0 interaction with cytochrome c oxidase, and/or water/H" exit (Hiither & Kadenbach, 1988; Ferguson-Miller & Babcock, 1996). Both subunit VIa and VIb interact with the other monomer of the oxidase dimer; thus, they are likely involved in stabilizing the dimeric structure (Yoshikawa et al., 1998). 2. Pt oxide know rm; 21 2. Proton Transfer Pathways 1n the respiratory chain of eukaryotes and many prokaryotes, electrons are transferred fiom NADH to 02 through NADH dehydrogenase, ubiquinol cytochrome c oxidoreductase (bc, complex) and cytochrome c oxidase. These three enzymes are also known as complexes 1, III and IV, respectively. As electrons flow through, protons are pumped across the membrane by these complexes. Proton translocation across the membrane can have three forms. One is called the “carrier” mechanism, where a proton will attach to a carrier molecule and be transferred. Another one involves conformational change; a transmembrane protein will pick up a proton and then go through a conformational change, which takes the proton to the other side of the molecule and releases it. A third form is transfer through a proton channel, which is the most favored model in cytochrome c oxidase . A hydrogen bonding network of amino acid side chains may be involved in such proton translocation, and water molecules may also be involved (Nagle & Morowitz, 1978). 2.1. Proton Wire Theory In the 19603, Onsager suggested that the side chains of some amino acid could form a hydrogen bonded chain (HBC) for the transport of ions through membranes (Onsager, 1969). These amino acids include Ser, Thr, Tyr, Glu, Asp, Gln, Asn, Lys, Arg, His; water molecules can also participate. The HBC theory is based upon dielectric and conductivity studies in ice, which is a most thoroughly studied hydrogen bonded crystal. Morowitz first made the connection between hydrogen bond chain and the proton translocation mechanism, and proposed a “proton wire” theory (Morowitz, 1978). “Proton wire” theory uses the simple example of a hydrogen bonded chain made of serine residues (Nagle & Morowitz, 1978). The theory includes hop/tum two alternating processes of proton transfer. It was proposed that a long hydrogen bond would make proton transfer step (“hop”) difficult, but would make it easy for the proton 10 r: 2.2. on pum £51311 H250 emal 22 to reorient away from the bond (“turn”). When the bond is short, it is easy to “hop” but difficult to “turn”. 2.2. Bacteriorhodopsin - A Model Proton Pumping System for Cytochrome c Oxidase Cytochrome c oxidase uses redox energy from the oxygen reaction to drive proton pumping (Babcock & Wikstrom, 1992), while bacteriorhodopsin uses conformational change induced by light absorption to drive proton pumping. Bacteriorhodopsin (bR) is a 26 kDa protein found in the purple membrane of Halobacterium halobium. Upon illumination with visible light, a retinal molecule that is covalently attached to the protein through a protonated Schiff base linkage to Lys216 goes through a conformational change, which triggers proton translocation from the cytoplasmic side to the extracellular side of the membrane, providing a chemiosmotic gradient that drives ATP synthesis (Rothschild, 1992; Gai et al., 1998). A high resolution structure of bacteriorhodopsin has been determined based on electron cryo-microscopy (Henderson et al., 1990) and extensive mutational studies were carried out to identify the residues that might be involved in proton pumping (Krebs & Khorana, 1993). Rothschild and coworkers proposed a comprehensive model of the proton pumping mechanism involving a hydrogen-bonded network (Rothschild, 1992), based on the coordinates from a 3.5 A two-dimensional crystal structure (Henderson et al., 1990) and on F TIR studies. Several hydrophilic residues and water molecules (in order: Asp96, H20, Thr46, Thr89, H20, Tyr185, Asp212, Schiff base, Asp85, Arg82 and H20) are considered to be connected by hydrogen bonding interactions and to play critical roles in proton translocation from the cytoplasm to the extracellular space. The photocycle consists of six states, bR57o, K630, L550, M412, N550 and 0640 (the subscript numbers refer to spectral characteristics). Light causes isomerization of the retinal chromophore from all-trans to 13-cis; this conformational change induces the transition from bR57o state to K630 state. lht in t has: [OW Tyr trans retin.c AspS result item 23 The hydrogen bond between Tyr185 and Asp212 breaks due to a conformational change in the Schiff base. In the transition from L550 to M412, Asp85 is protonated by Schiff base, leaving the deprotonated Schiff base free to form a hydrogen bonded network toward the inside of the membrane in the order Asp96 : Thr46 : water : Thr89 : water : Tyr185 : Asp212 : Schiff base. Toward the outside of the membrane, the protonation of Asp85 releases Arg82 from its charge interaction with Asp85. A water at the exit site releases a proton and stabilizes the Arg82 with a hydroxyl ion. From M412 to N550, Asp96 indirectly reprotonates the Schiff base via the hydrogen bonded network. In the transition from N550 to 0640, Asp96 is reprotonated from the cytoplasmic surface and the retinal is converted from l3-cis to all-trans form. During the 0640 to bRs-m transition, Asp85 reprotonates hydroxide to water and all the residues are reset to the ground state. Recently, high resolution two-dimensional (3.0 A (Kimura et al., 1997)) and higher resolution three-dimensional (2.5 A (Pebay-Peyroula et al., 1997)) structures of bR have become available, which gives hope for further refining the proton translocation mechanism in transmembrane pumps. Recent studies suggest the step-by-step mechanism (Rothschild, 1992) of the photocycle needs modification. It was found by spectroscopic titration that the pK, of Asp85 is linked to that of Glu204 (Brown et al., 1995; Balashov et al., 1996; Richter et al., 1996a; Richter et al., 1996b). Photoisomerization of retinal causes a conformational change, which makes the pKll difference between the Schiff base and Asp85 narrow from 5 to <1. In the early phase, as Asp85 is protonated by the Schiff base, the linkage between Asp85 and Glu204 causes the pK. of Glu204 to drop from 9 to 5. Based on kinetics of absorbance changes, it seems that Glu204 protonates Glu194, and the latter releases a proton to the extracellular surface. This, in turn, causes the pKll of Asp85 to increase and makes it unfavorable to transfer a proton from Asp85 back to the Schiff base. Next, the Schiff base enters into a hydrogen bond with Asp96, probably through water molecules; this is induced by change in local geometry. A further protein conformational change causes the pK, of Asp96 to is n phat malt one fc (Fergu- subsm 24 decrease from 11 to 7, thus it becomes a proton donor to the Schiff base. Finally, retinal is reisomerized followed by proton transfer from Asp85 to Glu204, completing the photocycle. The pK. of Asp85 and Glu204 are returned to 2.5 and 9, respectively, making the reaction irreversible (Lanyi, 1997). This mechanism clearly shows that the proton loading and unloading are stringently regulated at the active site, and the changes in ApK, between proton donor and acceptor groups ensure a cytoplasmic-to-extracellular direction for proton translocation. 2.3. Models of Proton Pumping Coupled to Electron Transfer Two mechanisms of coupling of electron transfer and proton translocation have been proposed: a direct coupling model and an indirect coupling model. In the direct coupling model, two separate proton input channels were pr0posed, one for substrate protons and one for pumped protons, as well as one proton exit channel (Ferguson-Miller & Babcock, 1996). The substrate proton channel would translocate substrate protons to the binuclear center to reduce 02 to H20; in the meantime, the pumped proton channel would also transfer protons to the binuclear center, after which they would leave the binuclear center through a proton exit channel. The product, H20, was proposed to use the proton exit channel to leave the active center (Ferguson-Miller & Babcock, 1996; Tsukihara et al., 1996; F lorens et al., 1998). In the indirect coupling model (Tsukihara et al., 1996), oxidase was proposed to have a substrate proton channel, similar to the direct coupling model, but two “remote” proton exit routes away from the binuclear center. Therefore, the pumped proton channel would not be coupled to the 02 reaction directly at the binuclear center. The conformational change caused by the 02 reaction would be the driving force for proton pumping in the indirect coupling model. 2.3.1. Direct Coupling Models Chan and co-workers have suggested that a redox active metal could accomplish proton pumping by undergoing ligand exchange, wherein the protonation state of the {lwat there mecia Wiksti “Fake “0.903 pro1305 M561 313163, 25 ligand changes (Chan & Li, 1990). Based on this idea, Woodruff et al. have suggested a “ligand shuttle” mechanism upon binding of exogenous ligands at the binuclear center (Woodruff, 1993). Experiments need to be carried out to test this model. Rousseau er al. have also proposed a “ligand exchange” model, and have suggested an exchange of the proximal heme a3 histidine ligand (His419) with a tyrosine (Tyr422) (Rousseau et al., 1993). However, mutagenesis studies demonstrate that Tyr422 is not a critical residue in the electron transfer and proton pumping mechanism in cytochrome c oxidase (Mitchell et al., 1996), although its location is close to heme 03, as confirmed by the crystal structure (Iwata et al., 1995). Recently, several well-developed models are proposed, all based on the redox-linked ligand exchange theory, but different residues have been proposed to be important in each model. More details are discussed. ’- 2.3.1.1 “Histidine Cycle” Model The original model for the “histidine cycle” is a directly coupled proton pumping mechanism (DCP model) proposed by Wikstrom and Krab (Wikstrbm & Krab, 1978; Wikstrbm & Krab, 1979). It suggests that proton binding and the directionality of proton uptake and release is controlled by redox-linked ligand exchange at one metal center as also proposed by Chan and coworkers (Gelles et al., 1986; Chan & Li, 1990). Based on the DCP model, a more refined “histidine cycle” model has been proposed by Morgan et al. (Morgan et al., 1994). His284 (Rhodobacter sphaeroides numbering) was proposed to cycle between imidazolium (Ime) and irnidazolate (Im') states, and two protons would be pumped in each cycle. Two separate proton uptake channels are proposed, a substrate proton channel and a pumped proton channel. First, His284 in the irnidazolate form binds to CuB. Then CuB is reduced, and a negatively charged oxygen intermediate binds to heme a 3, To balance the negative charge, two pumped protons are taken up from the matrix, and Hi3284 is converted to the imidazolium state. It dissociates from CuB but retains some electrostatic interaction with the negatively charged oxygen intermediate. At this point, two substrate protons are mien neede lCllll'l' “lllSllt denirr ('His32 pumpi seen ll H1531 and in binucle htstidir State 0] [0 the . P‘t’nper *9 list to HOWEVQ “Chan. “Wm l 26 taken up and a water molecule is formed. The electrostatic stabilization is no longer needed; therefore, two protons are released to the other side of the membrane, and Hi5284 returns to the imidazolate state and binds to CuB as a ligand. Then this cycle begins again. Electroneutrality at the binuclear center is the key in this model. Iwata et al. have revised the “histidine cycle” model and have proposed a “histidine cycle/shuttle” mechanism based on their crystal structure data on Paracoccus denitrificans cytochrome c oxidase (Iwata et al., 1995). They propose that His333 (His325 in Paracoccus denitrificans numbering) may play a critical role in the proton pumping mechanism, because the electron density of the side chain of this residue is not seen in their structure, which could indicate multiple conformations. They propose that His333 shuttles between two conformations, and cycles through irnidazolate, imidazole and imidazolium states. The mechanism also obeys electroneutrality at the heme a3-CuB binuclear redox center. However, the scheme is slightly different from that of the histidine cycle model. It first takes up two pumped protons to form the imidazolium state of His333. Then a third pumped proton is picked up, and two protons are released to the outside of the membrane. Two substrate protons are then taken up, and the last pumped proton is bound to the imidazole form of His333 to form imidazolium. Finally, the last two substrate protons are taken up and the final pumped protons are released. However, Tsukihara et a1. resolve His333 in both the oxidized and reduced bovine oxidase structures (Tsukihara et al., 1996; Yoshikawa et al., 1998). In addition, 2.7 A Paracoccus denitrificans oxidase structure also resolves this residue, arguing against the two conformation theory at this position (Ostermeier et al., 1997). It is also now clear that the first proton taken up is not a pumped proton, but a substrate proton, according to the mutagenesis studies on Lys362 (Hosler et al., 1996; Konstantinov et al., 1997). However, the concept of electroneutrality as the driving force seems to be the common mechanistic feature for direct coupling of electron transfer and proton translocation. This concept must also be considered in indirect coupling, but has received little attention. of St cyioc mun E36 {Verk hum unfit ADVOIVI fight hl'dmg remain 27 2.3.1.2. Glu286 and Glu286 Flipping Model Glu286 is the only highly conserved acidic residue in the transmembrane domain of subunit I (Figure 4). Based on the crystal structure of Paracoccus denitrificans cytochrome oxidase, Glu286 is buried in the membrane and has been proposed to be involved in proton translocation (Iwata et al., 1995). Mutagenesis studies show that E286A and E286Q both block electron transfer activity and proton uptake in E. coli (Verkhovskaya et al., 1997) and in Rhodobacter sphaeroides (Adelroth et al., 1997). Interestingly, replacing Glu286 with cysteine yielded 15% of wild type electron transfer activity but a loss of proton pumping ability. E286D can translocate protons with the same efficiency as wild type (Verkhovskaya et al., 1997). This strongly suggests that Glu286 is a critical residue in proton translocation. There is also strong evidence for involvement of Aspl32, Asn121, Asn139 and Glu286 in the proton pumping channel (Figure 6). Although the crystal structures of bovine and Paracoccus denitrificans can trace a hydrogen bonded network from Aspl32 to Ser201 between helix II and helix III, a gap remains in the region between Ser201 and Glu286 (Figure 6). Water molecules have been proposed to connect this region (Iwata etal., 1995). Recently, several bound water molecules have been predicted using a "Potentials of Mean Force" (PMF) calculation in conjunction with energy minimization based on the coordinates from the bovine heart cytochrome oxidase crystal structure (Riistarna et al., 1997). Three water molecules have been found to connect Ser201 at the top of D channel to Glu286 with PMF approach, consistent with the model Iwata et al. propose. If electron transfer is directly coupled with proton pumping, and if the "histidine cycle" model is true, one would expect to see a network between Glu286 and one of the Cu]; ligands. FTIR results on E286D and E286C suggest that Glu286 is protonated, and there is connectivity between Glu286 and one of the CuB histidine ligands (Puustinen et al., 1997). They suggest that critical histidine ligand is His334, since it is the closest ' ”u:- 28 Figure 6. Proposed proton input channels in the direct coupling model. In the putative pumped proton channel, a hydrogen bonded network can be traced from Aspl32 to Ser201, but there is a gap between Ser201 and Glu286. In the putative substrate proton channel, Ser299, Lys362 and Tyr288 are shown. The two channels are predicted to be directly coupled at the heme a3-Cul3 binuclear center. This figure was created with the bovine cytochrome c oxidase coordinates at 2.8 A resolution (Tsukihara et al., 1996). 29 pumped [1+ channel substrate 11+ channel Figure 6 16514 Glu hydi (Yo: still 1 mole [0 [Cf water 30 residue to Glu286. PMF calculations have predicted three water molecules connecting Glu286 and His334 (Riistama et al., 1997). Although both high resolution crystal structures could see Met107 (Met71 in bovine and Met99 in Paracoccus denitrificans) hydrogen bonded to Glu286 in reduced and oxidized form of oxidase in bovine (Yoshikawa et al., 1998) and in Paracoccus denitrificans (Ostermeier et al., 1997), they still could not resolve these water molecules at this region, nor could they see the water molecules between the D channel and Glu286. Higher resolution structure will be needed to resolve all the water molecules in this region. Another molecular dynamics simulation method has been used to predict internal water sites using coordinates for both oxidases (Hofacker & Schulten, 1997). Two channels involving water molecules were predicted to be able to transfer protons. One is in the putative pumped proton channel (or D channel), where a chain of water molecules is seen to complete the hydrogen bonding network between Aspl32 and Glu286, similar to predictions by Riistama et al. (Riistama et al., 1997); the other one is the putative substrate proton channel (or K channel), in which from Ser299 to Tyr288 via Lys362, they only predict two water molecules leaving a gap between Ser299 and Lys362. They suggest that the Lys362 side chain is flexible and its head group can move downwards to close the gap between Lys362 and Ser299. Since the pathway ends close to the oxygen binding site, their results are consistent with a viable K channel and its assignment as the substrate proton channel. Interestingly, they do not see water molecules beyond Glu286. They propose that Glu286 could deliver a proton to the binuclear center or a heme a3 propionate group by flipping its side chain. The proposed mechanism is that two water molecules dissociate from CuB during the reduction of oxygen and hydrogen bond to a propionate group of heme a3. A proton is transferred from the D channel to the propionate group of heme a3 through flipping of the Glu286 side chain. Uptake of a substrate proton from the K channel drives the pumping of a proton from Glu286. This mechanism can transfer 0an mec Hm (Yo 199i COUp grou- be de shutt beset mull binue tom] 23.1. Elite D132 4 elemc 53am TEL-Hit; Egjdui mp1] “isfe 31 only one proton at a time, as opposed to two protons for “histidine cycle/shuttle” mechanism, the latter being consistent with experimental observations (Wikstrtim, 1989). However, the crystal structures so far obtained only resolve one conformation of Glu286 (Yoshikawa et al., 1998). So far, there are at least four hypotheses for the direct coupling model. The original “histidine cycle” model suggests that Hi3284 is involved (Morgan et al., 1994); Michel's “histidine cycle/shuttle” model supports the involvement of His333 (Iwata et al., 1995); Wikstrom's group now suggests that His334 might be critical (Puustinen et al., 1997); and finally, Hofacker and Schulten propose that Glu286 is the key residue in the coupling mechanism, by-passing histidine and directly utilizing a heme a3 propionate group (Hofacker & Schulten, 1997). Whether any of these models are correct remains to be determined. There is, as yet, no experimental evidence to support either the histidine shuttle or the Glu286 flipping mechanisms. The strength of the histidine shuttle model is based on the fact that the histidines that coordinate the heme a 3-Cu3 center are the only totally conserved residues in all the cytochrome oxidases. Other residues close to the binuclear center may also play a role, especially with the discovery that His284 is covalently linked to Tyr288 (Ostermeier et al., 1997; Yoshikawa et al., 1998). 2.3.1.3. Putative Proton Input Channels in the Direct Coupling Model In the direct coupling model, there are two proposed proton input channels. One is the so-called pumped proton channel, also referred to as the D channel, since the D132A mutation eliminates proton pumping activity but allows retention of some electron transfer activity (Thomas et al., 1993b; Fetter et al., 1995). The crystal structures also shows Asp132 to be in a position to feed protons to a hydrogen bonded pathway within the protein (Iwata et al., 1995; Tsukihara et al., 1996). Two more residues in the same loop between helix II and III, Asn121 and Asn139, also lose proton pumping activity when altered in the nonconservative mutations, but retain electron transfer activity (Garcia-Horsman et al., 1995). The hydrogen bonded network leading 3 rttfi putati \lll. . studie oxida 32 from these residues on the inside of the membrane via several assumed waters has been found to lead to Glu286, another critical residue in the D channel. The route of protons ‘ past this site is not clear from the crystal structure. The other channel is the so-called substrate proton channel, which is also referred to as the K channel, due to the fact that mutation of a lysine residue, Lys362, results in complete blockage of activity and proton pumping (Thomas et al., 1993a; Fetter et al., 1995; Hosler et al., 1996; Konstantinov et al., 1997), and the position of this residue is in a region of the structure that is a plausible route for protons to the active site. The putative substrate proton channel involves several highly conserved polar residues in helix VIII, in the vicinity of the heme a3-Cu3 binuclear center. From mutagenesis and FTIR studies in the E. coli quinol oxidase (Thomas et al., 1993b), and Rhodobacter sphaeroides oxidase (Fetter et al., 1995), it has been proposed that the highly conserved polar residues in helix VHI, Thr3 52, Thr359 and Lys362, are essential for substrate proton uptake. Pro358, although very highly conserved, does not appear to play a critical role in catalytic function. It causes a kink in the transmembrane helix that may be important for positioning. Mutation of Lys362, the only conserved basic residue in the transmembrane domain in subunit I, leads to complete loss of enzyme activity when replaced with Leu, Met, and Glu, suggesting that K3 62 plays an important role in substrate proton uptake. Lys362 is about 15 A away from the heme a 3 Fe. Resonance Raman and FTIR studies on K362M suggest two different conformations of this mutant, one with a disturbed heme arCuB binuclear center (Hosler et al., 1996). The assignment of the roles of the two proton input channels is controversial. Recently, Konstantinov and coworkers suggested that the D channel is likely to take up both substrate and pumped protons, while the K channel may be responsible for loading only substrate protons (Konstantinov et al., 1997). Their data suggest that Lys362 is involved in proton uptake in the early steps of oxygen reduction, but the oxidase can function without this channel in the later stages. However, evidence from Ferguson- Millet Millet 23.1.1 binuch is, boil center. crystal both 51 Asp412 litigm lfllheu Etude; "551211 2 Spectra lhe men that A3 1997113 33 Miller and coworkers raises questions regarding this interpretation (Mills & Ferguson- Miller, 1998). 2.3.1.4. Proton Exit Channel and Water Channel in the Direct Coupling Model A proton exit channel would be expected to be located above the heme a3-CuB binuclear center if electron transfer and proton translocation are coupled “directly”, that is, both electron transfer and proton pumping events occur at the heme a3-Cu3 binuclear center. A water channel is observed at the interface of subunit 1 and subunit II in the crystal structure of the mammalian oxidase, including several hydrophilic residues from both subunit I and subunit 11 (Figure 7). The Mg ion, along with its ligands His411 and Asp412, and three water molecules are at the starting point of the apparent water channel (Ferguson-Miller & Babcock, 1996; F lorens et al., 1998), and are proposed to be involved in the water channel (Tsukihara et al., 1996; Yoshikawa et al., 1998). Asp407, a highly conserved residue located immediately above heme a 3-CuB binuclear center, was suggested to have a role in proton exit (Iwata et al., 1995), based on crystal structures. However, D407A, D407N and D407C mutants show that the visible spectra, electron transfer activity, proton pumping activity, and the environment of all the metal centers are almost identical to wild type enzyme. Thus the data strongly argue that Asp407 is not critical for proton translocation, or for Mn binding (Qian et al., 1997)(See Chapter IV). The MnfMg site has been proposed to play an important role in stabilizing the structure at the interface of subunit I and subunit II, organizing the water, and controlling the efficiency of proton pumping. Therrnostability studies on the Mn/Mg ligands His411 and Asp412 demonstrate the significant role Mn/Mg plays in stabilizing heme and Cu. Deuterium isotope, ESEEM, and EPR studies on the wild type enzyme and these two mutants indicate that the Mn/Mg site is accessible to water in a time-scale of minutes at least, consistent with a role in the structure of the proposed water channel. pH dependence studies indicate that the Mn/M g site has some influence on pKa values that 34 Figure 7. The structure of the putative HZO/proton exit channel in cytochrome c oxidase. Several hydrophilic residues in subunit I and subunit II at the interface of these two subunits are proposed to be involved in this channel. The underlined names and numbers denote the subunit 11 residues. The coordinates are from 2.8 A bovine cytochrome c oxidase (Tsukihara et al., 1996). 35 H O/H+ channel s22_p__9/ Asp271 Thr337 Figure 7 contra that th an im] 19%: ‘ 2.3.2. 1 1976; t centers possibl chemis drives 1 1998). direct it no p0. helix )Q (Tsukth, Pupil], COIESpo; Rhodabc £11030] 1C punlped 36 control overall activity of the oxidase. The data are all consistent with the hypothesis that the proposed water channel is the proton exit channel as well, and the Mn/Mg site is an important element in it, structurally and functionally (Ferguson-Miller & Babcock, 1996; Florens et al., 1998). 2.3.2. Indirect Coupling Model An indirect or “conformational” mechanism involving redox Bohr effects (Papa, 1976; Chance et al., 1977) requires long-range protein mediated coupling between redox centers and the proton carrying residues. Yoshikawa and coworkers identified two possible indirect proton channels, supporting the idea of indirect coupling, i.e., oxygen chemistry might cause a conformational change at a distance from the active site, which drives proton translocation across the membrane (Tsukihara et al., 1996; Yoshikawa et al., 1998). They suggest that the proton pumping channel may start at Asp132, as in the direct model, but not go through Glu286 and not access the heme a3-CuB site. Instead, two possible hydrogen-bonded networks involving residues in helix III and helix IV, or helix XI and X11 in subunit 1, were initially proposed to facilitate the proton translocation (Tsukihara et al., 1996). The first hydrogen bonded network involves conserved residues Asp91, Asp98, SerlOl, Ser156, Ser157, Ser142 and Ser115 (bovine numbering, coresponding to Asp132, Asn139, Serl42, Ser200, Ser201, Ser186 and Ser156 in Rhodobacter sphaeroides) in helix III and helix IV, connecting the matrix side and cytosolic side of the oxidase (the lower part of this channel is the same as the proposed pumped proton channel in Figure 6). The second proposed proton pumping network consists of Glu407, His4l3, and Thr424 (bovine numbering, coresponding to G1u450, His456, and Tyr467 in Rhodobacter sphaeroides) in helix XI and Ser461, Ser454 and Asn451 (bovine numbering, coresponding to Ser504, Ser497, and Asn494 in Rhodobacter sphaeroides) in helix XII. Arg38 and Thr443 (bovine numbering, coresponding to Arg51 and Tyr486 in Rhodobacter sphaeroides) as well as heme a are also involved in the network. More recently, a higher resolution fully reduced and fully oxidized bovine heart oxid hydr strut Ser4- lle48 rphat AspS “hen buriet netwo filly} 37 oxidase structures were obtained (Yoshikawa et al., 1998). The new structures extend the hydrogen bonding network between helix XI and helix XII which they saw in the previous structure, and connect the network to the cytosolic side via Tyr371, Tyr54, Tyr440, Seer and Asp51 (bovine numbering, coresponding to Tyr414, Trp95, Tyr483 and lle484 in Rhodobacter sphaeroides; bovine Asp51 is not conserved in Rhodobacter sphaeroides, it is a glycine). The loop between helix 1 and helix II of subunit 1, containing Asp51, is seen to be the only region of the protein to undergo a conformational change when the fully reduced and fully oxidized structures are compared. Asp51 is completely buried in the fully oxidized state and has access to the matrix through a hydrogen bonded network, but it moves toward the intermembrane space upon full reduction of the enzyme. The mechanism appears to be reasonable, especially when the direct coupling model lacks structural evidence to support it, i.e., no proton transfer path is seen from Glu286 to Hi3333 or His334, the proposed ligand shuttles (see section 2.3.1). However, one drawback to the indirect model is that Asp51 (bovine numbering) is not conserved in plant and bacterial oxidases, so the model in bovine oxidase is not suitable for bacteria. The authors suggest that the differences are due to an evolution in the proton pumping mechanism. In summary, there are two proposed pumped proton channels in the indirect mechanism. One involves Asp132, which has been shown to be important for proton pumping (Thomas et al., 1993b; Fetter et al., 1995; Garcia-Horsman et al., 1995), and the other involves a different entrance and exit via Asp51, which undergoes a conformational change in the transition from fully oxidized to fully reduced states. How and whether they function in the coupling of electron transfer to the proton pumping is still under investigation. 2.4. E acids A hig cytocl them a the me water 1 Ser299 carbon only ts Channel resoluti 38 2.4. Bound Water Water molecules can participate in hydrogen bonding networks, just like amino acids (Nagle & Tristram-Nagle, 1983), as seen in bacteriorhodopsin (Rothschild, 1992). A higher resolution (2.7 A) crystal structure of two subunit Paracoccus denitrificans cytochrome c oxidase resolves 52 water molecules (Ostermeier et al., 1997). Only four of them are located in the cytoplasmic half or below the heme groups toward the inside of the membrane, while most of them are at the interface of subunit I and subunit 11. Two water molecules have been identified in the K channel and the D channel. One connects Ser299 and Lys362 in the K channel and one is hydrogen bonded to the backbone carbonyl oxygen of Met107, which hydrogen bonds to Glu286. It is surprising to see only two water molecules in the channels, since a water chain is seen in the proton channel of the photosynthetic reaction center (Ermler et al., 1994). However, the resolution of the oxidase structure is still much less than the reaction center. Another important discovery from the high resolution crystal structure is an extended hydrogen bonded network including seven water molecules at the interface between subunit I and subunit 11. Each propionate group of heme a and heme a 3 is hydrogen bonded to a water molecule. His334 shares a water molecule with the propionate side chain of heme a3. The Mn/Mg site has four ligands, Glu254 (subunit II), Asp412, His411, and a water molecule, which hydrogen bonds to Asp229 (subunit 11). Two more water molecules are resolved at the Mg site in the 2.3 A bovine oxidase structure (Yoshikawa et al., 1998) although they are not seen in the Paracoccus denitrificans structure. This extended hydrogen bonded network may play an important role in maintaining the structure at the interface, and/or facilitate electron transfer from Cu to heme a, and/or control the proton/water exit. lnvol' since 1 The it there a in the t trperir. bacteria annn mutager lherefo meets Studies ( c3931316 ( mleflOr h mittared' X anOlhe A; rel'calet 9‘. . ”mortal halted by .' 39 2.5. Site-Directed Mutagenesis Studies on Aspl32 and Asp407 to Study Their Involvement in Proton Pumping Subunit I is the largest and the most important subunit in cytochrome c oxidase, since both the oxygen reaction and the proton pumping reaction occur in this subunit. The mechanism of how electron transfer and proton pumping are coupled is not clear; there are debates about direct coupling and indirect coupling models, and which residues in the direct coupling model play key roles is still not clear. However, at present, great experimental advantages are available in the form of the structural information from bacterial and mammalian three-dimensional crystal structures of cytochrome c oxidase, the availability of the sequences of more than 75 homologous oxidases, powerful site-directed mutagenesis technology and an array of various biochemical and biophysical analyses. Therefore, mutational studies were used to test models of the proton pumping mechanism. Highly conserved residues with hydrophilic side chains were chosen for studies of their involvement in the proton pumping mechanism, since these residues are capable of forming hydrogen bonding interactions. One focus of this study (see section 3.4.4 of this chapter for the second focus) was on two highly conserved regions in subunit 1. On the putative proton input site in an interior loop between helix II and helix III, a conserved acidic residue, Asp132, was mutated. In the proposed proton exit site on the exterior loop between helix IX and helix X, another conserved acidic residue, Asp407, was mutated to test its involvement in proton exit or Mg binding. All the mutants were analyzed in terms of structural changes, as revealed by the spectral characteristics of heme a, heme a 3-Cu3, Cu, and Mg sites, and functional changes in overall electron transfer activity, and proton pumping activity caused by mutations. 40 3. Electron Transfer Pathways 3.1. Electron Entry Site Four carboxylates on subunit II of cytochrome c oxidase (Zhen et al., 1997) and several lysine residues surrounding the exposed heme edge of cytochrome c (Ferguson- Miller et al., 1978) form electrostatic interactions, guiding complex formation between these two redox partners. Bruce Hill first provided evidence that Cu», is the initial electron acceptor in cytochrome c oxidase in his transient absorption spectroscopy study (Hill, 1991). He showed that the electrons are rapidly transferred from cytochrome c to Cu, then to heme a, finally to the heme a3-CuB binuclear center. Mutations of the Cu ligands (Hi3260 and Met263) result in alteration of the unique EPR signal due to the dinuclear copper center, indicating the coppers are decoupled, but not lost. Dramatically decreased overall electron transfer activity supports the hypothesis that the Cut, is the sole electron entry site in cytochrome c oxidase (Zhen et al., 1997). A decoupled CuA center is also observed in the equivalent mutant, M2271 in Paracoccus denitrificans, supporting its role in maintaining the unique character of the Cu», site that is important for efficient electron input to oxidase (Zickermann et al., 1995). Millett's group has designed a photoactivatable cytochrome c that makes it possible to measure the electron transfer rates from cytochrome c to CuA, and from Cu to heme a. Ru(II) that is covalently attached to cytochrome c, can be photoexcited to a metal-to-ligand charge-transfer state, Ru(II*), which can transfer an electron into the heme group of cytochrome c rapidly. Many different ruthenium cytochrome c derivatives have been made to validate the method. All the derivatives interact with cytochrome c oxidase in a similar way as native cytochrome c. This useful technique has now been applied to Wild type and many Rhodobacter sphaeroides cytochrome c oxidase mutants, and is essential for determining the electron transfer characteristics (see Chapter V for applying ruthenium kinetics assay in Arg482 mutants). oxide and tr 3.3. ( 41 3.2. Cytochrome c Docking Algorithms The precise nature of the interaction of cytochrome c and cytochrome c oxidase is still not established and mutagenesis approaches are likely to be more successful with guidance fi'om computation predictions. Roberts and colleagues developed a computer software program TURNIP to study the protein-protein interaction between plastocyanin and cytochrome c (Roberts et al., 1991). An even more powerful refined method has subsequently been devised, DOT (Ten Eyck, 1995). Electrostatic and van der Waals forces between the proteins can be evaluated and probable docking sites can be determined. With the availability of atomic level crystal structures of both cytochrome c oxidase and cytochrome c, the docking of cytochrome c to the oxidase can be predicted and tested by mutational analysis. 3.3. Cytochrome c - Cytochrome c Oxidase Interaction Chemically modified cytochrome c was used to study the cytochrome c binding domain of cytochrome c oxidase (Smith et al., 1977; Ferguson-Miller et al., 1978; Osheroffet al., 1980). Several lysine residues (LysS, 13, 72, 87) on the front surface of horse cytochrome c, close to the exposed heme edge, were found to profoundly influence electron transfer and binding of cytochrome c to cytochrome c oxidase. Consistent with these studies, it revealed a similar interaction site for both yeast and horse cytochrome c (cc) with cytochrome c peroxidase (CCP) (Pelletier & Kraut, 1992). The crystal structure of yeast cytochrome c peroxidase: horse cytochrome c (CCP:cc(H)) complex reveals that three out of four lysine residues (Lys8, 72, 87) on horse cytochrome c interact with cytochrome c peroxidase via hydrogen bonds, but they are all at the interface of CCP:cc(H) complex. This complex has been confirmed (Geren et al., 1991) and single-site binding with a 1:1 electron transfer complex mechanism has been proposed for the CCchc reaction based on their structure. In the case of cytochrome c oxidase, there is evidence for two cytochrome c binding sites, a low affinity site and a high affinity site, based on kinetic and binding 42 studies (Ferguson-Miller et al., 1976; Ferguson-Miller et al., 1978). The biphasic kinetic plot (Km. = 5 x 10'8 M, and sz = 0.35 to 1 x 10'6 M) of horse cytochrome c with cytochrome c oxidase suggested two active sites. Similarly, high affinity and low affinity sites have been suggested for the 1:1 cytochrome c and cytochrome c peroxidase complex based upon crosslinking and kinetic experiments (Wang & Margoliash, 1995). For both these enzymes it is still not clear what the significance of a second interaction is (Cooper, 1989) A series of carboxylic acid residues on the surface of cytochrome c oxidase subunit II are proposed to play an important role in binding the substrate cytochrome c by interaction with the positively charged lysine groups in cytochrome c. Mutants have been generated at the cytochrome c binding site and the Cu, site, and characterized by spectroscopy, EPR (electron paramagnetic resonance spectroscopy), RR (resonance Raman spectroscopy), fast kinetics analysis and meta] analysis. The results suggest that highly conserved carboxyls including Glu157, Glu195, Glu148 and Asp214 are involved in cytochrome c binding (Zhen et al., 1997). 3.4. Electron Transfer Pathway Models Electron-transfer (ET) reactions are critical in many biological systems, from photosynthesis to respiration; this fact has stimulated numerous experimental and theoretical investigations. In cytochrome c oxidase, electrons are transferred from substrate cytochrome c to the primary electron acceptors, Cu and heme a, and finally to the heme a3-Cu3 binuclear center to reduce 02 to water. This process provides a driving force for proton translocation. However, the mechanism of electron transfer from electron donor (D) to electron acceptor (A) has remained controversial. Several models have been proposed, including the “through space” or distance dependent model (Hopfield, 1974; Moser et al., 1992; Moser et al., 1995), and the “through bond” or pathway model (Beratan et al., 1992; Onuchic et al., 1992; Curry et al., 1995). The detelot in meta 3.4.1. 1‘ ein: 9. Di (/) E H there 1. rate fror mantel state (H lhe fret 1361“?te Want factor (1 item d0 “Only 1 43 development of ruthenium-modified proteins has made it possible to measure the ET rate in metalloproteins by surface labeling the redox active molecules (Gray & Winkler, 1996). 3.4.1. Marcus Theory The starting point for analysis of most long range electron transfer reactions in proteins is the semiclassical expression formulated by Marcus and Sutin (Marcus & Sutin, 1985): 4n3 (AG°+AY Er hZ/lkBT A” 6% 41kg" ] ( ) where h is the Plank’s constant, and k3 is the Boltzrnan constant. It predicts that the ET rate from a D to A depends on free energy difference (-AG°), a nuclear reorganization parameter (2.), and the electronic-coupling strength between D and A at the transition state (HA3), which generally decreases exponentially with increasing D-to-A distance, r. The free energy difference (-AG°) is the driving force for the electron transfer reaction between two redox centers. The reorganization energy (7t) is defined as the energy of the reactants at the equilibrium nuclear configuration of the product. The electronic coupling factor (HA3) reflects how strongly the protein allows electrons to leak across the protein from donor to acceptor. When (-AG°) = 2., the ET rate reaches its maximum value, which is only limited by the electronic coupling strength (HA3). In Marcus theory, the nature of the intervening material through which the electron travels is not considered. 3.4.2. “Through-Space” Model The simplest view of long range electron transfer applies the one—dimensional . square barrier (lDSB) tunneling model, which predicts that the rate drops exponentially with distance (Hopfield, 1974). Dutton and coworkers (Moser et al., 1992) have used the biological system photosynthetic reaction centers in Rhodobacter sphaeroides and Rhodobacter virdis, w} anl I31 the elet COl'l Streng @0161}- With 44 whose three—dimensional crystal structures are known (Deisenhofer et al., 1984; Allen et al., 1986; Chang et al., 1986), semisynthetic system including ruthenium-cytochrome c and ruthenium-myglobin (Cowan et al., 1988), and synthetic system to investigate the ET rate relative to the distance between electron donor and acceptor. A linear relationship has been found between the distance and the logarithm of electron transfer rate. Based on these results, they propose that in proteins, the difference in free energy between an electron donor and an acceptor (-AG°) for any reorganization energy (A) does not contribute to the electron transfer rate as much as the distance; i.e., distance seems to be the primary factor that can influence electron transfer rate up to 10‘2-fold in biological electron transfer reactions (Moser et al., 1992). It implies that the nature of the intervening medium does not affect the coupling between two redox sites, i.e., the pathway between D and A is not the factor determining ET rate. This is consistent with the lDSB model that treats protein as a homogenous square tunneling barrier (Hopfield, 1974) 3.4.3. “Through-Bond” Model Although Dutton’s rate/distance correlation studies fit reaction center coupling strength, extensive experimental and theoretical work clearly shows that the intervening protein structure must be taken into account to understand distant D-A coupling in other proteins (Beratan et al., 1991; Onuchic et al., 1992). Beratan and Onuchic have developed a tunneling-pathway model to describe electron coupling between redox sites. The model reduces the complex interaction between atoms into three: covalent bond, hydrogen bond and through-space contacts (Beratan et al., 1991; Onuchic et al., 1992). Electrons move the fastest through covalent or hydrogen bonds, while through-space jumps have the highest tunneling barrier. Analysis of ET rates from different Ru-binding sites in Ru-modified cytochrome c and azurin demonstrates that the coupling strength is extremely sensitive to the detailed composition of the medium bridging the redox sites (Bjerrum et al., 1995), and the struc elect role . the r« medi critic 3.4.-1 Orid 45 structure and composition of the intervening medium are critical in detenrrining distant electron coupling (Langen et al., 1995). A critical aspect of the pathway model is the vital role of hydrogen bonds as tunneling mediators (Curry et al., 1995). More studies upon the relationship of ET rate and protein secondary structure show that [3 sheets appear to mediate coupling more efficiently than or helical structure, and hydrogen bonds play a critical role in both (Gray & Winkler, 1996). 3.4.4. Testing “Through-Space” and “Through-Bond” Theory in Cytochrome c Oxidase The crystal structures of the bacterial and mammalian cytochrome oxidases show that a direct route connecting Cu), and heme a could involve 14 covalent bonds and two hydrogen bonds. This pathway has been considered a possible “through-bond” electron transfer pathway (Ramirez et al., 1995) that would account for the very fast rate of transfer from Cu to heme a (19 A apart) compared to that between Cu and heme a3 (22 A apart). Alternatively, Brzezinski (Brzezinski, 1996) calculates a different reorganization energy for heme a (0.3 eV) and heme a3 (0.8 eV), and concludes that the difference in electron transfer rates is due to the large differences in the reorganization energies at heme a and heme (13 sites. The authors suggest that heme a3 is in a more polar environment than heme a, and proton uptake events must occur at heme (13, which inhibit rapid electron transfer. In order to distinguish whether the rapid ET rate from Cut, to heme a is due to the protein structure between these two redox centers (“through-bond”) or the distance between them (“through-space”), mutagenesis efforts were made at the Arg482 position in subunit I of Rhodobacter sphaeroides cytochrome c oxidase; one of a highly conserved arginine pair located on the proposed “through-bond” pathway between Cu, and heme a (Ramirez et al., 1995). If the bonds are important for the ET rate, mutation at the critical position Arg482 would be expected to change the ET rate. Four mutants were designed 46 to disrupt the bonding network between Cu and heme a; intrinsic ET rates of these mutants using ruthenium labeled cytochrome c were measured, and other biochemical and biophysical analyses were carried out (Qian etal., 1998) (see Chapter V). REFERENCES Adelroth, P., Ek, M. S., Mitchell, D. M., Gennis, R. B. & Brzezinski, P. (1997) Glutamate 286 in cytochrome aa3 from Rhodobacter sphaeroides is involved in proton uptake during the reaction of the fully-reduced enzyme with dioxygen. Biochemistry 36: 13824-9. Allen, J. P., Feher, G., Yeates, T. 0., Rees, D. C., Deisenhofer, J ., Michel, H. & Huber, R. (1986) Structural homology of reaction centers from Rhodopseudomonas sphaeroides and Rhodopseudomonas viridis as determined by x-ray diffraction. Proc Natl Acad Sci U SA 83: 8589-93. Babcock, G. T. & WikstrtSm, M. (1992) Oxygen activation and the conservation of energy in cell respiration. Nature 356: 301-309. Balashov, S. P., Irnasheva, E. S., Govindjee, R., Sheves, M. & Ebrey, T. G. (1996) Evidence that aspartate-85 has a higher pK(a) in all-trans than in 13- cisbacteriorhodopsin. Biophys J 71: 1973-84. Beratan, D. N., Betts, J. N. & Onuchic, J. N. (1991) Protein Electron Transfer Rates Set by the Bridging Secondary and Tertiary Structure. Reports : 1285. Beratan, D. N., Onuchic, J. N., Winkler, J. R. & Gray, H. B. (1992) Electron-tunneling pathways in proteins [comment]. Science 258: 1740-1. Bjerrrun, M. J., Casimiro, D. R., Chang, I. J., Di Bilio, A. J., Gray, H. 8., Hill, M. G., Langen, R., Mines, G. A., Skov, L. K., Winkler, J. R. & et al. (1995) Electron transfer in ruthenium-modified proteins. J Bioenerg Biomembr 27: 295-3 02. Brown, L. 8., Sasaki, J., Kandori, H., Maeda, A., Needleman, R. & Lanyi, J. K. (1995) Glutamic acid 204 is the terminal proton release group at the extracellular surface of bacteriorhodopsin. J Biol Chem 270: 27122-6. Brzezinski, P. (1996) lntemal electron-transfer reactions in cytochrome c oxidase. Biochemistry 35: 5612-5. Buse, G., Steffens, G. J. & Steffens, G. C. M. (1978) Studies on Cytochrome c Oxidase, 111. Relationship of Cytochrome Oxidase Subunits to Electron Carriers of Photophosphorylation. Hoppe-Seyler's Z. Physio]. Chem. 359: 1011-1013. Calhoun, M. W., Thomas, J. W., Hill, J. J., Hosler, J. P., Shapleigh, J. P., Tecklenburg, M. M. J., Ferguson-Miller, S., Babcock, G. T., Alben, J. O. & Gennis, R. B. (1993) Identity of the Axial Ligand of the High-Spin Heme in Cytochrome Oxidase: Spectroscopic Characterization of Mutants in the bo-type Oxidase of Escherichia coli and the aa3-type Oxidase of Rhodobacter sphaeroides. Biochemistry 32: 10905-10911. 47 48 Casey, R. P., Chappell, J. B. & Azzi, A. (1979) Limited-Turnover Studies on Proton Translocation in Reconstituted Cytochrome c Oxidase-Containing Vesicles. Biochem. J. 182: 149-156. Casey, R. P., Thelen, M. & Azzi, A. (1980) Dicyclohexylcarbodiimide Binds Specifically and Covalently to Cytochrome c Oxidase While Inhibiting Its H+-Translocating Activity. J. Biol. Chem. 255: 3994-4000. Caughey, W. S., Smythe, G. A., O'Keeffe, D. H., Maskasky, J. E. & Smith, M. I. (1975) Heme A of cytochrome c oxicase. Structure and properties: comparisons with hemes B, C, and S and derivatives. J Biol Chem 250: 7602-22. Chan, S. I. & Li, P. M. (1990) Cytochrome c Oxidase: Understanding Nature's Design of a Proton Pump. Biochemistry 29: 1-12. Chance, B., Leigh, J. S. J. & Waring, A. in Structure and Function of Energy-Transducing Membranes van Dam, K. & van Gelder, B. F., Eds. (Elsevier/North-Holland, Amsterdam, 1977) pp. 1-10. Chang, C. H., Tiede, D., Tang, J ., Smith, U., Norris, J. & Schiffer, M. (1986) Structure of Rhodopseudomonas sphaeroides R-26 reaction center. F EBS Lett 205: 82-6. Cooper, C. E. (1989) The Steady-State Kinetics of Cytochrome c Oxidation by Cytochrome Oxidase. Biochim. Biophys. Acta 1017: 187-203. ' Cowan, J. A., Upmacis, R. K., Beratan, D. N., Onuchic, J. N. & Gray, H. B. (1988) Long-range electron transfer in myoglobin. Ann N Y Acad Sci 550: 68-84. Curry, W. B., Grabe, M. D., Kumikov, I. V., Skourtis, S. S., Beratan, D. N., Regan, J. J ., Aquino, A. J., Beroza, P. & Onuchic, J. N. (1995) Pathways, pathway tubes, pathway docking, and propagators in electron transfer proteins. J Bioenerg Biomembr 27: 285-93. Deisenhofer, J., Epp, 0., Miki, K., Huber, R. & Michel, H. (1984) X-Ray Structure Analysis of a Membrane Protein Complex: Electron Density Map at 3 A Resolution and a Model of the Chromophores of the Photosynthetic Reaction Center from Rhodopseudomonas viridis. J. Mol. Biol. 180: 385-398. Einarsdéttir, O. (1995) Fast Reactions of Cytochrome Oxidase. Biochim. Biophys. Acta 1229: 129-302. Ermler, U., Fritzsch, G., Buchanan, S. K. & Michel, H. (1994) Structure of the Photosynthetic Reaction Centre from Rhodobacter sphaeroides at 2.65 A Resolution: Cofactors and Protein-Cofactor Interactions. Structure 2: 925-936. 49 Espe, M. P., Hosler, J. P., Ferguson-Miller, S., Babcock, G. T. & McCracken, J. (1995) A Continuous Wave and Pulsed EPR Characterization of the Mn2+ Binding Site in Rhodobacter sphaeroides Cytochrome c Oxidase. Biochemistry 34: 7593-7602. Ferguson-Miller, S. & Babcock, G. (1996) Heme/Copper Terminal Oxidases. Chemical Reviews 96: 2889-2907. Ferguson-Miller, S., Brautigan, D. & Margoliash, E. (1976) Correlation of the Kinetics of Electron Transfer Activity of Various Eukaryotic Cytochrome c with Binding to Mitochondrial Cytochrome c Oxidase. J. Biol. Chem 251: 1104-1115. Ferguson-Miller, S., Brautigan, D. L. & Margoliash, E. (1978) Definition of Cytochrome c Binding Domains by Cherrrical Modification. J. Biol. Chem. 253: 149-159. Fetter, J. R., Qian, J., Shapleigh, J., Thomas, J. W., Garcia-Horsman, J. A., Schmidt, E., Hosler, J., Babcock, G. T., Gennis, R. B. & Ferguson-Miller, S. (1995) Possible Proton Relay Pathways in Cytochrome c Oxidase. Proc. Natl. Acad. Sci. USA. 92: 1604-1608. Florens, L., Hoganson, C., McCracken, J., Fetter, J., Mills, D. A., Babcock, G. T. & Ferguson-Miller, S. in The Phototropic Prokaryotes Preschek, G., Eds. (Plenum Press, 1998), vol. in press,. Frank, V. & Kadenbach, B. (1996) Regulation of the H+/e' Stoichiometry of Cytochrome c Oxidase from Bovine Heart by lntramitochondrial ATP/ADP ratios. F EBS Lett 382: 121 -124. Gai, F., Hasson, K. C., McDonald, J. C. & Anfinrud, P. A. (1998) Cherrrical dynamics in proteins: the photoisomerization of retinal in bacteriorhodopsin. Science 279: 1886-91. Garcia-Horsman, J. A., Puustinen, A., Gennis, R. B. & Wikstrom, M. (1995) Proton Transfer in Cytochrome b03 Ubiquinol Oxidase of Escherichia coli: Second-Site Mutations in Subunit I That Restore Proton Pumping in the Mutant Aspl35—)Asn. Biochemistry 34: 4428-4433. Gelles, J., Blair, D. F. & Chan, S. I. (1986) The proton-pumping site of cytochrome c oxidase: a model of its structure and mechanism. Biochim Biophys Acta 853: 205-36. Gennis, R. B. (1998) Cytochrome oxidase:one proton, two mechanism? Science 280: 1712-1713. Geren, L., Hahm, 8., Durham, B. & Millett, F. (1991) Photoinduced Electron Transfer Between Cytochrome c Peroxidase and Yeast Cytochrome c Labeled at Cys 102 with (4- Bromomethyl-4'-methylbipyridine)[bis(bipyridine)]ruthenium2+. Biochemistry 30; 9450- 9457. 50 Gray, H. B. & Winkler, J. R. (1996) Electron transfer in proteins. Annu Rev Biochem 65: 537-61. Gregory, L. C. & Ferguson-Miller, S. (1988) Effect of Subunit III Removal on Control of Cytochrome c Oxidase Activity by pH. Biochemistry 27 : 6307-6314. Haltia, T., Saraste, M. & Wikstrom, M. (1991) Subunit III of Cytochrome c Oxidase is not Involved in Proton Translocation: A Site-Directed Mutagenesis Study. EMBO J. 10: 2015-2021. Henderson, R., Baldwin, J. M., Ceska, T. A., Zemlin, P., Beckmann, E. & Downing, K. H. (1990) Model for the Structure of Bacteriorhodopsin Based on High-resolution Electron Cryo—microscopy. J. Mol. Biol. 213: 899-929. Hill, B. C. (1991) The Reaction of the Electrostatic Cytochrome c-Cytochrome Oxidase Complex with Oxygen. J. Biol. Chem. 266: 2219-2226. Hill, B. C. (1993) The Sequence of Electron Carriers in the Reaction of Cytochrome c Oxidase with Oxygen. J. Bioenerg. Biomembr. 25: 115 -120. Hofacker, I. & Schulten, K. (1997) Oxygen and Proton Pathways in Cytochrome c Oxidase. Proteins: Structure, Function & Genetics 30: 100-107. Hopfield, J. J. (1974) Kinetic proofreading: a new mechanism for reducing errors in biosynthetic processes requiring high specificity. Proc Natl Acad Sci U S A 71: 4135-9. Hosler, J. P., Espe, M. P., Zhen, Y., Babcock, G. T. & Ferguson-Miller, S. (1995) Analysis of Site-Directed Mutants Locates a Non-Redox-Active Metal near the Active Site of Cytochrome c Oxidase of Rhodobacter sphaeroides. Biochemistry 34: 7586-7592. Hosler, J. P., Ferguson-Miller, S., Calhoun, M. W., Thomas, J. W., Hill, J ., Lemieux, L., Ma, J ., Georgiou, C., Fetter, J ., Shapleigh, J. P., Tecklenburg, M. M. J ., Babcock, G. T. & Gennis, R. B. (1993) Insight into the Active-Site Structure and Function of Cytochrome Oxidase by Analysis of Site-Directed Mutants of Bacterial Cytochrome M3 and Cytochrome bo. J. Bioenerg. Biomembr. 25: 121-136. Hosler, J. P., Fetter, J., Tecklenburg, M. M. J., Espe, M., Lenna, C. & Ferguson-Miller, S. (1992) Cytochrome aa3 of Rhodobacter sphaeroides as a Model for Mitochondrial Cytochrome c Oxidase. J. Biol. Chem. 267: 24264-24272. Hosler, J. P., Shapleigh, J. P., Mitchell, D. M., Kim, Y., Pressler, M., Georgiou, C., Babcock, G. T., Alben, J. 0., Ferguson-Miller, S. & Gennis, R. B. (1996) Polar Residues in Helix VIII of Subunit I of Cytochrome c Oxidase Influence the Activity and the Structure of the Active Site. Biochemistry 35: 10776-10783. H1111 RCCt Dem Iwat‘ cm Kade CF10! Elm Kadel Ex‘olu K, M high-r Konst Roles Splice; Eltctrc Krebs. Transh 51 Hilther, F.-J. & Kadenbach, B. (1988) Intraliposomal Nucleotides Change the Kinetics of Reconstituted Cytochrome c Oxidase from Bovine Heart but Not from Paracoccus Denitrificans. Biochem. Biophys. Res. Commun. 153: 525-534. Iwata, S., Ostermeier, C., Ludwig, B. & Michel, H. (1995) Structure at 2.8 A resolution of cytochrome c oxidase from Paracoccus denitrificans. Nature 376: 660-669. Kadenbach, B., Jarausch, J., Hartrnann, R. & Merle, P. (1983) Separation of Mammalian Cytochrome c Oxidase into 13 Polypeptides by a Sodium Dodecyl Sulfate-Gel Electrophoretic Procedure. Anal. Biochem. 129: 517-521. Kadenbach, B., Shroh, A., Hfither, F.-J., Reimann, A. & Steverding, D. (1991) Evolutionary Aspects of Cytochrome c Oxidase. J. Bioenerg. Biomembr. 23: 321-334. Kimura, Y., Vassylyev, D. G., Miyazawa, A., Kidera, A., Matsushima, M., Mitsuoka, K., Murata, K., Hirai, T. & Fujiyoshi, Y. (1997) Surface of bacteriorhodopsin revealed by high-resolution electron crystallography. Nature 389: 206-11. Konstantinov, A. A., Siletsky, S., Mitchell, D., Kaulen, A. & Gennis, R. (1997) The Roles of the Two Proton Input Channels in Cytochrome c Oxidase from Rhodobacter sphaeroides Probed by the Effects of Site-directed Mutations on Time-resolved Electrogenic Intraprotein Proton Transfer. Proc. Natl. Acad. Sci (USA) 94: 9085-9090. Krebs, M. P. & Khorana, H. G. (1993) Mechanism of Light-Dependent Proton Translocation by Bacteriorhodopsin. J. Bacteriol. 175: 1555-1560. Langen, K, Chang, I. J., Germanas, J. P., Richards, J. H., Winkler, J. R. & Gray, H. B. (1995) Electron tunneling in proteins: coupling through a beta strand. Science 268: 1733- 5. Lanyi, J. K. (1997) Mechanism of ion transport across membranes. Bacteriorhodopsin as a prototype for proton pumps. J Biol Chem 272: 31209-12. Marcus, R. A. & Sutin, N. (1985) Electron transfers in chemistry and biology. Biochim. Biophys. Acta 811: 265-322. Merle, P. & Kadenbach, B. (1980) Eur. J. Biochem. 105: 499-507. Mills, D. A. & Ferguson-Miller, S. (1998) Proton uptake and release in cytochrome c oxidasezseperate pathways in time and space? Biochim.Biophys.Acta 1365: 46-52. Mitchell, D. M., Adelroth, P., Hosler, J. P., Fetter, J. R., Brzezinski, P., Pressler, M. A., Aasa, R., Malmstrbm, B. G., Alben, J. O., Babcock, G. T., Gennis, R. B. & Ferguson- Miller, S. (1996) A Ligand-Exchange Mechanism of Proton Pumping Involving Tyrosine- 422 of Subunit I of Cytochrome Oxidase Is Ruled Out. Biochemistry 35: 824-828. Mitt resp 208: Mar Mod Bum Xkrt $516 Mosc biolo Mose J Bic Nagle lleml Nagle condu CMSag 1364. Onucl PTOteii OShen dOmai Wbm 51. 03mm c‘33‘1p11 Pa}- 1 4610 4: g_. ‘ Way. EWCtu ‘, ti:- . «\QS 52 Mitchell, P. & Moyle, J. (1965) Stoichiometry of proton translocation through the respiratory chain and adenosine triphosphatase systems of rat liver mitochondria. Nature 208: 147-51. Morgan, J. E., Verkhovsky, M. I. & Wikstrom, M. (1994) The Histidine Cycle: A New Model for Proton Translocation in the Respiratory Heme-Copper Oxidases. J. Bioenerg. Biomembr. 26: 599-608. Morowitz, H. J. (1978) Proton semiconductors and energy transduction in biological systems. Am J Physiol 235: R99-1 l4. Moser, C. C., Keske, J. M., Warncke, K., Farid, R. S. & Dutton, P. L. (1992) Nature of biological electron transfer. Nature 355: 796 - 802. Moser, C. C., Page, C. C., Farid, R. & Dutton, P. L. (1995) Biological Electron Transfer. J. Bioenerg. Biomembr. 27: 263-274. Nagle, J. F. & Morowitz, H. J. (1978) Molecular Mechanisms for Proton Transport in Membranes. Proc. Natl. Acad. Sci. USA. 75: 298-302. Nagle, J. F. & Tristram-Nagle, S. (1983) Hydrogen bonded chain mechanisms for proton conduction and proton pumping. J Membr Biol 74: 1-14. Onsager, L. (1969) The Motion of Ions: Principles and Concepts. Science 166: 1359- 1364. Onuchic, J. N., Beratan, D. N., Winkler, J. R. & Gray, H. B. (1992) Pathway analysis of protein electron-transfer reactions. Annu Rev Biophys Biomol Struct 21: 349-77. Osheroff, N., Brautigan, D. L. & Margoliash, E. (1980) Definition of enzymic interaction domains on cytochrome c. Purification and activity of singly substituted carboxydinitrophenyl-lysine 7, 25, 73, 86, and 99 cytochromes c. J Biol Chem 255: 8245- 51. Ostermeier, C., Harrenga, A., Ermler, U. & Michel, H. (1997) Structure at 2.7 A resolution of the Paracoccus denitrificans two-subunit cytochrome c oxidase complexed with an antibody F v fragment. Proc. Natl. Acad. Sci. USA 94: 10547-10553. Papa, S. (1976) Proton translocation reactions in the respiratory chains. Biochim Biophys Acta 456: 39-84. Pebay-Peyroula, E., Rummel, G., Rosenbusch, J. P. & Landau, E. M. (1997) X-ray structure of bacteriorhodopsin at 2.5 angstroms from microcrystals grown in lipidic cubic phases [see comments]. Science 277: 1676-81. 53 Pelletier, H. & Kraut, J. (1992) Crystal structure of a complex between electron transfer partners, Cytochrome c peroxidase and Cytochrome c. Science 258: 1748-1755. Proshlyakov, D. A., Pressler, M. A. & Babcock, G. T. (1998) Dioxygen activation and bond cleavage by mixed-valence cytochrome c oxidase. PNAS in press: . Puustinen, A., Bailey, J. A., Dyer, R. B., Mecklenburg, S. L., Wikstrom, M. & Woodruff, W. H. (1997) Fourier transform infrared evidence for connectivity between CuB and glutamic acid 286 in cytochrome bo3 from Escherichia coli. Biochemistry 36: 13195-200. Qian, J., Geren, L., Hoganson, C., Pressler, M., Looney, A., Babcock, G. T., Millett, F. & Ferguson-Miller, S. (1998) Arg482 is important for stabilizing the interface of subunit I and subunit II of Rhodobacter sphaeroides cytochrome c oxidase. in preparation : . Qian, J., Shi, W., Pressler, M., Hoganson, C., Mills, D., Babcock, G. T. & Ferguson- Miller, S. (1997) Aspartate-407 in Rhodobacter sphaeroides cytochrome c oxidase is not required for proton pumping or Mn binding. Biochemistry 36: 2539-2543. Ramirez, B. B., Malmstrom, B. G., Winkler, J. R. & Gray, H. B. (1995) The currents of life: The terminal electron-transfer complex of respiration. Proc. Natl. Acad. Sci. USA 92: 11949 - 11951. Richter, H. T., Brown, L. S., Needleman, R. & Lanyi, J. K. (1996a) A linkage of the pKa's of asp-85 and glu-204 forms part of the reprotonation switch of bacteriorhodopsin. Biochemistry 35: 4054-62. Richter, H. T., Needleman, R. & Lanyi, J. K. (1996b) Perturbed interaction between residues 85 and 204 in Tyr-l85-->Phe and Asp-85-->Glu bacteriorhodopsins. Biophys J 71: 3392-8. Riistama, S., Hummer, G., Puustinen, A., Dyer, R. B., Woodruff, W. H. & Wikstrom, M. (1997) Bound Water in the Proton Translocation Mechanism of the Haem-Copper Oxidases. FEBS Lett. 414: 275-280. Roberts, V. A., Freeman, H. C., Olson, A. J., Tainer, J. A. & Getzoff, E. D. (1991) Electrostatic orientation of the electron-transfer complex between plastocyanin and cytochrome c. J. Biol. Chem. 266: 13431 - 13441. Rothschild, K. J. (1992) F TIR Difference Spectroscopy of Bacteriorhodopsin: Toward a Molecular Model. J. Bioenerg. Biomembr. 24: 147-167. Rousseau, D. L., Ching, Y.-c. & Wang, J. (1993) Proton Translocation in Cytochrome c Oxidase: Redox Linkage through Proximal Ligand Exchange on Cytochrome a3. J Bioenerg. Biomembr. 25: 165-176. Smith. to loce 497M Stellar behavi Suarez Miller, Oxidas Equilil Taanm lnterac 11841 . Ten E Moleo. COIthr Thoma mung uleIlii 1mm llmma SUhstit Lliiqui 32'. 109 T1103]; Billet: the Hit laugh. K, Na Glidizt imkih; {it Na} “Dunn 54 Smith, H. T., Staudenmayer, N. & Millett, F. (1977) Use of specific lysine modifications to locate the reaction site of cytochrome c with cytochrome oxidase. Biochemistry 16: 4971-4974. Steffens, G. C., Soulimane, T., Wolff, G. & Buse, G. (1993) Stoichiometry and redox behaviour of metals in cytochrome-c oxidase. Eur J Biochem 213: 1 149-5 7. Suarez, M. D., Revzin, A., Narlock, R., Kempner, E. S., Thompson, D. A. & Ferguson- Miller, S. (1984) The Functional and Physical Form of Mammalian Cytochrome c Oxidase Determined by Gel Filtration, Radiation Inactivation, and Sedimentation Equilibrium Analysis. J. Biol. Chem. 259: 13791-13799. Taanman, J.-W., Turina, P. & Capaldi, R. (1994) Regulation of Cytochrome c Oxidase by Interaction of ATP at Two Binding Sites, One on Subunit VIa. Biochemistry 33: 11833- 11841. Ten Eyck, L. F., Mandel], J., Roberts, V. A., and Pique, M. E. (1995) Surveying Molecular Interactions With DOT. Proceedings of the1995 ACM/IEEE Supercomputing Conference :. Thomas, J. W., Lemieux, L. J ., Alben, J. O. & Gennis, R. B. (1993a) Site-directed mutagenesis of highly conserved residues in helix VIII of subunit I of the cytochrome bo ubiquinol oxidase from Escherichia coli: an amphipathic transmembrane helix that may be important in conveying protons to the binuclear center. Biochemistry 32: 11173-80. Thomas, J. W., Puustinen, A., Alben, J. O., Gennis, R. B. & Wikstrom, M. (1993b) Substitution of Asparagine for Aspartate-l35 in Subunit I of the Cytochrome bo Ubiquinol Oxidase of Escherichia coli Eliminates Proton-Pumping Activity. Biochemistry 32: 10923-10928. Thompson, D. A., Gregory, L. & Ferguson-Miller, S. (1985) Cytochrome c Oxidase Depleted of Subunit III: Proton-Pumping, Respiratory Control, and pH Dependence of the Midpoint Potential of Cytochrome a. J. Inorg. Biochem. 23: 357-364. Tsukihara, T., Aoyama, H., Yamashita, E., Tomizaki, T., Yamaguchi, H., Shinzawa-itoh, K., Nakashirna, R., Yaono, R. & Yoshikawa, S. (1995) Structure of Metal Sites of Oxidized Bovine Heart cytochrome c Oxidase at 2.8 A. Science 269: 1069-1 074. Tsukihara, T., Aoyama, H., Yamashita, E., Tomizaki, T., Yamaguchi, H., Shinzawa-Itoh, K., Nakashima, R., Yaono, R. & Yoshikawa, S. (1996) The Whole Structure of the 13- Subunit Oxidized Cytochrome c Oxidase at 2.8 A. Science 272: 1136-1144. van der Oost, J., de Boer, A. P. N., de Gier, J.-W. L., Zumfi, W. G., Stouthamer, A. H. & van Spanning, R. J. M. (1994) The Heme-Copper Oxidase Family Consists of Three Distinct Types of Terminal Oxidases and is Related to Nitric Oxide Reductase. FEMS Microbiol. Lett. 121: 1-10. 55 Verkhovskaya, M. L., Garcia-Horsman, A., Puustinen, A., Rigaud, J. L., Morgan, J. E., Verkhovsky, M. I. & Wikstrom, M. (1997) Glutamic acid 286 in subunit I of cytochrome bo3 is involved in proton translocation. Proc Natl Acad Sci U S A 94: 10128-31. Wang, y. & Margoliash, E. (1995) Enzymic Activities of Covalent 1:1 Complexes of Cytochrome c and Cytochrome c Peroxidase. Biochemistry 34: 1948 - 195 8. Wikstrom, M. (1989) Identification of the Electron Transfers in Cytochrome Oxidase that are Coupled to Proton-Pumping. Nature 338: 776-778. Wikstrom, M. & Krab, K. (1978) Cytochrome c oxidase is a proton pump: a rejoinder to recent criticism. FEBS Lett 91: 8-14. Wikstrbm, M. & Krab, K. (1979) Proton-pumping cytochrome c oxidase. Biochim Biophys Acta 549: 177-22. Wikstrom, M., Krab, K. & Saraste, M. Cytochrome Oxidase - A Synthesis (Academic Press, New York, 1981). Wilrnanns, M., Lappalainen, P., Kelly, M., Sauer-Eriksson, E. & Saraste, M. (1995) Crystal structure of the membrane-exposed domain from a respiratory quinol oxidase complex with an engineered dinuclear copper center. Proc.NatI.Acad. Sci. (USA) 92: 1 1955- 1 1 959. Woodruff, W. H. (1993) Coordination Dynamics of Heme-Copper Oxidases. The Ligand Shuttle and the Control and Coupling of Electron Transfer and Proton Translocation. J Bioenergetics and Biomembranes 25: 177 -188. Woodruff, W. H., Einarsdottir, O., Dyer, R. B., Bagley, K. A., Palmer, 0., Atherton, S. J., Goldbeck, R. A., Dawes, T. D. & Kliger, D. S. (1991) Nature and Functional Implications of the Cytochrome a3 Transients after Photodissociation of CO- Cytochrome Oxidase. Proc. Natl. Acad. Sci. USA 88: 25 88-2592. Yoshikawa, S., Shinzawa-Itoh, K., Nakashima, R., Yaono, R., Yamashita, E., Inoue, N., Yao, M., Fei, M. J., Libeu, C. P., Mizushima, T., Yamaguchi, H., Tomizaki, T. & Tsukihara, T. (1998) Redox-coupled crystal structural changes in bovine heart cytochrome c oxidase. Science 280: 1723-9. Zhen, Y., Mills, D., Hoganson, C. W., Lucas, R. L., Shi, W., Babcock, G. & Ferguson- Miller, S. in Frontiers in Cellular Bioenergetics: Molecular Biology, Biochemistry and Physiopathology Papa, S., Guerrieri, F. & Tager, J. M., Eds. (Plenum Press, New York, 1997) pp. in press. Zickermann, V., Verkhovsky, M., Morgan, J., Wikstrom, M., Anemuller, 8., Bill, E., Steffens, G. C. & Ludwig, B. (1995) Perturbation of the CuA site in cytochrome-c oxides Biochc 56 oxidase of Paracoccus denitrificans by replacement of Met227 with isoleucine. Eur J Biochem 234: 686-93. CHAPTERII ASPARTATE 132 IS ESSENTIAL FOR PROTON PUMPING PROCESS IN RHODOBAC T ER SPHAEROIDES CYTOCHROME c OXIDASE 57 ofsub proton as pre Panza Two rr purifier 0133‘ with D charm pIOIOn wing-r malOr 31 ABSTRACT Aspartate 132, a highly conserved residue in the loop between helix II and helix III of subunit I in Rhodobacter sphaeroides cytochrome c oxidase, is in a good position for proton uptake from the cytoplasm in bacteria or the matrix of mitochondria in mammals, as predicted from the sequence and shown by the crystal structure of bovine and Paracoccus denitrificans cytochrome c oxidase (Iwata et al., 1995; Tsukihara et al., 1996). Two mutants were generated at this position, D132A and D132E, and the enzymes were purified and their properties were compared to those of wild type. Initial studies with D132N in Rhodobacter sphaeroides cytochrome c oxidase were confirmed and extended with D132A and the conservative mutation to glutamate. The D132E mutant shows characteristics similar to wild type (unpublished data), while D132A demonstrates loss of proton pumping ability and inhibition of electron transfer activity (3% of wild type activity retained) (Fetter et al., 1995). As in the case of D132N, D132A does not have major structural changes at the heme sites, reflected by a native visible spectrum with only a slightly lower CO binding efficiency (85% of wild type efficiency). These results support the role of the carboxyl moiety of Asp132 as an important player in proton pumping activity (Fetter et al., 1995). In addition, D132A, like D132N, has abnormal respiratory control responses. Instead of stimulation of oxygen consumption activity of the reconstituted enzyme upon addition of the ionophore valinomycin or the uncoupler CCCP, inhibitory effects are observed. Interestingly, addition of the long chain fatty acid, arachidonic acid, resulted in stimulation of the activity of D132A by up to seven fold, and 58 59 restoration of normal respiratory control. However, no restoration of proton pumping was observed. A reverse proton uptake model will be discussed (Fetter et al., 1996). C) protons (s the hemet translocatt residues t2 opposite s tater may afldNagle, In homologou mumgenesi to inVeStigz have deem; “illitalem PWnping a1 Wiping ac he import-u Mme of 5 Later analys DBSN lest Shows the It INTRODUCTION Cytochrome c oxidase accepts four electrons from cytochrome c, takes up four protons (substrate protons) from the matrix, and reduces one oxygen molecule to water at the heme a3-Cu3 binuclear center. Meanwhile, four more protons (pumped protons) are translocated across the membrane. In order to pump protons, it is necessary that some residues take up protons on one side of the membrane while others release protons on the opposite side. A hydrogen-bonded network including amino acid side chains and bound water may be involved in mediating proton translocation between the two sides (Nagle and Nagle, 1983). In subunit I of cytochrome bo3 ubiquinol oxidase from E. coli, which is homologous to M3 cytochrome c oxidase from Rhodobacter sphaeroides, site-directed mutagenesis of the five most highly conserved acidic residues of subunit 1 has been done to investigate their role in proton translocation (Thomas et al., 1993). All of the mutants have decreased oxygen consumption activity. In one mutant, D135N (E. coli numbering, equivalent to D132N in Rhodobacter sphaeroides cytochrome c oxidase), proton pumping appears to be decoupled from electron transfer activity, i.e., it lost proton pumping activity but retained 45% wild type activity. Therefore, this residue seems to be important for proton pumping. However, in the original studies, the bo3 quinol oxidase of E. coli could not be assayed for proton pumping in the reconstituted state. Later analysis by reconstitution of bo, quinol oxidase into proteoliposomes confirms that D135N lost proton pumping ability but retained 45% electron transfer activity, and shows the respiratory control of D135N similar to that of wild type (Verkhovskaya et 60 d, l canl more Oink trans lfighl ornia (Shap Pdetr Rhoai “min Rhnul c0“Cor l 61 al., 1997). In contrast, the aa3-type cytochrome c oxidase from Rhodobacter sphaeroides can be purified, reconstituted independently, and assayed for proton pumping. Thus more definitive evidence may be obtained with this system. Fetter and coworkers generated D132N in Rhodobacter sphaeroides cytochrome c oxidase, and also observed loss of proton pumping but retention of only 3% of electron transfer activity (Fetter et al., 1995). Rhodobacter sphaeroides cytochrome c oxidase has three subunits, which are highly homologous to the three largest subunits encoded by mitochondrial DNA in mammalian oxidase. The subunit I gene of Rhodobacter sphaeroides cytochrome c oxidase has been cloned, deleted from the bacterial genome, and reintroduced on a plasmid (Shapleigh & Gennis, 1992). This plasmid complements the subunit I-deleted strain and produces a fully active enzyme. Therefore, site-directed mutagenesis is feasible in Rhodobacter sphaeroides cytochrome c oxidase, which serves a good model to study the mammalian version of the enzyme. Previous studies show that the long chain fatty acid, arachidonic acid, can stimulate oxygen consumption activity of D132A mutant four fold at 250 pM concentration. Further characterization of the effect of arachidonic acid on D132A electron transfer activity was carried out. She- menu (Kunl lbali conufi dgefl mouni MATERIALS AND METHODS Site-Directed Mutagenesis Site-directed mutagenesis at residue Aspl32 was performed by the Kunkel method using MUTA-GENE M13 IN VITRO MUTAGENESIS kit from BIO-RAD (Kunkel, 1985). The primers used for mutagenesis are shown in Figure 1. A 470 bp XbaI/Kpnl fragment of cytochrome c oxidase subunit I gene (COXI gene) was cloned into M13mp18 and used as a template for site-directed mutagenesis. The XbaI/Kpnl fragment containing the mutation was subcloned into plasmid pJ S3. The plasmid pJS3 was digested with EcoRI/Hindlll, and the 4.2 Kb EcoRI/HindIII fragment containing the mutation was subcloned into an expression vector, plasmid pRK415-l, followed by transformation into the Ecoli host strain 817-], which is competent to conjugate into Rhodobacter sphaeroides strain J 8100, a COXI-deleted strain. The base pair change in each mutant was confirmed by sequencing using the dideoxy chain termination method (Sanger et al., 1977) with Taq DNA polymerase kit, (TaqTrak Sequencing System, from Promega), and no secondary mutations were found. Growth of Rhodobacter sphaeroides The cells were grown on Sistrom’s plates, pH 7.0, with antibiotics: streptomycin (50 ug/mL), spectinomycin (50 ug/mL), and tetracycline (1 ug/mL) at 30°C for 2-3 days after conjugation. A single colony was picked and restreaked on the new Sistrom’s plates with the same antibiotics. Cells were grown at 30°C for 3 days, then inoculated into three 100 mL volumes of Sistrom’s media with the same antibiotics and grown at 30°C for 62 63 Figure 1: Wild type amino acid sequence (upper) and oligonucleotide sequences (lower) of wild type and mutants at position 132 in subunit I of Rhodobacter sphaeroides cytochrome c oxidase. 64 AAP_I_)_MAF WT 5’- GC GCG CCG GAC ATG GCC TT -3’ D132A 5’- GC GCG CCG GCC ATG GCC TT -3’ D132E 5’- CC GCG CCG GAA ATG GCC TT -3’ Figure 1 about grown man supplt inediu when 1 min at afimu chba. New B: Memb, 65 about 2 days until the absorbance reading at 660 nm was 1.5 (exponential phase of cell growth). This starter culture was transferred to a 12 L fermentor containing Sistrom’s media and antibiotics. The cells were grown under vigorous stirring, and with air supplemented to 30% 02. As the cells metabolized the succinic acid in Sistrom’s medium, the pH was adjusted back to 7.0 for complete growth. Antifoarn was added when necessary. Cells were harvested by centrifuging in a GS3 rotor at 9000 rpm for 20 min at 4°C. The cell pellets were kept at -80°C. Since a very low yield of cytochrome M3 for the D132A mutant was obtained by growing in the fermentor, cells were grown again in 2800 mL Fembach flasks with the same growth condition except that no pH adjustment was made and no antifoarn was added. The aeration conditions in the Fembach flasks are probably better than that in fermentor, due to vigorous shaking in a New Brunswick shaker. Membrane Preparation Frozen cells were resuspended in 50-70 mL of 50 mM KHzPO4, pH 7.2, and then homogenized, after which the volume was brought up to 100 mL with the same buffer. To the cell suspension, 2 ug/mL pepstatin, 2 ug/mL leupeptin, 1 mM EDTA, 50 ug/mL DNase I, and 1 mM PMSF were added. The cells were broken by two passages through a French pressure cell at pressures >12,000 psi for the first time, and pressures >18,000 psi for the second time. Whole cells and debris were removed by centrifugation in HB-4 tubes at 11,000 rpm for 20 min at 4°C. Following the addition of 0.4 mM PMSF, 1 ug/mL pepstatin and leupeptin, membranes were centrifuged in the ultracentrifuge at 45000 rpm 1 sucrose, 50r lmli PMS} 15% sucrose EDTA in Ti- gradient. and interface harm on 50 mM hours at 4°C Hl'droxyapat 66 45000 rpm for 1.5 hours at 4°C twice. The pellets were resuspended in 15 mL of 5% sucrose, 50 mM KHZPO4, pH 7.2, 1 mM EDTA, 2 ug/mL pepstatin, 2 ug/mL leupeptin, 1 mM PMSF and homogenized. A sucrose gradient was prepared by adding 8 mL of 25% sucrose over 12 mL of 60% sucrose (w/v) in 50 mM KH2P04, pH 7.2, 1 mM EDTA in Ti-70 tubes. The homogenized sample was loaded on the top of the sucrose gradient, and spun at 38,000 rpm for 4 hours at 4°C to remove the outer membrane. The interface band, which contains cytoplasmic membranes, was unloaded and diluted 3 times with 50 mM KHZPO4, pH 7.2, 1 mM EDTA, and centrifuged at 45,000 rpm for 1.5 hours at 4°C. The pellets were stored at -80°C. Hydroxyapatite Chromatography The membranes were resuspended in 10 mM KHZPO4, pH 7.0, 0.2 M KCl, 1 mM EDTA, and protease inhibitors, and dissolved in 4% lauryl maltoside for 10 min, then dissolved in 4% CHAPS for 10 min. The solution was centrifuged at 18000 rpm for 20 min at 4°C, and the supernatant was loaded onto a Bio-Gel HTP (Bio-Rad) column, which had been resuspended in 10 mM KHZPO4, pH 7.2 and equilibrated with 10 mM KHZPO4, pH 7.0, 100 mM KCl, 1 mM EDTA, 0.2% lauryl maltoside, and 0.6% CHAPS. The column was washed with above bufi‘er until the yellow pigment was completely removed, then eluted with 300 mM KHZPO4, pH 7.0, 100 mM KCl, 1 mM EDTA, 0.2% lauryl maltoside, and 0.6% CHAPS. The cloudy lipid material was first eluted. The colored fractions containing cytochromes (aa3, bc,, c) were pooled and lauryl maltoside was added to a final concentration of 1%, followed by dialysis against 4 L of 10 mM 67 KHZPO4, pH 7.2, 1 mM EDTA in well washed Spectra-Por tubing for 4 hours. The cytochrome fraction was stored in 3% lauryl maltoside at -80°C. Purification by FPLC The cytochrome fraction was filtered and loaded onto two tandem DEAE-SPW Fast Performance Liquid Chromatography (FPLC) columns (Tosoh Corp.). The columns were pre-equilibrated with 10 mM KH2P04, pH 7.6, 1 mM EDTA, 0.2% lauryl maltoside, and bound proteins were eluted with 52 mL of linear gradient from 0-1 M KCl in the same buffer at a flow rate of 0.5 mL/min. Peak fractions determined from the dithionite-reduced minus ferricyanide oxidized spectrum were pooled and lauryl maltoside was added to 3% final concentration. The sample was applied to DEAE-SPW columns again, and the bound proteins were eluted with same gradient as above at 0.6 mL/min. Peak fractions were collected and stored at -80°C. Reconstitution of Cytochrome c Oxidase All glassware was rinsed with ethanol, and then with distilled water to remove any residues of detergent. Asolectin was suspended to 40 mg/mL by sonication in 1% cholate, 75 mM HEPES, pH 7.4 at 0°C under argon. A Heat Systems-Ultrasonics sonicator (Model W225) was used at a power setting of 5 for intervals of 30 s on, 30 s off until clarity was reached. The suspension was centrifuged at 10,000 rpm for 15 min at 4°C to remove titanium particles. 4% cholate was added to the purified cytochrome c oxidase. After this, the enzyme was added to the above resuspension buffer to a final concentration of 0.65 nmol/mL and dialyzed at 4°C with rapid stirring in 100 volumes of 68 75 mM HEPES, 14 mM KCl, 0.1% cholate, pH 7.4 for 6 hours; 100 volumes of 75 mM HEPES, 14 mM KCl, pH 7.4 for 12 hours; 100 volumes of 50 mM HEPES 24 mM KCl, 15 mM sucrose, pH 7.4 for 12 hours; and 500 volumes of 1 mM HEPES, 44.6 mM KCl, 43.4 mM sucrose for 12 hours. Visible Spectroscopy In a solution containing 50 mM KHZPO4, pH 7.2, 1 mM EDTA, and 0.2% lauryl maltoside, dithionite-reduced and air-oxidized absolute spectra of the purified enzyme were recorded. To obtain CO difference spectra, 1 mL of CO was bubbled over 5 min into 0.5 mL of dithionite-reduced cytochrome aa3 at 25°C. This results in the complete saturation of the oxidase with CO. Extinction coefficients used were A8606-650 = 40 cm'l mM'1 for dithionite-reduced spectra and A8606-650 = 24 cm'1 mM'I for dithionite-reduced minus ferricyanide oxidized difference spectra. Oxygen Consumption Assay In 1.8 mL of 50 mM KHZPO4, pH 6.5, containing 2.8 mM ascorbic acid and 0.05% lauryl maltoside, 4 uL of purified enzyme was added, followed by addition of 1.11 mM TMPD, and 30 uM cytochrome c. Turnover number (molecular activity) was calculated as electrons/second. Proton Pumping Assay The assay was followed at 556.8 minus 504.7 nm on an Aminco DW2a dual wavelength spectrophotometer. Cytochrome c was fully reduced by adding dithionite, and purified by Sephadex G-75 column in 1 mM HEPES, and 44.6 mM KC1, pH 7.4. 69 0.58 uM (final concentration) cytochrome c was used in each assay. Cytochrome c was tested on the blank vesicles to adjust the pH so that no pH change occurred upon addition. 50 uM NaHCO3, 44 mM sucrose, and 45 mM KCl, pH 7.4 was used as the assay buffer. Vesicles containing 0.1 nmol cytochrome oxidase were added to the stirred 3 mL cell in an Aminco DW2a dual wavelength spectrophotometer, followed by equilibration with 3.2 uM valinomycin and 0.4 nM CCCP. Rapid acidification occurred upon addition of 2 uL of 0.8 uM cytochrome c. To completely dissipate the pH gradient, 5 uM CCCP was added. A second addition of cytochrome c at the same concentration resulted the net alkalinization. 1.5 uL of 0.5 mM HCl acid standard was used in each assay. Respiratory Control Ratio (RCR) Assay The assay was done in 10 mM HEPES, 41 mM KCl, 38 mM sucrose, pH 7.4 on an oxygen polarograph. Ascorbate (5.5 mM) and TMPD (0.3 mM) were added to fully reduce 30 uM cytochrome c. 3.2 uM of valinomycin and 5 uM CCCP were added. The uncoupled rate (afier addition of valinomycin and CCCP) was divided by coupled rate (before their addition) to give the RCR. RESULTS Visible Spectra, Electron Transfer and CO Binding Assay The activity of the enzyme was examined polarographically using horse heart cytochrome c as substrate. Ascorbate and TMPD were added to keep the cytochrome c fully reduced throughout the assay. D132A and D132E have electron transfer activity of 43 s'I and 1300 s", respectively, compared to 1700 s'l for the wild type (Table 1). These results indicate that a carboxyl group at position 132 is critical for overall enzyme activity. Visible spectra of D132A and D132E appear to be highly pure of cbb3 oxidase (Figure 2). The reduced spectra of D132A and D132E are similar to wild type at Ct (606 nm) and Soret (444 nm) region, suggesting that the heme sites of these mutants are unaltered. Heme a3 is five coordinated, and has a labile distal coordination site available to bind ligands. Both O2 and CO can bind at this site. Under conditions of exposure to CO that give 100% conversion of wild type cytochrome c oxidase to a CO bound form, D132A mutant is 85% saturated and D132E is 96% saturated (Table 1), suggesting that the heme a3-CuB binuclear center is little changed for D132E and only slightly changed for D132A, possibly due to the presence of some partially denatured form. Reconstitution of Enzyme and Respiratory Control Assay To test for proton pumping activity, the purified mutant enzyme was reconstituted into phospholipid vesicles. The activities of reconstituted wild type, D132E and D132A are 900 s", 1000 s", and 50 s", respectively (Table 1). The respiratory control assay is used to test for intactness of vesicles and effective insertion 70 mm o ed ow mv +V BEES coHmecooB 0850 vowing mfivfin 00 ..\° 3+: mum .«o £38m mo bréom emetic SunfloEm wcmvamm 00 Ba .xocomoEm 9:955 885 .EUMV cued Sacco boufiamom .oEEm voHBamcoooM can Bur—5m .«o bm>uo< Hammad; Eben—m ”8%me o 0805830 noumoamexma begohoxx 35:2 was 093. 25» mo nomfimafioo A 03mg. 72 Figure 2: Comparison of visible spectra of purified wild type and mutant Rhodobacter sphaeroides cytochrome c oxidase. Dithionite-reduced absolute spectra of the purified wild type, D132A, and D132E oxidases were recorded with a Perkin-Elmer Lambda 6 UV—visible spectrophotometer at 25 °C. 0.3-0.5 uM oxidases in 100 mM KHZPO4, pH 7.2, 0.2% lauryl maltoside were used in each assay. 73 :[ 0.008 A.U. Absorbance m D132A D132E I I I I I I I I I I I I I I I I l I I I J I I 400 450 500 550 600 650 700 Wavelength (nm) Figure 2 74 of the oxidase into the membrane. If the vesicles are intact and the oxidase is inserted properly, addition of cytochrome c to the reconstituted oxidase causes immediate production of a proton gradient, which inhibits electron transfer activity; therefore, the oxygen consumption rate decreases. Addition of valinomycin (potassium ionophore) and CCCP (uncoupler) releases the electrical and proton gradient, respectively, and stimulates the oxygen consumption rate. The respiratory control ratio is the rate of Oz consumption in the presence of uncoupler divided by that in the absence of uncoupler. Wild type and D132E have RCR of 5 and 7, respectively, while D132A shows an RCR less than 1 (Table 1, Figure 3), which indicates inhibition instead of stimulation by addition of valinomycin and CCCP. If the vesicle is leaky or the oxidase is not inserted properly, there should be no respiratory control, i.e., no proton gradient is established after the addition of cytochrome c; therefore, no response can be seen when valinomycin and CCCP are added. In the case of D132A, the fact that it does respond to the releasing of electrical and proton gradient proves that the oxidase inserts properly and maintains an intact membrane. About 90% of the D132A oxidase was inserted in the correct orientation, according to the ratio of the absorbance at 605 run when reduced by the water soluble reductant, ascorbate, and by the lipid penetrating reductant, TMPD. Proton Pumping Assay Proton pumping was assayed using the pH-sensitive dye, phenol red, to measure extravesicular pH changes in response to the addition of reduced cytochrome c. For wild type enzyme, external acidification is followed by alkalinization, which results from proton uptake on the inside of the vesicles to convert 02 to water, causing leakage of 75 Figure 3: Respiratory control assay of reconstituted wild type and mutant Rhodobacter sphaeroides cytochrome c oxidase. The left panel shows the effect of adding valinomycin (4.4 uM) followed by CCCP (5.6 uM), and the right panel shows the effect of adding CCCP followed by valinomycin. Rates of oxygen consumption were measured polarographically at 25 °C in 10 mM Hepes, 41 mM KCl, 38 mM sucrose, pH 7.4, with 5.5 mM ascorbate, 0.3 mM TMPD, and 30 “M cytochrome c. 76 m 8sz .rl. i. .fi K @000 @000 Asn. Biochemistry 34: 4428-4433. Hosler, J. P., Ferguson-Miller, S., Calhoun, M. W., Thomas, J. W., Hill, J., Lemieux, L., Ma, J., Georgiou, C., Fetter, J ., Shapleigh, J. P., Tecklenburg, M. M. J ., Babcock, G. T. & Gennis, R. B. (1993) Insight into the Active-Site Structure and Function of Cytochrome Oxidase by Analysis of Site-Directed Mutants of Bacterial Cytochrome aa3 and Cytochrome b0. J. Bioenerg. Biomembr. 25: 121-136. Iwata, S., Ostermeier, C., Ludwig, B. & Michel, H. (1995) Structure at 2.8 A resolution of cytochrome c oxidase fiom Paracoccus denitrificans. Nature 376: 660-669. Kunkel, T. A. (1985) Rapid and Efficient Site-Specific Mutagenesis Without Phenotypic Selection. Proc. Natl. Acad Sci. USA 82: 488-492. Sanger, F ., Nicklen, S. & Coulson, A. R. (1977) DNA sequencing with chain-terminating inhibitors. Proc. Natl. Acad. Sci U. S. A. 74: 5463-5467. Shapleigh, J. P. & Gennis, R. B. (1992) Cloning, Sequencing, and Deletion from the Chromosome of the Gene Encoding Subunit I of the aa3-type cytochrome c oxidase of Rhodobacter sphaeroides. Mal. Microbial. 6: 635-642. Takahashi, E. & Wraight, C. A. (1991) Small Weak Acids Stimulate Proton Transfer Events in Site-Directed Mutants of the Two Ionizable Residues, GluL212 and AspL213, in the QB-binding Site of Rhodobacter Sphaeraides Reaction Center. F EBS Lett. 283: 140- 144. Thomas, J. W., Puustinen, A., Alben, J. O., Gennis, R. B. & Wikstrom, M. (1993) Substitution of Asparagine for Aspartate-135 in Subunit I of the Cytochrome bo 87 88 Ubiquinol Oxidase of Escherichia coli Eliminates Proton-Pumping Activity. Biochemistry 32: 10923-10928. Tittor, J., Soell, C., Oesterhelt, Butt, H.-J. & Bamberg, E. (1989) A Defective Proton Pump, Point-Mutated Bacteriorhodopsin Asp 96 —) Asn is Fully Reactivated by Azide. EMBO. J. 8: 3477-3482. Tsukihara, T., Aoyama, H., Yamashita, E., Tomizaki, T., Yamaguchi, H., Shinzawa-Itoh, K., Nakashima, R., Yaono, R. & Yoshikawa, S. (1996) The Whole Structure of the 13- Subunit Oxidized Cytochrome c Oxidase at 2.8 A. Science 272: 1136-1144. Verkhovskaya, M. L., Garcia-Horsman, A., Puustinen, A., Rigaud, J .-L., Morgan, J. E., Verhovsky, M. I. & Wikstrom, M. (1997) Glutamic Acid 286 in Subunit I of Cytochrome b0 3 is Involved in Proton Translocation. : . Zhen, Y., Mills, D., Hoganson, C. W., Lucas, R. L., Shi, W., Babcock, G. & Ferguson- Miller, S. in Frontiers in Cellular Bioenergetics: Molecular Biology, Biochemistry and Physiopatholagy Papa, S., Guerrieri, F. & Tager, J. M., Eds. (Plenum Press, New York, 1997) pp. in press. CHAPTERIII CONSTRUCTION OF OVEREXPRESSION STRAIN AND IMPROVEMENT OF PURIFICATION OF RHODOBAC T ER SPHAEROIDES CYTOCHROME c OXIDASE 89 ABSTRACT Rhodobacter sphaeroides cytochrome c oxidase has been an excellent system for characterization by mutational studies with biochemical and biophysical analyses. Recent three-dirnensional crystal structures of highly homologous eukaryotic and prokaryotic oxidases have made interpretation of these data even more meaningful. However, a limitation has been that wild type strains of Rhodobacter sphaeroides yield only low levels of oxidase, and the purification procedure was inefficient. A new construct was made by combining the operons containing the subunit I and subunit II/III genes of Rhodobacter sphaeroides oxidase and introducing them into an expression vector, yielding up to a four-fold increase in the level of oxidase expression. In addition, the purification of histidine-tagged Rhodobacter sphaeroides cytochrome c oxidase with NTA-Ni chromatography has been refined to avoid denaturing effects of high concentration of imidazole, resulting in higher purity and a reproducible procedure. The mutant D132A was overexpressed and purified with the improved system. This improved expression and purification of Rhodobacter sphaeroides oxidase should allow more massive spectral and biochemical analysis of mutants, as well as sufficient production of wild type enzyme for crystallization efforts. 90 INTRODUCTION Crystal structures of mammalian and bacterial cytochrome c oxidase show that they share similar structural features at metal centers as well as in the overall protein fold (Iwata et al., 1995; Tsukihara et al., 1995; Tsukihara et al., 1996; Ostermeier et al., 1997). Rhodobacter sphaeroides cytochrome c oxidase only has three subunits, while mammalian oxidase is more complex and has thirteen subunits; the three subunits in Rhodobacter sphaeroides oxidase are highly homologous to the three largest subunits in mammalian oxidase. A site-directed mutagenesis system in Rhodobacter sphaeroides oxidase has been established (Cao et al., 1992; Hosler et al., 1992; Shapleigh & Gennis, 1992) (also see Chapter II, introduction). Therefore, we are in a good position to study the mechanisms and coupling of electron transfer and proton translocation by a site-directed mutational approach. However, various existing Rhodobacter sphaeroides strains express cytochrome c oxidase at a low level even under ideal growth conditions, and levels are even lower for some mutants. In order to carry out various assays and ultimately to crystallize the wild type and mutant oxidases, large amounts of pure enzyme are required; thus, an overexpression system is desirable. The construct in this study increases expression of oxidase to up to four-fold, while maintaining the normal enzyme characteristics. A modified rapid protein purification procedure is also reported. With improved expression and purification, obtaining purer and larger amounts of oxidase is feasible. Overexpression and his-tagged strains of D132A were constructed and purified for further characterization, such as by stopped-flow proton pumping assays. Other 91 92 mutants, at the Asp407 and Arg482 positions, were also purified with the improved protocol (see Chapter IV and Chapter V). MATERIALS AND METHODS Construction of Overexpression Plasmid The procedure used to create the overexpression plasmid pJQ200 is shown in Figure 1. First, a 2281 bp HindIII/Smal fragment from plasmid pJS3 containing the COXI gene, was blunted ended at the HindlII site and inserted into the Smal site in plasmid pUC18. Then, another 4718 bp PstI/PstI fragment from pYJ 100 containing the COXII/III genes was subcloned to the PstI site in the plasmid pUC18. Finally, a 7050 bp EcaRI/HindIII fragment containing the COXI and COXII/III genes was subcloned to the broad-host expression vector, pRK415-1, followed by transformation into the E. coli strain 817-], which acts as a donor in the conjugation with Rhodobacter sphaeroides strains. Both COXI-deleted Rhodobacter sphaeroides strain J 8100 and COXII/III- deleted strain YZZOO were used for conjugation. Small Scale Cytoplasmic Membrane Preparation All the conjugants were grown as described previously (Chapter II, Materials and Methods). The Rhodobacter sphaeroides cytoplasmic membranes were prepared as before (Chapter 11, Materials and Methods) with modifications as described below. Frozen cells from a 200 mL growth were resuspended in 15 mL of 50 mM KHZPO4, pH 7.2, and then homogenized, after which the volume was brought up to 25 mL with the same buffer. The cells were broken by one passage through a French pressure cell at pressures >12,000 psi. Whole cells and debris were removed by centrifugation in HB-4 tubes at 10,000 x g for 20 min at 4°C. After which membranes were pelleted by 93 94 Figure 1: Construction of the overexpression plasmid pJQ200 by subcloning of the COXI and COXII/III operons of Rhodobacter sphaeroides cytochrome c oxidase into expression vector pRK415-1. The HindIII/Smal fiagment from p183 containing the COXI gene was blunted ended at the HindIII site and inserted into the SmaI site in plasmid pUC18, followed by insertion of the PstI/PstI fragment from pYJ 100 containing the COXII/III genes into the PstI site on pUC18. The EcaRI/HindIII fragment containing the COXI and COXII/III operons was subcloned to expression vector, pRK415-1. 95 El 35%. an 8.2 883 Wk. 0 a use Em... .5 aging use. boo am ~ 8ng Tavmfi Illlr ESEESM Em HEM. 28% ESE 96 centrifugation at 200,000 x g for 1.5 hours at 4°C. The membrane pellets were solubilized in 2% lauryl maltoside, 50 mM KH2P04, pH 7.2, 1 mM EDTA, and were ready to take visible spectra. Visible Spectroscopy In a solution containing 50 mM KHZPO4, pH 7.2, 1 mM EDTA, and 2 % lauryl maltoside, dithionite-reduced minus ferricyanide-oxidized spectra of the dissolved membranes were recorded on a Perkin-Elmer Lambda 40P UV-visible spectrometer at 25°C, and an extinction coefficient at 606 nm of 24 cm'I mM'1 was used (Vanneste, 1966) Purification of His-Tagged Cytochrome c Oxidase Mutants were purified with Ni-NTA agarose affinity chromatography (Mitchell & Gennis, 1995) with modifications. All steps were carried out at 4 °C. 40 g of wet cells were resuspended in 200 mL of 50 mM KHZPO4, pH 7.2, followed by two passages through a French pressure cell at 20,000 psi. Whole cells and debris were removed by centrifugation at 20,000 X g for 20 min. The membranes were pelleted from the supernatant by centrifugation at 200,000 x g for 1.5 h. The membranes were homogenized in 10 mM Tris and 40 mM KCl, pH 8.0 at 6 mg protein / mL, and solubilized in 3% lauryl maltoside for 1 h, followed by centrifugation at 200,000 X g for 30 min to remove the unsolubilized membranes. The amount of cytochrome M3 was estimated from the dithionite-reduced minus ferricyanide-oxidized spectrum. Imidazole was added to the solubilized membranes (at 0.7 mg of cytochrome aa3/mL) to a final it "'n I " 97 concentration of 10 mM to prevent nonspecific binding. Packed resin (0.7 mL / mg of oxidase) was added. The mixture was stirred for 1 hr and loaded onto a gravity-flow column. The column was washed with 10 column volumes of 10 mM imidazole, 0.1% lauryl maltoside, 10 mM Tris and 40 mM KCl, pH 8.0, followed by 5 column volumes of 0.1% lauryl maltoside, 10 mM Tris and 40 mM KCl, pH 8.0. The oxidase was eluted with 100 mM histidine, 0.1% lauryl maltoside, 10 mM Tris and 40 mM KCl, pH 8.0. After the Ni-NTA column, all the enzymes were further washed with 0.1% lauryl maltoside, 10 mM Tris and 40 mM KCl, pH 8.0 for three times by centricon filtration to remove Ni-histidine. Construction of Overexpression Plasmid for D132A Mutant Plasmid pJQ100 with the whole construct of the subunit I and [VIII genes was digested with BglII to completion, followed by partial digestion with XbaI. The 906 bp BglII/Xbal fragment from pJQ100 was replaced with the 906 bp BglII/Xbal fragment containing the D132A mutation. The following subcloning and conjugation steps are the same as before (see Construction of overexpression plasmid in Materials and Methods in this chapter). Construction of His-Tagged Plasmid for D132A Mutant A 2281 bp of SmaI/HindIII fragment containing the D132A mutation was subcloned to the plasmid pUC18. Another 300 bp SphI/HindIII fiagment containing the his-tag from the plasmid pJS3-X6H2 (Mitchell & Gennis, 1995) was subcloned to the same plasmid pUC18 to get a his-tagged construct for the D132A mutant. This construct has the short COXI gene, which contains the entire coding region for subunit 1 and two 98 small flanking regions. It is 1883 bp shorter than the regular construct at 5’-end. The following steps to put the his-tagged D132A mutant subunit 1 into the overexpression plasmid are the same as before (see page 93). RESULTS Overexpression System Development in Rhodobacter sphaeroides The overexpression construct pJQ200 has two operons, one containing the COXI gene, and the other containing genes coding for COXII, COXIII, ORFl, ORF2, and ORF3. These two operons are in the same orientation determined by restriction enzyme digestion. ORFl and ORF2 are involved in the formation of the famesyl group and the formyl group of heme A, respectively (Tzagoloff et al., 1990; Tzagoloff et al., 1993). An E. coli Sl7-l strain with the overexpression plasmid pJQ200 was conjugated with the COXI-deleted and COXII-deleted Rhodobacter sphaeroides strains J 8100 and YZ200, and the JQ100 and JQ200 strains were obtained, respectively. Small scale preparations of cytoplasmic membrane of each conjugant were made and visible spectra of solubilized membrane were measured. The expression level of cytochrome c oxidase was calculated by comparing the absorbance of cytochrome b at 560 nm with that of cytochrome a at 606 nm. Wild type Rhodobacter sphaeroides strains Ga, 2.4.1, and CY91 have a b/a ratio of 4.8, 5.0, and 2.9, respectively (Hosler et al., 1992; Zhen et al., 1998). JQ100 and JQ200 have a b/a ratio of 1.3 and 2.0, respectively, both lower than that of wild type strains, demonstrating that both strains yield higher expression of aa, than wild type; i.e., 3.8 fold for JQ100 and 2.5 fold for JQ200 when compared with Ga, and 2.2 fold for JQ100 and 1.5 fold for JQ200 when compared with CY91 (Figure 2). Therefore, the construct with the genes for all the subunits did give overexpression of aa, cytochrome c oxidase. 99 100 Figure 2: Comparison of the expression of Rhodobacter sphaeroides cytochrome c oxidase from the overexpressed construct with regular wild type strains, 2.4.1, Ga and CY91. Small scale cytoplasmic membranes of each strain were prepared and dithionite reduced minus ferricyanide oxidized visible spectra were recorded on a Perkin-Elmer Lambda 40P UV-visible spectrophotometer at 25°C. 101 JQ100 b/a=1.3 Q) 8 CU g JQ200 b/a = 2.0 m '2 CY91 b/a = 2.9 Ga b/a = 4.8 2.4.1 500 550 600 650 700 Wavelength (nm) Figure 2 Hhfidu (llosler room: (NTA); shfi(2 umghu imodm tried and Wfi¢uok Chomam cmwnuk Constru( A; JMOQIH helofp reElllar co Rimuhn Tlr Wiened 102 Histidine-Tagged Cytochrome c Oxidase Purification The conventional method of protein purification has a low yield and is inefficient (Hosler et al., 1992) (Chapter II, Materials and Methods). Mitchell and coworkers fused a six-histidine sequence to the C-terrninus of subunit 1, and used a nitrilotriacetic acid (NTA) affinity column to purify it (Mitchell & Gennis, 1995). However, a slight blue shift (2 nm, data not shown) at the or-band region was observed when their procedure using high imidazole was used to elute the oxidase, probably because imidazole can insert into the protein and bind to hemes, causing denaturation. A new eluant, histidine, was tried and the visible spectra were normal. Histidine is bigger and more hydrophilic than imidazole, thus less likely to bind the hemes in the oxidase. The Ni-NTA affinity chromatography method is faster and gentler and has higher recovery than the conventional method (Figure 3). Construction of Overexpressed and His-Tagged Plasmids for D132A After construction of the overexpressed plasmid for D132A and conjugation with J 8100, membranes were prepared and visible spectra were measured. The expression level of D132A with the overexpression plasmid increased to 1.8 fold compared to the regular construct (b/a ratio of 2.1 vs. 3.8, Figure 4), confirming that the construct is useful for mutants as well as wild type. The His-tagged D132A plasmid with the short subunit I Operon was also transferred into JSlOO. The visible spectra of solubilized membrane showed that the expression level was increased 1.4 fold compared to the regular construct (b/a ratio of 2.7 103 vs. 3.8). The purification procedure was very rapid and the protein had high purity (Figure 4). 104 Figure 3: Purification of histidine-tagged wild type Rhodobacter sphaeroides cytochrome c oxidase with a Ni-NTA affinity column. The upper spectrum (500-700 nm) indicates the solubilized membranes before loading the column, the middle spectrum (500- 700 nm) indicates the runthrough of the column, and the bottom spectrum (400-700 nm) indicates the purified the oxidase after the column. 105 solubilized membrane Absorbance runthrough purified oxidase I I I I I I I I I I I 400 450 500 550 600 650 700 Wavelength (nm) Figure 3 106 Figure 4: Comparison of the visible spectra of the regular construct of D132A with that in the overexpressed construct pJQ200 and in the short COXI operon. The samples were prepared as indicated in the figure legend of Figure 3. 107 Q) 0 5 '8 D132A in pJQ200 O _. .8 (b/a—2. 1) < D132A in short coxI (b/a=2.7) regular D132A (b/a=3.8) j I I I I I I 500 550 600 650 700 Wavelength (nm) Ifigun34 Overexpr ln manipulat and two sr a15'-end, express ox lni the Opposi oxidase w; COXI and 000336 We ”'0 Opero that either ngesting ha503 on DISCUSSION Overexpression of Rhodobacter sphaeroides Cytochrome c Oxidase In order to reduce the size of the overexpression plasmid and make it easy to manipulate, a short subunit I operon which contains the entire coding region for subunit 1 and two small flanking regions was used. It is 1883 bp shorter than the regular construct at 5’-end, suggesting that a 208 bp fragment upstream of the COXI gene is sufficient to express oxidase. Initial studies suggested that a construct with the COXI and COXII/III operons in the opposite orientation did not overexpress oxidase when transferred into JSlOO, and no oxidase was expressed when transferred into YZ200. Therefore, a new construct with the COXI and COXII operons in the same orientation was generated and overexpression of oxidase was observed in both J 8100 and YZ200 strains. The same orientation of these two operons seemed important for expressing oxidase. However, recent studies show that either orientation gives overexpression to the same extent (Zhen et al., 1998), suggesting that the orientation is not critical. Each of the COXI and COXII/III operon has its own promoter region. Therefore, overexpression of the oxidase is most likely dependent upon the presence of these two promoter regions as well as the genes coding for all the subunits, but not the orientation. The original observation of little or no expression with the wrong orientation were probably due to the selection of colonies that had partially deleted the large plasmid and thus were more rapidly growing. It is now Clear that slower growing (small colonies) need to be selected to insure that the large plasmid is intact. 108 of 011 0“ dif in 1. WE CXF cor Ni. 109 For each overexpression strain JQ100 and JQ200, several single colonies were grown and cytoplasmic membranes were prepared. It has been observed that different expression levels of oxidase can be obtained fiom the same strain. Overexpression plasmid pJQ200, a derivative of the tetracycline resistant broad-host-range plasmid pRK415-1, is very large (16150 bp). When the E. coli strain 817-1 containing plasmid pJQ200 is conjugated with JSlOO and YZ200, deletions can happen due to the large size of the plasmid. Conjugants carrying the smaller plasmids tend to grow faster. Therefore, only when the whole plasmid is transferred into Rhodobacter sphaeroides cells can overexpression of oxidase occur. The different expression level from the same strain but different conj ugants may be due to the various mixtures of deleted and undeleted plasmids in the colony. Visible spectra of conjugants with the highest expression level of oxidase were compared with regular wild type (Figure 2). Recent studies show that high pH and maximal aeration gives even higher experssion of oxidase, and a his-tagged wild type overexpression version was also constructed (Zhen etal., 1998). Ni-NTA Chromatography Purification of His-Tagged Cytochrome c Oxidase Rhodobacter sphaeroides cytochrome c oxidase is a large protein associated with the cytoplasmic membrane. Many other proteins are also associated with the membrane. Some of them even have similar cytochrome c oxidase and / or proton pumping ability (e.g. cbb,, bc,, etc.). Removing these contaminant proteins is critical for purifying the oxidase. The conventional purification method for cytochrome c oxidase combines hydroxyapatite and DEAE anion exchange chromatography (Hosler et al., 1992). It takes for the 1116 al chr pur the Ger intr uitl E; ' L.,.) at 0 him" imid risit there 0056; 110 four days to prepare the membranes and purify the protein, while the recovery is low and the purity is not consistent. The overall yield is approximately 30% from cytoplasmic membranes to final product (Hosler et al., 1992). Mitchell and Gennis (1995) deveIOped a histidine-tagged cytochrome c oxidase strain and took advantage of metal-affinity chromatography to purify the oxidase. This new method greatly simplified the purification procedure and higher purity was obtained, while the structure and function of the enzyme appears identical to that purified with the conventional protocol (Mitchell & Gennis, 1995). They .fused a six-histidine tag to the C-terminus of subunit I by first introducing a XhaI site at the stop codon and then replacing the XhaI-Hindlll fragment with a sequence of six histidine codon followed by HindIII site. The resultant construct is 316 bp shorter than the original construct. Since the C-terminus of subunit 1 is located at the cytoplasmic side of the membrane, and far away from the binuclear center, the histidine-tag is not likely to affect the oxidase activities or properties. However, using imidazole as the eluant in their protocol introduced a blue shift in the or-region of the visible spectra of oxidase, since it can replace histidine ligands of heme a and / or heme a,, thereby denaturing the protein. With the new eluant histidine, no spectral shift is observed at the at region, suggesting that the structure at the heme a/a, site is not changed. However, nickel-histidine has absorption in the 520 to 560 nm region. Therefore, it is necessary to wash the sample with buffer for three times to remove the histidine. With the modified protocol, the overall yield is 70 to 90% and the whole process takes only 1 to 2 days. terr Call 18.1 Cher P101 :V'Sl Wu and 111 Applying the Overexpression Construct and His-Tagged Plasmid to D132A Mutant The purpose of developing the overexpression and metal-affinity purification system is to overexpress wild type or mutant cytochrome c oxidases and purify them efficiently for various analyses. D132A is an important mutant for elucidating the proton pumping mechanism. With a new stopped-flow spectral analysis, more quantitative, time-resolved measurements can be made of proton pumping. The new D132A strain in the overexpression plasmid gives almost two fold more oxidase than the regular D132A strain. A second construct with D132A in a short COXI operon with a his-tag in the C- terminus, also gives 1.4 fold higher D132A oxidase expression, while a Ni-NTA column can be applied for rapid purification with much higher yield. With the his-tagged plasmid, 18.6 mg D132A oxidase were obtained from a 10 L growth. Due to the unique characteristics of D132A mutant, the construct allows studying the involvement of the proton exit route in the regulation of proton pumping efficiency (Zhen et al., 1997). This system appears very powerful in the site-directed mutational studies in Rhodobacter sphaeroides cytochrome c oxidase with the potential for time-resolved resonance Raman and crystallography. REFERENCES Cao, J., Hosler, J., Shapleigh, J., Gennis, R., Revzin, A. & Ferguson-Miller, S. (1992) Cytochrome aa3 of Rhodobacter sphaeroides as a Model for Mitochondrial Cytochrome c Oxidase. J. Biol. Chem. 267: 24273-24278. Hosler, J. P., Fetter, J., Tecklenburg, M. M. J., Espe, M., Lerma, C. & Ferguson-Miller, S. (1992) Cytochrome aa3 of Rhodobacter sphaeroides as a Model for Mitochondrial Cytochrome c Oxidase. J. Biol. Chem. 267: 24264-24272. Iwata, S., Ostermeier, C., Ludwig, B. & Michel, H. (1995) Structure at 2.8 A resolution of cytochrome c oxidase from Paracoccus denitrificans. Nature 376: 660-669. Mitchell, D. M. & Gennis, R. B. (1995) Rapid Purification of Wildtype and Mutant Cytochrome c Oxidase from Rhodobacter sphaeroides by Ni2+-NTA Affinity Chromatography. FEBS Lett. 368: 148-150. Ostermeier, C., Harrenga, A., Ermler, U. & Michel, H. (1997) Structure at 2.7 A resolution of the Paracoccus denitrificans two-subunit cytochrome c oxidase complexed with an antibody Fv fragment. Proc. Natl. Acad. Sci. USA 94: 10547-10553. Shapleigh, J. P. & Gennis, R. B. (1992) Cloning, Sequencing, and Deletion from the Chromosome of the Gene Encoding Subunit I of the aa3-type cytochrome c oxidase of Rhodobacter sphaeroides. Mal. Microbial. 6: 635-642. Tsukihara, T., Aoyama, H., Yamashita, E., Tomizaki, T., Yamaguchi, H., Shinzawa-itoh, K., Nakashima, R., Yaono, R. & Yoshikawa, S. (1995) Structure of Metal Sites of Oxidized Bovine Heart cytochrome c Oxidase at 2.8 A. Science 269: 1069-1074. Tsukihara, T., Aoyama, H., Yamashita, B., Tomizaki, T., Yamaguchi, H., Shinzawa-Itoh, K., Nakashima, R., Yaono, R. & Yoshikawa, S. (1996) The Whole Structure of the 13- Subunit Oxidized Cytochrome c Oxidase at 2.8 A. Science 272: 1136-1144. Tzagoloff, A., Capitanio, N., Nobrega, M. P. & Gatti, D. (1990) Cytochrome Oxidase Assembly in Yeast Requires the Product of COX] 1, a Homology of the P. denitrificans Protein Encoded by ORF3. EMBO J. 9: 2759-2764. Tzagoloff, A., Nobrega, M., Gorman, N. & Sinclair, P. (1993) On the function of yeast COXIO and COXll gene products. Biochem Mol Biol Int 31: 593-598. Vanneste, W. H. (1966) The Stoichiometry and Absorption Spectra of Components a and 03 in Cytochrome c Oxidase. Biochemistry 5: 83 8-848. 112 113 Zhen, Y., Mills, D., Hoganson, C. W., Lucas, R. L., Shi, W., Babcock, G. & Ferguson- Miller, S. in Frontiers in Cellular Biaenergetics: Molecular Biology, Biochemistry and Physiapatholagy Papa, S., Guerrieri, F. & Tager, J. M., Eds. (Plenum Press, New York, 1997) pp. in press. Zhen, Y., Qian, J ., Follmann, K., Hosler, J ., Hayward, T., Nilsson, T. & Ferguson-Miller, S. (1998) Overexpression and purification of cytochrome c oxidase from Rhodobacter sphaeroides. Protein Expression and Purification in press: . CHAPTER IV ASPARTATE 407 IN RHODOBACT ER SPHAEROIDES CYTOCHROME c OXIDASE IS NOT REQUIRED FOR PROTON PUMPING OR MANGANESE BINDING (based on published work: Qian et al., (1997) Biochemistry 36: 2539-2543) 114 ABSTRACT Several pathways for proton transport in cytochrome c oxidase have been proposed on the basis of mutational analysis and X-ray structure: at least one for moving “pumpe ” protons from the interior to exterior of the membrane and a separate route for transporting “substrate” protons from the interior to the binuclear metal center to combine with oxygen to make H20. According to the crystal structures of cytochrome c oxidase, Asp407 (Rhodobacter sphaeroides numbering) is at the interface of subunit I and subunit II of the oxidase, in a negative patch proposed to be the proton exit site in a pumping pathway, as well as a possible ligand to Mg [Iwata et al. (1995) Nature 376, 660-669]. Three mutants at the Asp407 position of R sphaeroides cytochrome oxidase, D407A, D407N, and D407C, have been purified and characterized. All showed electron transfer activity, and pH dependence of activity, similar to that of the wild type enzyme and no major structural changes, as evidenced by visible, EPR, and resonance Raman spectroscopy. When reconstituted into artificial vesicles, the purified mutants pumped protons with normal efficiency and responded to the membrane pH and electrical gradients in a manner similar to that of wild type. Furthermore, the EPR spectra and Mn quantitation analysis of mutants grown in high Mn indicated no significant alteration in the Mn/Mg site. These results suggest that Asp407 does not play a critical role in proton translocation or in Mn/Mg binding. 115 INTRODUCTION The mechanism of coupling of electron transfer to proton pumping is a major unsolved problem in understanding cytochrome c oxidase. A model for a proton pump is provided by bacteriorhodopsin (Krebs & Khorana, 1993), in which a hydrogen-bonded relay system is suggested to be important for proton movement, from theoretical considerations (Nagle & Morowitz, 1978) and from structural and spectral analysis (Henderson et al., 1990; Rothschild, 1992). Thus residues that can participate in hydrogen bonding are of particular interest for mutational analysis of proton pathways. In previous studies, Asp132 in an interior loop between helix II and helix III of subunit 1 of Rhodobacter sphaeroides oxidase was mutated to alanine and asparagine. The mutants showed no proton pumping but retained some electron transfer activity, leading to the proposal that this residue provided the entry site in a proton pumping pathway (Fetter et al., 1995). Similar results were obtained with mutants of the homologous cytochrome b0 3 in Escherichia coli (Thomas et al., 1993; Garcia-Horsman et al., 1995). In agreement with this idea, two possible proton pathways have been identified in the crystal structure of Paracoccus denitrificans cytochrome c oxidase (Iwata et al., 1995), which shows that Aspl32 is on the interior side and in a good position to load protons for pumping. Iwata and colleagues also propose that another acidic amino acid, Asp407, located in an exterior loop between helix IX and helix X of subunit 1, could be part of a proton exit path. In the crystal structure of the mammalian enzyme, Yoshikawa and colleagues (Tsukihara et al., 1996) also identify Asp132 as a possible entry site for protons to be pumped but do not invoke Asp407 as an exit site. However, both the P. denitrificans and mammalian oxidase crystal structures show Asp407 at the interface of subunit I and subunit II in a negative patch near a proposed Mn/Mg site, between Cu, and CuB (Figure 1). In this study we examined the effects of three different mutations at the Asp407 position on electron transfer, proton pumping, Mn/Mg binding, and CuA structure. The results suggest that a 116 117 Figure 1: Structure of the interface between subunit I and subunit II of beef heart cytochrome c oxidase [in R. sphaeroides oxidase numbering; coordinates are from Tsukihara et al. (1996), with permission of the authors]. The solid lines indicate the bonds between metals and their ligands. The dash lines indicate the hydrogen bonds. Asp407, His 411, and Asp412 are on the loop between helix IX and helix X in subunit 1. 118 Figure 1 119 carboxyl group at this position is not critical for any of these structural or functional properties of cytochrome c oxidase. MATERIALS AND METHODS Site-Directed Mutagenesis The D407A/N/C mutants were made by PCR overlapping extension methods (Ho etal., 1989). Oligo primers of 18-21 bp were synthesized (Figure 2). For each mutant, four primers were used for the polymerase chain reaction, and plasmid DNA pJ S3, which contains the COXI gene, was used as the template. The 523 bp PCR products containing mutations were digested with BgIII/Sphl, and the 352 bp BglII—Sphl fragment was subcloned to plasmid pJS3-X6H which has six histidines tagged to the C-terminus of COXI gene (Mitchell & Gennis, 1995). The BglII—Sphl region of pJS3-X6H of all the mutants was sequenced to confirm the mutation, and no secondary mutations were found. The pJS3-X6H with the mutation was subject to digestion with EcaRI/Hindlll. This fragment was subcloned to plasmid pRK415—1 , followed by transformation into the E. coli strain 817—], which is capable of conjugating with R. sphaeroides strain JSlOO, a COXI-deleted strain. Single conjugants were grown on Sistrom's plates with antibiotics spectinomycin (50 mg/mL), streptomycin (50 mg/mL), and tetracycline (l mgmL) at 30 °C for 2-3 days. Enzyme Purification All mutant R sphaeroides cells were grown in Sistrom’s medium as described (Hosler et al., 1992) except different Mn and Mg concentrations are indicated in the figure legends. Mutants were purified with Ni-NTA agarose affinity chromatography (Mitchell & Gennis, 1995) with modifications. All steps were carried out at 4°C. Cells were resuspended in 50 mM KH2PO4, pH 7 .2, followed by two passages through a French pressure cell at 20,000 psi. Whole cells and debris were removed by centrifugation at 20,000g for 20 min. The supernatant, which contains the membranes, was pelleted by centrifugation at 200,000g for 1.5 h. The membranes were homogenized in 10 mM Tris and 40 mM KCl, pH 8.0, and solubilized in 3% lauryl maltoside. The solution was 120 121 Figure 2: Wild type amino acid sequence (upper) and oligonucleotide sequences (lower) of wild type and mutants at position 407 in subunit I of Rhodobacter sphaeroides cytochrome c oxidase. WT D407A D407N D407C 122 A s v Q R Y Y 5’- GCG AGC GTC 95c; coc TAT TAT -3’ 5’- GCG AGC GTC @ CGC TAT TAT -3: 5’- GCG AGC GTC _AA_C CGC TAT TAT -3: 5’- GCG AGC GTC TGC CGC TAT TAT -3’ Figure 2 123 stirred for 1 h. The solubilized membranes were centrifuged at 200,000g for 30 min. The amount of cytochrome aa 3 was estimated from the dithionite-reduced minus ferricyanide- oxidized spectrum. Imidazole was added to the solubilized membrane (at 0.7 mg of cytochrome aa3/mL) to a final concentration of 10 mM. Packed resin (0.7 mL)/ mg of oxidase was added, and the mixture was stirred for l h. The mixture was loaded into a gravity-flow column. The column was washed with 10 column volumes of 10 mM imidazole, 0.1% lauryl maltoside, 10 mM Tris and 40 mM KCl, pH 8.0, followed by 5 column volumes of 0.1% lauryl maltoside, 10 mM Tris and 40 mM KCl, pH 8.0. The oxidase was eluted with 100 mM histidine, 0.1% lauryl maltoside, 10 mM Tris and 40 mM KCl, pH 8.0. All the enzymes after Ni—NT A column were further washed with 0.1% lauryl maltoside, 10 mM Tris and 40 mM KCl, pH 8.0 for three times to remove histidine. Reconstitution of Cytochrome c Oxidase Reconstitution of oxidase into phospholipid vesicles was conducted as described (Fetter et al., 1995). EPR Spectroscopy For the mutants grown in high Mn (700 mM) and low Mg (50 mM), adventitiously bound Mn was removed by incubating the purified enzymes in 50 mM EDTA for 30 min at 4 °C (Hosler et al., 1995), and the oxidase was separated from EDTA by running through a G-75 column. Electron transfer, proton pumping, CO binding, and EPR spectroscopy were measured as described (Hosler et al., 1992). Resonance Raman Spectroscopy Resonance Raman spectra were measured as described (Mitchell et al., 1996), except that the excitation wavelength is 438.4 nm, and an average of six scans is shown. Mn Content Determination Quantitation of the bound Mn was performed as described (Hosler et al., 1995), except that the spectra were taken on a Bruker ESP300E. RESULTS Visible Spectra, Electron Transfer, and CO Binding Assays Visible spectra of these mutants show that they are highly pure of cbb 3 oxidase and have spectral characteristics similar to those of wild type, although there is a slight blue shift in the a band and Soret band (less than 1 nm; see Figure 3). Purified D407A, D407C, and D407N mutants have 1100, 650, 1400 5‘1 activities, which are 85%, 50%, and 100% of wild type oxidase activity, respectively (Table 1). The activities of reconstituted mutants are 700, 550, and 700 3'1 for D407A, D407C, and D407N, respectively, compared to 1200 S'1 for wild type. The activities after the reconstitution are lower than the activities before the reconstitution, partly because the assay conditions for the reconstituted enzyme are not optimal for maximal activity and, in this case, the mutants may have lesser stability than wild type during the long dialysis. Other possibilities to explain the loss of activity include a different sensitivity to pH or less of the enzyme reconstituted in the correct orientation. The pH dependence of the activity of purified D407A was compared to wild type over the pH range of 6.0—9.0, and a similar pH profile was found (Figure 4). After bubbling 1 mL of CO into dithionite-reduced cytochrome M3, 100% of D407A, 70% of D407C, and 90% of D407N are converted to the CO-bound form, compared to 100% for wild type under these conditions. This indicates that no major conformational change has occurred at the binuclear center (Table 1). D407C has the lowest activity before and after reconstitution and the lowest CO binding among all the mutants, suggesting that a cysteine at this position may have additional detrimental effects unrelated to the loss of a carboxyl, such as seeking a more hydrophobic environment or forming a disulfide bond with another cysteine residue, either of which could lead to structural instability. 124 125 Figure 3: Comparison of visible spectra of purified wild type and mutant cytochrome aa3 of Rhodobacter sphaeroides. Dithionite-reduced absolute spectra of the purified enzyme were recorded with a Perkin-Elmer Lambda 6 UV—visible spectrophotometer at 25°C. The assay solution includes 100 mM KHZPO4, pH 7.2, 0.2% lauryl maltoside, and (A) 0.6 uM D407A, (B) 0.3 uM D407C, (C) 0.7 uM D407N, and (D) 2 uM wild type cytochrome aa3. 126 Absorbance WT D407A / D407N D407C ff 1 I I I i I I I I I I j 1 I I I I 400 430 560 550 600 650 700 Wavelength (nm) Figure 3 127 on 0.3.0 o own one 2.on 8 ed a 2K 8: 283 8. not; a 2K 2: a <85 2: 3.3 a 8082 8382 25 25 A25 880 + Ara Are 23» 8 339208 E> +v 0850 3933332 2:50 3E5:— mqeaa 8 s or: mom a 558 a 558 as? esteem waeam oo 28 538.... mean? :95 .956 Seam .380 bocfiaom .oaaam eaauasom as 3E2 .«o bm>uo< SHEEP 5.32m ”egg 9 080.3830 noemoamgmu auaoeaomoafi 35:2 28 out; ES, mo acmtaaoo : 2an 128 Figure 4: pH dependence of activity of purified wild type (D) and D407A (O) cytochrome c oxidase. Activities were assayed polarographically using 0.002 uM D407A and 0.006 [1M wild type in 50 mM potassium phosphate for pH 6.5-7.5 and in 50 mM Tris for pH 7 .5-9.0, including 2.8 mM ascorbate, 0.056% lauryl maltoside, 2 mg of cholate- solubilized soybean phospholipid, and 1.1 mM TMPD. Relative activities were calculated as a percent of the highest turnover measured: for wild type (1050 s") and for D407A (1200 s"). 129 mAw vnzswrm mm 3 3w 2 as nmv A20 <5.on 0 BB 0 [1. I coo I com I com I com I ooofi I oofifi I oomfi J 82 ocvfi KirAnov Re pro or} 130 Respiratory Control and Proton Pumping The respiratory control ratio (RCR) is a measure of the completeness of reconstitution and intactness of the artificial membrane vesicles, as well as the innate properties of the mutant oxidase. In the reconstituted oxidase vesicles, an increase of oxygen consumption is expected after addition of the ionophore valinomycin or uncoupler CCCP, due to their ability to release the inhibiting effect of an electrical or proton gradient which is formed when activity is stimulated by addition of cytochrome c. The RCR's (activity after ionophore addition divided by activity before) of D407A, D407C, and D407N are 7, 9, and 7, very close to that of wild type (Table 1). When proton pumping is measured using the pH-sensitive dye phenol red, acidification is observed, followed by alkalinization due to consumption of protons to make H20 in the vesicle interior (Figure 5). The H+/e' ratio ranges from 0.4 to 0.8 for the Asp407 mutants, while the wild type has an H+/e' ratio of 0.5—0.8, indicating that removal of the carboxyl causes no significant change in proton pumping efficiency, within the accuracy of the assay. Mn and CuA Binding Sites Cytochrome c oxidase contains two copper ions at the CuA site and, under certain growth conditions (Hosler et al., 1995), close to stoichiometric Mn. Both metals give unique, quantifiable EPR signals. When R. sphaeroides is grown with high Mn (700 mM) and low Mg (50 mM), Mn appears to be inserted into what is normally a Mg site, and the strong Mn EPR signal in the g = 2.0 region masks the CuA signal. However, under growth conditions of high Mg (1200 mM) and low Mn (0.5 mM), the EPR spectrum of oxidase shows only the Cu signal, due to the fact that Mn content of the purified oxidase is below the detection level of EPR spectroscopy and the site is presumably occupied with EPR-silent Mg. When R. sphaeroides is grown with high Mg (1200 mM) and medium Mn (27 mM), both Mn and CuA signals show in the EPR spectrum of the purified oxidase. Wild type and mutants were grown under the three different conditions, and EPR spectra at 110 K and 10 K are shown in Figure 6 and Figure 7, 131 Figure 5: Proton pumping activity assay of reconstituted wild type and mutant cytochrome c oxidase. The extravesicular pH change was measured at 556.8 minus 504.7 by phenol red at 22°C with an Aminco DW2a spectrophotometer. Wild type and mutant cytochrome c oxidase (0.065 nmol) were reconstituted into soybean phospholipid vesicles, which were added in 2.5 mL of assay solutions containing 50 IIM NaHCO,, 45 mM KCl, 44 mM sucrose, and 50 IIM phenol red, pH 7.4, followed by equilibration with 3.2 ItM valinomycin and 0.4 nM CCCP. Rapid acidification was detected upon addition of 0.8 nmol of cytochrome c (0.32 IIM in final concentration). After addition of 5 IIM CCCP and the same amount of cytochrome c, pure alkalinization was detected. HCl standard acid (0.5 nmol) was added before (Hf) and after (Hf) addition of concentrated CCCP. The measurements of Ho”, HE”, and the pure alkalinization were used to calculate the electron transferred (e') and the H+/e' ratio can be calculated from the electron transferred (e') and acidification (H'). _ ALIA‘ 4 "I.4II‘7 A 132 133 Figure 6: EPR spectra of purified wild type and mutant Rhodobacter sphaeroides cytochrome c oxidase measured at 110K. Cells were grown in (A) high [Mn] (700 M) and low [Mg] (50 M), and (B) high [Mg] (1200 M) low [Mn] (0.5 uM). EPR spectra were recorded on a Bruker ESP300E series spectrometer by using a TE102 cavity, at X-band at 110 K. 134 fiflflfi 6C 25 ozone: coon ocnm cavm cwnm oa~m c9_m oocm _ <5.on ZSVQ 0“.on 0 22mm EV omen 6v 23a 25%“: o@VM aomm econ hr 053D Zbon <85 2V 135 Figure 7: EPR spectra of purified wild type and D407A mutant Rhodobacter sphaeroides cytochrome c oxidase measured at 10K. Cells were grown in (A) high [Mg] (1200 IIM) and low [Mn] (0.5 IIM), (B) high Mg (1200 IIM) and medium Mn (27 IIM), and (C) high [Mn] (700 IIM) and low [Mg] (50 IIM). EPR spectra of Mn were shown at g = 2.0 region. The spectra were recorded on a Bruker ESP300E series spectrometer by using a TB“); cavity, at X-band at 10 K. Purified mutants (0.3 mL) were used for each EPR spectrum (Data of D407N and D407C are not shown). 136 WT A M D407A ' I ' I f I ‘ I ' I I I T I r I i I ' I ' 2150 2350 2550 2750 2950 3150 3350 3550 3750 3950 4150 4350 B C WT W W J D407A D407A . . . . , 2700 3100 3500 2700 3100 3500 Magnetic Field (Gauss) Magnetic Field (Gauss) Magnetic Field (Gauss) Figure 7 [C 1 cy 137 respectively. At 110 K, all the EPR spectra measured on wild type and mutant cytochrome c oxidase show similar signal at Mn site as well as Cu A site (Figure 6). Under low Mn/high Mg growth conditions, no Mn is bound in wild type or mutant D407A, revealing a normal CuA signal for the mutant (Figure 7A). In the high Mg/medium Mn grth condition, both wild type and mutant D407A reveal similar signals for CuA and Mn (Figure 7B), although a slight broadening of the Mn lines is observed in the mutant compared to wild type, leading to an attenuated Mn signal. At high Mn/low Mg growth conditions, both the D407A mutant and wild type show the typical six-line hexaqua Mn2+ spectral characteristics in the g = 2.0 region (Figure 7C), suggesting that mutation at the 407 position does not change the Mn ligand structure. Slight broadening is also observed in this signal, indicative of some minor increase in disorder in the Mn binding site. The Mn content was quantified by acidifying and precipitating the protein and comparing the Mn signal to a set of quantitation standards. Under high Mn/low Mg conditions, purified wild type and mutant D407A oxidases showed similar Mn incorporation, 88% and 85% of stoichiometric, respectively. Resonance Raman Spectroscopy The resonance Raman spectra of D407A and D407N are identical to wild type in both the high and low frequency region, indicating undisturbed heme a and a3 environments (Figure 8). However, D407C has about a 50% decrease in intensity at 214, 365, and 1662 cm", revealing a significant effect on the heme a3 environment that may relate to partial denaturation and lower activity of this mutant. 138 FIGURE 8: Resonance Raman spectra of purified wild type and mutant cytochrome c oxidase. High frequency region (A) and low frequency region (B) are shown for all the mutants. The mutants were reduced with dithionite, and the spectra were measured at 43 8.4 nm excitation. The indicated modes are 214 cm", Fe—NHi, stretch of heme a3; 365 cm", ring bending of heme a3; 1611 cm", formyl stretch of heme a; 1624 cm", vinyl stretch of heme a; 1662 cm", formyl stretch of heme a 3. 139 8 A. _ A783 aim 583— 80— w 255E p.53 55. 5.5 a J _ d d l- 8: 8: 82 - 51.1 L _ a . Arrsualul 4. fi__.a.‘ acti‘ cytc posi 61 a exit 010.1 proc 00115 clus‘ 0407- DISCUSSION Previous data demonstrated that aspartate-132 is critical for proton pumping activity in both R sphaeroides cytochrome aa 3 (Fetter et al., 1995) and E. coli cytochrome ba 3 (Thomas et al., 1993). It is located on the interior of the membrane in a position that makes it a good candidate for an entry site for protons to be pumped (Iwata et al., 1995; Tsukihara et al., 1996). For understanding the mechanism of coupling between proton translocation and the oxygen chemistry, it is also important to define the exit site for protons. An exit site in close proximity to the binuclear center where the oxygen chemistry occurs would support the idea of direct coupling between these processes, whereas proton efflux at a distance from that region might encourage consideration of an indirect, conformational coupling model. Aspartate-407, an acidic residue in subunit 1, is located in a negatively charged cluster at the interface between subunit I and subunit 11, immediately above the heme a 3— CuB center (Figure 1). In the context of a direct coupling mechanism, Iwata et al. (1995) suggest Asp407 as a possible proton exit site. The crystal structure also shows that Asp407 is close to the Mn/Mg ligands, His411 and Asp412, identified by mutational analysis (Hosler et al., 1995) and by analysis of the mammalian crystal structure (Tsukihara et al., 1995). In the bacterial enzyme structure, Iwata and colleagues suggest Asp407 as a possible Mg ligand itself. If either of these suggested roles is correct, substitutions at this position might be expected to alter either proton pumping or Mn/Mg binding, or both. The visible spectra of mutants D407A, D407C, and D407N show that the a band of all the mutants has a maximum at 605 nm, similar to wild type oxidase, except for D407C (604 nm). In addition, CO binding analysis shows that all substitutions except cysteine retain high efficiency of CO binding. These results suggest that replacement of a negatively charged residue with a neutral residue at this position does not alter the 140 611) 1101 Pro bet tha‘ ISL cha bin dist Ulll‘ “it 1111 03m Ofr 61111 141 environment at the heme a 3—Cu3 redox center and imply that the charge at this position is not critical for stabilizing the protein structure. The oxygen consumption assay of the purified mutants shows that they all have high activity, approaching that of wild type, except D407C, which is only ~50% active. A cysteine at this position may interact in some way to destabilize the protein, by seeking a more apolar environment or forming an inter- or intra-molecular disulfide bond. From the crystal structure, possible candidates would be Cys64, Cys88 in the loop between helix 1 and helix II of subunit 1, or a cysteine in subunit II that ligates CuA. All mutant pump protons with efficiency similar to that of wild type, indicating that Asp407 is not required for proton translocation. A new model proposed by Tsukihara et al. (1996) suggests two possible pumped proton channels, in addition to a channel for substrate protons. Since neither of these pumping pathways accesses the binuclear center, they imply an indirect coupling of electron transfer to proton translocation involving a conformational change driven by the oxygen chemistry, but at a distance fiom the active site (Tsukihara et al., 1996). The lack of effect of the Asp407 mutations on proton pumping efficiency, or pH dependence of activity, is consistent with this model in that this residue is not predicted to be a part of either of the proposed “indirect” channels. However, these results certainly do not rule out the histidine cycle/shuttle model (W ikstrom etal., 1994; Iwata et al., 1995) since any one of a number of residues, or bound water in the region above the active site, could be involved in proton efflux. EPR data show that Asp407 is not required for Mn binding, although it is close to the two proposed Mg/Mn ligands, His411 and Asp412. The similar spectral characteristics and content of bound Mn in wild type and the D407A mutant grown in high Mn indicate that this carboxyl does not play a critical role at the Mn/Mg site. The observed broadening of the Mn signal could be related to removal of its hydrogen bond to the His411 ligand. The spectra also show that the Cup, site is not changed by sul sul bor of l C() desr 500.- mm subs (Tsu isstr 011d; SUbu hidrt there 18 C0 under 142 substitution, suggesting that this residue is not critical for maintaining the interface of subunits I and II where the CuA and Mg are located. The heme environments are not changed by replacing aspartate with alanine or asparagine, which confirms that neither the carboxylate function nor the hydrogen- bonding capacity of this residue is structurally critical. The altered heme a3 environment of D407C seen by resonance Raman is in agreement with shifted visible spectra, lower CO binding, and lower activity and suggests that the presence of cysteine causes some destabilization of the heme a y—CuB center, such that some proportion of the enzyme, 5— 50% depending on the assay conditions, can become denatured. In conclusion, all the data in this study show that aspartate-407 is not a critical residue in cytochrome c oxidase for proton pumping, for Mn binding, or for maintaining protein structure at the subunit I/II interface. Other residues may be capable of substituting for its function in a proton (Iwata et al., 1995) or water exit channel (Tsukihara etal., 1996), but the lack of effect on Mn spectral characteristics and binding is strong evidence that it is not a Mn/Mg ligand. Recently, higher resolution (2.3 A) crystal structures from bovine cytochrome c oxidase was obtained and it shows a slightly different structure at the interface of subunit I and subunit H (Yoshikawa et al., 1998) (Figure 9). The residue Asp407 is hydrogen bonded to His411 (Mn/Mg ligand) and a prOpionate group of heme a3, therefore, replacing aspartate with alanine might destabilize the Mn/Mg center. This is consistent with the observation of the broadened EPR spectra of D407A grown under high Mn low Mg (Figure 7C). 143 Figure 9: Structure of the interface between subunit I and subunit II of beef heart cytochrome c oxidase at 2.3 A resolution [in Rhodobacter sphaeroides oxidase numbering; coordinates are from Yoshikawa et al. (1998)]. The dash lines indicate the hydrogen bonds. 144 411 ,xuor Figure 9 Fett Hos Rel: Iwata Cltoc Kiel). Tfans Mite} More: 422 O Mitch CHOQ Nagle’ Plitmb REFERENCES Fetter, J. R., Qian, J., Shapleigh, J., Thomas, J. W., Garcia-Horsman, J. A., Schmidt, E., Hosler, J., Babcock, G. T., Gennis, R. B. & Ferguson-Miller, S. (1995) Possible Proton Relay Pathways in Cytochrome c Oxidase. Proc. Natl. Acad. Sci. USA. 92: 1604-1608. Garcia-Horsman, J. A., Puustinen, A., Gennis, R. B. & Wikstrom, M. (1995) Proton Transfer in Cytochrome b03 Ubiquinol Oxidase of Escherichia coli: Second-Site Mutations in Subunit I That Restore Proton Pumping in the Mutant Asp135—>Asn. Biochemistry 34: 4428-4433. Henderson, R., Baldwin, J. M., Ceska, T. A., Zemlin, F ., Beckmann, E. & Downing, K. H. (1990) Model for the Structure of Bacteriorhodopsin Based on High-resolution Electron Cryo-microsc0py. J. Mol. Biol. 213: 899-929. Ho, S. N., Hunt, H. D., Horton, R. M., Pullen, J. K. & Pease, L. R. (1989) Site-directed mutagenesis by overlap extension using the polymerase chain reaction. Gene 77: 51-59. Hosler, J. P., Espe, M. P., Zhen, Y., Babcock, G. T. & Ferguson-Miller, S. (1995) Analysis of Site-Directed Mutants Locates a Non-Redox-Active Metal near the Active Site of Cytochrome c Oxidase of Rhodobacter sphaeroides. Biochemistry 34: 75 86-7592. Hosler, J. P., Fetter, J ., Tecklenburg, M. M. J., Espe, M., Lerma, C. & Ferguson-Miller, S. (1992) Cytochrome aa3 of Rhodobacter sphaeroides as a Model for Mitochondrial Cytochrome c Oxidase. J. Biol. Chem. 267 : 24264-24272. Iwata, S., Ostermeier, C., Ludwig, B. & Michel, H. (1995) Structure at 2.8 A resolution of cytochrome c oxidase from Paracoccus denitrificans. Nature 376: 660-669. Krebs, M. P. & Khorana, H. G. (1993) Mechanism of Light-Dependent Proton Translocation by Bacteriorhodopsin. J. Bacterial. 175: 1555-1560. Mitchell, D. M., Adelroth, P., Hosler, J. P., Fetter, J. R., Brzezinski, P., Pressler, M. A., Aasa, R., Malmstrom, B. G., Alben, J. O., Babcock, G. T., Gennis, R. B. & Ferguson- Miller, S. (1996) A Ligand-Exchange Mechanism of Proton Pumping Involving Tyrosine- 422 of Subunit I of Cytochrome Oxidase Is Ruled Out. Biochemistry 35: 824-828. Mitchell, D. M. & Gennis, R. B. (1995) Rapid Purification of Wildtype and Mutant Cytochrome c Oxidase from Rhodobacter sphaeroides by NiZ+-NTA Affinity Chromatography. FEBS Lett. 368: 148-150. Nagle, J. F. & Morowitz, H. J. (1978) Molecular Mechanisms for Proton Transport in Membranes. Proc. Natl. Acad Sci. USA. 75: 298-302. 145 146 Rothschild, K. J. (1992) F TIR Difference Spectroscopy of Bacteriorhodopsin: Toward a Molecular Model. J. Bioenerg. Biomembr. 24: 147-167. Thomas, J. W., Puustinen, A., Alben, J. O., Gennis, R. B. & Wikstrbm, M. (1993) Substitution of Asparagine for Aspartate-135 in Subunit 1 of the Cytochrome bo Ubiquinol Oxidase of Escherichia coli Eliminates Proton-Pumping Activity. Biochemistry 32: 10923-10928. Tsukihara, T., Aoyama, H., Yamashita, E., Tomizaki, T., Yamaguchi, H., Shinzawa-itoh, K., Nakashima, R., Yaono, R. & Yoshikawa, S. (1995) Structure of Metal Sites of Oxidized Bovine Heart cytochrome c Oxidase at 2.8 A. Science 269: 1069-1074. Tsukihara, T., Aoyama, H., Yamashita, E., Tomizaki, T., Yamaguchi, H., Shinzawa-Itoh, K., Nakashima, R., Yaono, R. & Yoshikawa, S. (1996) The Whole Structure of the 13- Subunit Oxidized Cytochrome c Oxidase at 2.8 A. Science 272: 1136—1144. Wikstrbm, M., Bogachev, A., Finel, M., Morgan, J. E., Puustinen, A., Raitio, M., Verkhovskaya, M. L. & Verkhovsky, M. I. (1994) Mechanism of Proton Translocation by the Respiratory Oxidases. The Histidine Cycle. Biochim. Biophys. Acta 1187: 106- 111. Yoshikawa, S., Shinzawa-Itoh, K., Nakashima, R., Yaono, R., Yamashita, E., Inoue, N., Yao, M., Fei, M. J., Libeu, C. P., Mizushima, T., Yamaguchi, H., Tomizaki, T. & Tsukihara, T. (1998) Redox-coupled crystal structural changes in bovine heart cytochrome c oxidase. Science 280: 1723-9. CHAPTER V ARGININE 48218 IMPORTANT FOR STABILIZING THE INTERFACE OF SUBUNIT I AND SUBUNIT II OF RHODOBACTER SPHAEROIDES CYTOCHROME c OXIDASE 147 imr hyd con mut elec met that alte: over Mn" prox Pum like] 1mm uh bu1 r. 50 3' ABSTRACT A highly conserved arginine pair, Arg481 and Arg482, were proposed to be important in direct electron transfer from Cu, to heme a, due to their involvement in a hydrogen bonding network at the interface of subunit I and subunit II. Arg482 was converted to alanine, lysine, glutamine and proline by site-directed mutagenesis. The mutants were purified and characterized in terms of oxygen consumption rates, internal electron transfer rates, proton pumping efficiency, pH dependence, heme content and metal center structures by EPR, UV/visible and optical spectroscopy. The results show that R482K is similar to wild type oxidase in all respects except that its Mn/Mg site is altered, compared to wild type. R482Q, R482A and R482P show significant effects on overall electron transfer activity, internal electron transfer rates, heme a 3 content, and Mn/Mg binding. Strikingly, these mutants lose proton pumping ability, which might provide evidence for the importance of at least one of the heme propionates in the proton pumping pathway. The hydrogen bonding network that this residue is involved in is likely critical in stabilizing the interface of subunit I and subunit II, but even more important for maintaining the stability of heme a 3, as well as the environment of the Mn/Mg site. Its importance in direct electron transfer from Cu, to heme a is suggested, but not established, by a dramatic reduction in that rate, from 1.2 x 105 s" in wild type to 50 s'1 in the proline mutant which breaks the predicted hydrogen-bonding pathway. 148 INTRODUCTION The two cytochrome c oxidase crystal structures reveal a complex hydrogen bonding network involving a number of water molecules, at the interface of subunit 1 and subunit II (Iwata et al., 1995; Tsukihara et al., 1996). In this region, several hydrogen bonds and the peptide backbone between a highly conserved arginine pair Arg481 and Arg482, are predicted to participate in a direct through-bond pathway connecting a histidine ligand of Cu), in subunit II to heme a in subunit 1 through its propionic acid side chain. This through-bond pathway is postulated to facilitate electron transfer between the two metal centers (Ramirez et al., 1995). Arginine pairs have been observed in other proteins to play an important role in structuring water (Magalhaes et al., 1994); thus, in cytochrome oxidase it is possible that they play a role in both electron and proton transfer events. Arg482 is located on a loop between helix XI and helix XII in subunit I of Rhodobacter sphaeroides cytochrome c oxidase (Figure 1). On the basis of the structural information of all the available crystal structures of oxidases, it is apparent that Arg482 not only is involved in a hydrogen bonding network between Cu, and heme a, but also is related to the Mn/Mg site through hydrogen bonding interactions with water molecules (Ostermeier et al., 1997) (Figure 1). Furthermore, its neighboring residue Arg48l is hydrogen bonded to the 6-propionate group of heme a 3 (Figure 1), close to the site for oxygen reaction. It is generally accepted that electron transfer from the electron donor (cytochrome c) to the final acceptor (molecular oxygen) goes through three redox active centers in the sequence: Cu), —9 heme a —> heme a 3 (Hill, 1993). However, the exact electron transfer route(s) are not clear, nor is the extent to which the precise protein structure is important in the rate and direction of electron flow. The present study is aimed at investigating the role Arg482 plays in oxidase function, especially electron transfer, in order to further elucidate the electron transfer mechanism. Four mutants (R482K, R482A, R482Q, and R482P) have been generated, 149 150 Figure 1. Structure of the interface between subunit I and subunit II of cytochrome c oxidase (with Rhodobacter sphaeroides oxidase numbering). The coordinates are from the Paracoccus denitrificans cytochrome c oxidase crystal structure. The residues in subunit I are colored in yellow, and the residues in subunit II are colored in red. c oxidase. 151 H217 Figure 1 HEME a3 H333 152 purified and characterized. The results demonstrate that non-conservative replacement of Arg482 causes dramatic changes structurally (in the heme, Mn/Mg and cytochrome c binding sites) and functionally (in overall electron transfer activity, in intrinsic electron transfer rates between Cu A and heme a, and in proton pumping efficiency). Maintenance of the positive charge by replacement of Arg482 with lysine results in minimal alteration of activity, ruling out the need for an arginine pair, or all aspects of the intricate hydrogen bonding pattern to satisfy the functional needs of the protein. MATERIALS AND METHODS Site-Directed Mutagenesis Site-directed mutants were constructed using PCR overlapping extension methods (Qian et al., 1997). All the oligonucleotide primers were synthesized by Michigan State University Macromolecular Structural Facility, East Lansing, MI. The primers used to create the mutants are shown in Figure 2. The 692 bp final PCR product was digested with SaII/Hindlll, and then the 519 bp fragment was subcloned into the plasmid pND38, a plasmid that contains part of the subunit I gene of cytochrome c oxidase. The pND38 plasmid with the mutation was digested with BglII/HindIII, and the 652 bp fragment was subcloned to pJS3-X6H2, a plasmid containing the entire subunit 1 gene with a 6-histidine tag at the C-terminus of the COXI gene (Mitchell & Gennis, 1995). The subsequent subcloning and conjugation were conducted as previously described (Qian et al., 1997). All the mutants were subject to DNA sequencing by Michigan State University Sequencing Facility, East Lansing, MI, and no secondary mutations were found. Enzyme Purification Mutant oxidases were grown in Sistrom’s media as described (Hosler et al., 1992), and purified by Ni-NTA affinity column (Qian et al., 1997) with some modifications. The cell pellets were resuspended in 50 mM KHZPO4, pH 6.5, and 1 mM EDTA, followed by two passages through a French pressure cell at 20,000 psi. The cell pellets were solubilized in 1% lauryl maltoside at 6 mg protein/mL. All the enzymes after Ni-NTA column were firrther washed with 0.1% lauryl maltoside, 10 mM Tris, and 40 mM KCl, pH 8.0 three times by centricon-filtration to remove Ni-histidine. Reconstitution of Cytochrome c Oxidase Cytochrome c oxidase vesicles were prepared as described (Hosler et al., 1992; Fetter et al., 1995) with 20 mg/mL asolectin and 2 uM oxidase enzymes. The vesicles were dialyzed against 100 volumes of 75 mM HEPES-KOH, pH 7.4, 14 mM KCl, 0.1% cholate for 6 h; 100 volumes of 75 mM HEPES-KOH, pH 7.4, 14 mM KCl for 12 h; 100 153 154 Figure 2: Wild type amino acid sequence (upper) and oligonucleotide sequences (lower) of wild type and mutants at position 482 in subunit 1 of Rhodobacter sphaeroides cytochrome c oxidase. c oxidase. WT R482K R482A R482Q R482P 155 R g Y I D Y 5’- CG CGG fl TAC ATC GAC TAT —3’ 5’- CG CGG AAA TAC ATC GAC TAT -3’ 5’- CG CGG (fl TAC ATC GAC TAT -3’ 5’- CG CGG 93.4. TAC ATC GAC TAT -3’ 5’- CG CGG CCC TAC ATC GAC TAT -3’ Figure 2 156 volumes of 50 mM HEPES-KOH, pH 7.4, 25 mM KCl, 15 mM sucrose for 12 h; and 500 volumes of 50 11M HEPES-KOH, pH 7 .4, 45 mM KC], 44 mM sucrose for 12 h. Measurement of oxygen consumption was determined for the reconstituted enzymes in order to determine the respiratory control ratio (RCR), which is a test of whether the vesicles are intact and the enzymes are inserted correctly (Fetter et al., 1995; Qian et al., 1997) Stopped-Flow Proton Pumping Assay All experiments were performed at room temperature. Before the proton pmnping assay, control experiments (i.e. without phenol red addition in the proton pumping buffer) were conducted to ensure no problems and establish time of data collection. Cytochrome c oxidase vesicles (COV) were diluted to 160 nM in aa3 in 50 [AM HEPES, 45 mM KC], and 44 mM sucrose, pH 7.4 (proton pumping buffer). Pre-reduced cytochrome c was desalted through a Sephadex G-75 column. 2.5 uM of cytochrome c (final concentration) was used for each assay (about 8 turnovers). Measurements were taken at 1000 scans/sec on an OLIS rapid-scanning stopped-flow spectrophotometer. Cytochrome c oxidation was followed at 550 nm and the reduced minus oxidized spectra were fit by Global analysis to first-order rate. The reactions of cytochrome c with cytochrome c oxidase were carried out under three conditions: 1) without ionophore (coupled); 2) with 2 11M valinomycin; and 3) with 2 uM valinomycin and 10 11M CCCP (uncoupled). The cytochrome c oxidation rate was derived by Global fitting analysis and an RCR was obtained. This experiment also decided the isosbestic point of cytochrome a reduced and oxidized spectra in each reaction, which was used for the proton pumping assay (Zhen et al., 1997). Experiments with the same conditions as above with addition of 200 uM phenol red were carried out for the detection of proton pumping. Both 550 nm and 557 nm (the isosbestic point of cytochrome c) were monitored for 1) without ionophore (coupled); 2) with 2 uM valinomycin; and 3) with 2 11M valinomycin and 10 uM CCCP (uncoupled). rm... .- 157 The acidification and alkalinization rates were obtained by using non-linear curve fitting in Microcal Origin program. The backleak in the kinetic trace was fitted and subtracted from the individual scan to eliminate the background pH change contribution. The difference scan was fitted and the rate was determined. Correction for an initial acidification artifact was also made by subtracting the coupled vesicle absorbance changes fiom the val and uncoupled changes. Protein Assay BCA (bicinchoninic acid) protein assay system from Pierce was used on wild type and mutant cytochrome oxidases (Smith et al., 1985). BSA (bovine serum albumin) ranging from 0 to 20 ug was used to create a standard curve. BSA and all the oxidase samples were diluted with 10 mM Tris, 40 mM KCl, 0.25% deoxycholate, pH 11.2, and then mixed with 2 mL of working buffer (a mix of reagent A (1% BCA'Naz, 2% NazCO3-H20, 0.16% Nay-tartrate, 0.4% NaOH, 0.95% NaHCO3, pH 11.25) and reagent B (4% CuSO4-5 H20 in ddeO) in 50:1 ratio), followed by incubation at 60°C for 30 min. The absorbance at 562 nm for the samples was measured on Perkin-Elmer MOP- UV/visible spectrometer. Pyridine Hemochromogen Assay 800 uL of mutant and wild type oxidases (final concentration is 5 IIM) was mixed with 0.2 N of fresh NaOH, followed by immediate addition of 10% pyridine. Dithionite- reduced minus ferricyanide-oxidized difference spectra were taken for all the samples, and heme A (total of heme a and heme a 3) concentration was calculated by using the extinction coefficient of 8533420 of 25 mM’lcm'l (Berry & Trumpower, 1987). It was noted that immediate measurements of spectra is critical (within 20 min). Ruthenium Kinetics Assay Ru55Cc was used to measure the electron transfer rate in Rhodobacter sphaeroides cytochrome c oxidase (CcO). Ru55Cc stands for the horse cytochrome c (Cc) with the inorganic complex Ru(bipyridine)3 covalently attached at the lysine-55 158 position(Pan et al., 1993). The purified oxidase enzymes were exchanged into 5 mM Tris, pH 8 buffer containing 0.05% lauryl maltoside in a concentrator to remove excess amounts of salt. For kinetics measurements, about 300 ILL cytochrome c oxidase sample solution containing 10 mM aniline and 1 mM 3-CP (3carboxyl-2,2,5,5-tetramethy1-1- pyrolidinyloxy free radical) was transferred into a semi-micro cuvette with an optical path of 1 cm. Aniline and 3-CP were electron donors to re-reduce Ru(III) and prevent the back reaction (Pan et al., 1993). The CcO:Cc ratio was maintained to be between 1 and 1.5 so that all cytochrome c was bound. After a laser flash reduced the cytochrome c at 6 x 105 s", the pseudo first-order rate constant for electron transfer from cytochrome c to the Cu A‘center of oxidase was measured by monitoring the decrease of absorbance at 550 nm for cytochrome c as well as 830 nm for Cu. The reduction of heme a was followed at 605 nm. The ionic strength of the sample solution was adjusted with a 5.0 M NaCl stock solution. Optical Spectroscopy F erricyanide oxidized minus dithionite reduced spectra of the purified enzyme were recorded with a Perkin-Elmer Lambda 6 UV-visible spectrometer at 25 °C. At least 30 ILM enzymes were used in each assay for measuring the 830 nm band. Other Assays Determination of visible spectra, CO binding, EPR, resonance Raman spectroscopy and electron transfer activity were measured as described (Qian et al., 1997). RESULTS Visible Spectra, Electron Transfer Activity, and CO Binding Assays The visible spectra of R482K and R482A appear to be similar to that of wild type; the spectrum of R482Q shows a slight blue shift of the or and Soret peaks (< 2 nm), while the spectrum of R482P reveals shifted and broadened or and Soret peaks (Figure 3). These results suggest that the environment at heme a and heme a 3 is normal for R482A and R482K, slightly altered for R482Q, but dramatically changed for R482P. Purified R482K, R482Q, R482A and R482P oxidases have overall electron transfer activity of 1400, 600, 480, and 60 s", respectively (Table 1), demonstrating a significant decrease of the activity in all the mutants except R482K, compared with wild type cytochrome oxidase (1500-2000 s"). Since turnover is calculated on basis of heme a content (605 nm), and from heme content studies, it shows that some portion of heme a3 is lost in mutant enzymes (see Table 2, column IX), the turnovers corrected for purity are 1591, 698, 527, 143 s", for R482K, R482Q, R482A and R482P, respectively. The pH dependence of R482P was compared with that of wild type (Figure 4). The pK,l of this mutant (8.0 i 0.2) is slightly different from that of wild type (8.2 i 0.4), but still within the error of the assay. There is a significant difference in the extent of inhibition of activity at high pH (Figure 48). At pH 9, wild type is 80% inhibited but R482P only 40% is inhibited. CO binding assays demonstrate that R482K, R482Q, R482A and R482P have 100%, 94%, 78-84% and 55-85% of the CO binding efficiency, respectively, compared with 100% for wild type (Table 1). This indicates some alteration of the heme a3-Cu3 binuclear center for R482A and R482P, perhaps related to changes in the hydrogen bond network involving the propionate of heme a3. The broad range of CO binding efficiency of R482P reflects the different degrees to which the heme a 3 of the mutant was reduced. The incomplete reduction of heme 03 in R482P is also seen in the second derivative of the reduced CO-bound cytochrome c oxidase spectrum (data not shown). 159 160 9:”va 0:0 Omwvm 00.0 80800 e 080: 00 00.30800 80800 <00 0030— 0 08000800 30080008 : 8000880 .0800 0 00: wow w88=mma 008300000 8: wow 00 0000—00 mm 8: on» 00 008000030 2F 0 088 #000 .00 802000 :09. 800 . .850 0.80 805 80>808~0> .00 802000 8090 000000600 05 .00 0082980 05 3.58800 .3 0000—00—00 mm -0 PI .0 o._ 3-3 o Na 3 o0 mmwvm ON 33:. c We cg owe Si m was 11211151116110 _R482P I g ‘ 51.5.. - 035134 p 20 < 11111041) - 225150.544 § 1500 - Finnl(A2) -12.1301.47 . Xuvsooto) - 302050.155 [— Widdex) - 0435170145 1 8 '0 II 2; I 1000 - SigmoidaKBoltzmau) fit a wt_wt 16 c1113." - 26886.435! I " lnit(Al) - 2290.9177 Final(A2) - 215.20717 xmsqu) - 520540.405 ‘4 500 _ Width(dx) - 0513910299 l ' l ' l 1 ' I V I ' l 12 6.0 6.5 7.0 7.5 8.0 8.5 9.0 pH SigmoidaKBoltzmu) fit to A.activity_R482P Chisqr - 4.80745 1111qu ) - 100,552.19 Final(A2) - 54.647534 XltY50(x0) - 793230.135 wanna) - 0449830114 I R482P 50 ' .. SigmoidaKBoltzman) fit to A.activity_wt Chisqr - 50.98854 40 "‘ .. lnit(Al) - 99.405772 FiuKAZ) - 9.9439300 30 d XatY50(x0) - “8740.395 Width(dx) - 0.5] l9l0.296 20 "' l ' I ' l l l ' l ' l 6.0 6.5 7.0 7.5 8.0 8 5 9.0 pH Figure 4 nu...-—--. av. R482P Activity 165 Respiratory Control and Proton Pumping Assay All the mutants except for R482P, as well as wild type oxidases, were reconstituted into asolectin lipid vesicles properly, shown by a normal respiratory control ratio (RCR) significantly greater than 4 (Table 1). The RCR for R482P is significantly lower than that of wild type, which is probably due to the low activity of this mutant. However, all the mutants show a lower activity in the reconstituted form than in the solubilized form (Table 1), which is typical since the assay conditions are not the optimal for oxidase and some mutants may be unstable during the long dialysis process. The proton pumping assay was conducted on an OLIS rapid-scanning stopped- flow spectrophotometer (Antonini et al., 1993), which is able to detect the oxidation rate of cytochrome c and the proton pumping rate of the oxidase at the same time. Proton pumping was detected by the change in absorbance at 557 nm, close to the maximal absorbance of phenol red and at an isosbestic point for cytochrome c. Alkalinization upon uncoupling with CCCP reflects the uptake of protons from the interior of the vesicles for 02 reduction, which can be seen on the exterior, because of rapid equilibration of protons into the interior of the vesicle via the uncoupler. The H‘Ve' ratio is derived by comparing the amplitude of acidification after adding valinomycin with that of alkalinization after adding CCCP. R482K has an H+/e' ratio of 1.04, similar to that of wild type (0.77). However, R482Q, R482A and R482P do not show proton pumping activity (Table 1, Figure 5). The very fast acidification seen in R482Q, R482A and R482P mutants is an artifact, since its time scale is too fast compared to cytochrome c oxidation, while in the case of wild type and R482K, the time scale for these two events is close (Figure 5). The results suggest that a positive charge at the Arg482 position and maintaining the correct position of Arg481 are critical for normal proton pumping activity in oxidase. A possible mechanism is discussed (see Discussion). 166 Figure 5: Stopped-flow proton pumping assay on wild type (top), R482K (middle) and R482Q (bottom) cytochrome c oxidase. Cytochrome c oxidation and proton pumping upon addition of 2 11M valinomycin were monitored at 550 nm and 557 nm, respectively. Cytochrome c oxidase vesicles were diluted to 80 nM in 50 1.1M NaHCO3, 45 mM KC], and 44 mM sucrose, pH 7.4, with 200 11M phenol red. Measurements were taken at 1000 scan/sec on the OLIS rapid-scanning stopped-flow spectrophotometer at 25 °C. Each trace was fitted with Microcal Origin program. R482A and R482P show proton pumping and cytochrome c oxidation traces similar to those of R482Q (data not shown). 167 Absorbance at 557 nm Absorbance at 557 nm Absorbance at 557 nm r 4 0.000 *- “ 0.000 r . 1 -0.002 . L‘, WTprotonpmnping " 47.005 1.l l J ‘l‘ l1 ' 1 11 ~ .0904 l- W ‘ . t ‘ , 4 .0010 ‘1‘. "I' H M”, 1‘ 1' ‘ ”1 .1 .l 111 ‘ 1‘V1‘l ‘ . 1 ., l , "1111111 1 '11 11' "'l" .0005 - '1‘, , .1 .0015 b 1" "H" {I r 4 1 I 4 '7 A l ‘1 WTCcoxidation " a! “ 1'" "11"" " ’ ‘ -0.008 ‘- 0.020 0.020 ‘ ‘ ‘ ‘ J g ‘ ' 0.020 0.0 0.2 0.4 0.6 0.8 llo t 1 0.015 " '1 0.015 b : 1 ‘ R482K Cc oxidation 0.010 - - 0.010 b . \fl'y\ 1 0005 - ' “‘5 «1 0005 1.7.." m, 1. , r l 1 I 1'- ‘ . . 1 '1' ‘JUIIl n ' {Til}? 1;; 4.44-1. . . , “. 0.“ .- l !.‘ If." '[Vhll’ll l'. "HHH‘ gil '.‘ T111141“ 01x” 1 . R482K protonpmnpmg .0005 v 0.005 1 . 1 . L 4 1 4 7 0.0 0.2 0.4 0.6 0.8 1.0 Time (sec) 0.014 0.014 P d 0.012 f - 0.012 1- 4 0.0l0 1' ' 0.010 r . . 4 0.004 r- ' 1, RmQC‘ mm” .. 0.008 P l". 1 0.006 b " 1... ~ 0.006 0.004 - V “a - 0.004 b I“ , l ‘ 1 cam P y ' .' u ‘ ‘ o-m I 1‘ y ‘ " m. ”V W111 o... ’ i -0.002 b . -0.002 1 11432015111101. umpi 3 ~01!” l- p 113 a «0.004 1- I -0.006 r 1 ‘ l . . 1 ‘ L . 1 . -‘ ~0.006 0.0 0.5 1.0 1.5 2.0 2.5 3.0 Time (sec) Figure 5 run 055 in aoueqaosqv um 095 in aauchosqv am 09918 aounqrosqv 168 Purity and Home Content The purity and homogeneity of the different mutant samples have been carefully assessed by examining their spectral properties and protein content. In the at band region (605 nm) of the fully reduced spectrum of cytochrome c oxidase, 80% of the absorbance is due to heme a, and 20% to heme a3. In the Soret region (444 nm), heme a and heme a 3 each contribute 50% of the absorbance (V anneste, 1966). Consequently, the A444/A605 ratio is correlated to the heme a3 content in the enzyme; the higher the ratio of A444/A605, the higher the heme 03 content. The pyridine hemochromogen assay is used to determine the total heme ([heme A] = [heme a] + [heme a 3]) content in the enzyme, and the [M3] value is determined using the extinction coefficient 8605 = 38 mM'1 cm'1 for the dithionite reduced spectrum, which actually is a measurement of heme a content only, and is doubled to represent the two hemes. Therefore, the ratio of [heme A]/[aa3] can also give information about the relative heme a3/heme a content. Theoretically, [heme A]/[aa3] is 2 for wild type oxidase due to the fact that there are two hemes (heme a and heme 03) per oxidase. The [heme A]/[aa3] ratio is 1.60, 1.58, 1.64 and 1.21 for R482K, R482Q, R482A and R482P, respectively, compared to 1.70 for wild type enzyme (Table 2), indicating that there is more heme a (605 run) than the heme A assay shows; therefore, there must be less heme 03. If all the loss is heme a3, then this would give an estimated heme 03 content of 88%, 86%, 92% and 42% for R482K, R482Q, R482A, and R482P, respectively, compared to 100% for wild type. Furthermore, all the mutants show a lower A444/A605 ratio, suggesting a selective loss of heme (13, as seen in other mutants (Hosler et al., 1994; Hosler et al., 1996), which is consistent with the results from the [heme A]/[aa3] data. The fact that some mutants might have lost both heme a and heme a3 (i.e., apo-enzyme) is indicated by the low heme A/mg protein and high Ana/Am ratio (see below), especially for R482P, in which case these mutants should have lost more heme 03 than heme 0. But the apo-enzyme will not affect any of the spectral assessments of the enzyme properties. .0850 A205 0380 .«0 3% 0880800 05 .8 088000 :0 wfifim 00 ma 5030000000 .mmem 000 5:28 580 02,8802 0me 05 <38 .0008 .0830 80 .< 2:8 .8 $8 000 $0 60> 300 08— 08088 05 0x02: 00 0080800 003 8033 >4 mm 038 38:70. 0805 80800 05 0:0 .3 080: m0 000— 00 0:0 mm < 080: .«0 08— 05 :0 w88=mm< .9; 882000 008 300:? 0805 05 80¢ 003000 80800 m0 080: 0088000 06 mm 5 08200 .0 169 oé NV 00:. gm; Q: 0A.: 0:0. mmd mAN o.m mevM ~.w Na {00¢ #04 N.NN {as md @wfi mém ”N .2 a: .3 A .8380 e 808 2a .033 .183... 080E .33 332%... 805 AESEE A< 080E £3.83 ”000005 0 08080080 0.030003%. L0~00¢030§ 8082 000 09¢. 055 .00 000800800 ”N 033. 170 UV/visible spectral analysis shows that all the four mutants have a higher A230/A424 ratio, in the oxidized spectrum of the enzyme, with a two-fold increase in the case of R482P (Table 2). This indicates that the ratio of protein (absorbs at 280 nm) to hemoprotein (absorbs at 424 nm) is higher in mutant enzymes. The amount of protein in each preparation of the mutants may vary depending on the contaminating proteins and the presence of apo-protein due to instability of mutants forms, leading to loss of heme A. The apo-protein, especially subunit 1, will copurify because it contains the histidine tag. The protein concentration of wild type and mutant cytochrome c oxidase samples was measured with the BCA assay system. Table 2 shows that R482K, R482Q, R482A, and R482P have [heme A]/[protein] ratio of 11.8, 10.5, 9.5, and 2.6 nmol/mg, respectively, compared to 15.7 nmol/mg for wild type, which gives the purity of 75%, 67%, 61% and 17%, respectivly. The lower [heme A]/[protein] value for mutants compared to wild type indicates that these mutants have lower heme contents, due to the presence of contaminating proteins or denaturation caused by loss of heme a 3. However, SDS-PAGE on the R482K and R482Q mutants indicates that the proteins appear intact compared to wild type, i.e., all three subunits are present, and no additional contaminant proteins were observed, arguing that the major impurity is apo-enzyme. Mn/Mg Site and Cu A Environment EPR spectroscopy was performed on all the mutants, grown under various conditions, and measured at 110 K and 10 K. EPR spectra for wild type and all the mutants grown under high Mn (700 MA) and low Mg (50 uM) are shown in Figure 6. R482P, R482A and R482Q show severe distortion in the Mn signals, suggesting that they lost substantial Mn binding ability. The spectrum of R482K shows a slight alteraion in the Mn signal, indicating structural change at the Mg site (Figure 6). Integration of the first peak of the spectra of R482K and of wild type measured at 110 K demonstrates that this mutant contains about 65% Mn (data not shown). EPR spectra of all the mutants and wild type grown in high Mg (1200 M and low Mn (0.5 uM) are shown in Figure 7; 171 Figure 6: EPR spectra of purified wild type and mutant Rhodobacter sphaeroides cytochrome c oxidase grown in high [Mn] (700 uM) and low [Mg] (50 uM). The spectra were recorded on a Bruker ESP3OOE series spectrometer at 10 K. In order to show all the signals clearly, the scales for R482A, R482Q and R482? are different from that for wild type and R482K. Amplitude 172 WT R482K EL R482A ww— ___,_/\ R482 ,_/L Q R482P I I I l l J I 2400 2800 3200 3600 Magnetic Field (Gauss) Figure 6 173 Figure 7: EPR spectra of purified wild type and mutant Rhodobacter sphaeroides cytochrome c oxidase grown in high [Mg] (1200 uM) and low [Mn] (0.5 uM). The spectra were recorded on a Bruker ESP300E series spectrometer at 10 K. 174 '3 NJ 302A ii (g=2.84 a Wu J g=2.85 ' ‘/R482P J , fl , 1 , . I . . . , , I , 2200 2600 3000 3400 3800 Magnetic Field (Gauss) Figure 7 175 these spectra show the spectral characteristics of CuA, no longer masked by Mn signals, which are similar to wild type but with some evidence of change in the heme a signals and the g = 2.19 band of Cu, whose significance is not understood (Hosler et al., 1992). Optical spectroscopic analysis of cytochrome c oxidase is another way to monitor the Cu environment. The ferricyanide oxidized minus the dithionite reduced spectrum of wild type enzyme shows a peak at 830 nm, and a 6330 of 2 mM'1 cm’1 is calculated from the absorbance of heme a at 605 nm with em = 38 mM'l cm". The optical spectra of all the mutants appear to be similar to that of wild type (Figure 8), although the extinction coefficient of R482Q and R482P are only 1.5 mM'1 cm’1 and 1.0 mM'1 cm", respectively, compared to 2.0 mM'I cm'1 for wild type, R482K and R482A (Table 1), assuming a normal 605 extinction, suggesting that there might be loss of CuA in R482Q and R482P, due to loss of subunit II. Analysis of Intrinsic Electron Transfer Rates Ru55(horse)Cc was used to measure the electron transfer rate between cytochrome c (cyt.c) and Rhodobacter sphaeroides cytochrome c oxidase. The electron transfer (ET) rate constant from cyt.c to the CuA site of cytochrome oxidase is measured by monitoring the decrease of absorbance at 550 nm as well as the decrease of absorbance at 830 nm. The rate constant for ET from CuA to heme a is measured by following the increase of absorbance at 605 nm. Scheme 1 shows the mechanism for the assay (Geren, 1995). At low ionic strength (0-55 mM), cytochrome c and cytochrome c oxidase form a 1:1 complex. Electrons are then transferred in an intramolecular fashion, and therefore, only one (fast) phase is observed (1 in Scheme 1). At intermediate ionic strength (55-105 mM), some cyt.c is uncomplexed and both fast and slow phases exist. At high ionic strength (above 100 mM), all the cyt.c is uncomplexed, and the reaction follows II in Scheme 1, so only the slow phase is observed, which represents the second order reaction of cytochrome c with oxidase from solution. For all the mutants, k, (the rate constant for ET between cyt.c and Cu), lq, (the rate constant for ET between Cu A 176 Figure 8: Comparison of optical spectra of wild type and mutant cytochrome c oxidase. 30-50 uM oxidases in 50 mM Tris, 40 mM KCI, 0.1% lauryl maltoside were used in each assay and 700-1000 nm spectra were shown. Ferricyanide oxidized minus dithionite reduced spectra were measured on a Perkin-Elmer Lambda 6 UV- visible spectrometer at 25°C. 177 Absorbance ' l ' l ' l ' I 1 I 700 750 800 850 900 950 1000 Wavelength (nm) Figure 8 178 H 052% 3 £539. A+ma+~<=uv+ +~o ‘4' fma+m<=uv+ +mo 80¢ cowomom d x 2: .2 3 5388 much A+~s+~<=uvn+mo T A+m~+_<:uvu+mo T A+.c..w+~,.‘=Uvu+~o T A+ma...~<=va..mo 55m? cowomom A a. flu— :— 179 and heme a) and kslow (the second-order rate constant measured from the concentration dependence of the slow phase) were measured. R482K has a k. similar to wild type (6 x 10‘ s"), while R482Q, R482A and R482P have rates of 8000 s", 4500 s" and 7 s", respectively, significantly lower than that of wild type, suggesting a slower ET rate from cyt.c to Cu for these mutants (Table 3). The fact that in R482P only a very slow phase (7 s") was observed for k, at low ionic strength (5 mM), but that a much higher k, (800 s’ l) was seen after an increase in ionic strength to 70 mM, suggesting that the cyt.c is able to reorient for more efficient electron transfer to Cu, or that a lot of cytochrome c is bound to apo-protein in which no rates can be measured. The kb rates for R482K and R482A are much faster than k” so the actual Cu to heme a rates are not measurable with this cytochrome c derivative. Importantly, the kb rates of R482Q and R482P were measured as 3500 s'1 and 50 s", respectively, compared to 1.2 x 10s 8'1 for wild type, indicating a much slower ET rate from Cu A to heme 0 than from cyt.c to CuA. The ionic strength at which there is the peak rate (kslow) (Figure 9) indicates the tightness of binding of cyt.c and cytochrome c oxidase, while the maximal km“, reflects the electron transfer capability between cyt.c and cytochrome c oxidase. All the mutants have lower maximal km“, compared to wild type (Table 3, Figure 9). The peak position of the second order phase for R482K and R482Q is similar to that of wild type, while R482A has a shifted peak and R482P has a broadened and shifted peak, suggesting that R482K and R482Q bind to cyt.c in a correct orientation, but R482A does not bind to cyt.c as tightly as wild type, and R482P might have more than one conformation in the sample due to denaturation (Table 3, Figure 9). 180 down—cm 89¢ o 0802038 no cogs»: mo 88 3:888 K 2:me 82m sov— mo 33> «8&2 . 3-3 cow 8 s mm: mm 25 83 A 83 £3 2 8mm 8mm 8% Guam fl 8? .2 x e A .2 x o MSE 8-? 88 “2 x S .2 x e 25 25 925 so} a oEonl<=O <=UTo§o as é aways gas .933; as :3 J 9-3 J .0330 a 080385 25:: 2a 25. 2:5 we 863m 885 can a 23. 181 Figure 9: Dependence on ionic strength of second-order rate constant of the reaction between Ru55Cc and cytochrome c oxidase. 182 com 0mm / o oSwE 9:5 fimfiam 2:3 com 09. 8.. cm I ooom I ooov (,-s)s“°>1 I coco ... ooom DISCUSSION Structural Characteristics of the Subunit [-1] Interface and Significance for Electron Transfer Since the discovery of a well organized hydrogen bonding network between subunit I and subunit II from the crystal structures of both bovine and Paracoccus denitrificans cytochrome c oxidase, the role that this interface plays in the enzyme has focused attention on the functional aspects of the network, particularly as it is close to the sites of electron input, proton output and the non-redox active Mg/Mn site. According to the crystal structures of bacterial and mammalian oxidase (Iwata et al., 1995; Tsukihara et al., 1996), the backbone carbonyl oxygen of Arg481 is hydrogen bonded to the e-nitrogen of the imidazole group on Hi3260, a CuA ligand . The backbone nitrogen of Arg482 is hydrogen bonded to the 7-propionate groups of heme a, and the positively charged guanido group of Arg482 is hydrogen bonded to the 6-propionate group of heme a (Figure 1). Recently, higher resolution crystal structures were obtained for both Paracoccus denitrificans and bovine oxidase (Ostermeier et al., 1997; Yoshikawa et al., 1998), in which more detailed structures were revealed at the interface. In the Paracoccus denitrificans oxidase structure, several water molecules are shown to be involved in hydrogen bonding interactions, such as between Arg481 and Asp412 (a Mn/Mg ligand), and between a side chain of Arg482 and Glu254 (a bridging ligand for Cu and Mn/Mg) (Figure 1). Computer graphics and semiquantitative studies on 41 proteins with arginine- arginine short range interactions show that water molecules can bridge the two guanidinium groups and stabilize the system (Magalhaes et al., 1994). In the case of cytochrome c oxidase, this arginine pair might be important for both electron transfer and proton pumping due to its central location in the protein. A through-bond pathway including several hydrogen bonds and residues Hi5260, Arg481, Arg482 and a propionate group of heme 0 appears to connect CuA to heme a and 183 184 would form a route for electron transfer between these two metal centers (Tsukihara et al., 1996). Ramirez et al. proposed that the ET rate between CuA and heme a is dependent upon this pathway of 14 covalent bonds and 2 hydrogen bonds in between them, i.e., “through-bond” electron transfer theory (Ramirez et al., 1995). In contrast, Marcus theory predicts that ET rate is determined by the difference in free energy (AG°), reorganization energy (A), and distance (r) between the electron donor and electron acceptor (Marcus & Sutin, 1985). Among these three factors, Dutton and coworkers conclude that the distance between the donor and acceptor is the most important factor in determining the ET rate (Moser et al., 1992; Moser et al., 1995), sometimes described as the “through-space” electron transfer theory emphasizing the relative lack of importance of the precise structure of intervening medium through which the electrons travel. In order to test whether the electron transfer reaction between Cu A and heme a is “through- bond” or “through-space”, Arg482 was converted to R482K, R482Q, R482A and R482P, in the order of most conservative to most dramatic change. If the hydrogen bonding system is critical for ET between CuA and heme a, breaking the hydrogen bonding network by introducing mutations at Arg482 would be expected to cause a big change in the ET rate. Overall Properties of Arg482 Mutants The oxygen consumption activity in R482A, R482Q, and R482P is different from wild type, unlike R482K (Table 1), suggesting that the positively charged side chain at Arg482 is important for electron transfer activity. Visible spectra demonstrate that the R482P mutant has a major shift in the a and Soret regions, suggesting a large structural change at the heme a and heme a 3 centers, probably due to loss of H-bonding interactions to the side chain and backbone and contaminating apo-enzyme with varying degrees of loss of heme a and heme a 3. A shift in position of the XI-XII loop caused by the ring structure of the proline residue in the R482P mutant, would be expected and likely contribute to the severe disruption of the protein. R482P is the most affected mutant of 185 all, because both the backbone and the side chain lost hydrogen bonding interaction with Hi5260 and the 6-propionic acid moiety of heme a, respectively. Instability of Mutants UV/visible spectral, heme assay and protein assay results all suggest that some heme (13 was lost in the mutants. This is not too surprising for two reasons. First, unlike C-type hemes, which are covalently linked to the apo-protein via a thioether bridge between a cysteine residue and the porphyrin ring, A-type hemes (e.g. in Rhodobacter sphaeroides cytochrome c oxidase) are noncovalently bound cofactors maintained in the protein by electrostatic interactions, H-bonds and coordination chemistry between the iron and the protein ligands. Second, in oxidase, heme a is six-coordinated and has two histidine axial ligands, while heme a3 is five-coordinated and has one histidine axial ligand; and both have propionic acid groups protruding into the solvent accessible subunit I-II interface. Therefore, these hemes tend to be unstable when the electrostatic interaction between the arginine pair and the propionate groups are broken and the interface disordered. Effect on Proton Pumping: Mechanistic Significance The most surprising result is that the proton pumping ability of all the non- conservative mutants is lost while maintaining substantial electron transfer rates and good respiratory control. According to the Glu286 flipping model (Hofacker & Schulten, 1997) (see Chapter 1, section 2.3.1.3), the 6-pr0pionate group of heme a3 is the site to which Glu286 transfers a proton by flipping its side chain upward. It is conceivable that when Arg482 is replaced with residues with shorter uncharged side chains (e. g. R482Q, R482A and R482P), the XI-XII loop backbone would shift and break the hydrogen bonding interaction between neighboring Arg481 and the 6-propionic acid of heme 03. This could cause a pK. change in the 6-propionate group of heme a 3, as seen in mutations of bacteriorhodopsin (Lanyi, 1997) (see Chapter 1, section 2.2), which would make it difficult to deprotonate, or more likely, unable to pick up a proton because of too low pKa 186 in the absence of the neighboring positive charge. Therefore, if this acidic group is involved in proton pumping, mutation of Arg482 might eliminate the proton pumping ability of cytochrome c oxidase. Since the mutants retain significant electron transfer activity (Table 1), it suggests that there is a loss of coupling of electron transfer and proton pumping. This could mean that the protons must go back out through the proton input channel (or D channel) to an exit channel, although there is no evidence to support this hypothesis. The Mn/Mg Site: Role of Arginines EPR spectroscopy suggests that the Mn/Mg site was altered in all the mutants. The structural change at the Mn/Mg site of R482K is probably due to disruption of the hydrogen bonding network between the side chain of Arg482 and the Mn/Mg ligands, which is normally mediated by water molecules (Figure 1), since the side chain of lysine is somewhat shorter than that of arginine. Mutants R482Q, R482A and R482P completely lost their Mn/Mg site, indicating that a long side chain is critical in maintaining the structure of Mn/Mg through a H-bonding network. The spectra for the CuA site for all the mutants were not grossly altered, suggesting that the CuA site has not been significantly affected, which is consistent with the normal 830 nm optical spectra (Figure 8). A heterogeneity in R482Q and R482P might explain the lower €330 for these mutants than for wild type (Table l). Intrinsic Electron Transfer Rates: Test of the Through-Bond Hypothesis The fast kinetics studies reveal two important aspects of the electron transfer reaction in these mutants. First, all the nonconservative mutants show lower ET rates between cyt.c and Cu (k. in Table 3). This result might imply that Arg482 is close to the cyt.c binding site and that mutations at this site might have reoriented the binding of cyt.c to oxidase, a possibility supported by the ionic strength dependence of the second order reaction which is broadened and has a lower peak than in the wild type or R482K. This, in turn could result in decreased electron transfer facility between cyt.c and oxidase, 187 as reflected by the decreased magnitude of the maximal kslow. Second, the most interesting result is that kb, the intrinsic rate constant from CuA to heme a, for R482Q and R482P is 3500 s’1 and 50 s", respectively, much slower than wild type (1.2 x 105 s"). In the R482P mutant, kb is the same as the overall electron transfer activity, suggesting that the electron transfer between CuA and heme a is rate-limiting, i.e., the hydrogen bond at the interface connecting the two redox centers is important for ET, consistent with the “through-bond” theory. Of course, the real structure of these interface mutants, in terms of whether the hydrogen bonding network is broken, could only be certain with three- dimensional crystal structures. Furthermore, the fact that the mutations induced a global structural change at the heme, Mn/Mg and Cu A sites could argue that the redox potential at the metal centers has changed, thus inhibiting electron transfer. In R482Q, the redox potential difference between heme a and Cu is 46 mV (E.° - ECqu = + 46 mV), close to the wild type (+50 mV), suggesting that the redox potential change would not be the major factor contributing to the slow kb. Conclusions This study demonstrates that Arg482 plays a critical role in maintaining the stability of the interface between subunit I and subunit 11, due to its central position in the hydrogen bonding network, i.e., it is important for maintaining the stability at heme a 3- CuB, Mn/Mg and cyt.c binding sites. The loss of proton pumping activity of the non- conservative mutants might suggest an important role for the 6-propionate group of heme 03 in the proton pumping mechanism. The rate-limiting, internal ET rate between Cu A and heme a for R482P might suggest a “through-bond” electron transfer mechanism between these two redox centers, but other possibilities exist. Therefore, a different mutant that does not affect the global structure but which still breaks the bonding network is desired. Towards this goal, another mutant, R481K, has been generated, and more characterization is underway (also see Summary and Perspectives). Ear" -' . REFERENCES Antonini, G., Malatesta, F., Sarti, P. & Brunori, M. (1993) Proton pumping by cytochrome oxidase as studied by time-resolved stopped-flow spectrophotometry. Proc Natl Acad Sci U S A 90: 5949-53. Berry, E. A. & Trumpower, B. L. (1987) Simultaneous Determination of Hemes a, b, and c from Pyridine Hemochrome Spectra. Anal. Biochem. 161: 1-15. Fetter, J. R., Qian, J., Shapleigh, J., Thomas, J. W., Garcia-Horsman, J. A., Schmidt, B., Hosler, J., Babcock, G. T., Gennis, R. B. & Ferguson-Miller, S. (1995) Possible Proton Relay Pathways in Cytochrome c Oxidase. Proc. Natl. Acad. Sci. USA. 92: 1604-1608. Hill, B. C. (1993) The Sequence of Electron Carriers in the Reaction of Cytochrome c Oxidase with Oxygen. J. Bioenerg. Biomembr. 25: 115 -120. Hofacker, I. & Schulten, K. (1997) Oxygen and Proton Pathways in Cytochrome c Oxidase. Proteins: Structure, Function & Genetics 30: 100-107. Hosler, J. P., Fetter, J ., Tecklenburg, M. M. J ., Espe, M., Lerma, C. & Ferguson-Miller, S. (1992) Cytochrome aa3 of Rhodobacter sphaeroides as a Model for Mitochondrial Cytochrome c Oxidase. J. Biol. Chem. 267: 24264-24272. Hosler, J. P., Shapleigh, J. P., Mitchell, D. M., Kim, Y., Pressler, M., Georgiou, C., Babcock, G. T., Alben, J. O., Ferguson-Miller, S. & Gennis, R. B. (1996) Polar Residues in Helix VIII of Subunit I of Cytochrome c Oxidase Influence the Activity and the Structure of the Active Site. Biochemistry 35: 10776-10783. Hosler, J. P., Shapleigh, J. P., Tecklenburg, M. M. J ., Thomas, J. W., Kim, Y., Espe, M., Fetter, J., Babcock, G. T., Alben, J. O., Gennis, R. B. & Ferguson-Miller, S. (1994) A Loop between Transmembrane Helices IX and X of Subunit I of Cytochrome c Oxidase Caps the Heme a-Heme a3-CuB Center. Biochemistry 33: 1194-1201. Iwata, S., Ostermeier, C., Ludwig, B. & Michel, H. (1995) Structure at 2.8 A resolution of cytochrome c oxidase from Paracoccus denitrificans. Nature 376: 660-669. Lanyi, J. K. (1997) Mechanism of ion transport across membranes. Bacteriorhodopsin as a prototype for proton pumps. J Biol Chem 272: 31209-12. Magalhaes, A., Maigret, B., Hoflack, J ., Gomes, J. & Scheraga, H. (1994) Contribution of Usual Arginine-Arginine Short-Range Interactions to Stabilization and Recognition in Proteins. Journal of Protein Chemistry 13: 195-215. 188 189 Marcus, R. A. & Sutin, N. (1985) Electron transfers in chemistry and biology. Biochim. Biophys. Acta 811: 265-322. Mitchell, D. M. & Gennis, R. B. (1995) Rapid Purification of Wildtype and Mutant Cytochrome c Oxidase from Rhodobacter sphaeroides by Ni2 +-NTA Affinity Chromatography. FEBS Lett. 368: 148-150. Moser, C. C., Keske, J. M., Warncke, K., Farid, R. S. & Dutton, P. L. (1992) Nature of biological electron transfer. Nature 355: 796 - 802.’ Moser, C. C., Page, C. C., Farid, R. & Dutton, P. L. (1995) Biological Electron Transfer. J. Bioenerg. Biomembr. 27: 263-274. Ostermeier, C., Harrenga, A., Ermler, U. & Michel, H. (1997) Structure at 2.7 A resolution of the Paracoccus denitrificans two-subunit cytochrome c oxidase complexed with an antibody Fv fragment. Proc. Natl. Acad Sci. USA 94: 10547-10553. Pan, L. P., Hibdon, S., Liu, R.-Q., Durham, B. & Millett, F. (1993) Intracomplex Electron Transfer Between Ruthenium-Cytochrome c Derivatives and Cytochrome c Oxidase. Biochemistry 32: 8492-8498. Qian, J., Shi, W., Pressler, M., Hoganson, C., Mills, D., Babcock, G. T. & Ferguson- Miller, S. (1997) Aspartate-407 in Rhodobacter sphaeroides cytochrome c oxidase is not required for proton pumping or Mn binding. Biochemistry 36: 2539-2543. Ramirez, B. B., Malmstrom, B. G., Winkler, J. R. & Gray, H. B. (1995) The currents of life: The terminal electron-transfer complex of respiration. Proc. Natl. Acad. Sci. USA 92: 11949 -11951. Smith, P. K., Krohn, R. 1., Hermanson, G. T., Mallia, A. K., Gartner, F. H., Provenzano, M. D., Fujimoto, E. K., Goeke, N. M., Olson, B. J. & Klenk, D. C. (1985) Measurement of protein using bicinchoninic acid [published erratum appears in Anal Biochem 1987 May lS;l63(1):279]. Anal Biochem 150: 76-85. Tsukihara, T., Aoyama, H., Yamashita, E., Tomizaki, T., Yamaguchi, H., Shinzawa-Itoh, K., Nakashima, R., Yaono, R. & Yoshikawa, S. (1996) The Whole Structure of the 13- Subunit Oxidized Cytochrome c Oxidase at 2.8 A. Science 272: 1136—1144. Vanneste, W. H. (1966) The Stoichiometry and Absorption Spectra of Components a and a3 in Cytochrome c Oxidase. Biochemistry 5: 838-848. - _ --_ .. 190 Yoshikawa, S., Shinzawa-Itoh, K., Nakashima, R., Yaono, R., Yamashita, B., Inoue, N., Yao, M., Fei, M. J., Libeu, C. P., Mizushima, T., Yamaguchi, H., Tomizaki, T. & Tsukihara, T. (1998) Redox-coupled crystal structural changes in bovine heart cytochrome c oxidase. Science 280: 1723-9. Zhen, Y., Mills, D., Hoganson, C. W., Lucas, R. L., Shi, W., Babcock, G. & Ferguson- Miller, S. in Frontiers in Cellular Bioenergetics: Molecular Biology, Biochemistry and Physiopathology Papa, S., Guerrieri, F. & Tager, J. M., Eds. (Plenum Press, New York, 1997) pp. in press. SUMMARY AND PERSPECTIVES 1. Asp132 Is Essential for Proton Pumping in Rhodobacter sphaeroides Cytochrome c Oxidase. Previous mutagenesis studies in E. coli cytochrome c oxidase suggested that D135N (E. coli numbering, corresponding to Asp132 in Rhodobacter sphaeroides) lost proton pumping activity but retained 45% of wild type electron transfer activity (Thomas et al., 1993; Garcia-Horsman et al., 1995). Fetter and coworkers generated an equivalent mutant in Rhodobacter sphaeroides, D132N , and also observed loss of proton pumping activity but retention of only 3% of wild type electron transfer activity (Fetter et al., 1995). Based on these studies, two more mutants were generated, D132A and D132E, to test the involvement of Asp132 in proton pumping in Rhodobacter sphaeroides cytochrome c oxidase. The D132E mutant exhibits characteristics similar to wild type. The D132A mutant appears to have completely lost proton pumping activity, but retains some electron transfer activity at a rate of 43 3", compared to about 2000 s" for wild type. Since the structure at the heme sites of D132A is unaltered, these results indicate that the carboxyl moiety of Asp132 is critical for proton pumping activity. When the D132A mutant was reconstituted into vesicles, it showed inhibition instead of stimulation of electron transfer activity upon release of membrane potential or pH gradient, which confirms that the enzyme was reconstituted correctly. The long chain fatty acid arachidonic acid was able to restore the electron transfer activity of D132A up to seven-fold and restored normal respiratory control, suggesting that the fatty acid compensates for the lost charge group at position 132, and facilitate proton uptake by the normal uptake route (Fetter et al., 1996). 191 192 It was concluded that the carboxyl group of Asp132 plays a critical role in proton pumping; its location at the entrance of a hydrogen bonded network has been shown in three-dimensional crystal structures of bacterial and mammalian cytochrome c oxidase (Iwata et al., 1995; Tsukihara et al., 1996). 2. Improvement of Expression and Purification of Cytochrome c Oxidase Is Important for Site-Directed Mutagenesis Studies. In order to get large amounts of cytochrome c oxidase for various biochemical and biophysical analyses and ultimately for crystallizing the enzyme, an overexpression construct encoding the subunit I and II/IH genes. on a single plasmid was generated. The oxidase expression level with the overexpressed strain was increased up to four-fold over wild type. In previous studies, a rapid purification system had been achieved by constructing a histidine-tagged Rhodobacter sphaeroides cytochrome c oxidase strain and using NT A-Ni affinity chromatography (Mitchell & Gennis, 1995). However, the imidazole eluant caused a spectral shift in the purified enzyme, reflecting a denaturation effect. A new eluant, histidine, was used and no spectral shift was observed, suggesting that histidine is a gentler eluant than imidazole (Zhen et al., 1998). The overexpressed construct was also applied to the mutant D132A, followed by purification with nickel chromatography, resulting in a faster procedure with higher yield and better purity. Mutants in the later studies of this work, including D407A, D407N, D407C, R482K, R482Q, R482A and R482P, were all attached to a histidine tag and purified with Ni chromatography with great success. 193 3. Asp407 Is Not Required for Proton Pumping or Mn Binding in Rhodobacter sphaeroides Cytochrome c Oxidase. The three-dimensional crystal structure of bacterial cytochrome c oxidase shows that a highly conserved aspartate residue, Asp407, is located at the interface of subunit I and subunit II, above the heme a J-CuB binuclear center; it was thus proposed to be involved in proton exit or manganese binding (Iwata et al., 1995). Based on these hypotheses, Asp407 was converted to asparagine, alanine and cysteine (Qian et al., 1997). All the mutants showed good electron transfer and proton pumping activity. The structures at the heme, Mn/Mg and CuA sites, as evidenced by EPR, visible and resonance Raman spectroscopy, were all similar to wild type. These results strongly demonstrate that Asp407 is not required for proton pumping activity. A slight broadening in the EPR spectrum of the D407A mutant grown under high Mn/low Mg condition was observed, indicating some minor disorder at the Mn binding site. However, the degree of Mn incorporation for the D407A mutant (85%) was found to be similar to wild type (88%), suggesting that Asp407 is not a Mn ligand. 4. Arg482 Is Important for Stabilizing the Interface between Subunit I and Subunit II, and Plays an Indirect Role in Proton Pumping. According to the bacterial and mammalian cytochrome c oxidase structures (Iwata et al., 1995; Tsukihara et al., 1996; Osterrneier et al., 1997; Yoshikawa et al., 1998), a highly conserved arginine pair are located at the interface between subunit I and subunit II and are involved in a through-bond pathway connecting Cu A and heme a, including fourteen covalent bonds and two hydrogen bonds between the CuA ligand His260 and the heme a propinonic acid. This path was proposed to be the facilitated 194 electron transfer route between these two redox centers (Ramirez et al., 1995). Alternatively, a distance-dependent electron transfer theory must also be considered (Moser et al., 1992; Moser et al., 1995), which proposes that the precise structure of the protein between two redox centers is not a critical determinant of rates of electron transfer. To test these two hypotheses (“through-bond” or “through-space”), four mutants, R482K, R482Q, R482A and R482P, were generated and characterized. They were designed to disrupt the bonding pathway to various extents. Except for R482K whose properties are very similar to wild type, all the mutants show decreased overall electron transfer activity and no proton pumping activity. It might suggest that the positive charge at the R482 position is critical for normal function, and that all other mutations introduce a backbone shift, breaking both the through-bond pathway and a hydrogen bond interaction between neighboring Arg481 and the 6-propionate group of heme a,. This propionate group was proposed to be involved in one version of a direct coupling model for proton translocation, a purely theoretical model so far. These mutants may provide the first suggestive evidence for the hypothesis. Unfortunately, the usefulness of the mutants is somewhat diminished by evidence of instability from visible spectroscopy, heme assays and protein assays suggesting loss of heme A and a selective loss of heme a, in most mutants. EPR spectroscopy studies also show that the Mn/Mg binding site of these mutants are altered. The internal ET rate from cytochrome c to CuA is slower than wild type, indicating some effect on the binding of cytochrome c to oxidase when mutations were introduced at Arg482. More strikingly, R482Q and R482P have much slower ET rates than wild type between CuA and heme a, suggesting that the bonding path between Cu A and heme a might indeed be important for electron transfer between CuA and heme a. However, the global structural changes at the heme a, heme a3 and 195 Mn/Mg sites prevent any firm conclusions to be drawn in this respect from these mutants (Qian et al., 1998). 5. Future Research It is of great interest to understand the mechanism of how electron transfer and proton pumping are coupled in cytochrome c oxidase due to its key function in energy conservation in both eukaryotes and prokaryotes (Mitchell & Moyle, 1965; Wikstrom etal., 1981). The results from the Arg482 mutants strongly suggest that this residue may be important for controlling the direct electron transfer rate from CuA to heme a, and indirectly involved in proton pumping activity. Overall structural changes following the mutagenesis also indicate that the positive charge is critical for maintaining the protein stability through the hydrogen bonding network. It is apparent that the arginine pair is critical in enzyme structure and function. More mutagenesis studies need to be done at this region. Although EPR, resonance Raman, UV/visible and optical spectroscopy can provide large amounts of information on the structure at Mn/Mg, heme and CuA sites, subtle changes in the protein structure are not easy to detect with these methods. Crystallization and X-ray crystallography of wild type and mutant Rhodobacter sphaeroides cytochrome c oxidase are essential to reveal the structural information. It would be especially useful for the Arg482 mutants, because whether the hydrogen bonding network is broken or not in these mutants cannot be known without the crystal structures. An ideal mutant to test the “through-bond” or “through-space” theory would meet the criteria that it would break the bonding pathway but not cause a global 196 structural change at the metal centers. Arg481, the residue neighboring Arg482, is hydrogen bonded to the 6-propionate group of heme a,. By changing Arg481 to a lysine, a positively charged residue with a shorter side chain, it is expected that the lysine side chain of R481K might be drawn toward the 6-propionic group of heme a 3 to maintain the hydrogen bond between R481K and the propionate of heme a 3. This could cause a backbone shift and breaking the H-bonding network between Cu A and heme a. Hopefully, this change would not cause substantial structural change at the heme a, heme a 3, Mn/Mg or Cu A sites. And if the bonding network is essential for determining the ET rate between CuA and heme a, a significantly decreased ET rate for R481K would be expected. REFERENCES Fetter, J. R., Qian, J ., Shapleigh, J., Thomas, J. W., Garcia-Horsman, J. A., Schmidt, E., Hosler, J ., Babcock, G. T., Gennis, R. B. & Ferguson-Miller, S. (1995) Possible Proton Relay Pathways in Cytochrome c Oxidase. Proc. Natl. Acad. Sci. USA. 92: 1604-1608. Fetter, J. R., Sharpe, M., Qian, J., Mills, D., Ferguson-Miller, S. & Nicholls, P. (1996) Fatty acids stimulate activity and restore respiratory control in a proton channel mutant of cytochrome c oxidase. FEBS Lett 393: 155-160. Garcia-Horsman, J. A., Puustinen, A., Gennis, R. B. & Wikstrom, M. (1995) Proton Transfer in Cytochrome b03 Ubiquinol Oxidase of Escherichia coli: Second-Site Mutations in Subunit I That Restore Proton Pumping in the Mutant Asp135—rAsn. Biochemistry 34: 4428-4433. Iwata, s., Ostermeier, C., Ludwig, B. & Michel, H. (1995) Structure at 2.8 A resolution of cytochrome c oxidase from Paracoccus denitrificans. Nature 376: 660- 669. Mitchell, D. M. & Gennis, R. B. (1995) Rapid Purification of Wildtype and Mutant Cytochrome c Oxidase from Rhodobacter sphaeroides by Ni2+-NTA Affinity Chromatography. FEBS Lett. 368: 148-150. Mitchell, P. & Moyle, J. (1965) Stoichiometry of proton translocation through the respiratory chain and adenosine triphosphatase systems of rat liver mitochondria. Nature 208: 147-51. Moser, C. C., Keske, J. M., Warncke, K., Farid, R. S. & Dutton, P. L. (1992) Nature of biological electron transfer. Nature 355: 796 - 802. Moser, C. C., Page, C. C., Farid, R. & Dutton, P. L. (1995) Biological Electron Transfer. J. Bioenerg. Biomembr. 27: 263-274. Ostermeier, C., Harrenga, A., Ermler, U. & Michel, H. (1997) Structure at 2.7 A resolution of the Paracoccus denitrtficans two-subunit cytochrome c oxidase complexed with an antibody Fv fragment. Proc. Natl. Acad. Sci. USA 94: 10547-10553. Qian, J., Geren, L., Hoganson, C., Pressler, M., Looney, A., Babcock, G. T., Millett, F. & Ferguson-Miller, S. (1998) Arg482 is important for stabilizing the interface of subunit I and subunit II of Rhodobacter sphaeroides cytochrome c oxidase. in preparation : . Qian, J., Shi, W., Pressler, M., Hoganson, C., Mills, D., Babcock, G. T. & Ferguson- Miller, S. (1997) Aspartate-407 in Rhodobacter sphaeroides cytochrome c oxidase is not required for proton pumping or Mn binding. Biochemistry 36: 2539-2543. Ramirez, B. B., Malmstrom, B. G., Winkler, J. R. & Gray, H. B. (1995) The currents of life: The terminal electron-transfer complex of respiration. Proc. Natl. Acad. Sci. USA 92:11949 -11951. 197 :I' 198 Thomas, J. W., Puustinen, A., Alben, J. O., Gennis, R. B. & Wikstrom, M. (1993) Substitution of Asparagine for Aspartate-135 in Subunit I of the Cytochrome bo Ubiquinol Oxidase of Escherichia coli Eliminates Proton-Pumping Activity. Biochemistry 32: 10923-10928. Tsukihara, T., Aoyama, H., Yamashita, E., Tomizaki, T., Yamaguchi, H., Shinzawa- Itoh, K., Nakashima, R., Yaono, R. & Yoshikawa, S. (1996) The Whole Structure of the 13-Subunit Oxidized Cytochrome c Oxidase at 2.8 A. Science 272: 1136-1144. Wikstrom, M., Krab, K. & Saraste, M. Cytochrome Oxidase - A Synthesis (Academic Press, New York, 1981). Yoshikawa, S., Shinzawa-Itoh, K., Nakashima, R., Yaono, R., Yamashita, B., Inoue, N., Yao, M., Fei, M. J., Libeu, C. P., Mizushima, T., Yamaguchi, H., Tomizaki, T. & Tsukihara, T. (1998) Redox-coupled crystal structural changes in bovine heart cytochrome c oxidase. Science 280: 1723-9. Zhen, Y., Qian, J., Follmann, K, Hosler, J., Hayward, T., Nilsson, T. & Ferguson- Miller, S. (1998) Overexpression and purification of cytochrome c oxidase from Rhodobacter sphaeroides. Protein Expression Puri in press: .