LIBRAM Michigan State University This is to certify that the dissertation entitled HYDROGEN BONDING INTERACTIONS IN CYTOCHROME A OF CYTOCHROME OXIDASE: A RESONANCE RAMAN STUDY. presented by Jose A. Centeno—Ortiz has been accepted towards fulfillment of the requirements for Ph.D. degree in Chemistry Major professor WM M], 01% '/ MS U is an Affirmative Action/Eq ual Opportunity Institution 0-12771 MSU LIBRARIES v RETURNING MATERIALS: Place in book drop to remove this checkout from your record. FINES will be charged if book is returned after the date stamped below. HYDROGEN BONDING INTERACTIONS IN CYTOCHROME A 0F CYTOCHROME OXIDASE: A RESONANCE RAMAN STUDY By Jose Antonio Centeno Ortiz A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree DOCTOR OF PHILOSOPHY Department of Chemistry 1987 ABSTRACT HYDROGEN BONDING INTERACTIONS IN CYTOCHROME A OF CYTOCHROME OXIDASE: A RESONANCE RAMAN STUDY. By Jose Antonio Centeno Ortiz Hydrogen bonding interactions at the formyl (-CHO) group of cytochrome a in cytochrome oxidase have been investigated by using spectroscopic and chemical methods. Ikme a model compounds that mimic the spectroscopic pro- tmrties of cytochrome a have been isotopically synthesized idth formyl (-CD0) and (-CHlBO) groups. With the aid of these isotopic substitutions we have identified essentially all the internal formyl vibrations, and compared them with similar isotopic and H—bonding effects in benzaldehyde. To identify resonance Raman enhancement of cytochrome a2t formyl related modes, we have obtained the resonance Raman spectra in the range from 150 — 1700 cm'l, under four different experimental conditions: visible excitation in resonance with the Q(0-0) ----> ‘ transition; BCidiC denaturing conditions; H/D exchange; and ligand deconvo- lution. From these spectra the hydrogen bond sensitive vibrations of cytochrome a2t can be identified and a comparison of these vibrations with those previously obs 0011 (flmerved for heme a and benzaldehyde under H-bonding conditions provide the assignment of -CHO internal modes. Because of the lower symmetry of the heme a macrocycle and the strong electron withdrawing character of the H-bonded CbO, we observed RR activity (and intensity changes) for symmetry forbidden Eu modes. The importance of the cytochrome a2t H-bonded structure in the oxidase proton pumping activity was examined. Our results reveal significant perturbations at the cytochrome 93* H-bond interaction as the adjacent protein protons are exchanged by deuterons during enzyme turnover conditions. The suggested involvement of the Cu;2+ atom in the oxidase redox-linked mechanism was also investigated by ENDOR spec— troscopy. Our results reveal the occurrence of exchangeable protons at sites near the CUA site; however, their involvement in the Ht—pump remains unclear. To Maria doselito iv and Waldemar I my l BHCOUI l readil educa' rese assi COmF Neg USe ACKNOWLEDGMENTS : I would like to thank Prof. Gerald T. Babcock for his nmny helpul suggestions to this project and for the encouragement he offered me when I most needed it. I am deeply grateful to Dr. George Leroi for critical reading and unconditional support during my graduate education. I would like to express my gratitude to all my family members and friends who, in their own way, motivated me mnfing my graduate career at Michigan State. So to my Imments, Amelia and Domingo; my brothers, Luis, Alex, and thawin; my uncles Papo and Adelina; and my friends, Jerry Babcock, Roberto Lopez, Ismael Scott, Luis Garcia, Benito, mmlJuan Lopez; I extend my heartfelt appreciation. My thanks are also extended to members of Prof. Babcock’s research group for their contribution and general technical assistance, in particular, Dr. Robert Kean, who helped me in the design of the Raman difference cuvette and the necessary- cmmputer programming; Mr. Dwight Lillie for computer {upgramming; and Tony Oertling for his instruction in the usecfi’the lasers (and for his occasional delicatessen). My thanks are also extended to Drs. Patricia Callahan, Patricia throney, and Juan Lopez, for their encouragement, many helpful discussions, and constant cooperation. My gratitude is also extended Mr. Hak-Hyum Nam and Mr. Dean Petersen for their would not f: Ingle maize Progl fro" 0m their assistance on the recording of the FTIR Spectra. I would still be doing the work in Chapters 3 and 5 if it were run.for the valuable assistance I received from them. I would like to thank Mr. Ron Hass, Marti Rabb, and Punfred Langer for their technical assistance. The design cfi'the Raman difference apparatus would not have been pmssible without the assistance of Marti and Manfred. Thank you guys !!! I would like to acknowledge research efforts with Bob Ingle and Prof. Sheilagh Ferguson-Miller on the study of nmize cytochrome oxidase developed while this project was in progress. The efforts of Mrs. Dianne Kazmierski and Maria Cbnteno, in the preparation of this manuscript are greatly appreciated. I also wish to acknowledge the Department of Chemistry, the National Institutes of Health, and the Office of Urban Affairs at Michigan State University and its Director Dr. (meandra Simmons, for financial support during the course of this project. Finally, I don’t have words that can compensate for all the love, understanding, and endurance that I have received from Maria and my kids during the course Of my graduate career. I couldn’t have done it without you: vi TABLE OF CONTENTS List of Tables . . . . . . . List of Figures . . . . . . . . . Chapter I . . . . . . . . . I. Cytochrome Q Oxidase: Structure and Functions A. Introduction and Overview of Thesis B. Brief History . . . . . . . . . . . II. Cytochrome Oxidase: Structure and Metal Active Sites. . . . . . . . . . . A. Subunit Composition and Protein Structure 0 I O O O O O 0 O 0 O 0 B. Metal Centers: Structure and Protein Environment . . . . . . . . . . . . . C. Chromophore Geometry and Ligand- Coordination . . . . . . . . . . . . III. Proton Pumping . . . . . . . . . . . . . IV. Concluding Remarks . . . . . . V. Aims and Strategy of this Thesis . . . Chapter II Materials and Methods . . . . . . . . . I. Introduction . . . . . . . . . . . . . . . II. Materials . . . . . . . . . . . . . . . . A. Isolation of Cytochrome c Oxidase . . . . B. Preparation of Cytochrome c Oxidase Derivatives . . . . . . . . . . . . . . . C. Alkaline and acidic pH-modification of Purizied Cytochrome Oxidase . . . . . D. Redox—cycled Cytochrome Oxidase . . . . E. Preparation of Heme a Derivatives . . . III. MethOdS o o o o o 0 0 ' ° ‘ ' ' ° ° ' vii 14 22 26 3O 32 34 34 35 35 36 37 38 39 44 A. Electronic Optical Absorption Spectroscopy . . . . . . . . . . . . . . 47 B. Resonance Raman Spectroscopy . . . . . . 52 1) Resonance Raman Difference Spec- troscopy . . . . . . . . . . . . . . 56 C. Electron Nuclear Double Resonance Spectroscopy (ENDOR) . . . . . . . . . . 67 D.\ Infrared (IR) Spectroscopy . . . . . . . . 68 Chapter III Formyl Isotopic Substitution and Hydrogen Bond Sensitive Modes in Heme a Model Compounds: Models for Cytochrome a . . . . . . . . . . . 69 Summary . . . . . . . . . . . . . . . . . . . . . 69 I. Introduction . . . . . . . . . . .‘. . . . . . 70 II. Results and Discussion . . . . . . . . . . . . 74 A. Effects of Hydrogen Bonding on the Absorption Spectra of Heme a Complexes. . 74 B. Effects of H-bonding and isotopic sub- stitution on the Raman spectra of benzaldehyde . . . . . . . . . . . . . . 82 C. Formyl modes in the RR and IR spectra of heme a and its Cu2* Porphyrin Ana- log 0 O 0 o o o o o o o o o o o o o o o o 86 D. H-bonding effects in Heme a Models. . . . 110 III. conClUSIODS o o o o o o o o o o o o o o o 119 Chapter IV Visible Excitation Resonance Raman Spectra of Cytochrome a 1n Cyto- 121 chrome Oxidase . . . . . . . . . - . . 1‘ IntrOdUCtiOD o o o o o O 0 0 . 121 11° Materials and Methods . . . . . . . . - ' ° ' 123 A' Materials 0 o o o o O 0 0 ° 0 . . 123 . 5 B. Instrumentation . . . - . . . . - ° 12 III. Results . . o- o . . . . . . . . . o 125 A- Visible Excitation Resonance Raman Spectra viii IV. V. of Cytochrome git in H20/D20 Buffers B. Effects of deuterium substitution for cyt azt modes above 1000 cm-1 . . . . . C. Low-frequency cytochrome 92+ vibration below 1000 cm‘1 . . . . . . . . . . . . Discussion . . . . . . . . . . . . . . . . A. High-frequency vibrational assignments of cyt §2+ . . . . . . . . . . . . . . B. Low-frequency vibrations of cyt a2* . . Conclusions . . . . . . . . . . . . . Chapter V Hydrogen Bond Sensitive Vibrations of Cytochrome a in Cytochrome Oxidase . . I. Introduction . . . . . . . . . . . . . II. Materials and Methods . . . . . . . . . . . III. Results 0 O O O O I O O O O O O O O 0 O O O A. Low pH Effects on the RR spectra of Radical Cytochrome Oxidase . . . . . B. Low pH Effects on the RR spectra of Cyanide inhibited Cytochrome Ox1dase. IV' DiSCUSSiOH. o o o o o a o o o o 0 O 0 ' ° A. Assignments of H-Bonding Sensitive Vibrations of Cyt azt . . . . . . . . B. Low frequency Vibrational Assignmenfis V ° ConClUding Remarks 0 o o o o o o 0 Chapter VI Evidence for cyt a Involvement on the cyt g oxidase Proton Pump Mechanism. A Resonance Raman and ENDOR study . . . 1' IntrOdUCtion o o o o o 0 0 0 ° ° ° ° . . 11' ReSUltS o o o o o o O 0 0 ° . ‘ . O . A. Resonance Raman Spectra of Resting Cytochrome Oxidase . ° ' ' ° 7 ' ° B. Resonance Raman Spectra of Redox— CyCled OXidaseO C O 0 0 O O 0 O 0 ix 0 125 130 136 143 143 147 150 152 152 153 153 153 162 174 181 184 187 188 188 191 191 197 C. EPR and ENDOR Studies and Redox- Activated Oxidase . . . . . . . . . . . . 216 III. Discussion . . . . . . . . . . . . . . . . . 227 A. Enhancement of Cytochrome a Formyl H-bonding Sensitive Modes: Slow H/D Exchange Process . . . . . . . . . . . . 227 B. Low Frequency Region . . . . . . . . . . 234 C. ENDOR Results . . . . . . . . . . . . . . 236 IV. Conclusions . . . . . . . . . . . . . . . . . 236 Chapter VII Summary and Future Work . . . . . . . 239 1. Summary . . . . . . . . . . . . . . . . . . . 239 II. Future Work . . . . . . . . . . . . . . . . . 241 A. Spectroscopy of Cyt Oxidase and Heme a Models . . . . . . . . . . . . . . . . 241 3- Hydrogen Bonding . . . . . . . . . . . . 245 References Cited . . 247 LIST OF TABLES Ease Chapter III. Tablgmng Spectroscopic Charac- teristics of Hydrogen Bonded Com- plexes of Heme a, Cu2+ a, and Cuzt Porphyrin . . . . . . . . . . . . . . 77 Table 3.2 Summary of Raman Fre- quencies (cm'l) for Hydrogen—Bond Modes in Benzaldehyde . . . . . . . . 84 Table 3.3 Formyl-Sensitive Modes in Heme a (N-MeIm): (-CHO/—CDO) and Cu2+ Porphyrin a (-CHO/-CDO). . . 92 Table 3.4 Resonance Raman (RR) and Infrared (IR) Frequencies (cm-1) Sensitive to 'CHIGO -—--> ‘CHIBO Substitution at the Formyl Group of (4—Vinyl—8-Formyl) Por- phyrin. . . . . . . . . . . . . . . . 102 Chapter IV. Table 4.1 Tentative Assignments for the Observed High-Frequency Modes in the Resonance Raman Spec- trum of Cyt alt. . . . . . . . . . . 131 Table 4.2 Tentative Assignments for the Observed Low-Frequency Modes in the Visible Excitation Resonance Raman Spectrum of Cyt 82+. 0 O 0 O 0 O O O O 0 0 O 0 0 0 0 139 Table 4.3 Correlation Table for Species with Dan. Dan. and Can Point Group Symmetry. . . . . . . O O 149 Cil'lapter 'V. Table 5.1 High Frequency Vibra- tional Assignments of Cytochromes g and E3 0 o o o o o o o o o o o . . 172 Table 5.2 High Frequency Alde— hyde Group Vibration Assignments and Shifts Observed for Benzal- dehyde in the Presence of Hydrogen Donors and Cytochrome alt at Neu- 173 tral and Acid pH. . . . . . . . . . . Table 5.3 pH Senstive Vibrations (cm-1) at Cytochrome a2+ Observed with 441.6 nm excitation . . . . 175 xi Chapter VI. Table 6.1 Tentative Raman Fre- quency Assignments for Cytochrome a Formyl Modes in Cytochrome g Oxidase . . . . . . . . . . . . . Table 6.2 Summary of Raman Fre- quency Assignments for Cytochrome a Vinyl Modes in Cytochrome c OXidase. .0000...OOOOOOIOOOOOOOOOOO xii 230 231 Chapter Chapter I. II. LIST OF FIGURES Fig“. 1. 1 The geometrical arrange- ment of the polypeptides (subunits) in cytochrome c oxidase in the mito- chondrial membrane ................ Fig4m442 The molecular structure of heme a and protoporphyrin IX .... Eig4wl4§ Electronic optical absorp- tion spectra of oxidized ( ) and reduced (----) cytochrome g oxidase in the Soret, visible, and near—IR regions. 0.0.00000000COCOCOOOOOOOCO E4g4m444 Geometry and coordination properties for cytochrome g, cyto- chrome aa, CUA, and Cu: in cyto- chrome g oxidase . ................ Eiggmlmg A simplified scheme of the redox- linked proton pumping activity on cytochrome c oxidase. .. Fig. 2.1 Structures of free base porphyrin a and its porphyrin analog 0 OOOOOOOOOOOOOOOOOOOOOOOO E1g. 2. 2 Absorption spectra of formyl- substituted derivatives of 2, 6- -dipentyl- -4— vinyl- -8— formyl Cu2+ porphyrin. OOOOOOOOOOOOOOOOOOOOOOO Fig.w 2. 3 Absorption Spectra of Free Base Porphyrin a and Guat- substituted Porphyrin a . .......... Fig. 2.4 Energy diagram for metal- loporphyrin absorption spectra...... Fig. 2.5 Diagram illustrating Raman and Rayleigh scattering and the resonance Raman effect ........ Flg. 2. 6 Block diagram of the resonance Raman difference appa— ratus. .0.0.000000000000000000COOCO xiii 15 17 24 28 43 46 49 51 53 57 Chapter III. Fig. 2.7 Schematic of the reso- nance Raman difference hardware and split spinning cuvettte . ........... Fig. 2.8 Schematic of the digital and timing logic for the resonance Raman difference apparatus. ......... Fig; 2.9 Resonance Raman spectra of cytochrome 92* in protonated and deuterated buffers (pH:pD=7.4), and the associated Raman differ- ence spectrum. OOOOOOOOOOOOOOOOOOOO. Fig44344 Proposed H-bonded struc- ture of cyt a in cytochrome g OXidase 0.00....OOOOOOOOOOOOOOOOOOOOO Fing§42 Absorption spectra of heme a (N-MeIm): under various hydrogen bonding conditions ......... £1g4m343 Plot of observed red- shift in the Soret and cx—band transition of Cu2+ porphyrin a as a function of the acidity (pka) of phenol H—donor derivative ........ §i§;"§;4 Raman spectra of ben- zaldehyde-CHO and its H-bonded derivatives with phenol-OH and phenol-OD as H-donors ............... Fig. 3.5 Resonance Raman (RR) spectra of oxidized and reduced detergent monomerized heme a (N-MeIm): in dry CHzClz. after treatment with KCN/HzO and KCN/DzO .......................3..... Fig. 3.6 Resonance Raman (RR) Spectrauof Cu2* porphyrin a with normal (-CH0) and deuterated (-CDO) formyl group substituent, in dry CH201. ....................... Egg. 3.7 Resonance Raman (RR) Spectra of 4-vinyl—8-formyl Cu2+ porphyrin with normal (—CH160) and 18O-labelled (-CH1°O) formyl group SUbStituent in dry CH3012 oooooooooooo Fig. 3.8 FTIR spectra of heme ERV:CITT Cu2+ (-CH1°O), and Cu2+ xiv 58 61 72 76 81 83 89 91 96 Chapter ChaPter IV. V. (-CH180) porphyrin derivatives in KBR pellets in the region from 1000 to 1800 cm-IOOOIOOIOO‘DOOIOUCODOOOOOO 98 EichEAE. FTIR spectra of heme a3i—Cl', Cu2+ (- -CH150) porphyrin, and the dimethylester derivative of protoporphyrin Feat—Cl- in carbon tetrachloride (CCla) ........ 101 Fig4_§44g Resonance Raman (RR) spectra of ferric nonhydrogen bonded heme a3i (N-MeImlz (in CH2C12) and ferric heme §3* (N— —MeIm)z (in 0014) hydrogen bonded to phenol— —OH and phenol— OD sssssssss 00.0000.- OOIOOCIUDIIIO.‘ 112 tra of ferrous heme a(N— —MeIm)2 H- bonded to phenol- -ODs in dry CHZClZ anuuooococo-cocooooon-oono-oo 113 Fig. 3_ _12 Resonance Raman (RR) spectra of Cu2* (— —CH150) porphy- rin in CHzclz and under H-bonding to phenol-OH and phenol-OD.......... 115 F_ig;m3w43 Low- frequency resonance Raman (RR) spectra of ferrous heme a2+ (N-MeIm); in CHzclz, in Hzo/detergent, and D20/detergent.... 116 §i§;_4'l Visible excitation (605 nm) resonance Raman spectra of reduced cytochrome oxidase from 1000 to 1300 cm-l.............. 127 FigLM4. 2 Visible excitation (605 nm) resonance Raman spectra of reduced cytochrome oxidase from 1250 to 1700 cm‘l.............. 129 EingéLg Low frequency visible excitation (605 nm) resonance Raman spectra of fully reduced cytochrome oxidase ................. 137 F1g_4.4 Low frequency visible excitation (605 nm) resonance Raman spectra of fully reduced CYtochrome oxidase. ................ 138 Fig. 5.1 Optical absorption XV spectra of reduced cytochrome Fig. 5.2 High frequency resonance Raman spectra of reduced cytochrome oxidase at the indicated acidic pH Fig. 5. 3 Low frequency resonance Raman spectra of reduced cytochrome Fig 5. 4 Low frequency resonance Raman spectra of fully reduced and partially reduced cyanide inhibited cytochrome oxidase obtained with 441.6 nm excitation at the indi- oxidase at neutral and acidic pH.... 154 level oooooooooo-Cocooooono-ooocoooo 157 oxidase at the indicated pH levles... 160 cated pH levels .................... 165 Fig ___ Optical absorption spec— tra of partially reduced, cyanide- inhibited cytochrome oxidase at several pH levels ................... 168 F4g44545 Low frequency resonance Raman spectra of partially reduced cyanide inhibited cytochrome oxi- dase at several pH levels obtained with 441.6 nm excitation. ........... 169 Fggwm541 High frequency resonance Raman spectra of mixed-valence cytochrome oxidase at the indicated pH levels. 10000-000ounioonovloololooo 170 Fig4_54§ Schematic of proposed structural changes of reduced and partially reduced plus cyanide heme at acid pH levels .................... 180 Chapter VI- FigLWQLL Optical absorption spec- tra of resting cytochrome oxidase in protonated and deuterated buff- ers. ................................ 192 FigLnng Resonance Raman spectra and the associated Raman differ— ence spectrum (RDS) of resting CYtochrome oxidase in protonated and deuterated buffers from 1000 to 1440 cm'1 and from 1300 to 1700 cm'1 ......... ................. 194 Fi . 6.3 Resonance Raman spectra and cytochrome oxidase under various turnover conditions as monitored with 406.7 nm excitation............. 199 gig. 6.4 Resonance Raman spectra ‘ of fully reduced cytochrome oxi- \ dase observed in the carbonyl ‘ stretching region of cytochrome ‘ "g2“ nanoun...oat-cocoooooa-oaoo-cooa. 202 Eigym§45 Plot of frequency of the cytochrome §2* carbonyl stretching frequency as a function of incuba— tion time in D200 Ono-ooQonooootoooo. 205 El§4~§i§ Resonance Raman spectra of redox-cycled oxidase as a func- tion of enzyme turnover cycles induced by increasing the concen- tration of ascorbate in the solu— tion noooooooocount-otto-co'ooooccoo 207 gig, 6.7 Resonance Raman spectra and the associated difference spectrum of redox—cycled mixed- valence cytochrome oxidase obtained with 441.6 nm (7A) and 406.7 nm (78) excitation frequencies............... 210 fig. 6.8 Low frequency resonance Raman spectra of reduced (§2*§32*) and redox—cycled (az‘ga3*-CN‘) cytoe chrome oxidase in protonated and deu— terated buffers ..................... 214 gig. 6L2 EPR spectra of cyto— chrome oxidase ..................... 219 El§;,6-10 ENDOR spectra of cyto- chrome aa' at the electronic g value of g=3.03 . ................... 221 gig, 6.11 Matrix ENDOR spectra of unz’ in cytochrome oxidase at the electronic g; value of 3:2.02 (perpendicular region)........ 224 Egg; 6.13 Matrix ENDOR spectra of CUA2+ in cytochrome oxidase at the electronic g value of . 811:2.1717 (parallel region)......... 226 Chapter VII. gig. 7.1 Comparison of the resonance Raman spectra of cyto— chrome g oxidase isolated from maize (corn seedlings) and bo- vine beef heart mitochondria. . b o o 244 CHAPTER I I. Cytochrome c Oxidase: Structure and Function A. Introduction and Overview of the Thesis: In the mitochondrial respiratory chain of eukaryotic cells the production of energy is based upon a sequence of electron transfer reactions through a set of redox proteins which culminates with the reduction of molecular oxygen to produce water (Wikstrom et al., 1982). The last enzyme in this mechanism is called cytochrome c oxidase. It catalyzes the reduction of dioxygen to water by transferring electrons from reduced cytochrome 92+ to oxygen. The overall reaction is described as follows: 4 cyt.92* + 02 + nH*(in)-—>4 cyt.g3‘ + 2H20 + (n-4)H*(out) The transfer of electrons in this reaction is believed to be coupled to the generation of an electrochemical proton potential gradient (AH?) across the mitochondrial membrane (Wikstrom and Krab, 1979; Casey et 31., 1979; Proteau et 81.) 1983; Wikstrom, 1984; Thelen et al., 1985; Malmstrom, 1985; Blair et al., 1986). This electrochemical proton Potential is composed of a membrane potential (AW ), which arises from vectorial electron transfer from the outside of the membrane, and of the consumption of H+ in the inside giving rise to a ApH. The formation of (A%+) is linked to the production of adenosine triphosphate (ATP), an energy source which is used in biochemical reactions which require energy. In addition to the conSumption of H+ ions in the inner side of the membrane during reduction of 02, the transfer of electrons through the oxidase is also known to be coupled to the translocation of at least one H+/per electron to the cytosol side (or outside) of the membrane (Wikstrom, 1977; Wikstrom et al., 1982). This coupling between electron transfer and proton translocation is essential for energy transduction during respiration. In this introduction I review recent research on the structural, functional, and physiological aspects of cytochrome c oxidase. In the following six Chapters, experiments will be presented which show the application of spectroscopic methods (Uv—vis, resonance Raman (RR), infrared (IR) electron paramagnetic resonance (EPR) and electron nuclear double resonance( ENDOR)) to study structural features of the enzyme and to provide insights into the in vivo structural/functional aspects of this important metalloprotein. The spectroscopic techniques and experimental proce— dures used during the course of this work to study cyto- chrome oxidase, its ligand-bound derivatives, and its isolated heme a model compounds are described on Chapter II. A brief introduction to Uv-vis electronic optical absorption and resonance Raman spectroscopy (RRS) will be presented. The design of a home—made Raman difference apparatus, which features the use of a double-compartment spinning cuvette and digital gating logic, will be described. Chapter III demonstrates the use of isolated heme a model compounds and the application of available spectros- copic techniques (UV—vis, infrared, and RR) to study the molecular structure and the sensitivity of Raman frequencies to porphyrin peripheral substituents. Results are presented from infrared (FTIR) and resonance Raman experiments on heme a model compounds with the aim of obtaining insights into the involvement of the n-conjugated peripheral substituents in the electronic states of heme a. I concen- trate my efforts on elucidating Raman frequencies sensitive to formyl vibrations and on describing the sensitivity of these bands to formyl-proton and oxygen isotopic substitu- tion in addition to hydrogen-bonding effects. In Chapter IV, visible excitation resonance Raman studies of reduced and mixed-valence cytochrome oxidase in protonated and deuterated buffers are presented. Since cytochrome a2+ is the dominant contributor to the 605 nm absorption band of reduced cytochrome oxidase (Vanneste, 1967), laser excitation in this Q(o-0) TY-—-—> nx transition is expected to be dominated by cytochrome §2* Vibrational modes. The strong enhancement of cytochrome a modes with visible excitation, and their sensitivity upon protein H/D-exchange and pH—denaturation, is sugested to arise from interactions of the cytochrome a heme moiety With the surrounded protein residues. A detailed study of the. dey of as 01‘ re vi ox id depolarization ratios of cytochrome a vibrational modes is also presented. Chapter V summarizes our results on the identification under Soret (B(o-o)—transition) excitation. The sensitivity of these cytochrome a modes to protein perturbations, such as acidic pH effects, was also studied. Since, upon raising or lowering the pH of the oxidase solution, the cytochrome a H—bonding structure is expected to be disrupted,a systematic resonance Raman study on the pH-induced effects in the vibrational spectra of reduced and mixed-valence cytochrome oxidase, is expected to provide additional evidence on the identity of the cytochrome a H—bonding modes. Soret excitation resonance Raman studies on the accessibility of labile protons, at the cytochrome a site, were also conducted, with the specific aim of gaining insight into the rate of H/D—exchange at the cytochrome a (-HC=O...H/D-protein) site and its relevance to enzymatic redox changes (as studied by repeated enzymatic cycles of reduction and reoxidation). This data is presented in Chapter VI. The results from these experiments are discussed in terms of the involvement of the cytochrome a formyl group (-HC=O) in the cytochrome oxidase proton pump mechanism. Soret excitation on the resting oxidase indicates the occurrence of exchangeable protons at the cytochrome g/ protein site, as well as spin—state transformations associated with the cytochrome as site. This chapter also deals with the application of electron paramagnetic resonance (EPR) and electron nuclear double resonance (ENDOR) spectroscopies to study the metal active centers of cytochrome oxidase. In this work, EPR spectroscopy was used to monitor enzyme integrity and thus quality of the preparation, as judged by the cytochrome a and CUA EPR- signals. To study H20/D20 exchange at the CuA2+ site (Gelles et al., 1986) ENDOR spectroscopy was used. Finally, future experiments and new approaches are suggested in Chapter VII. B. Brief History: The initial investigations on the active role of cytochrome oxidase in the cellular respiratory chain were reported in 1939 by Keilin and Hartree. In their original work Keilin and Hartree (1939) showed that cytochrome oxidase was composed of two functionally distinct components, which they called cytochrome a and cytochrome Q3. ~They proposed that cytochrome a was involved in the electron uptake from cytochrome 9 while cytochrome as was involved in the binding and reduction of dioxygen. Though Keilin and Hartree (1939) initially commented on the possibility of the participation of the copper atoms in the cYtochrome oxidase oxygen reaction, their involvement was not considered seriously until 1962 when Beinert et al. 185 the 196 com team 0in abs‘ cyt, in "en Shit 0X11 med (19 (1962)-first introduced the use of electron paramagnetic resonance (EPR) to study the electronic environment of the metal sites in the oxidase. The pioneering work by Greenwood and Gibson (1963, 1965, & 1967) on the photolability of the cytochrome a32*—CO complex and its subsequent reaction with dioxygen at low temperatures, made it possible to start studies on the oxidation of the reduced enzyme by oxygen. Using optical absorption spectroscopy techniques, Greenwood and Gibson (1963,1965, & 1967) were able to follow the route of electron transfer from cytochrome a via Cu; to the binuclear cytochrome as /Cua center by detecting short-lived enzyme intermediates. These short-lived transient intermediates were further characterized by using Uv-vis optical (Clore et al., 1980) and EPR (Karlson et al., 1981; Chance et al., 1975) spectroscopic techniques. The results from these studies indicated that the reaction of reduced cytochrome oxidase with dioxygen proceeds via at least three inter- mediates (termed compounds A, B, and C in the Chance et al., (1975) terminology; or compounds I, II, III in the Clore et al., (1980) terminology). Recent attempts to study these intermediates by resonance Raman sepectroscopy have been rePorted by Babcock et al. (1984 & 1985). By using flow- flash CO-photolysis and time-resolved resonance Raman spec— troscopy during the oxidation of reduced cytochrome oxidase by oxygen, Babcock et al. (1984 & 1985) were able to detect the initial occurrence of a photolabile oxy—intermediate as the precursor to oxygen reduction, with the subsequent formation of a more stable oxygen-intermediate at the cytochrome as /Cun site. Another important finding with respect to the existence of these enzyme intermediates has been the recent resolution of this "oxy" intermediate at room temperature by using CO-flash-photolysis accompanied by rapid-scanning spectrophotometry (Orii, 1984). Therefore, the available data accumulated on the oxidation products of reduced cytochrome oxidase by oxygen support the formation of an "oxy" intermediate during the first stages of the reaction, which is subsequently followed by fast inter- molecular electron transfer to dioxygen. Research on the proton pumping mechanism of cytochrome oxidase was initiated by Wikstrom (1977) and later by Sigel and Carafoli (1978). The proton pumping action of cytochrome oxidase has been investigated in whole mitochondria as well as in isolated cytochrome oxidase reconstituted into phos~ pholipid vesicles (Wikstrom and Saari, 1977). Recently, it was suggested that cytochrome oxidase depleted of subunit III did not show significant proton pumping activity even though the electron transfer capabilities were not signifi— cantly altered (Penttila, 1983). In addition, in order to maintain the obligatory coupling between electron transfer and proton translocation in cytochrome oxidase, two redox— linked proton pumping models have been suggested at the molecular level for cytochrome a (Babcock and Callahan, 1983) and CuA (Blair et al., 1986; Gelles et al., in press), res; and in. 11.1 198! 600| 9a. thr. by 1 sum se i 3111 0x11 03h thr 198 respectively. The mechanistic implications of both models and their relevance to the present study will be presented in Section III of this Chapter. II.Cytochrome Oxidase: Structure and Metal Active Sites. A. Subunit Composition and Protein Structure Cytochrome oxidase from bovine heart muscle is composed of more than 12 polypeptide subunits (Capaldi, 1982; Azzi, 1980; Buse et al., 1985) with molecular weights ranging from 6000 to 40,000 Da. It has a total molecular mass of ~200,000 Da. The subunit arrangement is shown in Figure I-l. The three largest subunits (subunits I, II, III) are coded for by mitochondrial DNA (Tzagoloff et al., 1979), while the smaller subunits are coded by nuclear DNA. Cytochrome oxida- se has been purified also from prokaryote sources with much simpler subunit composition. These range from an gas-type oxidase with two subunits (Yamanaka et al., 1979; Honami and Oshima, 1984; and Fukumori et al., 1985), to oxidase with three subunits (Sone and Yanagita, 1982; De Vrij et al., 1983), to the recent characterization of cytochrome oxidase from Thermus thermophilus with just one subunit (Yoshida et [al., 1984). As with bovine oxidase, these bacterial type oxidases have been reported to show proton pumping activity. Examples of these are Thermus thermophilus (Sone et al., 1984; Yoshiba and Fee, 1984), the Thermophilus bacterium P83 (Sone and Hinckle, 1982), and Paracoccus denitrificcans (Solioz et al., 1982). Subunits I and II are known to 1) 60A CytOSOI 40 A Membrane flSA hflatrix Figure 1.1: The geometrical arrangement of the polypep- tides (subunits) in cytochrome c oxidase in the mitochon- drial membrane. Subunits I and 11 contain the two pros~ thetic heme groups and the two copper atoms. The binding site for ferrocytochrome c is located on the outside of the membrane in close contact with the electron acceptor site, cytochrome a/Cu Adapted from Azzi (1980). A . 0011 cytl CUA exp: and dep. 10c: ret: Pen' oxic et : sub1 The pro a r. int, the Pro- lem‘ inn. 10 contain the four prosthetic groups, with cytochrome a, cytochrome g3, and Gus most likely located in subunit I and Cu; in subunit II (Winter et al., 1980). From labelling experiments with 510r2+, it has been suggested that subunit II contains the electron entry site from cytochrome 9 (Jones and Wilson, 1984). It has been recently suggested that the depleted subunit III oxidase does not exhibit proton trans- location even when 50% of the electron transfer activity is retained (Thelen et al., 1985; Nalecz et al., 1985; and Penttila, 1983). However, experiments with bacterial oxidase, with two subunits (I,II), from Paraccocus (Soliocz et al., 1982),'have made this interpretation of the role of subunit III on the oxidase proton pump somewhat uncertain. The general consensus about the role of subunit III in the proton pumping activity of cytochrome oxidase is that it has a regulatory function for the pump rather than being an integral part of the mechanism (Thompson et al., 1985). The structure of cytochrome oxidase has been studied by electron microscopy and image reconstitution of a two— dimensional crystalline array of the protein (Fuller et al., 1979 & 1982). Results from these studies revealed a three domain structure in the form of a Y shaped molecule, where the arms of the Y each span the membrane bilayer and protrude approximately 15-25 A0 on the matrix side of the membrane; the stalk of the Y extends from the mitochondrial inner membrane into the intermembrane space (see Figure 1.1) (Terrence et al., 1985; and references therein). The aggre- 01' 19 ea en hi 39 a1 1110 fo PEI st- St: pm 10' Otl cy‘ 101 Uh: ll gation state of the oxidase has been controversial, but it has been recently proposed from experiments with vesicle reconstituted oxidase that the enzyme from bovine heart 1982; and references therein). The monomeric form of cytochrome oxidase from rat liver (Thompson and Ferguson, 1983) and bovine (Nalecz et al., 1983; Suarez et al., 1984) materials have been also described. Monomers of bovine enzyme could be prepared by treatment of the oxidase with high detergent concentrations and alkaline pH, which appears to be correlated with the loss of protein subunits (Georgevich et al., 1983). Kinetics studies (Reinhard et al., 1985) of cytochrome oxidase activity in its dimeric and monomeric forms indicate biphasic kinetics for the dimeric form and monophasic behavior towards the oxidation of reduced cytochrome g2+ in the momeric form. Changes in ionic strength, enzyme concentration, and type of detergent used to suspend the oxidase appear to influence the association properties of the oxidase (Thompson and Ferguson—Miller, 1983; Suarez et al., 1984; Thompson et al., 1985). The dimer structure appears to be necessary for the enzyme proton pumping activity, whereas the monomer structure may only be active in electron transfer (Wikstrom et al., 1981). On the other hand, recent kinetics experiments on the monomeric cytochrome oxidase from shark have suggested that the monomer containing the four metals is the basic functioning unit for dioxygen reduction (Bickar et al., 1985; ar te ab ad ef 0h es 8m 91: st) on fine UM Georgevich et al., 1983). The determination of the heme a-copper (or heme-heme) distances within the isolated oxidase has been an active area of research for the last few years. Two physical techniques, electron paramagnetic resonance (EPR) and X—ray absorption spectroscopy (EXAFS), have been widely applied to address these questions. For instance, by studying the effect of cytochrome a on the EPR-spin relaxation properties of NO-bound ferrocytochrome as, a distance of 15-20 A0 between the irons of the two hemes was estimated (Boelens et al., 1984; Scholes et al., 1984; Brudvig et al., 1984; Ohnishi et al., 1981). Similarly, a distance of 8-12 A0 was estimated between CUA and cytochrome a, while between Cu; and cytochrome §3 this appears to be ”3—4 A0 (Stevens et al., 1982; Naqui et al., 1986). The interaction of cytochrome oxidase and its physiolo_ gical electron donor, reduced cytochrome 9, have been studied by fluorescence measurements (Dockter et al., 1978). Electron transfer is believed to occur through the heme edge of cytochrome c (Timkovich,1980; Britain and Mattews, 1986) to the electron acceptor pole of cytochrome oxidase, cytochrome g/CuA. The precise distance between the heme g and the oxidase electron accepting center is still uncertain. In addition to the mammalian mitochondrial and bacte- rial oxidases, cytochrome oxidase has been also studied in higher plant mitochondria (Denis, 1981; Dutch et al., 1986). 13 Initial results reported by Denis and Clore (1981) and Denis and Richaud (1982) suggested the possibility of differences in the heme aa-Cug environment in the plant system. Their results on the CO recombination studies in fully reduced cytochrome oxidase after low-temperature flash photolysis revealed a multiphasic rebinding of different conformers in the heme aa/Cun site (Denis and Richaud, 1982). Near infra- red studies on the components of plant mitochondria (Richaud and Denis, 1984) reveal a reduced minus oxidized difference spectrum assigned to Cun, whereas in the mammalian material no Cua signal is observed. Interestingly, recent EPR studies on this plant material (Denis et al., 1984) suggested that the antiferromagnetic coupling responsible for heme g3 and Gun being EPR silent in the resting state of bovine oxidase (Tweedle et al., 1978) was not apparently present in resting whole plant mitochondria. Recently, another plant oxidase, that from etiolated corn seedlings, was successfully isolated with high activity in the laboratory of Prof. Sheilagh Ferguson-Miller (Ingle, R. and Ferguson—Miller S., Michigan State University, unpu- blished results). The resonance Raman vibrational analysis of maize mitochondria and isolated corn protein was conduc— ted in our laboratory yielding surprising results on the cytochrome a heme structure (Dutch et al., 1986). The optical absorption spectra of this enzyme, as well as our RR analysis, indicate that the heme a environment of maize to cop oer. nit equ 0911 for sit CN- the mag Cau she "101* She as he l4 cytochrome a is altered as compared to the bovine heart enzyme, suggesting that this change is associated with the formyl substituent of the heme a porphyrin ring. B. Metal Centres: Structure and Protein Environment Cytochrome oxidase catalyzes the reduction of dioxygen to water by using two different metals, heme iron and copper, which are organized in three different redox centers: cytochrome a is low-spin and six-coordinate with nitrogen from histidine residues as its fifth and sixth ligands (Martin et al., 1985) and is in rapid redox equilibrium with the EPR visible CuA(S=1/2). The third redox center, thecytochrome aa/Cun binuclear site, is reduced by cytochrome a /CUA. This binuclear center is responsible for the binding and reduction of dioxygen and is the binding site for a number a small respiratory inhibitors such as CO, CN' , NO, HN3-, and HCOO'. Recently, it has been suggested that cytochrome oxidase preparations may contain zinc and magnesium in addition to Cu and Fe (Einarsdottir and Caughey, 1984 & 1985). The structure of the heme a porphyrin macrocycle is shown in Figure 1.2A, where it is compared with that of the more commonly occurring protoheme species. The protoheme species contains vinyl groups at positions 2 and 4, as well as propionic groups at positions 6 and 7, with the remaining Porphyrin positions occupied by methyl groups. In contrast, 15 .msm50pOLa Low umpmowu:_ mp mcaim553c use; m_OLLAQ c_Lx;aLoa .m mew: Low umpmoaecw m, mcprmnszc co_ . . .pvmoa ucmzpwpmnzm mew; chowpcm>coo we» .xH csizcaxoa . Opera Ucm m mec n+0 QLJHUJLHW LMZJUQFOE ass ”N.H mi:m_u oEmnoaoca 100w 100w NIH—v NIH—U ~10 ~10 mIU UnI role on: ”Io :0 Nrw \ \ 16 for heme a two major structural differences are present: the vinyl group at position-2 in protoheme has been replaced in heme a by a hydroxy—farnesyl group, and the methyl group at position 8 in protoheme has been oxidized to a formyl group in the heme a macrocycle. Owing to the more delocalized electronic distribution of the formyl and vinyl groups, and to the well known electron withdrawing capabilities of the 0:0 group (Vinogradov and Linell, 1972), it is expected that the optical absorption spectra of cytochrome oxidase will be significantly altered as compared to hemoglobin and myoglobin hemeproteins. Indeed, protoheme derivatives show cx—band maxima in the range from 545—565 nm while cytochrome oxidase heme a derivatives display a—bands which range from 588 nm to 598 nm. The electronic optical absorption spectra of oxidized and reduced cytochrome oxidase are presented in Figure 1.3 (Vanneste, 1967; Halaka, 1981). The uv- and visible regions of the spectrum are dominated by the heme chromophores of cytochrome a and cytochrome a3, and are dependent on changes on the redox and spin state of the heme iron. By assuming no optical interactions between cytochromes a and as and by using difference spectroscopy on ligated oxidized and reduced cytochrome as, Vanneste (1967) was able to separate the individual contributions of cytochrome a and cytochrome Q3 to the optical absorption spectrum of cytochrome oxidase. However, it was recognized by magnetic circular dichroism (MCD) (Palmer et al., 1976; Babcock et al., 1976) and l7 E, mM" cm" 0411411111111111:1_L 500 550 600 650 700 “-——-—-~ ~- ‘ llllllllLllJlllllilllJi 650 700 750 800 850 Wavelength, nm ON-bmm 1’47 Figure 1.3: Electronic optical absorption spectra of oxidized ( and reduced (----) cytochrome g_oxidase in the Soret, visible, and near-IR regions. Extinction coefficients are expressed per enzyme functional unit containing two hemes and two copper atoms. The broad band at 830 nm is attributed to CuA. Similarly, the stronger shoulder at ~ 662 nm is attributed to the Fea3-L-CuB binuclear center. Repro- duced from Halaka (1981). l8 resonance Raman (Babcock et al., 1981) experiments that strong redox interactions between cytochrome a and cytochrome as were present. These results came from anaerobic redox titrations followed by MCD measurements (Palmer et al., 1976; Babcock et al., 1976). Similarly, resonance Raman spectroscopy (Babcock et al., 1981) experiments showed that cytochrome a and cytochrome Q3 are indeed spectrally distinct in both the absence or presence of ligands, and confirmed Vanneste’s original work. -In addition to the heme a groups, the copper ions (mainly CUA) also contribute to the optical absorption spectrum of resting cytochrome oxidase. The near infrared region of beef heart cytochrome oxidase displays a broad and weak band centered at 830 nm (Aezz2 mM'1 cm'l) (see Figure 1.3), assigned to oxidized CUA (Wharton and Tzagoloff, 1964; Yong and King, 1972), which vanishes upon copper ion reduction. The contribution of Gun to the 830 nm band has been estimated to be “15% (Powers et al., 1979); however this falls within the range of experimental uncertainty (Beinert et al., 1980). The spectral properties of the four metal ions have been most widely studied by electron paramagnetic resonance (EPR) spectroscopy owing to the fact that both the heme groups and the copper atoms may be identified in the EPR spectra. In the case of oxidized resting cytochrome oxidase, the low-spin, six-coordinated cytochrome a contributions are observed with g-values at 3.03, 2.213, and 1.5. The CUA 19 contribution is seen as an intense signal with a g-value of ”2.0 (Hartzell and Beinert, 1976; Aasa et al., 1976). The EPR signal from CUA was found to be unique for cytochrome oxidase (Aasa et al., 1976; and references therein) in that it displays low g-values (g=2.18, 2.03, 1.99), no resolvable copper hyperfine structure, and unusual EPR-signal satura- tion properties (its EPR signal is easily saturated at 10K in contrast to other CuII complexes). The unusual Cur EPR properties were suggested to originate from a unique ligand coordination (Chan et a1, 1979). To account for these Cui spectral properties, Stevens et al. (1982) conducted elec- tron nuclear double resonance (ENDOR) experiments on 15N- isotopically substituted yeast cytochrome oxidase. Their findings reveal the presence of methylene protons of one or two cysteines and the presence of at least one nitrogeneous ligand from a histidine residue. Accordingly, CUA ligation With two cysteine-sulfur ligands and two histidine-nitrogen ligands was proposed (Stevens et al., 1982). A recent com- parison of the amino acid sequence of subunit II of cyto- chrome oxidase with subunit II of a series of blue copper proteins appears to indicate that CUA is structurally related to type I copper while Cu» is related to type 3 oxidases (Malmstrom, 1986). In the 0338 Of CUB: the ENDOR data reveal strong similarities with type-3 CuII as in the case of laccase, and a ligand coordination of three nitrogens was proposed (Cline et al., 1983). The binuclear center heme aa/Cun is EPR silent in 01 h. 20 the resting oxidized oxidase. This was initially interpreted by Van Gelder and Beinert (1969) as due to a strong anti— ferromagnetic coupling between the high spin heme as iron (S=5/2) and the Gun ion (8:1/2) resulting in a total 832 system. The magnetic interaction of this binuclear center is believed to be mediated by a bridging ligand (Van Gelder and Beinert, 1969). In the optical absorption spectrum of resting oxidized cytochrome oxidase, a relatively weak band, centered at 662 nm, has been attributed to the heme aa-L-Cua structure interaction (Beinert et al., 1976). For fully reduced cytochrome oxidase this band disappears, indicating the decoupling of the Feas—Cun interaction. Upon ligand binding to the cytochrome as site, the occurrence of the 662 nm band is observed for ligands that maintain the Feaa in a high-spin configuration (i.e, HCOO'), whereas for ligands inducing a low-spin transition, the 662 nm band vanishes (i.e., CN-). Partial reduction of cytochrome oxidase removes the antiferromagnetic coupling within the binuclear Feaa-Cun center rendering this site EPR detectable. This allows for the formation of the g=6 high—spin signal ascribed to ferric heme Q3 (probably with Cue in the reduced state). However, this signal is not observed during enzyme turnover (Wilson et al., 1982), which suggested that Cua might be optically and EPR silent in all redox states of cytochrome oxidase. To explain the EPR spectral properties of cytochrome E3/CUB in the resting enzyme, Seiter and Angelos (1980) and 21 latter Hagan (1982) suggested the possibility of a cytochrome aa/Cun pair in which the iron is in a high-spin ferryl (FeIV) state and Cua+ is in the reduced cuprous state. However, the accumulated data on magnetic suscepti- bility, MCD, RR and Mossbauer spectroscopies (Kent et al., 1982 & 1983; Babcock et al., 1979) strongly argue against this suggestion. All these data are consistent with the original pr0posal that cytochrome as is spin—coupled to the cupric copper ion as Feaa—L-Cua. Recently, X—ray absorption spectroscopy (Powers, 1982) was used to study metal atoms and their immediate environ- ment in biological membranes. The application of this technique to oxidized cytochrome oxidase was pioneered by Powers et al. (1981) and their results suggest that the iron of cytochrome a3 is coordinated at the proximal site by a histidine and at the distal site by a sulfur atom, this latter ligand being responsible for the antiferromagnetic interaction by bridging to One. (The interpretation offered by Powers et al. (1981) relies heavily on comparison with known model compounds and on assumptions that the CUB was comparable to the type-I "blue copper" (as in stellacyanin), and that Guy was responsible for the absorbance at 830 nm. TYPc-I blue—coppers are known for their strong optical electronic transitions on the 600 nm region, which upon reduction of CuII to CuI are abolished (Siiman et al., 1975). In reduced cytochrome oxidase, the largest contri~ bution (BO-100%) to the visible absorption d—band at 605 nm arises from cytochrome a2+ (Lemberg, 1969; Nicholls and Chanady, 1981). In addition, type-I coppers, such as in plastocyanin (Colman et al., 1978) and azurin (Adman et al., 1978), are known to be ligand coordinated by two histidine nitrogens, at least one cysteine, and one methionine sulfur atom, arranged in a near tetrahedral geometry. Resonance Raman studies in these blue—type—I COpper compounds indicate that copper-ligand stretches, ligand deformations, and metal—ligand deformations are strongly enhanced in the low- frequency ((550 cm'1 ) region of the spectrum (Woodruff et al., 1984; Musci et al., 1985; Bovil et al., 1986). However, resonance Raman spectra in the visible region of resting, oxidized plus cyanide, mixed—valence, and reduced cytochrome c oxidase (Bocian et al., 1979; Centeno and Babcock, unpublished results; see Chapter IV) fail to show any copper-ligand related modes in the low—frequency region. Furthermore, several authors have convincingly established that CuA2+ is the responsible metal for the 830 nm band (Greenwood et al., 1983). C. Chromophore Geometry and Ligand-Coordination The main geometrical disposition, proposed coordi- nation, and ligand binding parameters of cytochrome a /CuA and cytochrome aa/Cug are displayed in Figure 1.4 (Brunori and Wilson, 1982; Callahan, 1983). Cytochrome a, as previously mentioned, is always in its low—spin, six- coordinated configuration independent of the iron redox 23 Figure 1.4: Geometry and coordination properties for cyto- chrome a, cytochrome a3, CuA, and CUB in cytochrome g oxi- dase. Metal distances and Cu-ligand coordination geometries as suggested by Powers et al. (1981). Heme a Chromophore and CuA contributions to the optical absorption spectra of resting and reduced cytochrome g oxidase as estimated by Vanneste (1967) and Beinert et al. (1972), respectively. Electron paramagnetic resonance (EPR) and carbonyl vibra- tional parameters as investigated by Babcock and coworkers (1978, 1981, and 1983). The identity of the bridging 11- gand (L) within the Fea3-CuB binuclear center is unknown; however, a sulfur ligand has been proposed based on EXAFS studies (Powers et al., 1981). 445 rum ( rod.) _50 °/o Nx. u m mam oz Ana 5:3 3.960 pooconoxov E: wa\.\.oao 7...: z/ \ m: \a» o/ 35 z x. .7 m H. 4905v .0920 WWW— u N I/ \In no. mm I 3...... as. . 2.: zone a / 3.3 mam oz Ecmom ox. o. :0 E: n¢¢ oxoom .nm .30 IQ :3: ...5 so. :33 rec one. . 3...: \\ % N\_ n a o n a O l 2.2. s: new .\.omA .35 :— GUOLV EC mug O\O on W8 80 RP un 00‘ re 0)’ 1h: En! 25 state (i.e., ferric or ferrous Fe a). Its associated CuA atom appears to be ligated by two cysteine sulfurs ligands and two histidines nitrogens in the Cu2+ state (Stevens et al., 1982). However, the redox state and ligand-coordination geometry around the CUA in resting oxidase are matters of controversy since, on the basis of EXAFS (Powers et al., 1981) and MCD (Greenwood et al., 1983) studies, a Cu2+ state was preferred. On the other hand, results from the amino acid sequencing of subunit II and studies on the unusual CUA EPR properties (Steffens and Buse, 1983; Chan et al., 1979), propose a CuA* state. Cytochrome as in its resting oxidized state is six—coordinate, high-spin (8:5/2) and exchange coupled with its associated Cua partner through a bridging ligand. The identity of this bridging ligand is still unknown, although Powers et al. (1981) have suggested the ocurrence of a sulfur ligand based upon EXAFS studies on resting cytochrome c oxidase. In its reduced form cytochrome as is five—coordinate, high-spin (8:2) with histidine nitrogen as the proximal ligand and with the sixthéligand distal position available for further coordination. The carbonyl group at position—8 contributes to the optical absorption properties and vibrational spectrum of cytochrome a and cytochrome g3. Resonance Raman evidence from heme a model compounds has convincingly established that the formyl group of cytochrome a lies in a hydrophobic environment buried inside the protein surroundings (VanS cytoc hydro resid commu the p (0:0. stron ‘ ion, H-bon Calla oxida and M the a 88 an in th mecha i“die imide (Prod the 0: "won. with ‘ 26 (VanSteelandt-Frentrup et al., 1981). However, for cytochrome g the carbonyl group appears to be involved in a hydrogen-bonding interaction with a protein polypeptide residue which serves as a structural intermediary in communicating redox events originating at the heme iron to the protein surroundings. The strength of this formyl (0:0...protein) hydrogen-bond interaction was observed to strongly depends on the redox state of the cytochrome a iron ion, and accordingly, a proton pump model that relies on a H—bond geometrical change was proposed (Babcock and Callahan, 1983). III. Proton Pumping Despite some arguments about the function of the oxidase as a redox-linked proton pumping system (Mitchell and Moyle, 1983; Lorusso et al., 1979; Papa et al., 1983), the available data strongly support proton pumping activity as an integral part of this mitochondrial terminal enzyme. Subunit III has been reported to play an important role in the electron—transfer activity and proton pumping mechanism of oxidase from eukaryotic (mammalian) cells, as indicated by inhibitory experiments with dicyclohexylcarbo- imide (DCCD), whose binding site was located in subunit 111 (Prochaska et al., 1981). DCCD is a well known inhibitor of the cytochrome oxidase proton pumping activity (McGovern~ Moroney et al., 1984; Casey et al., 1979). In experiments with oxidase from eukaryotic cells depleted of subunit 111 gr su Th ox ba 80 Yo 0) tr 0)’ of mel DI" cy' 1‘61 ad: in as: ”he Rat 27 it was observed that the oxidase retains its electron transfer activity but was totally deprived of proton pumping activity (Penttila, 1983; Thelen et al., 1985; Suarez et al., 1984; Prochaska and Reynolds, 1986). However, other groups had reported a partial proton pumping activity from subunit III depleted bovine oxidase (Puttner et al., 1985). The suggestion of subunit III being an integral part of the oxidase proton pump is also contradicted by experiments with bacterial cytochrome oxidase, which apparently does not have subunit III but nevertheless displays full proton pumping activity (Solioz et al., 1982; Sone and Hinkle, 1982; Yoshiba and Fee, 1984). A simplified scheme of the proton pumping mechanism of cytochrome oxidase is depicted in Figure 1.5. Overall, the transfer of electrons from reduced cytochrome c to the cytochrome a/Cui metal active site is linked to the uptake of mitochondrial protons from the matrix side of the membrane and their subsequent release to the cytosol side, presumably triggered by conformational changes at the cytochrome a/CuA metal active sites. To complete the reduction of oxygen at the cytochrome ga/Cun binuclear site, additional protons from the interior of the protein are used in the subsequent formation of H20. Cytochrome a and its associated CuA have both been suggested to serve as the site where the oxidase redox link to the proton pump occurs (Wikstrom et al., 1981). This is supported by resonance Raman studies, EPR, and other spectroscopic approaches used O.“ 0 dbo J°m°h* MU 28 of . +1 35 z .osocm Locou-z :empocq a moHOCou m .mmecwxo o msoczoopxo :_ ape>wpom m:_asza ceposa noxce_uxouoc one we osocom umeme_aswm < ”m.H ensued +2 +1 1111\\ m.» +« .o xzscqs_ + 4050...)". to Cu Bl ha an pr Ra di he pr moi 0y. st: he] ad: ca: su; di: thl qui fit the su, 0n Of an: 29 to study the molecular protein environment of the heme g and Cu; sites in cytochrome oxidase and its derivatives (Callahan and Babcock, 1983; Babcock and Callahan, 1983; Blair et al., 1986; Gelles et al., in press). Two models have been proposed, one involving the heme a of cytochrome a and the other CUA, as individual molecular sites for the proton pumping mechanism. On the basis of resonance Raman spectroscopy, Babcock and coworkers had proposed a direct mechanism in which the peripheral formyl group of heme glis involved in a hydrogen bond interaction with a protein residue acting as the H—donor. In this molecular model, the hydrogen-bonded proton is pumped during the redox cycle, with the strength of the C=O...H(protein) H—bond strongly dependent upon the redox level of the Fea atom of heme a. Recently Copeland and Spiro (1986) presented additional evidence for the involvement of the cytochrome a carbonyl group in a H—bond interaction as initially suggested by Babcock and Callahan (1983). However, the direct participation of the cytochrome g carbonyl group in the oxidase redox-linked proton pumping mechanism was questioned by the above authors and recently by Argade et al. (1986). On the other hand, the molecular basis of the cytochrome oxidase proton pump had been recently suggested to occur at the CUA site rather that at the cytochrome a carbonyl site. This was proposed on the basis 0f extended x-ray absorption fine structure (EXAFS) studies and EPR data on the coordination and ligation properties of 01“ al. Che eve pro the vic pro met. use. cyt: doul Our inv: intl data inn IV. fOru pro) 0X16 3O CUA in cytochrome oxidase (Chan et al., 1979; Stevens et al., 1982; Blair et al., 1986; Gelles et al., in press). Chan and coworkers had suggested a model in which redox events at the CUA metal site could be coupled to the oxidase proton pump through a redox—linked ligand displacement at the Cur metal site with a protein H—donor in the immediate vicinity of the Cu; ligation sphere. Both proton pumping models predict the occurrence of protein exchangeable protons in the near vicinity of the metal center active sites. To test these two models I have used resonance Raman spectroscopy and H/D exchange at the cytochrome a...protein interaction site and electron nuclear double resonance (ENDOR) spectroscopy at the CUA metal site. Our results with resonance Raman spectroscopy support the involvement of the cytochrome a formyl group in an H-bonding interaction with an exchangeable protein H-donor; the ENDOR data reveal the existence of exchangeable protons in the CuA2+ protein vicinity, nevertheless, their identity and involvement in the oxidase Ht-pump remains unclear. IV. Concluding Remarks: Cytochrome c oxidase is a multi—metal enzyme responsi_ ble for the terminal reduction of molecular oxygen in the formation of H20 during respiration in eukaryotic and prokaryotic cells. The marked difference and complexity of oxidase from mammalian, bacteria, and plant materials is th re it OK ap ar av Cu an pr or tw en Do on 31 recognized by their differences in protein subunit composition. Thus bovine cytochrome oxidase contains 12 polypeptide subunits while bacteria cytochrome oxidase has been reported to contain one (Yoshida et al., 1983), two (Ludvig and Schatz, 1986), and three subunits (De Vrij et al., 1983). However, common features among these oxidases appear to be conserved, such as the following: (1) the enzymatic minimum functional unit contains two heme a iron-containing chromophores and two copper atoms; (2) all these oxidases catalyze the electron transfer reaction from reduced cytochrome g to molecular oxygen to produce H20; (3) it is widely accepted that electron transfer in cytochrome oxidase is linked to proton pumping activity. The magnetic properties of the heme g metal sites appear to be fairly well established. Cytochrome g and CUA are magnetically isolated and presumably separated by an average distance of >12 A0. Cytochrome g3 and its associated Cun atom are close to one another (<3—4 A0) and there is an antiferromagnetic interaction between the two. Considerable progress has been made in understanding the cytochrome a and cytochrome as ligand geometries. Cytochrome a is ligated by two histidines, and these remain bound throughout the enzymatic redox cycle, while cytochrome Q3 is known to possess one histidine ligand on the proximal side but lacks a ligand on the distal side in the reduced state. Research on the ligand and coordination geometry around the metal CE I‘E at pr Cl ch ox 00 ab 3P ch 0? he sy: sp. Vil Se] id: th( in W the 32 centers is still a matter of controversy, especially with respect to the ligand coordination around the CUA and Cu; atoms. The molecular basis of the oxidase redox—link to the proton pump is suggested to occur at the level of the cytochrome g /CUA metal active sites. In the following chapters, research on the spectrOSCOpy of cytochrome oxidase, its ligand bound derivatives, and heme a model compounds is described. The H20/D20 exchangeabilities of the g/Cui sites has also been studied. I have used optical absorption, resonance Raman spectroscopy, EPR and ENDOR spectroscopies in order to investigate the molecular characteristics and hence functional involvement of cytochrome a as well as CUA in the oxidase proton pumping mechanism. V. Aims and Strategy of This Thesis: The aims of this thesis are two-fold: (1) to develop systematic vibrational analysis of the heme a (N—MeIm); spectra with the goal of elucidating formyl related vibrational and deformation modes, and to establish their sensitivity to in vitro H—bonding effects; (2) identification of cytochrome a H-bonding sensitive modes and the structural relationship between the cytochrome a formyl group and the redox—linked proton pumping activity of cytochrome oxidase. To obtain the neccessary information on the heme a model compounds, results on the isotopic re fo gr CY mi CY Th in re be tr re SP 33 replacement at the ~HC=O group (i.e., deuteration of the formyl proton and substitution of 18O for 16O at the formyl group of heme a) will be discussed. Identification of cytochrome a H-bonding sensitive modes will be studied by mild denaturation of cytochrome oxidase (disruption of the cytochrome a hydrogen bond) and by H/D—exchange experiments. The redox dependent incorporation of D-atoms at the protein interacting site will be studied under turnover conditions. The ligand—structure around the copper atom and its relationship with the oxidase proton pumping activity will be studied by using electron nuclear double resonance spec- troscopy with the specific aims of investigating pathways of redox-linked proton pumping through the Gui—coordination sphere. Cl ha dr th to id (110 W8: he: ti: W thl Dr. 1111 at at as mil CHAPTER II MATERIALS AND METHODS I. Introduction: In order to obtain information on the structure of the cytochrome a in vivo active site in cytochrome c oxidase, we have conducted resonance Raman experiments on submitochon- dria particles, detergent-solubilized cytochrome oxidase, ligated cytochrome oxidase, and isolated, as well as syn— thetic, heme a model compounds. Experiments with intact mi— tochondria membranes were conducted with the specific aim of identifying structural effects from the cytochrome gas moiety in the mitochondria surrounded protein and comparing these to the spectra obtained for purified oxidase; in this way possible contributions from other mitochondrial compo- nents that might arise from heterogeneities in the prepara- tion can be eliminated. Our studies with isolated heme a model compounds are expected to improve our understanding of the structure of the in vivo active site of cytochrome a, to predict molecular influences that might affect heme-protein interactions, and to determine structural changes occurring at the heme a periphery that might be linked to redox events at the heme metal ion. To achieve our goal, we were involved in the preparation of a series of H-bonded heme a compounds as well as the preparation of isotopically substituted for- mYl compounds. This Chapter describes in detail these pre— 34 ho wi ex of Mo st EX de 0e. (11' We: (lei. 35 parative procedures and the experimental physical techniques used in the study. II. Materials: A. Isolation of Cytochrome c Oxidase Cytochrome Q oxidase from beef heart mitochondria was prepared according to the procedure of Hartzell and Beinert (1974), as modified by Babcock et al. (1976). The final precipitate was dissolved in a minimum volume of 50 mM HEPES (pH 7.4) buffer containing 0.5 % lauryl maltoside. The re- suspended oxidase was dialyzed against buffer of 50 mM HEPES- (pH 7.4), 0.5 % lauryl maltoside and 0.1 mM EDTA for three hours, and for two more hours against the same buffer but without EDTA. This dialysis results on the removal of excess (NH2)2804 and cholate. Enzyme prepared by the method of Yonetani (1960) was a generous gift from Dr. Patricia McGovern—Moroney. Both oxidase preparations were frozen and stored in liquid N2 until use. The enzyme concentration, expressed in terms of two molecules of heme a/enzyme, was determined from the optical absorption spectrum as the redu— ced minus oxidized difference spectrum ( A8 = €605 - €598 = 27 mM"1 cm‘l) (Van Gelder, 1966; Babcock et al., 1976). Mitochondria membranes used for resonance Raman experiments were resuspended in a buffer containing 0.25 mM sucrose, 10 mM phosphate (pH 7.4), and the concentration was determined at 605 nm using the reduced minus oxidized Is he he SD ab Va fiz 0y 00: n8; 36 difference spectrum with an extinction coefficient of 11.5 ml‘i'lcm"1 (Baker et al., 1987). B, Preparation of Cytochrome c Oxidase Derivatives Resting or native cytochrome oxidase is the enzyme as isolated. All of its redox metal centers exist in their oxidized, high—valence state (i.e.; aft/Cue2+ §a3*/Cua2*). The Soret maximum of resting oxidase is usually observed in the range from 420 to 424 nm (the position of the Soret maximum varies from preparation to preparation); however, from the photoinduced deconvoluted spectra of Vanneste (1967), it has been shown that cytochrome a3+ absorbs at 427 nm while cytochrome a33+ absorbs at 415 nm. The overall catalytic activity of isolated cytochrome oxidase appears to depend upon protein purification procedures (Suarez et al., 1984) and it has been shown that, for oxidized cytochrome oxidase, three different conformations can be detected (Brudvig et al., 1981). To simplify this apparent oxidase heterogeneity, we have made use of ligand-binding experi- ments with the specific aims of deconvoluting the different spectral contributions of each chromophore to the optical absorption (and resonance Raman spectra) of oxidized, mixed~ valence, and reduced cytochrome oxidase. We used cyanide, azide, and formate as ligands to cytochrome oxidase. While cyanide and azide convert the heme of cytochrome Q3 into a low-spin iron, formate maintains the 93 in a high-spin configuration. To form the cyanide and azide complexes, a near neutralized solution of NaCN (pH 8.0) and NaNa (pH 6.2) SI lo in sh co ca sh an sh an di 0X III] In Wi CY 80' [an 0x: in 0f acj 37 was added to the resting oxidized protein and the binding was followed by the characteristic Optical absorption spectral changes. The formation of the a3+§33*-CN was fol- lowed by increased intensity of the 428 nm Soret band, increased intensity of the 545 nm visible band, slight blue— shift of the cx-band from 599 nm to 597.5 nm, and the con- comitant disappearance of the 662 nm near-IR band. In the case of the a3+a33+—HN3- complex, while the Soret band is shifted to 428 nm, the cx—band is now red-shifted to 601 nm and the 662 nm absorption retains its intensity and red- shifts to 672 nm. In the fully oxidized enzyme plus formate and the mixed-Valence with formate, cytochrome aa3i-HCOO‘ display absorptions at 415 nm and 662 nm. The cytochrome eat and a2* transitions in these two formate—inhibited oxidase forms are observed at 427 nm and 600 nm and at 441 nm and 603 nm, respectively. C. Alkaline and Acidic pH—Modification of Purified Cytochrome g Oxidase pH Modification studies on oxidized (non-inhibited), mixed-valence, and reduced cytochrome oxidase were conducted with the goal of analyzing the vibrational spectrum of CYtochrome a under conditions in which the H-bonding inter- action has been disrupted. Modes sensitive to the cytochrome a formyl H-bonded to the protein H—donor in intact native oxidase (Callahan and Babcock, 1983) are expected to display intensity and/or vibrational frequency changes upon removal of the protein H-bonded proton. To study the enzyme under acidic conditions we used a buffer containing 50 mM MES fo no pl cu sa th he fo 30. 0X. WI) llei 00) 38 (2-N-morpholinoethanesulfonic acid). To achieve the desired pH level we used 0.1 N NaOH and 0.1 N HCl. The formation of precipitates (due to protein denaturation) at the extreme pH values (pH 4.0 and 11.5) was somewhat more problematic in the acidic treated sample. The precipitate formed was quickly removed by gentle filtration using 0.45 pM(Millipore Inc.) filters. The pH induced effects in oxidized, mixed— valence, and reduced protein derivatives were immediately established; no incubation time was required. The progress of the pH-modification effects in cytochrome oxidase was followed by the resonance Raman difference technique. The non—pH treated and the pH-modified protein samples were placed in separated compartments of a divided Raman spinning cuvette. The resonance Raman spectra of these two oxidase samples were simultaneously recorded and directly compared; the relative uncertainty in peak positions was calculated to be no greater than 0.4 cm'l. I).Redox-Cycled Cytochrome Oxidase Redox-activated cytochrome oxidase was prepared by the following method: concentrated cytochrome oxidase was redis- solved in 3—5 mls of deuterated or protonated buffer (final oxidase concentration ”80 uM), the solution was aerated for 15-20 minutes, and then reductant ascorbate (4.8 to 10 mM) /TMPD (20 uM) was added in the presence of excess dioxygen with constant stirring. After 45-60 mins of reaction, neutralized cyanide (pH(pD) 8.0) was added to a final concentration of 10 mM, and the resonance Raman spectrum of 39 the aaiga3i-CN‘ complex was immediately obtained. Redox— cycled oxidase used for the ENDOR experiments was similarly treated, with the following exceptions: the oxidase concen- tration was 0.6 to 0.8 mM; after the reduction and reoxida- tion cycles the reoxidized protein was reprecipitated by increasing the (NH4)2SO4 concentration from 0 to 43% and no cyanide was added to the redox—activated protein. We also studied the addition of cyanide to a resting (non-redox) activated protein. For this sample, cytochrome oxidase (80 u M) was added to a D20/ or H20 buffer already containing cyanide (10 mM), and the resulting sample was incubated for 45 to 60 mins. E. Preparation of Heme a Derivatives The chloride complex of ferric heme a (high—spin, pen- tacoordinated) was isolated by acid/acetone extraction from purified cytochrome oxidase (Babcock et al., 1979). Low— spin, six-coordinate and high—spin, six- and fivefcoordinate heme a model compounds were prepared according to Van Steelandt-Frentrup et al. (1981) and Callahan and Babcock (1983). The high—spin, six-coordinate ferric heme a (DMSO)2 complex was produced by small additions of DMSO-ds to heme a3+ Cl' and monitored by the optical absorption spectrum. To produce the low-spin, hexacoordinate derivative, 0.6 M N- methylimidazole was added to the freshly prepared heme a3+ Cl“ solution. Incorporation of deuterium at the heme a formyl proton 4O (i.e., —CHO -—> CDO) was conducted as reported by Chancellor et al. (1978). The purified heme a3t Cl" was dissolved in 4.0 to 6.0 mls of a buffer consisting of 10 mM phosphate (pD 11.4), 0.07 M cetryltetramethyl ammonium bromide (CTAB), 0.001 M EDTA, and 0.6N N-methylimidazole in D20. Solid KCN (in a molar ratio of 3:1 KCN/heme a) was added to the above heme solution and incubated (under constant stirring) at 4 0C. To avoid precipitation in the presence of cyanide, CTAB was used instead of other surfactants such as Brij-35 or sodium dodecyl sulfate (SDS). To monitor the progrees of the exchange reaction, small aliquots (~25 ults) were periodically taken, dissolved in the same buffer/detergent system, reduced with minimum amount of Nazs§04 and the RR was obtained. For a fully exchanged sample, the RR dis- played a slight shift in the high-frequency region from 1633 cm-1 to 1629 cm"1 and the appearance of a new band at 1083 cm‘l. After the exchange was completed, the bis-imidazole heme §3+(-CDO) was washed several times with a slightly acidic solution of 10 mM succinate (pD 5.8—6.0) in D20 and extrated in CH2Clz. Reduction of heme §3+(N-MeIm)2 in aprotic solvents was carried out by using the "Freeze-Pump—Thaw" (FPT) technique to remove oxygen, and a methanolic solution of 2,2,2— Cryptand solubilized sodium dithionite as the reducing agent in a molar ratio of 2:1 (cryptand/dithionite) (Minoey and TraYlor, 1978; Van Steelandt—Frentrup et al., 1981). The glassware used for these experiments consisted of a two—side ar cu di 53 sh re by 56 80 Ni st re Cu od ba ro f0 ut ne St Su‘ 0h: 41 arm apparatus with an attached optical (1 cm pathlentgh) cuvette. Reduced heme §2*(N-MeIm)z in aprotic solvents displays optical transitions at 435 nm (Soret), 510 nm, 532 nm, and 588 nm. Special care in the reduction of heme a should be exercised, since it is known that excess reductant results in the saturation of the peripheral aldehyde group, leaving a monovinyl heme macrocycle (Vanderkoii and Stotz, 1965). Chemical saturation of the formyl group was detected by a new emerging band in the optical absorption spectrum at 565 nm independent of the presence of aprotic or protic solvents. Metal removal (i.e., iron) and insertion (i.e., CuT*, Niz*) at the free base of porphyrin a was conducted by standard procedures (Fuhrop and Smith, 1975). The iron was removed by HCl/acetic acid treatment of heme a3* 01- and the Cu2* and Ni2+ metal ions were inserted by the acetate meth- od. The purity and separation of the metalated and free- base porphyrin of all the compounds used in this work was routinaly carried out by thin layer chromatography (TLC), followed by Sephadex G-10 column chromatography. The utilization of Cu2+ and Niz*-substituted porphyrin a was neccessary in order to study the formyl C=O H-bonded structure under strong H—bonding conditions. These metal substituted derivatives of heme a are known to mimic the charateristic vibrational properties of ferric heme a3+ st is we th pr f0 mi 9X D0 mi- H2: am. mi: eql So: Dis 38 42 (N-MeIm)2 (Callahan and Babcock, 1983), and in contrast-to heme a, they appear to be unaffected by the oxidizing potential of phenol-OH compounds. Synthetic porphyrin model compounds that mimic the trans- structural disposition of the heme a vinyl-formyl groups were also studied. The free base of 2,6-dipentyl—4- vinyl—8-formyl porphyrin was a generous gift from Prof. C.K. Chang’s research group (Michigan State University). The structure of this porphyrin is shown in Figure 2.1, where it is compared with the porphyrin a structure. Cu2+ and Ni2+ were inserted by standard procedures (Fuhrop and Smith, 1975). Replacement of 16O for 180 at the formyl group of this model compound was conducted by hydrolysis of the protonated Schiff’s base porphyrin derivative in the presence of 18O-labelled water (H2180, 98% 18O) by the following procedure. The Schiff’s base was prepared by mixing ”20-30 mg of Cu2* (-CH160) porphyrin with a four—fold excess of N-butylamine (Ward et al., 1983). The Cu3* porphyrin/amine solution was allowed to react for ~45—60 mins under an inert (Ar) atmosphere. Since any presence of H20 is expected to establish an equilibrium between the amount of 0:0 available and the -C=N formed, an azeotropic mixture of benzene/CH2C12 was used, thus shifting the equilibrium towards the formation of the Schiff base. The SOlution was refluxed for 6-10 hours resulting in the com- plete formation of the Cu2* (~C=N) Schiff’s base porphyrin, as evidenced by the absorption spectrum shown in Figure 2.2, T JQCTW N- h" MHWCOfi—LWQM 0+. $100 Ummm 00103k1¢3 1W «Hy WJQ wfim UCTUwFVWmS WDWdon «My- 43 m amcsgw m.H.u mnwcnncxmm om wwmm comm oo1o3k1a:.m AHV mad dam ooxurkwa: mamdoo Amy. a] on no re (0 st ph 1e Th in 0P p0 de 0&1 e1. “16' W1. [110! to 44 with Soret maximum at 408 nm and (x-band at 586 nm (Ward et al., 1983). Addition of HCl vapors to this solution resulted on the spectrum of the protonated Schiff base (-C=N+H) as noted by the split Soret (379 nm and 435 nm) and the large red-shift on the a—band to 632 nm. To reconstitute the formyl group with labelled— 180, the Schiff base (dissolved in highly dry CHzclz) was placed in H2190 under an Ar stream for 12-14 hours. After this, the aqueous and organic phases were removed by vacuum pumping at reduced pressure, leaving behind the reconstituted Cu2*(-CH150)-porphyrin. This material was redissolved in dry CH2C12, distributed into small vials, and stored under Ar until use. The optical absorption of the Cu2* (-CH160) and Cu2+ (-CH180)- porphyrins were identical (see Figure 2.2). III.Methods: To study cytochrome oxidase and its heme a porphyrin derivatives I have used electronic optical absorption, reso— nance Raman (RR), infrared (IR), and, to a lesser extent, electron nuclear double resonance spectroscopy. In the fol- lowing sections a brief description of the basic theory of metalloporphyrin absorption and resonance Raman spectrosc0py Will be presented, with the aim of reviewing some of the most important and relevant concepts used in both techniques to interpret hemeprotein spectrao 45 Figure 2.2: Absorption spectra of formyl-substituted deriv- atives of 2,6-dipentyi-4-vinyl-8-formyi Cu2+ porphyrin. Soi- vent is dry CH2C12, except in protonated Schiff’s base spec- trum which contains HC] vapors. The samples are: A) Cu2+ porphyrin with —CH16O; B) Cu2+ Schiff base (————) and pro- tonated Schiff‘s base (----); and C) Cu2+ porphyrin recons- tituted with 18 O at the formyl group. 46 1 rise -CH‘°o I 600 1 500 INPV l 400 mozUmoxUfiao: mumnawo om wsmm.oomm oowoskwa: m m+-mccmaaflcamq oowuykwa:_m A----v. mom_m>Fpooamwc wpwaom oefio wwwpoocmch .mc:m_t ecu co mowm new: agar; men so :socm mp mach m o nuxc esp m>wm op A.H.ov :owpoocmpcw cospoczmwwcoo or> “cosmos? m z . A w m cm 3H m_wpwwto ruwxcameww ov.moompm oopwoxo-ooxwe one .ocsmac ago to coocoo o:o.oo ooooopo:F ohm . 5H accomam covaLOmoo cacazacoooPFopws cos Eocmo_o zmcmcm ue.m.mcympd zaa>zaaoa mi A53: emNaca_oa->.x oesoco o Anemo-aw A vm scpoeesm-sm ooNlco_oa-N // xx 7% NM xllpma i: U llmfim /. II/ A vam F g: \\ . \\ A: x \ Tm Ilw N v 111 mH no cot<\x\ no a 6 pwmwmnlluwllllllll chrou all t canoe absor Perki cuvet 52 To interpret the optical absorption spectra of cyto— chrome oxidase and its heme a model compounds, we have taken all the above factors into consideration and their signifi- cance will be presented in the pertinent chapters. The absorption spectra reported here were obtained by using a Perkin-Elmer Lambda 5 UV/visible spectrophotometer. Optical cuvettes of 1 cm pathlength were routinally used. B,Resonance Raman Spectroscopy Raman spectroscopy is a powerful tool to study struc- tural perturbations of molecules by following their charac— teristic molecular vibrations. The fundamental principle of the Raman process relies on the Raman effect (Raman, C.VJ, 1928; Krishnan and Shankar, 1981), which is a phenomenon describing the inelastic scattering of light by molecules; that is, the scattering process will result in a gain or loss of energy by the scattered molecules. A simplified diagram describing the energy of the Raman process is depicted in Figure 2.5. Notice that the normal (or spontaneous) Raman process can occur with two types of transitions; one at lower energy than the incident energy at vo - Av(called Stokes transitions), or a transition to state of higher energy, v0 + Av(called anti—Stokes transitions). When the incident and scattered frequencies are equal, the scattering process is known as Rayleigh scattering (Av :0). In Figure 2.5 the resonance Raman effect is also illustrated as a scattering phenomenon in which the incident laser radiation is of sufficient energy to bridge the gap between Rama-n In tens i Cy -- - - Virtual 1[ Level ,./ hVo : ——( Stokes transition \)0 -A\) Jill I l _J "i”- l" _..._.... ‘/~//’hvas ’A//,hv /\/ h‘Jo \‘J hVo I—JI—D —I(—— — — :i __2 _3_2 __ 2 0 -" 0 h-----0 anti-Stokes Rayleigh Resonance transition Scattering Raman 00 + Av Av = 0 Effect Rayleigh Line l0-——— Stokes Raman Lines —’ G—Anti-Stokes Raman Lines—-I( Raman Intensity > T L. A A —1 V0 Figure 2.5: Diagram illustrating Raman and Rayleigh scattering, and the resonance Raman effect (Fig. 5.A). anti-Stokes Raman intensity is shown in 5.8. implies virtual (or intermediate) levels. to the continuum state. An schematic of the Stokes and The dashed line in 5.A The darker zone corresponds The ground and excited electronic states are labelled G and F, respectively. th st br de tr ti Ra ea of th Bo Cl tr on Sqi fo] ”he 30! f0] St; inc D0] 54 the virtual (or intermediate) states and the continuum states denoted by the darker zone. The electronic and vi- brational nature of the resonance Raman scattering will be described later in this section. Returning to the definition of Stokes and anti-Stokes transitions in spontaneous Raman, even when these transi- tions are equally separated from the Rayleigh line, their Raman intensities may not be equal owing to the fact that each observed transition is proportional to the population of the energy level from which the transition originates; the population at each energy level is given by the Boltzmann distribution function (Rousseau et al., 1981; Clark and Stewart, 1979). This is diagramatically illus— trated in Figure 2.5B. The intensity of the Raman scattering also depends on the fourth power of the scattering frequency and on the square of the polarizability tensor, as given by the following expression: 4 I = 8TTUJS IL S 9C4 g0 2 (2.1) H 0‘00 )GFl where fi-is the intensity of the incident laser, L% is the scattering frequency, and, d is the polarizability tensor for a transition from the ground IG> to the final excited state | F). The p and 0 describe the direction (in x,y, and z Cartesian coordinates) of the incident and scattered polarizations (Rousseau et al., 1981; Clark and Stewart, 1979). In Raman spectrOSCopy we are primarily interested on the 83: whe and men he ene ele am 2.2 den dif ( Ei due The D11: the din Ram: the 3601 Stu 3De< 55 the form of the polarizability tensor, which is described as: (spa) z % gf<9nfqulif><1fl Rglgn> (2.2) 1, El gn _ _ . ( E EL 1PM) where Ign) and (if) are the wavefunctions for the ground and excited state transitions, R is the electric dipole mo- ment operator, (Eiv — RE") is the difference in energy be— tween the ground state and excited state, EL is the laser’s energy, and Fif is the halfwidth at half-maximum of the electronic excited state transition and hence it represents a measure of the excited state ( (if>) lifetime (Spiro, 1983; Clark and Stewart, 1979). The denominator in equation 2.2 is known as the resonance Raman term, since it is evi- dent that as the energy of the laser (EL) approaches the difference in energy between the ground and final state (Eiv - E6"), the denominator will approach zero; however, due to the damping factor F the term does not vanish. Therefore, resonance Raman spectroscopy is simply accom- plished by tuning the laser frequency to match the energy of the electronic transition of the chromophore of interest. A direct consequence of the process is the dependence of the Raman intensity on the transition electric dipole moment, or to a first approximation, on the extinction coefficient of the transition of interest. As previously mentioned in the section of absorption spectroscopy, heme proteins have strong n -——-> mt electronic bands in their absorption - spectra, and hence they appear as one of the most suitable sy: re: ti< the an til Frz mec wil ope obi 1m; prc 1a: Cal tec unc Con 9t 56 systems to be studied by resonance Raman spectroscopy. Since resonance Raman is specific for the heme chromophore vibra- tions (Spiro, 1983), other modes, such as those arising from the protein amino acids can be nearly eliminated. Other aspects of resonance Raman spectroscopy are: 1) the Opera— tive mechanisms responsible for mode enhancement such as Franck-Condon and Herzberg—Teller (vibronic coupling) mechanisms; the former is responsible for modes observed with Soret (B-state) excitation, while the latter is the operative mechanism in the RR spectra of heme proteins obtained with Q-band excitation (Spiro, 1983); 2) a second important aspect in the resonance Raman spectra of heme proteins is the polarization properties of the heme molecu— lar vibrations (Rousseau et al., 1981; Ondrias, 1980; Callahan, 1983). 1) Resonance Raman Difference Spectroscopy Resonance Raman difference spectroscopy (RDS) is a technique which allows for the simultaneous detection of Raman spectra from two different samples, reducing the uncertainty in peak position from 2-3 cm'1 obtained with conventional spectrometers to as low as 0.1 cm-1 (Shelnutt et al., 1981). The technique was originally reported by Keifer and coworkers (Keifer, 1973; Keifer et al., 1975), and modified latter by Rousseau (1981) and Shelnutt et a1. (1981). A block diagram of our RDS setup is depicted in Figure 2.6, and in Figure 2.7 I outline the RDS hardware components. A custom made spinning cuvette (Presicion Cells and”- adUfi-un-fi— .H emf-9.30:0— .mspeceeam oocmcoetwo ensem mocmcomec one to smcmm_o xoo_m no.m mesmed a Lmeczoo xuofi mac. mmipo: qugvumEovdo m5. .omm 80....» .955 mucmgmwmm fl H. Lofovom hzm .umfi T“ meow Loeosocsuocoz .d:¢ mison all/ML 83 xmom r m m 8.8 j N Lmrczou iii—lBuImI L it... L l p .7 + v. pals 3:2! m Ti 2.... l Lo. Lmrsdcou I_W+ _m .2557 m (T $218.. L Qua zom 3U$QD w .mccosor a TL 58 A. MATL.: BLACK ACRYLIC (LUCITE) $QEA§¥§EI TUBE DELRIN BOX ' . INTERLOCKING ll CLAMP .. T W wINoow —~ .__ L:"E§====:P SLOTTED/ l. x COVER — SHELF 4— SUPPORT wALL ADJUSTABLE SENSOR wINoow DISC l - DISC 2 0130 2 SLOT IN DISC | LASER WINDOW SLOT IN GEAR ,‘ CENTER OF ,,/// CLAMP SCREW - 75,. y; GEAR MOUNT DISC l ‘ ’ To DISC 2 SLOT IN DISC I SUPPORT PIECE WITH BEARING \‘fl— I/8” ,_______ NOTE: SHOWN I g 1? 1: I IN CLOSED ' I DISC 2 I l C.C. DIST. TD MDTDR RAMAN DIFFERENCE Flgur‘e 2.7: Schematic of the Raman difference hardware (A) and Split spinning cuvette (B). 59 Material *Quortz Top View: CHAMBER 1 CHAMBER 2 1 2 High-Vacuum uground shape " 3 Side Views= 5”le ti...... 1 |+———OD=37mm 4 UV QUARTZ RAMAN DIFFERENCE CELL Figure 2.78 (Continued): Inc. aral ear: the meni elec elec puis part com; The Whic tube posi Spin set the Sync Prov PeSp 11/2 Subt obta ted from “Hen 60 Inc.) which is divided down the center to provide two sep— arate compartments is used. Figure 2.8 describes the neces- sary digital and timing logic used in the implementation of the RDS. As the cuvette spins (at ~32 RPS) the two compart- ments are alternatively illuminated by the laser beam and electronically detected as the cuvette rotates. Two opto- electronic modules on the motor mount are used to generate pulses which are 1800 apart, and synchronized with the partitionin the cuvette. The pulses are shaped in two comparators which subsequently trigger a Flip—Flop (F-F). The complementary outputs from the F-F control two gates which are fed by the count pulses from the photomultiplier tube (PMT) and the shaped optoelectronic pulses. The position of the adjustable windows, which rotate with the spinning cuvette to provide the optoelectronic pulses, is set so that counts associated with samples in each half of the cuvette are separated and detected separately._ A Synchronizing pulse is taken from one of the shapers to provide a reference. The signals are accumulated in their respective counters and subsequently stored in the DEC LSI- 11/2 computer. The two separate spectra can be plotted, subtracted, or ratioed (i.e., channel 1/ channel 2) to obtain accurate peak positions. In the RDS spectra repor- ted here, the monochromator was advanced in steps ranging from 0.1 to 0.50 cm-1 allowing a better determination of the magnitude of the shift in the difference spectrum, the dwell time was varied from 0.25 to 0.5 seC‘l, and the spectral 6 1 z z o —c + 5V o 0 Ir — 4 l6 2 N o I 3 I2 9 2 8 COUNTZ OUT \ '2 COUNT 2 24° 2 3 I OUT I ‘ a 9 l0 PMT m _\ ,3 coum " ‘63“ ’ > %: IN . '11" 3 1 ON' JL J 61 12 u COUNTI OUT \ COUNT 1 5 I3 OUT V (WT) §(°°‘,’”T) I7 I i I ‘ 240 I . II +5V SYNC OUT > IO SYNC RED ‘ ‘ Q; OUT IZOV {—9 Ac eon >-—/ c BLACK BOOINE AC (——> MOTOR BLUE me 24 c = MODEL 960 c 1.3 u 220 v 78L05 BOTZTOM . lRCUITS muse £33 LM 393 1 2 3 74 00 our GNDIN 74Lsn3 RAMAN SPINNING SAMPLE LOGIC l l I l I l l | I l l I II 1-1 “L: L: 1-7 U l [—— 2‘3 _J‘___-I_____J-——-—1___. 2'5 '1____J__——_L____I__- coumm nnnnn nnnn nnnn counnour nun nnnn COUNT2 OUT TIMI!) SYNC OUT U Li'— RAMAN SPINNING SAMPLE TIMING Figure 2.8: Schematic of the digital and timing logic for the resonance Raman difference Apparatus width on th latte width tiona param RDS e can b resul Sheln of th diffe I1980 simpl frequ Peak. glVen Where Rama“ Sity. LOPEn tiVel lines I0 th 62 width was chosen at 3.0 cm“1 as determined by the broadening on the high frequency Raman lines. The setting of this latter parameter is of particular relevance since the half width at half height ( F ) of the Raman line is propor- tional to the spectrometer resolution and is one of the parameters used to calculate the frequency difference in the RDS experiment. With the RDS technique, quantitative frequency shifts can be readily calculated from the derivative shape of the resulting difference spectrum (Rousseau, 1981; Keifer, 1980; Shelnutt et al., 1981). In this regard, detailed analysis of the determination of frequency differences by the Raman difference technique has been published by Laane and Keifer (1980), Rousseau (1981), and Shelnutt et al. (1981). For simplicity, we reproduce here the relationship between the frequency separation, Av , of the Raman lines, and the peak-to—peak intensity, 18 , in the difference spectrum, given as: 10 Av : __Q__ . (2.3) Y ID where F is defined as the halfwidth at half height of the Raman line under question, and ID is the Raman line inten- sity. The derivation of the above expression assumes a Lorentzian line shape for the Raman lines as well as rela— tively small frequency differences between the two Raman lines of interest. If Av is small, higher terms contributing to the difference intensity can be neglected (Rousseau, 1981) smalle that i the al tivel: accure peaks detail of err freque recent We hai n0 art ence 5 buffer exPeri Signal Contrc to 0,5 0.4 or throng A Study in Fig Chrome asSOci diffeI 63 1981).. In addition, when the frequency difference is much smaller than the halfwidth at half height of the Raman line, that is, if Av / I <<< 1, then it is more accurate to use the above intensity expression to calculate Av . Alterna— tively, if Av /F is greater than unity then it is more accurate to use the frequency separation between the two peaks to obtain a measure of the frequency difference. A detailed analysis of mechanical factors and various sources of errors that could influence the determination of small frequency differences in the RDS experiments has been recently presented by Rousseau (1981). In the present study we have taken all the neccessary precautions to ensure that no artifactual differences are contributing to the differ- ence spectrum. As a routine test, the RDS of solvents and buffers used here were scanned before and after each Raman experiment to monitor spectrometer resettability. When signal averaging was neccessary, special care was taken in controlling the spectrometer step size (usually between 0.1 to 0.50 cm-l) and the start of each scan to be within 0.1 to 0.4 cm'l. The laser output power was kept constant at 10 mW throughout the recording of the RDS. As an example of the application of this technique to Study the resonance Raman spectra of heme proteins we report in Figure 2.9 the resonance Raman spectra of reduced cyto— chrome g2+ in protonated and deuterated buffers and the associated Raman difference spectrum. A frequency difference shift of —2.27 cm-1 is clearly seen in the 64 Figure 2.9: Resonance Raman spectra of ferrocytochrome g2+ in protonated and deuterated buffers (pH = p0 = 7.4), and the associated Raman difference spectrum. Cytochrome 32+ was 3011M. Upon use of deuterated buffers tye 1362 cm-1 line was shifted down by 2.27 cm'1 as calculated from the RDS spectrum. RAMAN INTENSITY RAMAN INTENSITY 65 Av'2'27cn7f' C.(A-B) ,—I36l cytochrome g“ )‘exc- 406-? nm 3Hl77 B30 I400 I589 I I301 '33 [one I I49) .547 :LA I I36 2‘” I 4 ”40 I280 I420 I550 FREQUENCY SHIFT (cm") diffs reduc Resor ident oxida 1975; cytoc ty of pocke is di tion) repor equip IPMT) from I406. Coher laSer out 0 laser Chrom nm We pumpe at th exDer Ments 66 difference spectrum in the 1362 cm"1 resonance Raman line of reduced cytochrome 92* upon use of deuterated buffers. Resonance Raman spectra of hemeprotein model compounds have identified this band as a marker Raman line for heme oxidation-state and porphyrin core size (Kitagawa et al., 1975; Spiro, 1983). The detected sensitivity of this band in cytochrome 92+ to deuterium atoms indicates the accessibili- ty of protein exchangeable sites on the vicinity of the heme pocket of cytochrome c, The significance of these results is discussed elsewhere (Centeno and Babcock, in prepara- tion). The Raman instrumentation used in the experiments reported here consists of a Spex 1401 double monochromator equipped with a cooled RCA 310340 photomultiplier tube (PMT). Laser excitation in the Soret region was obtained from a Spectra Physics Model 164-11 Kr+ ion laser head (406.7 nm and 413.1 nm) equipped with a high field magnet, a Coherent Innova 90 Kri laser, and a He—Cd (441.6 nm) Liconix laser. To separate the 406.7 nm and 431.1 nm lines coming out of the Krt laser, a prism, positioned in front of the laser cavity, was used. Resonance Raman spectra of cyto— chrome oxidase in resonance with the visible cx-band at 605 nm were obtained using a dye laser with Rhodamine 60 dye pumped with a Spectra Physics Ari ion laser. Power incident at the sample was typically 10 mW (for Soret excitation experiments) and ~80-110 mW (for visible excitation experi- ments). Temperature at the sample was maintained at 4—6 00, cont the: neat tior fol] spec The izat (CH: CHzC tic enzy douh ture exch ENDO °0py prov ment Bruk ENDo SUpp 67 controlled by a constant flow of liquid N2 boil off. A- thermocouple, positioned inside the sample holder and con- nected to a digital meter, was used. To measure depolariza- tion ratios (i.e.JD: IL /I,‘), a polaroid disk was used, followed by a polarization scrambler which corrects for spectrometer response to different polarization directions. The polaroid disk was calibrated by obtaining known depolar- ization ratios of standard solvents in the low-frequency (CHzClz and 0014) and in the high-frequency (CsHs and CH2C12) ranges. C. Electron Nuclear Double Resonance Spectroscopy ' (ENDOR) ‘ In this thesis we have made use of electron paramagne- tic resonance spectroscopy with the purpose of monitoring enzyme integrity and preparation quality. Electron nuclear double resonance spectroscopy, on the other hand, was used to gain information on the CuA2+ —ligand coordination struc- ture and to investigate possible occurrence of protein H/D— exchangeable sites in the immediate vicinity of Guilt. ENDOR spectroscopy combines the sensitivity of EPR spectros— copy and the capability of nuclear magnetic resonance to provide structural information about the immediate environ- ment of paramagnetic metal centers. The EPR and ENDOR spectra were obtained by using a Bruker ERZOOD X—band spectrometer equipped with an ER152E ENDOR/TRIPLE accessory. Radio frequency (rf) power was supplied to an eighteen turn selenoid via a 100 W amplifier (EN syn sta ( 19 Oxf ava sub and pow thi reg Opt be - sub: tio: Perl Del. aWe: CCI ball the 00m 00m 68 (ENI 31002) driven at frequencies generated by a Wavetek synthesizer (model number 3000-446). The coil was of a free standing design similar to that described by Hurst et al. (1978), and slid into position on the quartz Dewar of an Oxford cryostat. The resonant cavity for ENDOR was that available commercially from Bruker. EPR spectra were subject to the following experimental conditions: field modulation approximately 1G at 12.5 kHz, and microwave powers typically 0.5 to 1.0 mW. ENDOR spectra were obtained in the region near the free proton precession frequency, 7 this region corresponding to that defined as the 'matrix" region by Hyde (1967). The applied radiofrequency power was 100 W, and this frequency was modulated at 15 kHz. The optimum microwave power/temperature combination was found to be 6-7 mW and 3.4—10 K, respectively. D. Infrared (IR) Spectroscopy Infrared (IR) spectra of heme a3+Cl‘ and the metal substituted porphyrin models, in KBr pellets and in solu- tion, were obtained using a BOMEM FTIR (Model DA3.02) and Perkin Elmer spectrophotometers. For the solidified KBr pellets spectra the resolution was 2.0 cm'1 with an average of 150—250 scans. For the solution spectra we used 0014 as the solvent. The contributions of 0014 were fully balanced and subtracted from the porphyrin spectra yielding the spectrum of the porphyrin derivative. Nitrogen gas was constantly purged into the IR sample chamber to avoid H20 contributions to the IR spectra. Sum pou in lab and abs inf sho mod sub wit f re Str deu rin whi. Obs. eff. of j for) and CHAPTER III FORMYL ISOTOPIC SUBSTITUTION AND HYDROGEN BOND SENSITIVE MODES IN HEME A MODEL COMPOUNDS: MODELS FOR CYTOCHROME A Summary: To elucidate formyl (-CHO) modes in heme model com— pounds that mimic spectroscopic properties of cytochrome g in cytochrome c oxidase, we have synthesized isotopically labelled Cu2+ porphyrin (-CH180), Cu2+ porphyrin a (—CD0) and heme a I—CDO), and characterized them by using optical absorption and resonance Raman (RR) spectrocopies. The infrared (IR) spectrum of the 18O-labelled Cu2 *porphyrin shows substantial frequency perturbations for IR formyl modes. Resonance Raman activity of Eu-type IR-allowed substituent modes is observed. Upon substitution of 16O with 18O at the carbonyl group, the Vc=o stretching frequency decreases by 32 cm-l, and the ring-formyl stretching ( Vcb-cao) upshifts by ~4 cm-l. Similarly, deuterium substitution at the formyl proton of Cu2* porphy— rin g results in a 5 cm‘1 downshift in the C=O stretching frequency and a 4 cm-1 upshift in the Vcb-cno vibration, which is in agreement with analogous isotopic effects observed in the Raman spectra of benzaldehyde models. The effect of hydrogen bond formation between the carbonyl group of heme a and phenol H-donors on Raman frequency shifts of formyl vibrational modes was also investigated. The vc=o and Au (max) values appear to be related to the pka of the o 69 ter ana spe hem int tra: is . con of j m imp. oxi. sub' Sig: enz; Th0] Pro] eve; Pep. whll III aIor 0111 70 H-donor, but are insensitive to H/D exchange at the OH terminal end of the phenol H-donor group. A comparison with analogous isotopic and H-bonding effects on the Raman spectra of benzaldehyde suggests that the H-bond in isolated heme g is a weak bond, which contrasts to the stronger interaction observed in cytochrome g. I. Introduction: Although the nature of the coupling between electron transfer and proton pumping activity in cytochrome g oxidase is a matter of some controversy (Wikstrom et al., 1982), considerable advances have been made towards the elucidation of plausible protein and/or chromophoric sites which could serve as structural/functional linkages between these two important aspects of energy transduction in cytochrome oxidase. For instance, the non-chromophore containing subunit III has been implicated on the basis of the significant diminution in proton pumping activity when the enzyme was depleted of this polypeptide (Penttila, 1983; Thompson et al., 1985). However, cytochrome oxidase from Pr0karyotic sources containing two (Solioz et al., 1982) and even one subunit (Sone et al., 1984) has been recently reported to exhibit appreciable proton pumping activ1ty, Which made the above interpretation of the role of subunit III in the oxidase H+-pump somewhat uncertain. The CuziA atom has also been proposed to be involved in proton pumping 0n the basis of a detailed interpretation of its unusual EPR spec recs lige oxid prot al., strc that user prot redo prob 3hif« nyr , 71 spectroscopic properties (Chan et al., 1979), as well as on recent results from chemical modification studies at the CuA ligand-coordination sphere (Blair et al., 1986). The heme a chromophore of cytochrome a in cytochrome oxidase has been suggested as the site for the redox-linked proton pumping activity (Wikstrom et al., 1981; Wikstrom et al., 1982; Babcock and Callahan, 1983). Based upon spec— stroscopic evidence, Callahan and Babcock (1983) concluded that the formyl I-CHO) group at position 8 of the heme a macrocycle is involved in a hydrogen bond interaction with a protein H-donor residue (see Figure 3.1), communicating redox events at the heme a iron atom to the rest of the protein surroundings. This conclusion was reached from the observed red-shift in the a —maximum and the downshifted frequency of the vczo of oxidized and reduced cytochrome a relative to the low-spin, six-coordinated heme g model in aprotic (non—hydrogen bonding) solvents (Callahan and Babcock, 1983). The strength of the H-bond interaction was observed to depend upon the redox state of the cytochrome a iron atom, and accordingly, a redox-driven proton pumping model was proposed in which the H—bonded proton is translo— located along a chain of hydrogen bonded residues, triggered by a redox—dependent conformational change in the cytochrome é (C=O...H) structure. Since the above assignments relied heavily on the red— Shifted absorption maximum and on the behavior of the carbo— nyl group stretching frequency in cytochrome a, it appears 72 X-Hmo\ Figure 3.1: Proposed hydrogen bonded structure of cytochrome a in cytochrome g oxidase. The protein residue is designated as X—H (Babcock and Callahan, 1983). of ] mode Vibl f ror Insi gror subs form IH-( H-b< pres sucl f iec al. 38g] tior appé (Var 73 of particular interest to investigate other formyl related modes and hence to separate the formyl H—bond sensitive from other heme macrocycle and/or protein vibrations. Insights into the internal vibrational modes of the formyl group of heme a were reported by Choi et al. (1983). By substituting the formyl-carbon hydrogen with deuterium, formyl modes involving the ring-formyl stretch and the (H-CO) bend were indentified and observed to be sensitive to H-bonding (Choi et al., 1983). Other ring vibrations, presumably sensitive to the presence of the formyl group, such as v1, (Cb—Cb) and v33 (Cb-Cb), were also identi— fied as H-bond sensitive modes. Nevertheless, the Choi et al. (1983) assignments are weakened by extensive heme a aggregation in their studies as well as by Chemical satura- tion of the formyl group double bond, as evidenced by the appearance of the 565 nm band in their absorption spectrum (Vanderkoii and Stotz, 1965; Babcock, 1987). To quantify these hydrogen bonding effects, we report here the resonance Raman (RR) and IR spectra of heme g and its 4-vinyl—8—formylporphyrin analog, in which the isolated -CHO group has been studied under the influence of hydrogen bonding effetcs and after isotopic labelling at the formyl group. The results obtained are compared with analogous isotopic and hydrogen bonding effects observed for formyl modes in simpler aldehyde-containing systems with the aim of establishing trends of formyl sensitivity to hydrogen bor moc' wit str phe H-d cha H-d the of spe and unc or Vin the 11. Mel in and the 74 bonding. The patterns of frequency shifts in the heme a .model compounds and in benzaldehyde are similar for modes with substantial formyl character, such as the formyl stretching frequency and the ring-CHO frequency. The use of phenol—OH and its deuterated derivative (phenol-ODe) as H—donors to the formyl group does not result in additional changes that could be attributed to H/D exchange at the H—donor; however, placing an electron withdrawing group at the phenol ring (i.e., Cl-), markedly increases the strength of the H—bond interaction. Comparison of the infrared spectra of heme a3+ Cl", Cu2+ porphyrin labelled with 180 and protoporhyrin IX Cl' in KBr pellets and in solution," uncovered most of the formyl modes that are not Raman active or that overlapped with porphyrin macrocycle vibrations. The vinyl stretching frequency was not possibly identified in the IR spectra of heme e and its Cu2* porphyrin analog, but it was observed at 1626 cm'1 for protoporphyrin IX—chloride. II. Results and Discussion: A. Effects of hydrogen bonding on the absorption spectra of heme g complexes: The absorption spectra of ferric and ferrous heme Q (N— MeIm): in aprotic (non—hydrogen bonding) solvents are shown in Figure 3.2A. The effect of hydrogen-bond formation, with H20 and p—Cl—phenol as the H-donors, is shown in Figs 3.2B and 3.20, respectively. Table 3.1 summarizes our results on the observed changes in the absorption spectra of heme a and 75 Figure 3.2: Absorption spectra of heme a (N-MeIm)2 under various hydrogen bonding condition: A) (---) heme e3+(N-Me1m)2/CHZCT2 ___ 2+ ( ) heme g (N-MeIm)2/CH2CI2 B) (———) heme 33+(N-MeIm)2/H20-detergent a2+( (---) heme ‘ N-MeIm)2/H20-detergent C) (———) heme 33+(N-Me1m)2/CHZCl 2 (--—) heme 13+(N-MeIm)2 + p-Cl-Phenol/CHZCIz (———-)heme 92+(N-MeIm)2 + P-CI'Phe”01/CH2C]2' The detergent used in all these samples was 0.07 M cetryltetrametryl ammonium bromide (CTAB). 76 14285.7 20.?00 I6666.6 25 .900 ABE: 38:93 muz_ewowmcmm mono: we :pmcmpm>e3 coppmowoxm .mce_ cmmu one An umpmowuca mam agate _sEEou flooo-v ampacabsaa new AQIU-V m :_c>:acoq +m :u we mLoUon :mEmm mocmCOmmm .NFUNIU see an . .E: N.©oe pcmzpapmnzm _esco: spas u©.m mcsmam 91 Amasov>< cot ONE 05%. cos. 0mm. 00m. 00m. gum. om: Om: coo. coo. _ _ _ _ _ . _ _ _ a u w E: 5.00? m 5:539. 23.30 0001 m 59 a _ . _ _ _ _ new _ Oh: _ s . _ _ _ m _ _ _ C 82 __ _ _ . one. _ Q... _ _ _hww. _ wOn. _ _ E. . ozoI _ _ 8.. _ _ _ Qnm. _ — — — — N5: _ _ Rs . in. . _ .2. 2 _ am Can. AlISNBlNI NVWVH Table hie—s ()2 Cb- ill-cw ‘Jll (VCb- Pyr. Table 3.3. Formyl—Sensitive Modes in Heme g (N-MeIm)z (—CHO/-CDO) and Cu2+ Porphyrin g (-CHO/-CDO) Modes: Species: Pa? " .t ........ Eh.:.912 Raf. * .1: ...... 22:99.. Cu 2 *a + Ph-OD (”:0 1672 1652 1645 1630 1672 1643 (1667)3 (1650) (1637) (1630) (1667) (1641) \n Cb-Cb 1586 1589 1585 1588 1590 1597 (1588) (1589) (1587) (1589) (1593) (1595) M-c=o (1399)b 1405 13950 1397c V13 1238 1241 1240 1243 1235 1243 eva-cn0)(1240) (1241) (1243) (1245) (1241) (1243) Pyr. fold 450 ( )d 453e ~ 461f 345 345 vs ( 0Cb-S) 335 348 339 350 a. Frequencies (cm-1) within parenthesis correspond to those observed upon formyl—proton deuteration. This mode was observed for heme a3*(N-Melm)z in CCla. This mode overlap with a porphyrin macrocycle mode at 1395 cm'l. . Decreases in Raman intensity upon H-bonding. Frequency observed for Cu2+ porphyrin. Frequency observed for Ni2+ porphyrin a. 00" r-erQ- subs the H/D ted ' grou; form; cont: ing - and I substituent. The Vc=o frequency shifts that we detect in the heme g and Cu2+ porphyrin g (-6 cm-l) derivatives upon H/D formyl exchange are in good agreement with those repor— ted by Williems and Bocian (1984) for the deuterated formyl group of formyl-substituted metalloporphyrin in which a 10 cm‘1 shift was reported. Other vibrational modes that appear to be sensitive to formyl proton substitution, and that contain substantial contribution from porphyrin peripheral substituents accord- ing to the normal coordinate analysis of Abe et al. (1978) and Choi et al. (1982) are the following: V2 (Cb-Cb), V13 (Ca-Cb), ( V5 + V9) (Cb-S) in the high-frequency region, and V3 ( 6(36.5) and v9 ( 60b-s) in the low-frequency range. In the RR spectrum of ferric heme a, the high- frequency vibrations v; and 013 are seen at 1586 cm"1 and 1238 cm'1 for heme a3+ (CHO), shifting up to 1588 cm-1 and 1241 cm'1 for heme g3+ (CDO), respectively. The frequency of the non-fundamental mode ( V5 + V9) expected at ~1250 — 1255 cm-1 is largely overlapped by a strong N- MeIm band at 1259 cm-1 and hence its frequency upon formyl H/D exchange could not be ascertained. The RR spectrum of Cuz‘ porphyrin a reproduced the frequency shifts observed for V2 and 013 in heme a. In the spectrum of Cu2+ porphyrin a (CHO), these latter modes are observed at 1590 0m‘1 and 1235 cm-1, respectively, shifting up to 1593 cm"1 and 1241 cm-1 (see Figure 3.6). The formyl hydrogen bend of ferric heme a is difficult to assess, owing to the strong. cont use dete excl forn the with al., heme dist spec assi 1418 agre Simi; Spec. contribution of CH2C12 to a line at 1423 cm-1; however, upon use of CC14 as the solvent, a weak line at 1399 cm"1 was detected which shifts up to 1403 0111'1 upon formyl H/D exchange (spectrum now shown). The RR enhancement of this formyl hydrogen bending mode is expected to be stronger in the spectrum of reduced heme a, but it strongly overlaps with the porphyrin v29 Bzg frequency at 1395 cm"1 (Abe et al., 1978; Choi et al., 1983). The infrared (IR) spectra of heme a3* C17 in KBr and in solution (Figures 3.8 and 3.9) display a band at 1410 cm‘l, which is not observed in the IR spectrum of protoporhyrin IX—Fe3* Cl‘, and accordingly it is assigned to the 5ac-o bend. Its upshifted frequency to 1418 cm-1 upon 18O-substitution (see Figure 3.8) is in agreement with a similar upshifts in benzaldehyde models. The high-frequency RR spectra of 2,6-dipentyl-4-vinyl— 8-formyl Cu2+ porphyrin with unsubstituted CH150 and with oxygen—18 labelled CHIBO, and the associated Raman differ- ence spectrum (RDS) are depicted in Figure 3.7. The low- frequency region from 150 to 600 cm-1 is also illustrated. Of interest in this low—frequency spectrum is the upshift frequency (+4 cm‘l) for a mode at 453 cm-1 to 457 cm‘l, which we attribute to the pyrrole ring—folding mode since a similar feature is observed in the RR of benzaldehyde at 441 cm'l. The infrared spectra of ferric heme a3* 01— and ..n its isotopically labelled Cuzt porphyrin analog in the solid state are shown in Figure 3.8. The solution state IR spectra of heme a3+Cl', Cu2+ (—CH150) porphyrin and the 95 2.1.5. QM9 :9 .s: m.ooe pm cpmcm_m>m3 coaompwoxm .cowpsp -rpmnsm 0 con: Hugo Nme Op a: mpewcm cacxgacoq AooHIUIV ms +mzu we Eztpomam mm esp cs H.Eo mme pm m:_~ Aocmzcmta zo_ < .mom ass 5624 ambasso_aa m_ 2-56 cams be mace me ms; esp Cw H.Eo w.m+ we mocmcmmmau Aocmscmce < .Umomcpm:__a H.Eo oomH-oooH Eon; Echommm mocmcmeean :msem .N_QNIU sis :_ newspapmazm agate tsetse flowszo-v cm_m ma one aa__asag-oma use Aoegzo-c _aEcoc £523 cacssacoa +Nso _xstow-m1_xca>ee 40 mcpomam cmsmm mocw:0mmm ”N.m mczmad ass Kim szaemE Om. 00¢ Onw ooh 8m. own. 8m. 8m. _llIITIILl\l_l\\T\II Tl“ a _ u m a a u a . _I w . m L3. 2—2 W “ 3n. _ I “ I fl 9010 W n M... u .m I _ V m _ u m N i W _ I % u % ~_wn. 8.9 oz: m m . a 0.010) " HNJ «so. . m . _ s . _ won. Hal—I m3. " IA 97 Figure 3.8: Infrared (FTIR) spectra of heme a3+Cl7 16 (-CH 0), Cu2+(-CH160) porphyrin, and Cu2+ (-CH180) porphyrin derivatives in KBR pellets in the region from 1 1000 to 1800 cm- Instrumental conditions: resolution 2 cm-l; sensitivity 1 second; number of scans, 250 TRANS M l TTANC E % TRANSMITTANCE % 98 + 1* i A.) Heme g3‘CI‘ "I38 I I I I. I |$48 I574 I5I2 I ' I I l I I I I I6l4 I | 536' I I309 ... 8.) Cu” Porphyrin I I (-CI-I"0) I a I I g I669 I l I I742 l l I : | I I I I I I I ' I380 I II574'I : I I58? I53 I I I6Io II5I3 I I I I I I I C.) Cuz' Porphyrin ‘268 IOI8 (‘CH'°O) t i I l 800 I 600' I400 I 200 I 000 din Fie ini In fox sul mi of al. of - sub Ilil ana wit it sub 37m 381‘ and VII) Sub IR. 99 dimethylester of protoporphyrin Fe3+ C17 are displayed in Figure 3.9. In Table 3.4 we summarize the observed RR and infrared frequencies sensitive to formyl 18O-substitution. In addition, frequency shifts caused by hydrogen bond formation to the C=O group in Cu2+ (CHIGO) and Cu2+ (CHISO) substituted porphyrins are also indicated in Table 3.4. While the RR spectra of hemeproteins and porphyrin model compounds appear to be more sensitive to enhancement of in-plane porphyrin skeletal modes (Spiro, 1983; Choi et al., 1982), additional information on the activity of out— of-plane porphyrin modes and internal vibrations of the C6— substituents should be apparent in the IR spectra. Recently, Willems and Bocian (1984 & 1985) carried out a comprehensive analysis of the IR and RR spectra of Ni2+ deuteroporphyrins with conjugated carbonyl substituents. From these studies, it was clear that unsymmetrical implantation of conjugated substituents at the porphyrin periphery lowers the effective symmetry to produce splitting of En modes in the IR spectra and selective activity for some of these modes in the RR spectra. They also concluded that Raman modes with substan— tial peripheral substituent character such as V2, V8, and V13 were strongly substituent dependent. Differences ob— served by Willems and Bocian (1984) in the frequency of \4 and V13 were attributed to kinematic effects in the Vibrational Hamiltonian, which are induced by changes in substituent masses. The above authors were able to identify IR-allowed Eu modes that exhibited large dependence on 100 Figure 3.9: Infrared (FTIR) spectra of heme §3+Cl_, Cu2+ 16 (-CH 0) porphyrin, and the dimethylester (DME) derivative 3+ of protoporphyrin IX Fe -Cl" in carbon tetrachloride (CCI4) solutions. Instrumental conditions: resolution 2 cm-la sensitivity 1 second, one scan. Solvent contributions were substracted. TRANSMITTANCE TRANSMITTANCE 101 8' Cu” F’Qrphyrin / CCI‘ (-C H ”0) W i 4. I62 0'— I660— A- Heme Q’- CI. CCI‘ 9..- 2"..— I363 '31—'— I23!— Im— nos—— "95—- "69- I459— I430— C- PP Fe’tI‘ (DME) CCI, I600 I400 I200 I000 FREQUENCY earn") Subs‘ II—oIiEE ‘Iaa Cb ill-C=O VII Ca I)” Ch ‘IICb. II 3 102 Table 3.4. Resonance Raman (RR) and Infrared (IR) Frequencies (cm'l) Sensitive to ~CH150 -——- -CH130 Substitution at the Formyl Group of Cu2+ (4—Vinyl-8—Formyl) Porphyrin. Modes:a Cu2+ (~CH160) Cu2+ (~CH180) -CH150 + Ph-OD 8.13. TB. 8.43: .123. BB VC:0 1672 1669 1643 02 C5-C6 1593 1595 1598 1530 1530 \ms Cb-Cb 1513b (”1480)b 03-c=o 1399c 1410 1418 1405C W1 Ca-N 1365 1356 1367 (”1370) 1365 \M2 Cb-S 1261 1268 \)(Cb-CHO) 1232 (~1260) V13 1238 1243 1248 960 964 963 754 744 734 Pyr- 453 457 (453) fold 350 350 UséCb-S 337 337 348 a. Porphyrin mode assignments follow those by Abe et al. (1978). b. Replacement of 150 by 180 result in the disappearance of the 1513 cm-1 IR line with the concomitant appearance Of a new shoulder at ~1480 cm-l. 0. Observed in CCl4 solutions. I— pe Th VII thI 211: UN 32 161 enl fre The Ni the sei tee st: feI at 001 Spg enI SpI 103 peripheral substituent symmetry lowering effects, such as \Qa (Cb-Cb; 1550-1560 cm‘l) and V41 (Ca-N; 1360-1366 cm'1). The increased splitting and hence, loss of degeneracy of W2 (Cb-S; 1260—1268 cm'l) and \aa (Cb—S; 1165-1172 cm-l) in the IR spectrum of Ni?t(2,4-diformyldeuteroporphyrin) was also suggested. The assignment of some of these IR-allowed Eu modes in the RR spectra of heme a derivatives was recently attempted by Choi et al. (1983). By comparing the RR spectra of heme a2+ (N—MeIm)2 obtained with two different excitation wave— lengths (514.5 and 592 nm), the above authors suggested the enhancement and splitting of \G7 (Ca—Cm) mode with RR frequencies at 1614 cm'1 and 1588 cm‘l, presumably through the X— and Y— electronic dipole transitions, respectively. The arguments advanced by Willems and Bocian (1984) for Ni2+ formyl substituted porphyrin, however, does not confirm these observations, since splitting of \hv was not ob— served. Splitting of V33 in heme g3* 01‘ was also sugges— ted by Spiro and coworkers from a comparison of the solid state IR spectra of ferric five coordinate heme a Cl' and ferric protoporphyrin IX Cl'; in their study a small feature at 1525 cm'1 in the IR spectrum of the latter compound was correlated with lines at 1544 cm‘1 and 1512 cm*1 in the IR spectrum of heme a3+ Cl'. Because of the confusion in the frequencies and appar— ent vibrational splittings of V37 and V33 modes in the IR spectra of heme a, and owing to the expected contributions of MI sol p0: thI ThI tie ThI st: ev er PP at is ti he: a1 ed be up IR en in th di 104 of the formyl group to these Eu modes, we took the task of solid and solution states. The Cu2+ (CHIGOI/(CHISO) porphyrin derivatives have also been useful and we assess the sensitivity of these Eu modes to formyl H-bond effects. The IR spectra of heme a3+ Cl- reported here differ substan- tialy from those reported by Spiro and coworkers (1983). These spectral differences are as follows: both the solid state (Figure 3.8) and solution spectra (Figure 3.9) show no evidence for enhancement of the vinyl stretching frequency expected at 1626 cm-1 which is evident in the IR spectrum of PPIXIDME)-Fe3+ Cl‘ (see Figure 3.9, trace C); a relatively strong band at 1614 cm"1 is detected in our IR spectra, but is absent in the Choi et al. (1983) spectrum; a new absorp- tion at 1410 cm'1 was reproducibly detected with all the heme a3t Cl- and conjugated carbonyl Cu2+ models that was also absent in the Choi et al. (1983) spectrum. As mention— ed above, we assign this latter mode to the formyl-hydrogen bend, and its upshifted frequency to 1418 cm"1 (+8 cm'l) upon 160 ————> 18O substitution supports our assignment. The IR spectra reported here were reproduced with three differ— ent heme a3+ Cl- preparations and with two different FTIR instruments. Isotopically substituted Cu2+ (CHIBO) porphyrin display thé Vc=o frequency at 1640 cm-1, as detected by the RDS difference Spectrum (see Figure 3.7). This 32 cm-l shift in dif mag suc (K2 Pir. 198 upc (se pla Iac shi 4 c the RR con IR POT Str Sol We IIEm Wit men com 00m 105 in cho is in good agreement with the 42 cm'1 frequency difference calculated from the reduced masses (U2/ U, = 1.025) of the C=O carbonyl group. Carbonyl shifts of this magnitude have been reported in the past for simple systems such as isotopically labelled diisopropylketones (Karabatsos, 1960), 18O—labelled benzophenone (Halmann and Pinchas, 1958), and 18O-labelled acetone (Kamoun and Mirone, 1980). Formyl modes previously observed to be perturbed upon —CHO deuteration in heme a and Cu2+ porphyrin a models (see Figures 3.5 and 3.6) are also seen to be altered upon placement of labelled—150. Upon substitution of 150 ---—> 190, the RR spectrum of Cu2+ porphyrin revealed a frequency shift in V2 at 1593 cm"1 shifting up to 1595 cm-1, and a 4 cm-1 upshift in the v13 mode at 1240 cm'1 as detected by the RDS difference spectrum. The sensitivity of this latter RR mode to formyleproton and oxygen—substitution indicates considerable contribution from the ring—formyl stretch. The IR spectra of heme a3+ Cl‘ and its normal Cu2+ (CH‘50) porphyrin model in KBr and solution states show a sharp strong band at 1260 cm’l, which is not observed in the solution spectra of PPIX—Fe3+ 01' (see Figure 3.9, trace C). We assign this IR band to the V42 Vch-s) Eu mode of heme gat Cl'. It appears to be split into two components With frequencies at 1260 cm”1 and 1230 cm-1. Upon replace— ment of 16O by 18O at the formyl oxygen, the higher energy component shifts up to 1268 cm‘l, while the lower energy component appears to move under the stronger 1260 cm'1 line. W of ti: al. 001 F 11 am Ni bu‘ rm f 01 08.! p01 (Se the Sir (Fi ide COT anc‘ at Mr am- has 106 The 160/180 induced upshift of the higher energy component of V42 (1260 ——— 1268 cm-1) is consistent with a substan- tial contribution from the Vcb—cao stretch to this IR- allowed mode. The direction of its frequency shift is consistent with that observed in the benzaldehyde model (see Figure 3.4 and Table 3.2). Our results are in good agreement with the recent analysis of the formyl modes of Ni2+ (2,4—diformydeuteroporphyrin), in which strong contri- butions from the ring-formyl stretch to the IR-active compo- nent V42 mode was concluded from deuteration of the formyl proton (Willems and Bocian, 1984). In this latter case, loss of degeneracy on the V42 was also detected. The oxidation state marker band region of Cu2+ porphyrin displays two bands at 1376 cm'1 and at 1364 cm-1 (see Figure 3.7). The former band is assigned to the v4 (Ca-N) vibration while the latter appears to correspond to the RR active component of the V41 IR-allowed Eu mode, since the IR spectra of Cu2+ porphyrin clearly indicate that W1 is split with frequencies at 1380 cm-1 and 1356 cm-1 (Figure 3.8). The second component of \MI cannot be identified in the RR spectrum since it falls within a rather congested region. In the case of heme §3+ Cl", the solid and solution state IR spectra display at strong absorption at 1380 cm'l, with a weak shoulder at 1366 cm'l. The former line is assigned to the V41 IR-allowed mode and the 1366 cm-1 might correspond to the second split component of 041; however, the RR spectrum of heme a does not show activity 31 as th 13 th ac at 3}) wi KB ba: f o: CIR 002 CIII tuI fre cm- The 107 for either of these V41 vibrations. Interestingly, the RR spectrum of Cu2+ porphyrin a (see Figure 3.6) does show a pronounced shoulder at ”1366 cm'1 which might well corre— spond to the second RR active component of the V41, in agreement with results from Cu2+ porphyrin. H/D exchange at the formyl group of Cu2+ porphyrin g appears to shift this 1366 cm"1 shoulder under the strong 04 band at 1377 cm-1 (see Figure 3.6) as noticed by the narrower bandwith at halfheight (i.e., F =8 cm-l) in the RR spectrum of Cu2+ (-CDO) porphyrin a. Similarly, upon 160/130 substitution, the 1356 cm'1 IR absorption apparently vanishes, but this is accompanied by a significant increase in intensity of a line at ”1370 cm'1 in the RR spectrum of Cu2+ porphyrin. This apparent upshift in this V41 component is also consistent with strong formyl character. The IR spectra of heme a3+ Cl- and Cu2+ porphyrin, in KBr pellets, in the region from 1500 to 1600 cm-1 display bands at 1574 cm-1 (weak), 1544 cm-1 (weak), and 1512 cm.1 for the former compound, while for the latter bands at ~1574 cm-1 (weak), 1530 cm"1 and 1513 cm'1 are observed. The corresponding solution spectra show a single band at 1513 cm-1 for heme §3+—Cl- and weaker absorptions at 1574 cm'l, 1530 cm-1, and 1513 cm-—1 for Cu2+ porphyrin. Upon substi— tution of isotopically labeled-180, the 1513 cm-1 IR line from Cu2+ porphyrin vanishes, while the 1530 cm‘1 and 1574 Cm-1 bands retain most of their IR absorption intensity. The IR spectrum of PPIX—Fe3+ C1‘ in solution illustrates IR ir 38 ri SE Th su qu he ex er Eu th 00 an th fit an 0b of Th br 108 absorptions at 1588 cm‘1 (weak), 1562 cm'l, and a weak shoulder at 1508 cm'l. We assign the strongest absorption in PPIX—Fe3+ 01' at 1562 cm'1 as due to Vas (Ch-Cb), in agreement with similar assignments made for Ni-protoporphy— rin (Willems and Bocian, 1984; Abe et al., 1978). For heme a3* Cl' and its Cu2+ (CHIGO) porphyrin derivative, this mode seems to split with frequencies at 1544 cm'1 and 1513 cm-1 (heme a) and at 1530 cm'1 and 1513 cm'1 (Cu2+ porphyrin). The disappearance of the 1513 cm'1 IR absorption upon CHISO— ——-) CHISO substitution is consistent with this mode having substantial contribution from the formyl group. The fre— quency differences between the 1544 cm-1 Vas component of heme a3i Cl— and 1530 cm-1 component of Cu2+ porphyrin is explained on the basis of the presence of different periph- eral substituents. Our observation of splitting of the V33 Eu mode in the IR spectrum of heme a is in agreement with that by Choi et al. (1983); however, we believe that the corresponding frequency in PPIX-Fe3+ Cl‘ is at 1562 cm'l. The 1614 cm-1 and 1610 cm-1 IR lines of heme a3* Cl- and Cu2+ porphyrin, respectively, appears to correspond to the V37 (Ca—Cm) vibrational mode frequency described by Abe et al. (1978). This mode is largely resistant to splitting, and under our experimental conditions, RR activity is not observed. It also appears to be insensitive to substitution 0f 18O at the —CH0 group oxygen, as is shown in Figure 3.8. The contribution to V37 is predominantly Ca-Cm methine bridge (Abe et al., 1978), although Spiro and coworkers have 81 It 3F fr pc (19 vi ma me tw pl; <11: of ViI are its tie Cu: the To 311 Reg 109 suggested that this assignment should be of C6—C6 character. It is our Opinion that the former assignment would be more appropriate in view of the isotopic evidence presented here. We have not been able to identify the vinyl stretching frequency of heme a3+ Cl‘ and its 4-vinyl—8-formyl—Cu2+ porphyrin model. This vinyl mode is expected around 1625— 1628 cm-1, by analogy with the 1626 cm-1 absorption of PPIX— Fe3+ Cl‘ (see Figure 3.9). In the RR Spectra of heme proteins enhancement of vinyl modes has been suggested to depend upon the extent of vibronic coupling between the vinyl W—orbitals with the excited state of the porphyrin macrocycle (Choi et al., 1982), and hence lack of enhance- ment might be an indication of a slight out-of—plane twisting of the vinyl group with respect to the porphyrin plane. Recently, Rousseau et al. (1983) attributed the discrepancies in frequencies observed for the vinyl groups of hemoglobin A and leghemoglobin as due to different vinyl-protein interactions resulting in different vinyl group conformation. In the isolated heme g chromophore and its Cu2+ porphyrin analog, it is likely that steric interac- tions between the heme g hydroxylfarnesil tail at position 2, or the pentyl (i.e., —(CH2)4CH3) group at position 2 in Cu2+ porphyrin, and the vinyl group hydrogens will not allow the vinyl n—bond system to assume a COplanar configuration. To address this interesting observation, crystal structure studies of heme a3+ Cl- and its 4-vinyl-8-formyl models, are needed. the den 3.2 in as fer H-b spe phe ape wit pro 3.1. (15 tiv phy and 81. add rev deg to SP9 110 D. H—Bonding Effects in Heme a Models. Raman spectra of H—bonded benzaldehyde indicate that the C=O stretching and deformation modes are insensitive to deuterium exchange at the H—donor (see Figure 3.4, and Table 3.2), and hence suggest a weak H—bond. This seems to hold in the RR spectra of heme a and its Cu2+ porphyrin models, as shown in Figures 3.10 to 3.12. In Figure 3.10, the RR of ferric heme a3* (N—MeIm); non—hydrogen bonded in CH2C12, and H—bonded to phenol—OH and phenol-OD, are compared. The RR spectra of ferrous heme a2* (N—MeIm)2 in the presence of phenol—OD as the H—donor are shown in Figure 3.11. The RR spectra of Cu2+ porphyrin under the influence of H—bonding with phenol—OH and phenol-0D are illustrated in Figure 3.12. The low—frequency RR of ferrous heme a in aprotic, protonated, and deuterated solvents are shown in Figure 3.13. Our interest in measuring the low—frequency regions (150—650 cm—l) of the RR spectra of heme a and its deriva— tives arise because of the expected contribution from por— phyrin peripheral substituents to V3 ( 6cb-s, 330-350 cm'l) and V9 ( 5gb_s, 265—275 cm'l) modes as suggested by Abe et al. (1978), and recently by Willems and Bocian (1984). In addition, our hydrogen—bonding studies with benzaldehyde reveal a low—frequency mode at 441 cm'1 for ¢-CHO WhiCh decreases in intensity upon H—bonding formation, and Shifts t0 lower frequencies upon formyl—proton deuteration. The RR '1 (see Spectrum of Cu2* porphyrin displays a band at 453 cm 111 Figure 3.10: Resonance Raman (RR) spectra of ferric non- hydrogen bonded heme 33+(N-MeIm)2 (in CHZCl2 ) (trace A) and ferric heme 33+(N-Me1m)2 (in CCl4) hydrogen bondEd t0 phenol-0H (trace B) and phenol-OD6 (trace C). The bottom trace (D) is the Raman difference spectrum between samples B and C recorded with the Raman difference appara- tus. Exitation wavelength at 406.7nm. Average of ten scans. RAMAN INTENSITY 112 A o. (B- C) Heme gym-Me Im)2 9X6. l375 406-7nm ”96 I l I602 I571 IOOO I ”40 I2 80 I420 FREQUENCY SHIFT (cm") I560 '00 >IP_mzmI—I2H zqzqm 113 A- Heme g2+(N- Me Im)2 c HZCIZ Ha I exc.44rsnm Islsa l585 I420 I560 I700 FREQU ENCY SHIFT (cm") >. t: (D 12 LLJ I— 12 H 12 ‘4 IE ‘1 (I B- H eme IfI'N-Melm)2 Phenol—006 l l 1000 H40 I280 Figure 3.11: (N-MeIm)2 Resonance Raman spectra of ferrous heme g H—bonded to phenol-0D6 in dry CHZCTZ. 2+ 114 1 2+ Figure 3.12: Resonance Raman (RR) spectra of Cu (-CH16O) porphyrin in CH2C12 (trace A) and under hydro- gen bonding to phenol-0H (trace B) and phenol—OD6 (trace C). The bottom trace illustrates the Raman difference spectrum between samples B and C. Excitation wavelength is 406.7 nm. RAMAN INTENSITY C u2+Porpnyri n x exc. 406-7nm | I594 (248 [3| 2 I I ‘__I._( l 1 I000 n40 I280 I420 I560 FREQUENCY SHIFT (cm") I7 00 RAMAN INTENSITY T T ISO 330 465 600 FREQUENCY SHIFT (cm") 2 .Figure 3.13: Low frequency resonance Raman spectra of heme g +(N-MeIm) in CHZCl2 and under hydrogen bonding conditions (traces C and D). The' RR spectrum of NiLT porphyrin a is also illustrated. Excitation wave- length is at 406/7 nm. 2 Figu a H- form Ni2+ obse assi form‘ vent appe: Figu: sing. 3.13 shif‘ with Vibr: pI‘EVI detee "Don SubsI 117 Figure 3.7) which decreases in intensity in the presence of a H-donor. Intensity changes in this line upon H-bond formation was also observed in the RR spectra of Cu2+ and Ni2+ porphyrin g (spectra not shown). Based upon these observations and the assignments made for benzaldehyde, we assign this mode to a pyrrole folding mode with substantial formyl character. For heme a2+ (N—MeIm)2 in aprotic sol— vents, two low-frequency modes at 335 cm-1 and 345 cm-1 appear to be sensitive to H-bonding effects (traces B and C, Figure 3.13) since upon use of hydrogen bonding solvents a single band is observed at ”348 cm-1 (traces C and D, Figure 3.13). This behavior is consistent with the 335 cm‘1 line shifting up to ”348 cm"1 and becoming vibrationally coupled with the 345 cm“1 line (assigned as 2 \’35) upon H-bonding. Vibrational coupling between V3 and 2 V35 modes has been previously suggested (Abe et al., 1978). In the high-frequency region, the formyl-related modes detected by isotopic substitution undergo frequency shifts uPon H—bonding to the carbonyl oxygen. Modes at 1586 cm-1 ( V2) and 1238 cm-1 are upshifted by 4 cm‘1 and 3 cm-l, respectively; and the formyl-hydrogen bend shifts up to 1405 cm'l. In the RR spectrum ofCu2+ porphyrin (Figure 3.12), a 10 cm‘1 upshift is detected in the frequency of V13. This Upshifted frequency agrees with the direction of shift observed for the IR—allowed V42 Eu mode upon 16O——-->180 substitution (Figure 3.8), and accordingly a marked contri— bution from the VCb-cao stretching motion to the V13 Rama The bone forn mode difi spec tior this to T aque pher is t equa and This hYdI H fc terc dout elon 118 Raman active mode is consistent with the Raman and IR-data. The sensitivity of v2 (Cb—Cb) to isotopic and hydrogen bonding effects at the formyl group of heme g and its formyl-containing derivatives stems from its pyrrole-ring mode composition. Choi et al. (1983) interpreted frequency differences in V2 between heme g and protoporphyrin as specifically due to the heme a formyl-vinyl trans disposi— tion. The data presented here appears to be consistent with this interpretation. -Of interest in the evaluation of formyl modes sensitive to H—bonding is the Vc=o stretching. For ferric heme g-in aqueous detergent, Vc=o shifts down by 17 cm-l, with phenol-OH/-0D the downshift is 21 cm"1 and when p-Cl—phenol is the donor the shift is 25 cm'1 (see Table 3.1). The equal shifts in the Vc=o frequency upon use of phenol-OH and phenol-OD indicate a weak H-bond (Singh and Wood, 1969). This interpretation predicts no isotopic perturbation to the hydrogen bond distance (R0...o) as a result of changing the H for D and indicate that the motion of the proton and deu— teron is approximately harmonic. For this type of H—bond, a double minimum potential surface (with high barrier) with an elongated RIO...O), has been proposed (Emsley, 1984). IV. fon 4-v. alh cal the: data subs and and Heme unde subs ~CHC Sine Obse inte acid in DPeS the Oxid P631 att 119 IV. Conclusions: The evidence presented here in the identification of formyl modes in the RR and IR spectra of heme a and its 4-vinyl-8-formyl Cu2+ derivatives indicates that several IR- allowed Eu-modes appear to be split, owing to the asymmetri- cal disposition of the vinyl-formyl groups, and that some of these vibrations become Raman active as a direct consequence of the strong asymmetry of the porphyrin macrocycle. Our data from the isotopic studies of the formyl modes indicate substantial coupling between the C=O stretching frequency and the \tb-cuo ring-formyl stretching. We have used RR and infrared spectroscopies to investigate this proposal. Heme a and Cu2+ porphyrin a formyl modes have been studied under the influence of H—bonding formation and isotopic substitution at the H—donor. Our results indicate that the -CHO...H/D H—bond interaction is a weak H-bond interaction since no isotopic effect in the formyl vibrations was observed. On the contrary, the strength of the H-bond interaction increases for a series of phenol donors with acidic pKa values. This was evidenced by the increase in Vc=o and red—shifted a -band. These observations are presented as factors that might influence the strength of the cytochrome a...protein interaction in cytochrome oxidase, since changes in the pKa of the formyl surrounding residue are eXpected to bring about conformational changes at the H-bonded site. For heme a, Cu2+ porphyrin a, and Cu2* acti‘ effe< This upshf benzz ring- 120 Cu2+ porphyrin, the V13 Raman mode displays considerable activity towards formyl isotopic and hydrogen-bonding effects as evidenced by its 10 cm'1 upshifted frequency. This upshifted frequency is consistent with a similar upshift observed in the Vcb-cno...u/n motion of benzaldehyde, and suggests a similar contribution of this ring-formyl stretch to the heme a V13 RR active mode. Vi CYt mum its a c oxi. H-b. iro: mecl H‘bl born of . to . I‘es: Chaj com: thrc dis: int: undc CHAPTER IV Visible Excitation Resonance Raman Spectra of Cytochrome a in Cytochrome Oxidase. I. Introduction: The unusual spectroscopic properties of cytochrome a in cytochrome oxidase, its red-shifted cx—band absorption maxi- mum and altered carbonyl stretching frequencies relative to its isolated heme a models, have been interpreted as due to a cytochrome a hydrogen bonded formyl group in the in situ oxidase (Babcock and Callahan, 1983). The strength of this H—bond interaction was observed to be dependent upon heme iron redox state and accordingly a redox-linked proton pump mechanism was proposed which relies on a change in the H-bond geometry. Due to the mechanistic implications of this hydrogen bonding interaction in the functional and structural aspects of cytochrome a in cytochrome oxidase, it is of importance to identify other cyt a formyl sensitive vibrations in the resonance Raman spectrum of cytochrome a2+. In the previous chapter we developed a systematic approach with heme a model compounds, with the aim of uncovering formyl sensitive modes throughout the Raman spectrum of heme a that can be used to disentangle other hydrogen bonding sensitive modes in the intact chromophore. We have studied the isolated heme a under the influence of formyl isotopic substitution with the 121 goal nal resu imid PUFF vibr memb the inte Rama 1983 agre. 0Xid: CYt ; to t] heme to 81 pH‘dt 122 goal of identifying and separating formyl modes from inter— nal porphyrin macrocycle vibrations. We have also discussed results obtained after placing the isolated heme a (bis— imidazole) model under hydrogen bonding conditions with the purpose of establishing trends of formyl hydrogen bond vibrations and the strength of this interaction in the membrane bound chromophore. In this chapter, we will use the results developed with these heme a model compounds to interpret the cytochrome a2t visible excitation resonance Raman (RR) spectrum, studied under H/D exchange conditions. To isolate the vibrations of cytochrome a from those of cytochrome as, we have used visible excitation resonance' Raman spectroscopy, in resonance with the Q(o-o) n——> nx transition of reduced cyt a2+ at 605 nm. With this laser excitation, the vibrations of cyt a2+ are isolated from those of cyt a32+, owing to the dominant contribution of cyt §2* to the 605 nm <1—band (Vanneste, 1967). Previous work in the visible excitation range of cytochrome oxidase has been reported (Bocian et al., 1979; Callahan and Babcock, 1983). The results from these two separate investigations agree with the prediction that the visible spectra of oxidized, reduced and mixed-valence oxidase is dominated by cyt a modes and that these appear to be largely insensitive to the ligation and spin—state transformations at the cyt a3 heme. Babcock and Callahan (1983) used these observations to study the sensitivity of the cyt a vibrations to alkaline pH-denaturation effects. Their results revealed significant pert rais for cyto vati heme solv depe mode beha vibr valu in t caus é an rest sens II. expe 15 u 123 perturbations of the cyt a visible spectrum as the pH was raised form 7.4 to 10.5, which was interpreted as evidence for the occurrence of a cyt a hydrogen bonded structure in cytochrome oxidase. Our present investigation extend these original obser- vations. We have studied the RR visible excitation spectra of cyt a2+ under the influence of small perturbations at the heme a hydrogen bond induced by the presence of deuterated solvents. We have also studied the depolarization ratio dependence of cyt a vibrational modes and found that a few modes exhibit characteristic anomolous polarization behavior, while the great majority of the visible excitation vibrations appear as polarized modes, approaching a (3:0.33 value. This is interpreted as due to significant lowering in the molecular symmetry of the heme a porphyrin macrocyle caused by the unusual peripheral substituent pattern of heme a and strong cyt a ... protein interactions, that appear to result in the additional enhancement of Eu substituent sensitive modes and low Q values. II. Materials and Methods: A. Materials: Cytochrome oxidase from beef heart muscle was isolated by a modified Hartzell-Beinert procedure (Babcock et al., 1976). The oxidase concentration for visible excitation experiments was maintained at 150 tJM (heme a basis), and 15 MM for the optical absorption. The buffer-detergent sys lau cor (Gl pro pro buf (NH got and oha M E was ven ted CYC add the prc thc Sig Che 124 system consisted of 50rnM HEPES (pH (or pD)=7.4) and 0.5% lauryl maltoside. The pD of the deuterated buffer was corrected to pD=pH + 0.4 prior to the Raman experiment (Glasoe and Long, 1960). Deuterium exchange in the resting protein was carried out by the following experimental protocol: cytochrome oxidase was reprecipitated in D20- buffer by increasing the concentration of saturated (NH4)2SO4 (pD 7.4), the pellets were obtained by centrifu- gation at 19K for ~30 minutes, resuspended in D20/maltoside, and dialyzed for six hours at 4° C against two volume changes of a medium consisting of 50 mM HEPES (pD 7.4), 0.01 M EDTA, 0.5% lauryl maltoside in D20. This entire procedure was repeated twice. To prepare the mixed-valence oxidase in deuterated sol- vents, the resting non—inhibited protein was either incuba- ted in D20—buffer for approximately 2—3 days and/or redox— cycled (reduced and reoxidized in D20—buffer), prior to the addition of 10 mM cyanide. We have previously observed that the addition of cyanide to the resting freshly prepared protein in DzO/buffer does not yield appreciable changes in those cyt a modes with substantial formyl character. The significance of these observations will be discussed on Chapter VI. dii Chz pr< Rhc pal anc abs 3 I 231 II. We 125 B. Instrumentation: The Raman spectrometer, detection system and Raman difference apparatus employed in this study are described in Chapter II. Excitation in the visible absorption band was provided by a Spectra Physics Model 375 dye laser with Rhodamine 66 dye pumped by a Spectra Physics Model 165 Ar+ ion laser. Power incident at the sample was typically 80— 105 mW. The formation and integrity of the reduced and partially reduced oxidase derivatives were monitored before and after the visible excitation experiment by using optical absorption and Soret excitation RR spectra. The Soret excitation lines were provided by a Spectra Physics Model 164-11 Kr+ ion laser equipped with a high—field magnet, a Coherent Kr+ ion laser Innova 90K, and the 441.6 nm line of a Lyconix He—Cd laser. The method used to measure depolari- zation ratios is outlined in Chapter II. II. Results: A. Visible Excitation Resonance Raman Spectra of Cytochrome a2* iN H20/D20 Buffers: The 605 nm excitation resonance Raman spectra from 1000—1700 cm'1 of reduced cytochrome oxidase in protonated and deuterated buffers are depicted in Figures 4.1 and 4.2. Figures 4.3 and 4.4 show the corresponding low—frequency RR spectra (150—1000 cm'l). The difference spectrum (i.e., 126 Figure 4.1: Visible excitation (605 nm) resonance Raman spectra of reduced cytochrome g oxidase in: A) H20 and 8) D20 buffers in the region from 1000 to 1300 cm_1. The Raman difference spectrum (H20 - D20) is shown in trace C. Oxidase concentration is 165 /uM (heme a basis). Fre- 1 quency shifts on resonance Raman lines at 1228 cm- and 1 1 1173 cm- of +5 cm- and +7 cm'l, respectively, were cal- culated from the Raman difference spectrum. Modes at 1045 cm'1 and 1218 cm‘1 are attributed to D o. 2 RAMAN INTENSITY 127 B. 020 1300 a-_ _u 5: 2&1 HE IS: - I 8: ET 3:1 - ...:T e 237 a m ..m 5 ..l x o l ) 8.: 0 6 B 0 d _ 331 e m M m e . \A c .w RI 1000 >._._mZM:ZH 2324‘”. FREQUENCY SHIFT (cm"> 128 Figure 4.2: Visible excitation (605 nm) resonance Raman spectra of reduced cytochrome g oxidase in: A) H 0 and 2 B) D20 buffers in the range from 1300 to 1700 cm-1. The Raman difference spectrum (H20 - 020) is shown in trace C. Oxidase concentration is 165 /“M (heme a basis). Frequency shifts in modes at 1310 cm'l, 1329 cm-1, and 1615 cm'1 are detected. Increased intensity in the 1545 - 1 . . . . cm Vibration 18 also eVident. 129 em “.11.. MM XG o 0 dx “O “\A d e R C.(A- B) B. I320 1700 l250 >.:w2m._.ZH 2524K FREQUENCY SHIFT 1cm") H20 or the sum deu mot f re beh whi cm- are con The Oval Str Spe Oha fre 130 H20—D20) is shown at the top of each figure. No smoothing or background (baseline) subtraction was performed in any of the RR spectra shown in this chapter. Tables 4.1 and 4.2 summarize the frequency changes detected upon protein deuteration and the corresponding suggested normal modes of motion; mode depolarization ratios are also included. High frequency vibrations exhibiting anomalous polarized (ap) behavior are seen at 1585 cm‘l, 1336 cm'l, and 1305 cm'l, while depolarized (dp) vibrations are at 1131 cm'l, 1173 cm'l, 1167 cm'l, and 1569 cm‘l. The remaining vibrations are polarized (~ 16 modes) with O values approaching 0.33, consistent with the lower symmetry of the heme a chromophore The anomalously polarized Azg modes, particularly V19, show values between 0.75 and 1.0 rather than infinite (Spiro and Strekas, 1974), a phenomenon also observed with heme a {spectra (Babcock et al., 1979; Woodruff et al., 1982). B. Effects of deuterium substitution for cyt al+ modes above 1000 cm'l. Deuterium induced frequency shifts and/or intensity Changes in cyt a2+ vibrational modes are observed for high- frequency visible excitation lines at 1068 cm'l, 1173 cm-1, 1228 cm-1, 1310 cm-1, 1329 cm‘l, and 1546 cm‘l. With the exception of the latter mode, which does not change in frequency position, the other listed vibrations shift to 1065 cm—1, 1180 cm-1, 1233 cm'l, 1312 cm‘l, and 71333 cm'l, respectively. The resonance Raman spectra of the mixed— Tabl Tent Mode 2V33 131 ble 4.1. sntative Assignments for the Observed High—Frequency Modes in the Resonance Raman Spectrum of Cyt a1*.(‘) de: Symmetry: Assignments:b (ll/I” ) a2*a32+ 0 H20 D20 cm-1 ——- A1g thb—S:38)° 0.25 1040 1040 ‘0— a. Aig équ-cb-cb;28) 0.41 1066 1068 +2 V(Cb-Cb;15) 0.32 1089 1089 0 4 Eu V(Cb-S;29) 0.25 1115 1113 —2 V(C —Cb;26) 2 A2g Vlcb—s;28) 0.53 1131 1133 +2 2 1151 1155 +4 1165(sh) 3 Eu 0(cb—ca=cB) 0.65 1173 1180 +7 0 Bzg V(cb-s;50) V 0.63 ——-~ 1168 . V(cb—s;30> 2 1185 1184 (sh) (sh) 3 B.g V(Cm—H;67) 0.40 1228 1233 +5 V(Cb-CHO) 1 Azg 5(cm-H;53) 0.97 1304 1307 +3 V(Cb-Cb;18) 6CH: dp 1310 1312 +2 68:08. ap 1329 1333 +4 1336 1337 +1 W1 ho“ W9 he a) R and thOS equa waVe inte blN SUgg and c) T Dare 00mp d) T This 0:0. anOm 132 En V(Ca-N;50) 0.34 1367 1367 O 1 Azg? \KCb-S;24) 0.40 1389 1391 +2 Bzg WOOL—05:47) 0.56 1397 1398 +1 V(C5-Cb:26) 0.36 1432 1431 —1 Bzg \4Ca-Cm;52) 0.32 1469 1467 —2 2 1490 1490 (sh) (sh) 1e) 8.. v(cb-cb;57) 0.45 1518 1520 +2 v(Cb—S;16) Eu \hcb—cb;53) 0.25 1545 1545 0 \hcb-s;16) A2. \4Ca—Cm;36) 0.70 1569 1569 0 Eu \4Ca-Cm;67) 0.65 1586 1586 0 8., ltcu-cb) (f) 0.46 1615 1612 —3 esults obtained with Q—state excitation at 590, 595, 602 F605 nm wavelengths. The data reported in this Table are e observed with 605nm excitation. These results were lly reproduced with the different excitation lengths; however, many of the modes display an nsity—wavelength dependence. ormal mode numbering and assignment followed those ested by Abe et al.(1978); Choi et al.(1982); Willems Bocian (1984 and 1985) and Lee et al.(1986). he calculated potential energy distribution (PED) (Abe l. 1978) for each of tabulated modes is shown within nthesis at the right—hand side of the normal mode osition assignment. he assignment of V20 at 1389 cm‘1 is still uncertain. is due in part to the fact that the proposed mode shows 40 instead of o>0.75 as would be the expected for an alously polarized mode (Azg). Cont e) l is i the as c gr01 Porl with f) l freq 133 inued Table 4.1: he assignment of‘dl at 1518 cm'1 with visible excitation n agreement with Choi’s (1983) under Soret excitation, lower frequency of this mode as compared with more (etric porphyrins (i.e.,NiOEP\q1=1576 cm‘l) was explained .ue to the asymmetric disposition of the vinyl-formyl (ps in heme a, which results is an alteration of the )hyriniT system and hence lowering of the‘fll frequency . respect to the more symmetric (D4h) porphyrin. 'his mode might overlapp with the carbonyl stretching (uency expected at 1612 cm‘1 . shoulder depolarized polarized too weak to be measured. -otsp : 134 ralence protein in HzO/Dzo buffers resulted essentially in Ldentical features as those observed for fully reduced protein, and therefore are not reproduced here. The highest frequency polarized band at 1615 0111'1 seems to shift down to 1610 cm'1 as judged by close examination of the RDS, which displays a negative trough at 1610 cm'1 and a weak (positive) peak at 1616 cm'l. The frequency position of this 1615 cm'1 mode is of considerable interest since it falls in the region where the following cyt §2* modes are expected: vcso at 1616-1612 cm-1 (Babcock and Callahan, 983); Vcsc at 1622-1625 cm'l; V10 (Ca-Cm) at 1620 cm"1 Callahan and Babcock, 1981; Choi et al., 1983) and 0.. (Ca-Cm) at "1607-1610 cm‘1 (Choi et al., 1983; Babcock, 1987). Our current data cannot provide an unambiguous issessment of the origin of this high frequency vibration; tonetheless, it is noticed that in our series of experiments rith the mixed—valence oxidase, a 4 cm-1 downshifted fre- ,uency was also observed upon H/D solvent exchange, which mplies a possible formyl contribution to this mode. The 1546 cm-1 line appears to increase in intensity for euterium exchanged protein. This mode has been assigned to s, an Eu mode with substantial contribution from porphyrin aripheral substituents (Abe et al., 1978; Babcock, 1987). ,sible excitation spectra of low-spin heme §(bis-imidazole) mplexes also displayed a similar band at 1544 cm"1 abcock et al., 1979; Kitagawa et al., 1977). A comparison our data with those obtained by Callahan and Babcock 983) on the alkaline pH treatment of reduced cytochrome idase appears to indicate that cyt a2+ modes observed to altered upon raising the pH are also sensitive to euterium solvent atoms, and hence to H—bonding effects at 1e cyt a H-bonded structure. Additional cyt a2+ hydrogen bonding sensitive modes are pected to appear in the mid-frequency range of the visible citation RR spectrum, from 1200—1260 cm-l, as evidenced by e heme a model compound data discussed in Chapter III. In rticular, the ring-formyl stretch in heme a(bis-imidazole) Cu2+ porphyrin a has been assigned at ~1235-1238 cm'l, ifting up to 1243 cm-1 upon H—bond formation (see Figure 11, Chapter III). In the RR spectrum reported here, we served a band at 1228 cm-1 for the reduced protein in H20, ich shifts up to 1233 cm'1 upon use of deuterated buffers. RDS illustrates these changes by displaying peaks at 6 cm'1 (+) and 1238 cm"1 (-). A shoulder at 1218 cm"1L is sistently observed upon use of D20-buffer solutions; ever, this line may be properly assigned to a D20 tribution since control experiments from the D20-buffers w a strong and broad D20 peak at 1207 cm-1 as well as a k at 1043 cm-1. The remaining high frequency cyt a2+ deuterium sitive modes are observed modes at 1068 cm’l, 1173 cm-1, 0 cm'l, and 1329 cm'l. In cytochrome oxidase, the arized mode at 1068 cm-1 is only active with Q—state itation, and its suggested assignment (i.e., 2 V33 2g VCa-Cb-Cb) implies contribution from the porphyrin eripheral substituents. Its RR enhancement has been uggested to originate from porphyrin low-symmetry nvironments (Willems and Bocian, 1984). The shifted requency of the 1173 cm-1L band in the deuterated protein ppears to be accompanied by the enhancement of a new mode t 1168 cm-1, presumably V30 B2g\%Cb—S) (Choi et al., 982; Lee et al.. 1987). Comparison of these results with vailable RR assignments of heme model compounds with vinyl nd formyl groups as the peripheral substituents (Tsubaki et 1., 1980; Choi et al., 1982a,b), suggest the enhancement of inyl modes in this region. Low—Frequency Cyt a’* Vibrations Below 1000 cm'1 The low frequency visible excitation RR spectra of cyt * are shown in Figures 4.3 and 4.4. In Table 4.2 we mmarize these vibrational modes along with their suggested signments and measured depolarization ratios. No anom- ously polarized Raman lines were observed in these low equency spectra, but depolarized behavior for lines at 298 -1’ 747 cm-1, 789 cm-1, and 981 cm"1 were detected. Upon e of deuterated buffers, frequency changes are as follows: 5 cm"1 shifts down to 242 cm'l, 265 cm"1 to 262 cm‘l, 298 ‘1 to 295 cm-1, 344 to 342 cm‘l, 439 to 432 cm“1, 659 to 6 cm-l, 789 cm-1 to 785 cm*1, 955 to 951 cm-1, and 981 to 6 cmsl. Intensity differences are also noted for modes at 137 Reduced 0x idase an. 605 nm 441 A.H20 Ln q- T 432-- __--—- -—298 —N2____ B. 020 -—262 --44l ——181 -—295 372 .7531. .. - - 150 550 FREQUENCY SHlFT 16m") igure 4.3: Low frequency visible excitation (605 nm) resonance Raman ectra of fully reduced cytochrome oxidase in A) H90 and B) D 0 buffers the range from 150 cm‘1 to 550 cm‘1 . The top thace corres ond to C Raman difference spectrum. 138 C.) (HzO‘Dzlel 1 66) l 962 g‘Igg” 951 U A exc. 605 nrn ) A.) 11,0 5‘,” 633 659 l 786 — - r l j 50 640 730 820 910 1000 FREQUENCY SHIFT (cm") lgurwa 4.4: Low frequency visible excitation (605 nm) RR spectra reduced cytochrome oxidase obtained with the Raman difference Ltrumentation. le 4.2. 139 ntative Assignments for the Observed Low—Frequency Modes 1 the Visible Excitation Resonance Raman Spectrum of Cyt am2+(a) e Symmetry Assignments”)C (II/IN) Alg Eu Blg Alg Eu Alg <9Cb-S;23) (hob-s;43) 1233 cm-1 bands of cyt a2+ have also been recently investigated by Soret excitation on reduced and mixed-valence cytochrome oxidase (Argade et al., 1986; Centeno and Babcock, in preparation; see also Figure , Chapter VI). With B-state excitation, we noted upshifts for modes at 1228 cm‘1 and 1247 cm'1 to 1233 cm-1 and 1250 cm'l, respectively. With Soret-excitation, the 1247 cm-1 mode is assigned to a 1 05 + V9) combination mode which might not be strongly enhanced with Q—band state excitation owing to its totally symmetric character (i.e., A12). The 1228 cm"1 line, on the other hand, is assigned to V13 Big (Abe et al., 1978), and accordingly, its strong activity with Q~band excitation is expected (Spiro and Strekas, 1972; Shelnutt, 1978). The 147 derivative shape of the Raman difference spectrum shown in Figure 4.1 (C) is in fair agreement with a recently published RDS spectra obtained with 441.6 nm excitation (Argade et al., 1986) that shows a positive peak at 1236 cm'1 and negative troughs at 1223 cm'1 and 1249 cm'l. Our cyt a2+ frequencies reported here might differ from those of Argade et al. (1986) owing to the selective enhancement of cyt a2+ vibrations with visible excitation as discussed B. Low Frequency Vibrations of Cyt a2+ Low frequency resonance Raman vibrations of heme proteins are particularly sensitive to heme structural changes and enviromental interactions, porphyrin macrocycle deformation modes (Choi et al., 1983), metal—ligand stretches (Spiro, 1983), and peripheral substituent deformation motions (Choi et al., 1983). In addition, RR enhancement of IR-allowed Eu-type modes, which attain :resonance Raman activity due to the low-symmetry molecular (environment of the porphyrin ring, has also been observed (Willems and Bocian, 1985). This symmetry reduction in heme model compounds and heme proteins has been attributed to the asymmetric disposition of peripheral substituents as well as to protein—chromophore interactions (Choi et al., 1983; Valance and Strekas, 1982; Callahan et al., to be submitted). 148 The effects of symmetry lowering on the RR vibrational modes of metalloporphyrins have been well documented (see Woodruff et al., 1985; Choi and Spiro, 1983; and Ozuki et al., 1979). For metalloporphyrins exhibiting D42 symmetry, the expected active vibrations are those contained in the cross product of Eu x Eu : Alg + A2g + Blg + B2g symmetry; see Table 4.3 (Spiro and Strekas, 1974). Reduction of the heme molecular symmetry to either D22 or Czn point group symmetry is expected to transform the above symmetry representations according to Table 4.3 (Wilson et al., 1955). Therefore, the appearance of totally symmetric RR bands (9g 1/3) can be predicted with any of the point group symmetry transformations described in Table 4.3. The fact that the low frequency visible excitation spectrum of cyt a2+ (Figures 4.3 and 4.4) diSplays strongly polarized modes at 343 cm-1, 681 cm-1, 960 cm‘l, as well, as the unique doublet with frequencies at 735 cm“1 and 747 cm'l, is certainly an indication of the lower heme g ring symmetry in cyt a. In addition, the strong electron withdrawing capabilities of the H-bonded carbonyl group in OYt g2f is also expected to affect the symmetry of the heme ring, allowing the above spectral effects to be observed. The low frequency cyt §2* modes sensitive to protein deuteration effects lie at 245 cm'l, 265 cm‘l, 298 cm'l, 439 cm‘l, 659 cm-1, 789 cmrl, 955 cm'l, and 980 cm'l. Their estimated potential energy distribution (PED), as revealed by the normal coordinate analysis of Abe et al. (1978), 149 Table 4.3: Correlation Table for Species with D4h, D2h, and Czh Point Group Symmetrya C211 D411 D211 Alg >Ag Blg A2g >31. B2g B2g Eg < BBg Eg --W~~ — s _, }B2g,B3g Ag o(- N W Soret A2u Blu Alu Alu D4h D2h Eu x Eu : Alg + A2g +B1g +BZg a) From Wilson et al. (1955). 150 suggest a large contribution from the heme peripheral substituents, such as the vibrations at 298 cm'1 and 789 cm-1 modes, which exhibit an estimated 84% and 50% Cb-S character, respectively. Structural information on the cyt a peripheral formyl group can therefore be inferred by monitoring vibrational changes in the above listed hydrogen bonding sensitive modes. V. Conclusions: In this Chapter we have studied the resonance Raman spectra of reduced cytochrome oxidase obtained with visible excitation wavelengths that are in resonance with the cyt ng Q(0-0) n -——-> hi transition at 605 nm. Under these resonance conditions the spectra obtained are attributed to the cyt azt heme moiety. The identification of cyt a hydrogen bonding sensitive modes has been established by studying the in situ cyt a chromophore under the influence of deuterated buffers. Frequency shifts on many of these cyt a (formyl) H—bonding sensitive vibrations agree with previous visible excitation studies on the pH—dependent spectral shifts of reduced cytochrome oxidase, and confirm the H-bonded structure of cyt a. The enhancement and frequency perturbations on many of the cyt a low frequency modes implies considerably lower vibrational symmetry, which in part arises from the strong electron withdrawal capabilities of its H~bonded C=O group. Finally, this study is presented as an alternative approach to be used in the 151 interpretation of the structure and enviromental conditions of cyt g in cytochrome oxidase from different prokaryotic and eukaryotic systems. CHAPTER V Hydrogen Bond Sensitive Vibrations of Cytochrome g in Cytochrome Oxidase I. Introduction: Formyl (—CHO) hydrogen bonding sensitive vibrations of cytochrome a (cyt a) in cytochrome oxidase have been identi- fied in reduced and mixed-valence oxidase derivatives by controlled denaturation of the enzyme under acidic condi- tions.- The effect of lowering the pH of reduced cytochrome oxidase is to transform the cytochrome g3 (cyt g3) heme to a six-coordinate low spin state and to disrupt the hydrogen bond structure to the cyt g formyl peripheral substituent, resulting in diminished resonance Raman intensity for some cyt g low frequency modes. A comparison of these cyt a hydrogen bonding vibrations with those studied for heme a (N-MeIm)2 and/or benzaldehyde in Chapter III, under H—bonding conditions, results in a nearly complete assess- ment of the cyt a aldehyde vibrations. As with the visible excitation spectra (Chapter IV), we have also detected sig- nificant intensity enhancement of substituent sensitive IR- allowed Eu modes in the Soret excitation spectra of H—bonded cyt a. This is interpreted as a heme a symmetry lowering effect in cyt a induced by the strong electron withdrawing capabilities of its -HC:O group at position 8, and its hydrogen bonding interaction with the protein polypeptide backbone. Our results provide further information about the molecular and protein environment of cyt a and support the hydrogen bonded structure proposed for this chromophore in cytochrome oxidase (Babcock and Callahan, 1983). II. Materials and Methods: The isolation of cytochrome oxidase and its inhibitor complexes is described in Chapters II and IV, respectively. The buffer—detergent system and the procedures to obtain the desired pH level are given in Chapter II. The Raman instru- mentation and optical absorption spectrophotometer used are described in Chapter II. Some of the resonance Raman spec- tra reported here were recorded with Raman difference instrumentation (see Chapter II for details). III. Results: A. Low pH Effects on the RR Spectra of Reduced Cytochrome Oxidase The optical absorption spectra of reduced cytochrome oxidase at different pH levels are shown in Figure 5.1. The resonance Raman spectra of these pH—treated oxidase samples in the high— and low—frequency regions are depicted in Figures 5.2 and 5.3, respectively. As the pH was lowered from 7.4 to 4.5, the following optical absorption changes were distinguished: 1) in the Soret region, the absorption maximum was blue shifted from 443 nm to 434 nm, i.e., to higher energy than the Soret maximum of heme §2*(N—Melm)2 in protic solvents (see Figure 3.1 Chapter III); 2) in the Visible absorption region, we observed a blue shift of the 1 l L | 6'03 2+ 2+ - 93 59:.) ‘04 r r 599 8 ._ pH 7-4 ' c ....-pHs.5 / (l o / ' .9 ° _ _ pH4-5 I )1 - 0'3 0 I \ .8 I ‘ <1 %% .l \1 ,1, / l '02 .//\°—-.’__,é 1) ., .1 ( IO . -O-l \'\o 1 g l L J 1 \~d 400 450 500 550 600 650 Wavelength (nm) Figure 5.1: Optical absorption spectra of reduced cyto- chrome g oxidase at neutral and acidic pH. Enzyme concen- tration was approximately 25 uM (heme 3 basis) for all samples. 155 maximum from 603 nm to 593 nm, as well as a shift of the 518 nm band to 510 nm. These absorption changes produce a species that resemble low-spin, six-coordinate bis-imidazole ferrous heme a (absorption maxima at 438 nm (Soret), 518 nm, and 594 nm (cx-band)). The high-frequency resonance Raman spectra of these pH treated oxidase samples, obtained under 441.6 nm and 406.7 nm excitation are depicted in Figure 5.2 along with the RDS spectrum for 441.6 nm excitation. Vibrations that decrease in intensity upon lowering the pH of the reduced oxidase solution are: bands at 1665 cm"1 (cyt §32* v2.8), 1612 cm'1 (cyt 22+ vcso), 1569 cm'1 ( V33, cyt g2+ or V2, cyt gazi), 1468 cm“1 ( V29, cyt a32+ Ca—Cm), 1289 cm‘l, 1228 cm'1 ( V13), and 1115 cm-1. As the intensities of the 1665 cm'1 and 1612 cm'1 lines diminish, the corresponding Raman intensity of the 1638 cm’1 and 1586 cm"1 lines, increases. Frequency shifts are also noted for modes at: 1396 cm‘1---> 1393 cm'l, 1330—-—> 1334 cm'l, 1246—--> 1243 cm-1, and 1228— --> 1226 cm-1. These changes are readily evident from inspection of the RDS spectrum (RDS : pH 7.4 — pH 5.3) shown in the bottom trace of Figure 5.2. No frequency changes were evident in V4 at 1354 cm'1 as revealed by the RDS (trace 5.2D), suggesting that the acid induced changes in Either cyt a and/or cyt 93, may be confined to structural alterations at the heme periphery and in the spin state. The difference spectrum in the carbonyl region of cyt a and cyt Q3 display positive peaks at 1612 cm‘1 and 1665 cm"1 and 156 Figure 5.2: High frequency resonance Raman spectra of re- duced cytochrome oxidase at the indicated acidic pH level, and under the indicated excitation wavelengths. The Raman difference instrumentation was used to record the pH 7.4 - 5.35 spectrum under 441.6 nm excitation. Enz)’me concentration was approximately 50 /uM (heme 3 basis). -1 Instrumental conditions: resolution , 5 cm ; time constant, 1 second; scan rate, 50 cm-l/minute. 157 xoxc 406.7 nm 32) «ST 82) oampll one... moo.) _ ) (cam— 82 «ET can. .. m 4 5 mm .... p w 39) one.) 39.. moo.) % 12 2&7 . I I u m m 38.. near. ..va (new. a“ 1” mm m M1 c 5:. - «w ... h «x (a Cl. . A o m >._._m2m._.ZH 23242... I FREQUENCY SHIFT (cm') 158 a negative trough at 1638 cm-1, indicative of a hydrophobic environment for both heme a formyl groups at pH 5.3 (Van Steelandt-Frendrup et al., 1981; Babcock and Callahan, 1983). In the mid—frequency region, the RDS has positive peaks at 1226 cm'1 and 1249 cm“1 with a negative trough at 1236 cm‘l. The assignment of heme substituent normal modes of motion contributing to this mid—frequency region is of importance since it is in this region where the ring—formyl stretching mode of benzaldehyde and heme §2*(N-Me1m)2 models have been assigned (1206 cm'1 and 1230 cm'1 respectively; Zwarich et al., 1971; Choi et al., 1983). Our isotopic substitution and H-bonding studies on heme 3 models indie cated that upon formyl H—bond formation this mode will shift up by ~6—8 cm'1 (see Figure 3.10 Chapter III). Therefore, removal of this H—bonding interaction should drop the fre- quency of this mode, provided that no significant coupling with other formyl modes occurs (see Discussion). This is indeed shown by the RDS (trace 2D) in Figure 5.2, which displays a negative trough at ~1236 cm'1 and a positive peak at 1225 cm-1. Figure 5.3 shows the low—frequency RR spectra of redu— ced oxidase under the same pH conditions. Low-frequency modes at 215, 268, 367, 395, 438, 633, 714, and 850 cm-1 are seen to decrease in intensity, while modes at 342 cm*1, 750 cm-1, and 683 cm-1 shift down to 337 cm-1, 746 cm-1, and 679 cm-1, respectively. The former two modes are correlated with spin-state transitions at the cyt as iron as the g3 159 Figure 5.3: Low frequency resonance Raman spectra of reduced cytochrome oxidase at the indicated pH levels. The bottom trace was obtained with the Raman difference instrumentation and 441.6 nm excitation. Experimental conditions as in Figure 5.2. .138 6.45 pH 5 35 m (i a... .W «nal e a i x O a n m ....s .. 1 «a ... mm) tT I Y M New as .... .2 mm 9 ecu... .m \A R «3.1 «we 1 man! «3) (7.4 - 5.35) «B I «no I >._._me._.ZH 26.3.49. 3:. FREQUENCY SHIFT (cm") 161 changes from high—spin five-coordinate with lines at 215 cm'l, ”277 cm'1 (RDS), 327 cm‘1 (RDS), 365 cm‘l, 404 cm"1 (RDS), 683 cm-1, and 757 cm-1 (RDS) to low—spin six- coordinate with lines at ~243 cm-l, 344 cm-1, and 678 cm-1, and 746 cm-1. The low-frequency resonance Raman difference spectra of reduced and mixed-valence cytochrome Oxidase (Figure 5.3 trace D, and Figure 5.4 trace C) reveal all these cyt as changes as the pH is lowered to 4.5. Further indication of spin transformations at the cyt g; are given by the shift of its vcso stretch from 1665 to 1638 cm'l, and decreased intensity for modes at 1469 cm"1 and 1569 cm"1 (Babcock and Salmeen, 1979). Low-frequency cyt §2+ vibra- tions sensitive to acid pH denaturation are apparent in the RDS (Figure 5.3) at 344 cm'l, 397 cm‘l, 438 cm'l, 633 cm-1, 715 cm‘l, 850 cm'l, and 981 cm'l. A disruption of the hydrogen bonded formyl group struc- ture of cytochrome a should manifest itself as an increase in frequency of Vcao and an absorption blue-shift (Babcock & Callahan, 1983). These spectral shifts are indeed ob- served; the cytochrome a formyl stretching frequency at 1610 cm'1 (Callahan & Babcock, 1983) shifts to 1640 cm'1 at pH 4.5 and the Soret and G—band absorption maxima blue-shift by approximately 10 nm under the same conditions. Since the a—band absorption blue—shift and the Vc=o frequency UPShift are linearly related (Babcock & Callahan, 1983), this is consistent with the presence of weakly hydrogen bonding 162 environment for both heme g species at acidic pH. A struc- tural schematic describing the coordination change of cyt 332* and the formyl environment shift of cyt §2*, based upon the carbonyl frequency shifts of each chromophore, is given in Figure 5.8A. B. Low—pH Induced Effects in the RR Spectra of Cyanide Inhibited Cytochrome Oxidase The results described in Figures 5.1 through 5.3 clearly indicate that structural perturbations at both cyto- chromes a and as, induced by acidic pHs, are actually taking place in fully reduced cytochrome oxidase. To assign the cyt g aldehyde vibrations with confidence, it is desirable to deconvolute its RR spectrum from that of cyt 23° In Dhapter IV we demonstrated that such spectral separation is indeed possible by employing laser excitation in the visible region of reduced cytochrome oxidase in resonance with the -band of cyt a2+ at 605 nm. Since the contribution of cyt 12* to the (x-band of reduced oxidase is estimated to be "90% (Vanneste, 1967), nearly complete spectral separation )f cyt a2+ vibrations is observed. On the other hand, when Soret excitation is employed, as in the present investiga- ;ion, an alternate method to separate the contributions of :yt g and cyt a3 requires the use of respiratory inhibitors, luch as CO, CN—, HCOO-, etc., that bind to the heme as iron, .hereby removing its absorption maximum from resonance with .he 441.6 nm laser excitation (Salmeen et al., 1978). almeen et al. (1978) and more recently Ching et al. (1986) 163 made use of this ligand—binding deconvolution method to~ investigate the vibrational properties of the mixed—valence, cyanide bound protein. In the present study, we have used the mixed—valence cyanide inhibited enzyme (i.e., af+ a33+- CN') to isolate the pH induced spectral changes of cyt a2+ (B max.= 443 nm) from those of cyt a33i—CN‘ (B .22.: 427nm). With 406.7 nm excitation, approximately equal enhancement of Vibrational mode for cyt a33+-CN and cyt a2* will occur (Babcock et al., 1981). In contrast, the 441.6 nm excita— tion RR spectrum of a2* a33+—CN' is expected to enhance essentially cyt a2i modes. Accordingly, a resonance Raman difference spectrum of (a2+ §32+ - §2* §33*-CN' = a32*) would yield the cyt 932* spectrum (Salmeen et al., 1978; Ching et al., 1986) and uncover the acid pH—induced changes at this chromophore. This is illustrated by the Raman difference spectrum displayed in Figure 5.4, where fully reduced and mixed-valence oxidase samples, at pH 6.35 t 0.01, are simultaneouly compared. At this pH the acid— induced changes at cyt as are not fully developed; . nonetheless, this Figure serve as useful indicators of those mode expected to be perturbed by acidic pHs at the as heme . site. Cyt a32* vibrations are seen at 1665 cm'l, 1569 cm‘l, ' and 1225 cm‘1 (Figure not shown) (Babcock et al., 1981) and i at 215 cm‘l, ”278 cm'l, 330 cm'l, 365 cm'l, 404 cm‘l, 682 cm'l, and 752 cm-1 (Figure 5.4). Similar cyt gazt frequen— : cies were observed in Figures 5.2 and 5.3 to be sensitive to ‘ acid pH denaturation effects. 164 Figure 5.4: Low frequency resonance Raman spectra of fully reduced (A) and partially reduced cyanide inhi- bited cytochrome oxidase (B) obtained with 441.6 nm excitation at the indicated pH levels. The difference spectrum (bottom trace) corresponds to the deconvoluted 2+ cytochrome g3 spectrum at pH 6.35. Oxidase concentra- tion was 45 ,uM (heme 3 basis). Cyanide concentration was 10 mM. The spectra are the average of 6 scans. exc .1 441.6 nm )t (w N) anmFZH z. 1.... (To Z 1.).) p. 2 H z < E < o: N m 2 l 2 pH 5.0 2'4. 6 In in M N a s: "F T a a ' GD M d d l I pH 4.5 f i— i‘ F 5 ‘fi F 6 fl I? 0 I500 l600 I400 I200 I000 FREQUENCY SHIFT (cm") Figure 5.7: High frequency resonance Raman spectra of mixed-valence cytochrome oxidase at the indicated pH levels. samples (see Table 5.3). Therefore, the acid induced shifts in the cyt §2* carbonyl frequency (1612 cm'1—->1640 cm'l) observed in both sets of data support the occurrence of the hydrogen bonded structure in cyt a. The direction of 06.0 shift in cyt a2* upon H—bond formation is the same as those observed for heme g (N—MeIm)2, Cu2* porphyrin g, and benzal- dehyde (see Table 5.1 and Figures 3.1 and 3.8 in Chapter III). The remainder of the cyt a2* hydrogen bond sensitive Vibrations can be assigned based upon our previous (-CHO) H-bonding assignments in benzaldehyde and heme a models. In these model compounds we observed upshifts for the in—plane formyl-proton ( 6.-co) vibration at 1388 (to 1395 cm-1) and in the Vcb-cuo stretching mode at 1206 (to 1212 cm'l) upon H—bond formation (see Table 5.2). The downshifts in modes at 1228 cm'1 (to 1225 cm’l) and 1396 cm‘1 (to 1393 cm-1) upon removal of the cyt a formyl H-bonding perturbation are consistent with our model compound data. Similarly, in the low frequency region, benzaldehyde and Cu2+ porphyrin dis— play modes at 441 cm'1 and 453 cm‘l, respectively (see Chap— ter III). In Figure 5.7 intensity changes are noticeable for cyt a low—frequency modes at 633 cm‘l, 438 cm'l, and 397 Cm‘l, which is observed upon disruption of the formyl H— bonding structure. These modes are assigned to out-of—plane vibrations (Abe et al., 1978; Choi et al., 1983) with contributions from the ring—peripheral substituents. Therefore, it appears reasonable to assume that one immedia- te structural effect of the cyt g formyl hydrogen bonded 172 Table 5.1: High Frequency Vibrational Assignments of Cytochromes g and g3 RED CYT OX pH 7.4 PRED CN— pH 7.4 ASSIGNMENT a2+(b) a32+(3) a2+ a33+-CN' v(0:0) 1665 1672 AV(0:0...11) 1612 1612 V10, Blg 1620 1610 1620 1640 V2, Alg 1586 1575 1586 1589 Vaa(1), Eu 1569 1569 V3, Alg 1491 1467 1493 1507 V4, Alg 1356 1356 1356 1373 a. While the cyt §32+ frequencies are enhanced with both 406.7 nm and 441.6 nm excitations, those of cyt §33*-CN' are observed with 406.7 nm (see text). b. High frequencies vibrations of cyt g2+ are enhanced with both 406 .7 nm and 441.6 nm. Table 5.2: High Frequency Aldehyde Group Vibration Assignments and Shifts Observed for Benzaldehyde in the Presence of Hydrogen Donors and Cytochrome g2+ at Neutral and Acid pH. Cyt $212 Ph-CHO C6H60 cracoon pH4.5 pH7.4 Assigna (cm—1) (cm-1) (cm-1) (cm-1) (cm-1) Assigg; v(C=O) 1707 1692 1681 1638 1612 V(C:O) (-15) (—26) I-26) V9,, ring 1588 1585 1585 ' (-3) (-3) 611.00 1388 1395 1403 1394 1396 68-00 (+7) (+15) (+2) (wk) 1312 ‘ 1315 1318 (+3) (+6) VPh—CHO 1206 1212 1214 1225 1230 \)Cb—CHO (+6) (+8) (+5) 1178 1185 \)Cb-C -C (+7) a 5 60:0 649 642 2 (-7) Ring Fold 441 441 438 438 Pyr. Fold 6(Ph-CHO) 225 2 aZwarich et al., 1971 174 formation is to cause a lowering in the effective molecular symmetry of the heme macrocyle (Woodruff, 1982), providing new routes by which IR—allowed Eu-type out—of—plane modes could become resonance enhanced. A description of these modes in the RR spectra of cyt a2* is given in Table 5.3. IV. Discussion: Our previous assignments of the hydrogen bonded formyl structure of cyt a have been based on the absorption red- shift of this heme center relative to low—spin heme g model compounds, and on its downshifted formyl stretching fre— quency. Because of this frequency downshift, the formyl group frequency overlaps with several porphyrin ring vibra- tions which makes small frequency shifts or perturbations difficult to observe. The band assignments given in the Tables are made with reference to the previous studies of Abe et al. (1978), Choi et al. (1983), and Willems and Bocian (1984). The pattern of vibrational shifts and their assignments confirm the hydrogen bonded structure of cyt a and allow small perturbations at the cyt g formyl group to be identified. To identify the cyt g hydrogen bond sensitive vibra- tions unambiguously,it is necessary to follow the vibrations and frequency shifts caused by changes in acidity for both cyt a and as or to enhance selectively the vibrations of cyt a. We have used both of these tactics by following the acid Table 5.3: pH Sensitive Vibrations (cm‘l) 175 ....u With 441.6 nm excitation. at Cytochrome a2+ Observed cyt §2* cyt a2+ (pred CN-) (red cyt ox) pH pH heme a2+ Mode Assign 7.4 4 5 7 4 4.5 (NmeIm)2 (aq) 263-—>265 272-—>262 342-->334 342—->335 335 08(41g) 6CbS 355-—>350 345-—>348 350 2v35 395—->392 394-->392 YCbs 420-—>415 420——>415 pyr fold 438-->440 438-—>440 pyr fold 564——> D 564——> D V49(Eu) CaCbe 583-->586 583-->586 585 V..(Eu) 6obs 632-->638 633-->636 500 710-->715 712—->715 V.7(Eu) 5056 785--> D 788-—> D V32(Bzg) écbs 822--> D 820-—> D V6(A1g) CaCmCa 849—-> D 850——> D YCmH 1113—> 0 1115-> 0 1113 w v.2(422) CbS 1185—>1178 1184-)1178 VaoIBzg) VCb-s 1228->1225 1220 (sh) VPh—CHO 1290—>1288 1293—>1289 1287 V42(Eu) 1330->1334 1330—>1334 —-- 53=CH2 1396->1394 1397->1393 1390 V20 +‘29IA22) 5H-CO 1569-> D 1569—> D 1555->1558 v3.,Eu 1612—>1638 1612—>1638 1645—>1630 V(C=O) Continued Table 5.3: 1622-) D 1622-) D -- v(C=O) a) Mode assignments from Abe et al. (1983). I -— Intensity increase D -- Intensity decrease W -- Weak (1978) and Choi et al., 177 This page have been omitted. 178 pH dependence of the RR spectrum of the fully reduced enzyme and of the partially reduced cyanide inhibited complex. The optical absorption and the RR spectra of reduced cytochrome oxidase established that cyt a32* undergoes a high— to low— spin conversion as the pH is lowered from 7.4 to 5.0. At more acidic pH levels, both heme chromophores display spec— tra typical of low—spin ferrous heme a in a nonaqueous en— vironment. This indicates that the native cyt a hydrogen bonded formyl group interaction has been disrupted (see schematic, Figure 5.8A). In this low pH titration of the a2+ a33*—CN' complex, cyt a3 remains in the low—spin ferric form throughout, but its carbonyl frequency downshifts from a position indicative of of a non—hydrogen—bonding environ— ment at 1672 cm-1 to a frequency which suggests a hydrogen— bonding environment (1655 cm'l). While the spectrum of cyt §33*—CN' remains essentially unchanged at acidic pH, cyt §2+ displays the absorption blue—shift and Vc=o frequency upshift indicative of its formyl environment shifting from a hydrogen bonded to a weakly or nonhydrogen bonded configura— tion (Figure 5.8B). Similar spectral shifts for cyt a were also observed in an alkaline pH titration of cytochrome oxidase (Callahan and Babcock, 1983). The fact that both acidic and alkaline pH are able to disrupt the hydrogen bonded form of cyt a implies that neither denaturation system titrates a specific amino acid, but rather results in a gradual unfolding of the polypeptide backbone, as discussed previously (Callahan and Babcock, 1983). 179 Figure 5.9: Schematic of proposed structural changes of reduced (A) and partially reduced plus cyanide (B) heme a chromophores of cytochrome oxidase at acid pH levels. See text for further discussion. 180 A. Structurol Schematic 2 + (Zyrt <1 £1035) + ___Fle_c=u....H—x— pH 7.4 \ NChls) hfo'lIS) H\ pH 4.5 —Fe-— C=O X- H NW5) Hydrophoblc Non-hydrogen bonded B. Mixed-Valence 2+ (Zy"t <1 NChls) 1 2+ -——Fe———-—--C\=EJ~~~H-X- pH 7.4 H N(hls) thls)‘ H 2+ PH 4 5 Fe ——Q=D \X- ' l H N(hls) hydrophobic Non-hydrogen bonded Cyrt <1 thls) l 3* __Fe l - ‘H cu \l/ Nails) .H-D\ i 3+ :° H F e C50 l - H; H CN H-o/ (aqueous) 181 A. Assignments of Hydrogen Bonding Sensitive Vibrations of Cyt §2* in the High Frequency Region: For both reduced and mixed-valence cytochrome oxidase, the major cyt a band shifts are: decreases at 1612, 1570, 1545, 1396-->1393, 1330-—>1334, and 1118 cm'1 (Table 5.3). There has been some disagreement over the assignment of the 1612 cm-1 Vibration to Vc=o of cyt a. Other workers have assigned this band to \ho, a Big core-size marker band (Ogura et al., 1984; Choi et al., 1983). The V10, Blg mode of high-spin cyt a32+ should also occur in this approximate frequency region, and it is assigned as a weaker, overlap— ping band at 1612 cm-1 based on correlations with appropri— ate heme §2+ model compounds (Van Steelandt-Frentrup et al., 1981) and core-size marker correlations (Choi et al., 1983). The fact that the cyt a2+ formyl stretching frequency is shifted down to 1612 cm-1, at lower frequency than V10 and chc (see below) without coupling to these vibrations can be explained as follows. The 1612 cm"1 ‘Vc=o band is a polarized mode and therefore is defined as a totally symme- tric vibration (Callahan and Babcock, 1981; Choi et al., 1983; Willems and Bocian, 1984). Because of the different symmetry of V1. (Blg; Abe et al., 1978) at 1623 cm-1, no vibrational coupling of vc=o with this mode is expected. For the vinyl stretching frequency, the physical disposition of the vinyl and the carbonyl groups across the heme a ring diminishes the likelihood of vibrational coupling. 182 The remainder of the H-bond sensitive formyl group frequency vibrations are seen to shift in cyt 92* as follows: 1396--> 1393 cm-1, 1228——>1225 cm'l, 1247-->1243 cm‘l, 1185-->1179 cm'l, 981-->975 cm‘l, and 395-->392 0111'1 (Table 5.3). The 1396 cm-1 band is assigned as the in—plane ( dn—co) bending vibration (Choi et al., 1983). There are two overlapping bands at this position, V20, A2g, and ‘Vzg, Bzg, seen with visible excitation at 1389 cm'1 and 1397 cm'l, respectively (Callahan and Babcock, 1983; Choi et al., 1983; see Figure 4.2 Chapter IV), which have ~25% (Cb—S) stretching character. With either assignment, a frequency shift is predicted. Our visible excitation studies (see Chapter IV) reveal a +2 cm-1 upshift on the 1389 cm‘1 peak. In addition, our heme a model compound studies reveal a mode at 1399 cm-1 (see Chapter III, Figure 3.10) which shifts up to 1405 cm-1 upon formyl hydrogen bond formation. Accord- ingly, the sensitivity of this mode in reduced and CN- inhibited protein to pH—denaturation is consistent with the models and with the in situ enzyme visible excitation studies. The 1228 cm"1 band ( VCb-cno) has been previously discussed (Choi & Spiro, 1983; Willems and Bocian, 1984). Visible excitation on intact cytochrome oxidase reveals a +5 cm-1 upshift in this band upon deuterium incorporation (see Figure 4.2 Chapter IV). Similarly, hydrogen bonding studies in Cu2* porphyrin a resulted in +6 0111'1 shift for this Raman band upon H—bond formation. Our observation of a downshifted frequency in this line upon pH denaturation (and 183 disruption of the hydrogen bond to the a2+ formyl) of redu- ced and cyanide inhibited protein is consistent with these previous data. The cyt a2+ band shift observed at 1185 cm-1 and described as V30, Bag (Willems and Bocian, 1984) has a potential energy distribution which contains ~50% (Cb—S) stretching character. The pH sensitive vibration of cyt azi having vinyl character are identified as: decrease in ‘Jc=c intensity at 1622 cm“1 and a shift in 65=cnz at 1330 cm‘l. The Soret excitation spectra seem to indicate a shift of the 1330 cm-1 band to 1334 cm-1. However, visible excitation spectra clearly show two bands at 1334 and 1329 cm'l, with the 1329 cm"1 band decreasing in intensity as the cyt a formyl hydrogen bond is disrupted (acidic pH data not shown, alkaline pH - Callahan and Babcock, 1983), and upshifts to 1333 cm‘1 in the presence of deuterated buffers (see Figure 4.2 Chapter IV). The nature of the 1334 cm‘1 band is not clear at this time, but we assign the 1329 cm'1 vibration to the vinyl internal mode 63:032 (Choi et al., 1982), and it is seen to decrease in intensity in parallel with Vc=c at 1622 cm-1 as the formyl hydrogen bond is removed. Neither of these vinyl internal modes are observed for heme §2+ (N-MeIm)2 (Table 5.3). They have been reported previously for ferrous heme a model compounds (Choi and Spiro, 1983), but were present owing to overreduction with NazszO4 (Babcock, 1986). The selective enhancement of the vinyl modes in cyt a2+ may be caused by the strong hydrogen 184 bonding perturbation along the y—axis of the heme a ring, which contains both the formyl and vinyl peripheral substituents. Choi et al. (1983) have suggested that the enhancement of the vinyl modes in protoheme RR spectra is caused by significant coupling to porphyrin modes. Lacking such vinyl—skeletal mode coupling in heme a, the hydrogen bond perturbation of cyt a allows the vinyl mode to couple into the heme a RR spectra. Alternatively, the presence of the vinyl bands in the RR spectrum of cyt a2i may indicate that there is a structural perturbation at the vinyl peripheral substituent in the enzyme as well as at the formyl group. B. Low Frequency Vibrational Assignments The remainder of the vibrational shifts fall into three categories: 1) strictly D4h allowed modes, Azz symmetry vibrations that have substantial Cb—S character, such as vs, A1: which shifts from 343—334 cm'l; 2) D4h modes which are strictly forbidden with Soret excitation but which become allowed under low symmetry conditions such as V20 A2; and V29 B2: which shift from 1396-1393 cm-1 and V22 A2g at 1115 0111'1 and v32, B2; at 785 cm—1, both of which decrease in intensity as the pH is lowered; and 3) low symmetry allowed modes such as out—of—plane vibrations V45 at 981 cm-1 which shifts down to 976 cm‘l, Yams at 850 cm-1 and hb_s at 395 Cm'l, and Eu modes, V33, V42, V47, V43 (Table 5.3), all of 185 which have considerable Cb-S stretching or bending character. In this work, V3 of cyt a2+ is identified at 342 cm‘l. This band shifts to 334 cm?1 at pH 4.5 and a band increases in intensity at ~348 cm-l. This pair of vibrations is assigned to V3 and 2 V35, respectively. These vibrations both have considerable (Cb-S) in-plane bending character and are believed to couple via Fermi resonance (Abe et al., 1979). The extent of their coupling apparently depends upon the nature of the iron axial ligands (Choi and Spiro, 1983) and formyl interactions (Copeland and Spiro, 1986). The effects of symmetry reduction on metalloporphyrin spectra have been discussed by Woodruff et al. (1985), Ozuki et al. (1979), and Choi and Spiro (1983). On reducing the heme molecular symmetry from D4h to Czh, the predicted effects are those observed here; that is, Alg, A2g, Blg, B2; modes transform as Ag (Wilson et al., 1955) and would be expected to be observed with both Soret and visible excita— tion because A-term RR scattering is dominant, depolariza- tion ratios approach 1/3, and the IR allowed Eu modes may split and acquire Raman intensity by symmetry reduction (Bocian et al., 1986). Protein—induced symmetry lowering in cyt 92* has been used to explain the presence of Eu modes in its low frequency RR spectrum (Valence and Strekas, 1982). The symmetry lowering effect of the hydrogen bonded formyl group of cyt a allows all of the above effects to be ob- served. The cyt a2+ RR spectrum does not display the ...... 186 extreme symmetry lowering effects of a metallochlorin (Andersson et al., 1986), but rather, lies intermediate between the spectra observed for M2+ octaethylchlorin and protoheme. D20 incorporation studies have been carried on beef heart cytochrome oxidase (Argade et al., 1986; Copeland & Spiro, 1986; Centeno et al., unpublished results) to probe the exchangeability of the cyt a hydrogen bonded proton. Small frequency shifts were observed for Vc=o...u in D20 vs. H20 buffers, as would be expected if deuterium were substituted for hydrogen in the cyt a formyl hydrogen bonded structure. Because of the several overlapping bands near 1612 cm-1, it is difficult to clearly assess the effect of D20 on cyt a. However, with reference to the vibrations reported here and in Chapter IV, it can be seen that several of the hydrogen bond sensitive vibrations are affected, i.e., 1611, 1330, 1230, 1183, 436 and 340 cm-1 (Argade et al., 1986) and 1610, 1330, 1230, 350 and 343 cm-1 bands (Copeland & Spiro, 1986), thus supporting the above authors’ conclusions that deuterium is substituted for hydrogen in the native structure of cyt a. A further examination of the vibrations reported here may be useful in determining the role of cyt a in energy transduction. This will be the subject of Chapter VI. _7—w—_—:;_.__— ' 187 V. Concluding Remarks: In the present study we have attempted to probe the cyt a2+ aldehyde vibrations by controlling an acidic pH denaturation of the reduced and mixed-valence oxidase derivatives. Lowering the pH of the reduced oxidase solution brought about a dual chomophoric effect as spin state changes in cyt as and disruption of the hydrogen bonded structure of cyt a were observed. The cyt a formyl hydrogen bonding assignments made here are suggested as a useful tool to gain insights into the structure and environmental conformation of cyt a2+ in cytochrome oxidase from other eukaryotic and prokaryotic species. In the following chapter we will use these formyl assignments, together with those suggested in Chapter IV, to study the proposed role of the cyt a2+ formyl hydrogen bonding structure in the oxidase proton pump action. CHAPTER VI EVIDENCE FOR CYTOCHROME A INVOLVEMENT ON THE CYTOCHROME C OXIDASE PROTON PUMP MECHANISM. A RESONANCE RAMAN AND ENDOR STUDY. I. Introduction: Since the original report by Wikstrom (1977) of the proton pumping action of membrane-bound cytochrome oxidase, experimental evidence on the redox—linked proton pumping activity has accumulated that supports the View of the H+- pump as an integral part of this terminal enzyme. Two mechanistic models for proton pumping activity in cytochrome oxidase have been proposed. The first model involves con— formational changes at the periphery of the cytochrome a heme induced by H-bonding interactions with a protein residue (Babcock and Callahan, 1983); the second model proposes redox-linked ligand exchange reactions at CuA2+ as the basis for the proton translocation. In the work reported here, we present spectroscopic results for cytochrome a and CuA2+ that were obtained in order to probe the feasibility and mechanistic implications of these two proposals. The cytochrome a site has been studied by resonance Raman (RR) spectroscopy, while the Cu».2+ site was studied by electron nuclear double resonance (ENDOR) spectroscopy. The cytochrome a heme model proposed by Babcock and Callahan (1983), involves the formyl group in a H-bonding interaction with an acid-base ionizable protein donor as the basis for the redox—linked proton pump mechanism. From 188 changes in the cytochrome a carbonyl stretching frequency, Babcock and Callahan (1983) suggested that the strength of this H-bonding interaction was dependent upon the redox state of the cytochrome a heme-iron. The carbonyl stretching frequency of ferric cytochrome a was observed at 1650 cm'1 shifting down to ~1612 cm-1 upon iron ion reduction. The proposed mechanism predicted the occurrence of exchangeable protons at the protein H-donor site and an enhanced H/D— exchange rate upon catalytic cycles of reduction and re- oxidation. Resonance Raman spectroscopy provides a suitable tool by which these two proposals may be examined owing to the sensitivity of resonance Raman frequencies to heme structure and immediate chromophore-protein environment. Indeed, recently reported Soret-excitation Raman spectra of reduced (Copeland and Spiro, 1986) and mixed-valence cytochrome oxidase (Argade et al., 1986), as well as a visible-excitation resonance Raman study (Centeno et al., in preparation), have provided further evidence for labile protons at the cyt a protein vicinity. In this Chapter we report the resonance Raman results of experiments on the H/D-exchange rate at the cytochrome a H-bonded site as studied during enzyme turnover. The results presented here show that, for a freshly prepared sample of reduced protein in D20, the H/D-exchange at the protein H-donor is complete after ~10 hours. For "acti— vated" (reduced and reoxidized) oxidase, the rate of deu- terium incorporation is dramatically accelerated. Comparison 190 of the resonance Raman spectra of cytochrome oxidase under different reductant concentrations and varying D20- incubation times allows us to distinguish between fast- and slow- changes in the vibrational spectra of reduced and mixed-valence oxidase. Slow—changes are identified as specific H—bonding effects and are detected for modes with substantial carbonyl contribution. On the other hand, fast— changes involving H/D exchange or non-specific protein effects are detected by resonance enhancement of vinyl modes and a low-frequency pyrrole folding mode. In the CUA based proton-pump model, Chan and coworkers proposed that redox-linked ligand-displacement at the CUA ligand-coordination serves as the driving force for the translocation of protons to the cytosol side of the mem- brane. This model also predicts the occurrence of exchange- able protons in the immediate protein vicinity of the CUA ligand-coordination site. To test this proposal, we have used electron nuclear double resonance (ENDOR) spectroscopy on the CUA EPR signal of resting and redox-cycled cytochrome oxidase in deuterated solvents. ENDOR spectroscopy is capa— ble of detecting weakly interacting protons in the vicinity of paramagnetic centers (e.g., O‘Malley and Babcock, 1986) and has been recently used to study H/D-exchange kinetics in the ubiquinone component of the bacterial photosynthetic system (Okamura and Feher, 1986). This technique has also been used to establish the identity of the ligands around the CuA2+ metal center (Stevens et al., 1982). Results from 191 these studies have demonstrated that ENDOR spectroscopy is a sensitive technique to study differences in protein ligation at the metal centers. Our ENDOR spectra of native and redox cycled oxidases in protonated and deuterated buffers reveal the existence of exchangeable protons near CUA. II. Results: A. Resonance Raman Spectra of Resting Cytochrome Oxidase: The optical absorption spectra of resting cytochrome oxidase in protonated and deuterated buffers, in the Soret region, are shown in Figure 6.1. For oxidized oxidase incubated in H20, the Soret maximum was observed at 422.6 nm, while in D20 the Soret maximum was shifted to 420 nm. This transition in the Soret absorption maximum required 10— 12 hours, at ice bath temperatures, for completion. Shifts in the Soret absorption maximum towards shorter wavelengths are consistent with increased population of high—spin heme iron, presumably at the heme of cytochrome a3 (Vanneste, 1967). The resonance Raman spectra of resting cytochrome oxidase and the associated difference spectrum are shown in Figure 6.2. The cytochrome a3* carbonyl stretching fre— quency, observed at 1650 cm‘1 in the sample in H20 (Babcock and Callahan, 1983), is shifted down to 1647 cm-1 in D20, as illustrated by the derivative shape of the Raman difference spectrum (RDS) with a negative peak at 1645 cm"1 and a posi— tive peak at 1652 cm'l. This observation confirms work by ABSORBANCE 192 WAVE NUMBER (cm") 2857|.4 25300.0 2222122 20000.0 2.20- ' r —d '— L76 I32 - — _. -- 0.88 0.44 - h" (-——) Resting / H20 (---) Resting /020 \ o it + o 350 400 450 500 WAVE LENGTH (nm) Figure 6.1: Optical absorption spectra of resting cytochrome oxidase in H20 and 020 buffers. Oxidase samples were incubated for 12 hours in the resting state. Oxidase concentration was 6011M. .E: m.©oe pm Aocmzcmtc cowpmpaoxo immmn .Amwmmn m meocv z: ow >_mpmeax0taam mm; coapwipcmocoo wechm .mtzo; NH toe Eco; acacmmc we“ :_ ccpmnzocr was mccmewcmaxm mmmcc co com: mmmcwxo meatcoocsc .Amv H-EU ooas or oomH sort new Aqv H-Eo sees 193 Op ooofi Echo mimcczn cmcmtmpzmv use couscopOLa ea cmmcaxo oeoccoOpxo m:_pmot co “waxy szccomam mocmnccwac :msmm cop -wwoommm one use micomam Ammv :mEmm mocmcommm ”m.o wismwd 194 as: Kim >ozmoowmu ooe owe. 8: 82 81 8.2 ow: ooo. - q d Ono 6N ...... - _ a on: v _ km: W 03. «.8. V N J E: a. woe I N .m «82.6 8:31 .0 mom .4 n N 2.2 B .ue—o. ( . i '0: 50: IA . _ _ _ — _osUNdoq ANl-v .-.-.93 n q 195 Copeland and Spiro (1987) who used conventional resonance Raman instrumentation to detect the shift in the Vc=o band of cytochrome a3+ upon H/D exchange. In the present work, we used Raman difference techniques to obtain these shifts more quantitatively. The expression used to obtain the relationship between the frequency separation of the lines, Av , and the peak-to—peak intensity (In°) of the difference spectrum is given by Rousseau (1981): Av :0.38F (In°/Ip) where F is the bandwidth at half-maximum, and In' is the peak intensity. Using the above expression, we obtain a shift in the carbonyl stretching frequency of —3.85 cm-l. This downshift in \)c=o of cytochrome a3+ was interpreted by Copeland and Spiro (1986) as indicative of stronger H— bonding in D20, and hence a lower C=O stretching frequency. A resonance Raman (RR) line at 1398 cm"1 assigned to cytochrome a3+ (Callahan and Babcock, 1981) loses intensity and shifts up by +1.2 cm"1 upOn use of deuterated buffers. Resonance Raman studies of heme a model compounds (Choi et al., 1983) indicate that the in-plane bend of the formyl group ( én-co) occurs at ~1390 cm'1 and overlaps strongly with a porphyrin mode ( \hg Bzg Cb-Cb) expected at 1395 Cm‘l. Raman spectra of benzaldehyde compounds display this formyl H—bending motion at 1388 cm-1 (Zwarich et al., 1971) which, Upon H—bonding, shifts up to 1395 cm“1 ( A :8 cm’l) (Centeno et al., in preparation; see Figure 3.3, Chapter III). Similarly, our previous RR study on the formyl modes of bis-imidazole heme a and its Cu2+ porphyrin a models 196 indicates that a mode at 1399 cm-1 shifts up to 1403 cm“1 upon H—bond formation. Conversely, the small but reproduc- ible shift in the corresponding frequency of the isolated protein upon H/D exchange, can be indicative of an unusual hindered formyl motion in the cytochrome g heme chromophore. This might be due to stronger protein interactions and/or to weaker coupling of the H—bending mode and other formyl- related motions. As shown in Figure 6.2, H/D exchange effects in resting cytochrome oxidase are also detected in the 1170—1184 cm"1 region. The RDS, in Figure 6.2 indicates that lines at 1157 cm‘1 and 1174 cm-1 shift into two new vibrations at 1166 cm'1 and 1184 cm-1 upon H/D exchange. In contrasts to the vibrational changes in the carbonyl stretching region which are observed to take ~10 hours for full completion, the above shifts were quickly established in D20. Significant contribution from Tr-conjugated porphyrin peripheral substituents to Raman modes in this region have been documented and indeed normal coordinate analysis indicates substantial ring-substituent stretching character in the region from 1167-1172 cm-1 (Abe et al., 1978; Lee et al., 1986). Resonance Raman studies on vinyl—containing protOporphyrin model compounds (Choi et al., 1982; Choi et al., 1983) have identified the ring—vinyl stretching frequency at 1172 cm‘l. Our observations of a 10 cm-1 upshift on this line upon protein deuteration is consistent with its vinyl assignment. 197 B. Resonance Raman Spectra of Redox-Cycled Oxidase: Raman spectra above 1000 cm'1 of redox—cycled oxidase obtained with 406.7 nm excitation are depicted in Figure 6.3. Traces 6.3A and 6.3B correspond to the redox—cycled oxidase in protonated and deuterated buffers, respectively. Several vibrational changes are detected between these two samples. At first glance, the broad feature centered at 1619 cm‘1 in trace 6.3A splits into two clear bands at 1608 cm-1 and 1625 cm‘1 for the redox—cycled oxidase in D20. In addition, in the mid—frequency region, bands at 1186 cm'l, ”1228 cm'l, 1246 cm-1, 1306 cm'l, and 1330 cm“1 to shift in frequency and/or change in relative intensity upon reduction and reoxidation in D20. A weak peak at ~1043 cm‘1 and a shoulder centered at ~1218 cm'1 consistently appear upon use of deuterated buffers. Since D20 itself displays a weak peak at 1043 cm-1 and a strong and broad peak at "1207 cm'l, we infer that the 1218 - 1220 cm-1 mode in the spectra of Figure 6.3 most likely arises from a D20 contribution and not from cytochrome a as recently suggested (Argade et al., 1986). Oxidase samples used to record traces 6.30 and 6.3D were inhibited in the resting (§3+§33*) form immediately after suspension of the oxidase in D20-buffer by addition of 6 mM CN— (pD 8.0). After formation of the a3*g33*CN derivative (as monitored by the optical absorption spec— trum), the reducting system, ascorbate/TMPD, was added and Raman spectra were obtained (trace 6.3C). The resulting 198 Figure 6.3: Resonance Raman (RR) spectra of mixed valence cytochrome oxidase under various turnover conditions as mo- nitored with 406.7 nm excitation. Oxidase samples A, B. and C were freshly prepared. Oxidase used on trace 8 was - freshly prepared and immediately reduced with 10 mM ascor- bate/20 UM TMPD in the presence of excess dioxygen, and allowed to redox-equilibrate for 45 minutes, then 6-10 mM cyanide was added to form the 22+E33+' CN‘ derivative. Oxidase used on trace C was freshly prepared and immediate- ly inhibited in the resting form by the addition of 6-10 mM cyanide; then after 15-25 minutes of incubation, reduc- tant asc/TMPD was added to produce the mixed-valence deriv- ative resulting in spectrum C. Trace D was the observed Ra- man spectrum of sample C after an incubation period of 24 hours. Trace E show the Raman spectrum of the D .buffer. T.0.= turnover. 20. 199 406.7 nm 1 {93:- CN-/Asc:‘ A.) H20(T.0.) I223 ll3l I I040 I089 l E: .... (7)- 8.) 020 (120.) l2|32| z ”64 3 E- I I Z y—o 2 E q C.) 020 (NO TO.) (I 1213 D.) 020 (Inc) 1213 :on k“ E.) 020 WWW IOOO l I40 |280 I420 l 560 IGOO FREQUENCY SHIFT (cm“) 200 Raman spectrum displays a broad feature centered at 1619 cm-1, and non-shifted modes at 1228 cm'1 and 1246 cm'l. Nevertheless, shifts in modes at 1184 cm‘l, 1305 cm'l, and 1330 cm“1 and the D20 shoulder at ~1218-1220 cm‘l, were detected. Comparison of traces 6.3A and 6.30 suggest that the addition of cyanide to the resting (non—redox cycled) oxidase appears to lock the cytochrome a site into an initially non—exchangeable conformation as noted by the similarities between these two spectra. The resonance Raman spectrum of sample 6.3C taken 24—hours after incubation is shown on trace 6.3D. The spectral similarities between this latter spectrum and that of trace 6.3B (redox-cycled oxidase) are obvious; that is, splitting of the 1619 cm'1 line into two new lines at 1608 cm'1 and 1625 cm'l, as well as the apparent upshift in modes at ”1228 cm“1 and 1246 cm-1. The resonance Raman spectra of fully reduced cyto- chrome oxidase in the carbonyl stretching region, is shown in Figure 6.4. As with the mixed-valence protein, we observed a clear splitting of the 1619 cm-1 line to 1608 cm“1 and 1625 cm-1 after 12 hours of incubation of the ascorbate/TMPD reduced oxidase. Dithionite-reduced protein failed to reproduce the frequency Shift in the cytochrome a2+ Vc=o line over the same period of incubation (spectra not shown, but see Figure 6.5). These apparent reductant dependence differences in Vc=o were further studied Figure 6.5 shows a plot of the observed frequency shift ' ' ‘ ' ' e reduced In Vc=o as a function of incubation time for th 201 Figure 6.4: Resonance Raman (RR) spectra of fully reduced cytochrome oxidase observed in the carbonyl stretching region of cytochrome a2+. Trace A is the corresponding RR spectrum in protonated buffers. Trace 8 show the RR spectrum of a freshly prepared sample in deuterated buff- ers; traces C and D are the corresponding spectra after 6 and 12 hours of incubation, respectively. Instrumental conditions: laser excitation at 406.7 nm with a nominal output power of 15 mw (at the sample); resolution 5 cm-l; scan rate 20 cm‘l/minute; time constant 0.5 sec. Oxidase sample conditions: concentration 80 p M(heme 3 basis); reductant ascorbate/TMPD; Soret maximum observed at 443 0m; temperature at the sample was maintained at 4 0C. RAMAN INTENSITY 202 426.7;nm I586 '38 .558 o 03 /A$C‘ ' I625 - - I570 I I | I I l I468 ' ‘ ’ '5'8 I549 I 1 i I474 l I I I I A.) . ; I : : I472 : I I l5l8 I ' I I I I l B.) : I I I I ' I I ' I I I I 2* 2’ I Ag 93 /Fresh/H20 : I I . B. gagg/Fresh/Dzo I I l C. Sample (8) (6 hrs.) C) i I 0. Sample (a) II2 hrs.) ' I I I I ' I I ' I I472 I5I8 I I I547 D.) r l I I 1— I450 I500 I550 I 600 I650 I700 FREQUENCY SHIFT (cm") —_”—*7 ...... 203 oxidase in D20. In these experiments, two individual freshly prepared samples (non—inhibited) of resting oxidase dis- solved in D20 buffer were placed in separated compartments of the divided spinning cell. Reductant, ascorbate (10 mM)/TMPD (20 uM), was added to one compartment, while a few grams of solid Na28204 were placed in the other compartment to produce the fully reduced oxidase derivatives. Both sam- ples were capped and their Raman spectra in the cytochrome a carbonyl stretching region were simultaneously recorded at different times of incubation. As shown in Figure 6.5, the dithionite reduced protein does not display any considerable shifts in the cytochrome a carbonyl mode even after one day of incubation. However, the ascorbate/TMPD reduced protein demonstrates a shift of the brOad band at ~1619 cm-1 to 1608 cm'1 at the end of six hours of incubation. The downshifted frequency of the cytochrome a carbonyl stretching mode has been followed as a function of the number of electrons donated from ascorbate to cytochrome a, with the aim of revealing the number of enzymatic turnovers needed to trigger full incorporation of deuterium atoms at the C=O...H protein site. Our results are illustrated in Figure 6.6. Individual oxidase samples were freshly prepared for each Raman measurement. The number of enzyme catalytic cycles were estimated from the ascorbate concentration and the oxidase concentration by assuming complete consumption of the former. The spectrometer resettability was checked after each individual measurement by recording the Raman Lw>o:L:pn.o.H .corpmnsoc_ mo mwaap cognates? wcp pm H-Eo QQNH 0p coma anw Cowman mgb cw mzpmamqam mom mco ;p_3 z_m:omcmo_:ewm quLoowL mam; mioomam one .FFoc mo:mam%w_v cmsmm mcp c_ A flu v mp_cow:oau sawUOW Uv_0m new AAU V oazk\omm :p_3 vwozbmt mam; mm_qum mmmu_xo .omo c? 204 wasp corpmnzoca to covuoczc a ma Aocmzcmaw m:_copmtpm _A: -onamu +Nm mEOLSoOpxo as“ to Accwzqmam to pofia .m.m mtzmwd 205 A 3.5 we: Qm QV ON I I 1 I I I 63 of mew”... E 60.5 umO\Dn_2.—. O mEE. 8:835 .m> 3339 55 35:08.“. L _N_®_ 0.9 mow. C=O/z (.-UJQ) 206 Figure 6.6: Resonance Raman (RR) spectra of redox-cycled oxidase as a function of enzyme turnover catalytic cycles induced by increasing the concentration of ascorbate in the solution. An individual oxidase sample was freshly prepared for each separate measurement. The oxidase con- centration was maintained at 80 u M (heme a basis), Whlle the ascorbate concentration was monotonical increased to 0.6 mM, 1.2 mM, 1.8 mM, 3.0 mM, 4.2 mM, and 4.8 mM, res- pectively. The number of turnovers (T O.) was calculated as the ratio of concentration of ascorbate used per oxi- dase. Laser excitation wavelength is at 406.7 nm. 207 I370 ‘58? ' 406.7nm , 9.2 9? CN- '55 l672 I572 | l I , I358 A.) H20 ISIS l I... l I=j° I397 ‘ '5‘!” I l4l3 B.) 5 T.O./Dzo C.) 25 T.0./DgO RAMAN INTENSITY D.) 35 T.O./020 E.) 40-45 T.O./DzO 4 I300 I350 I460 I550 I6EO I700 FREQUENCY SHH?r(cm“) 208 spectra of protonated and fully exchanged oxidase deriva— tives. The resonance Raman spectrum of the control protein sample in H20 is shown in trace 6.6A. After the end of the first five catalytic cycles (in D20) (trace 6.6B), the carbonyl stretching frequency appears as a "buried" vibra— tion under the broad feature at 1619 cm‘l, indicating lack of deuterium incorporation. For the oxidase following 25 (trace 6.6C) and 35 turnovers (trace 6.6D), cho is seen at 1614 cm"1 and 1610 cm-1, respectively. Full develOpment of spectral changes in the cytochrome a2+ carbonyl frequency were achieved after 35—40 cycles of reduction and reoxi- dation, as monitored by the shift of this band to 1608 cm-1. In addition to the differences observed in the cytochrome a carbonyl stretch, modes at 1228 cm'1 and 1248 cm-1 appears to shift slowly as it was noticed by the relative intensity of the Raman difference spectrum of reduced and mixed valence oxidase obtained at different times of sample incubation. Representative RR difference spectra of redox-cycled cytochrome oxidase with 441.6 nm and 406.7 nm excitations in the high-frequency region are shown in Figure 6.7A and 7B, respectively. The RDS spectrum shown in these Figures has been normalized to the intensity of the V4 oxidation state marker bands at 1370 cm-1 (Figure 6.7B) and 1355 cm‘1 (Figure 6.7A). We have noticed that the use of different multiplication factors (in the range of 0.98 to 0.65) resulted in a difference spectra with altered relative 209 Figure 6.7: Resonance Raman spectra and the associated Ra- man difference spectra of redox-cycled mixed-valence cytO- chrome oxidase obtained with 441.6 nm (7A) and 406.7 nm (78) excitation frequencies. Freshly prepared cytochrome oxidase (60 MM) in H20 and 020 was reduced with 4.8 mM ascorbate/ 20 MM TMPD in the presence of excess subs- trate, and allowed to redox-equilibrate for 45 minutes, then 10 mM CN' was added to form the mixed-valence (22+333+-CN‘) derivative. The spectra are the sum of 10 scans. Other experimental conditions as describe in Figure 6.2. 21.0 Figure 6.7A: ¢¥4L6Ivn RDS of 9213" CN' I236 l ' | "I . (I-2) : I I620 l I606 l l I l I .. H84 I >_ I I: I I (n l 23 LIJ mo 5 ms l249 I355 H 2.) 020 I586 E; '365 lIsfs H32 I625 I250 I '553 a: I I I I IBIOBSIS .455 '49' we : m4 IZTII |39|9 ' I file : I2 . '043 |O|89 I "a"? l2l6l : ' I I I I 1 I62 625 ' I.) H20 QT? VI I35I I228 I306 l I396 I I l___i 1 iL 1 T T J; r J I000 “40 I2 80 I 20 |56O I700 FREQUENCY SHIFT (cm") RAMAN INTENSITY 211 Figgre 6.78: AUC'O 3 4.3I Cm” RDS 0f gz‘g? CN"/T.O. I62I I606 I370 I567 ' I I572 I I I I I672 I .470 .5051 I550 I I I642 I I397 I'479 I I I I I4I3 I I I ....I II I I B.) DzO/TO. I F I I F F I300 I380 I460 |54O I620 I700 FREQUENCY SHIFT (cm") intensities but frequency positions that were not affected. The most prominent changes are detected in the broad feature at 1619 cm'1 with 406.7 nm, and at 1612 cm'1 with 441.6 nm, both shifting down to 1608 cm"1 upon H/D-protein exchange. With 441.6 nm excitation, the intensity of the 1608 cm-1 increases considerably in intensity, an observation previously noted by Argade et al. (1986). The RDS at this wavelength demonstrate a strong negative peak at 1608 cm-1 and a positive peak at 1621 cm'l, from where we calculated a shift of -3.31 cm'1 corresponding to the 1612 our1 line. With 406.7 nm excitation, the calculated frequency shift from the RDS, is -4.30 cm'l, which represents an increase of 0.45 cm'1 as compared to the resting oxidase. Therefore, the H-bond interaction appears to be stronger in the reduced protein in the presence of deuterated buffers. The low-frequency resonance Raman spectra of reduced and redox-activated protein (with ascorbate as the reduc- tant) are shown in Figures 6.8A and 8B. Of particular interest are bands at 439 cm'l, 661 cm-1, and 981 cm"1 on the RR spectrum of the protonated protein. These modes are shifted to 432 cm-1, 656 cm'l, and 975 cm'l, respectively. The downshifted frequency of the 439 cm‘1 line was previously observed by Argade et al. (1986) and recently in the RR of intact mitochondria in deuterated buffers (Centeno and Babcock, unpublished observations). In Figure 6.8B, the mid—range (600-1000 cm'l) spectra have been expanded with an improved signal—to—noise ratio 213 Figure 6.8: Low frequency resonance Raman spectra of re- duced (8B&)and redox-cycled (8A and 88-2) cytochrome oxi- dase in protonated and deuterated buffers. The respective Raman difference spectra are also indicated. Sample con- ditions as described in Figure 6.7. Laser excitation wave- length at 441 6 nm; resolution 5 cm‘l. Figure 8B-1 and 8B-2 show the mid-range (600-1000 cm'l) resonance Raman spectra of reduced and redox-cycled cytochrome oxidase, respectively. 214 Figure 6.8A: 439 (A‘B) 43 C 265 396 "25 II_.I Qz’flgfw' E Z A.) H20 < E 26l2 I3: 826660 I I 937 I 633660 I I B.) 020 a I a I I50 320 490 660 830 I000 FREQU ENCY SHIFT (cm") RAMAN INTENSITY I 600 680 760 21 5 Figure 6.88: I 9.269%» A.) H20/T0. 8.) 020/130. 22's? CN- A.) H20/ Inc. B.) 020/ Inc. f FREQUENCY SHIFT (cm") I 840 920 IOOO 6.88-1 6.8B-2 216 (16 scans total). The spectra shown in this Figure illus— trate the downshifted frequency detected in the 661 cm-1 to 656 cmtl. C. EPR and ENDOR Studies of Resting and Redox-Cycled Oxidase EPR and ENDOR spectroscopy are two magnetic resonance techniques widely applied to study paramagnetic centers in biological systems containing transition metal ions or organic radicals. While EPR spectroscopy measures the energy needed to change the spin orientation of the unpaired electron relative to the direction of the applied magnetic field, ENDOR spectroscopy uses radio frequency (rf) excitation to monitor the intensity of the EPR signal. Therefore, the ENDOR technique combines the capabilities of a nuclear magnetic resonance (NMR) experiment with the greater sensitivity of EPR. Owing to its inherently higher resolution relative to EPR, ENDOR can provide information about metal~protein interactions. In this work, EPR spectroscopy has been used to monitor enzyme quality and integrity. ENDOR spectroscopy was used to study the ligand coordination around the CuA2+ center of cytochrome oxidase by investigating the accessibility of protein exchangeable sites at the vicinity of this metal center. Our work has been conducted with the aim of testing the redox—linked proton pumping model proposed by Chan and coworkers (1979 & 217 1986), who have suggested the CuA2+ ligand-coordination site. D. EPR Spectra of Resting and Redox—Activated Oxidase The EPR spectra of fully oxidized and redox-activated oxidase in protonated and deuterated buffers, respectively, are shown in Figure 6.9. EPR signals with g-values at gz=3.03, gy=2.21, and gx=1.45 are attributed to the low-spin cytochrome a3+ (8:1/2) component (Hartzell and Beinert, 1976; Aasa et a1., 1976). The EPR signal from CuA2+ (8:1/2) metal center is seen as an intense signal with g=2.02 and g =2.17. The EPR spectrum of resting oxidase and redox— cycled protein in protonated and deuterated buffers are essentially identical. The ENDOR spectra of resting and redox—cycled protein in protonated and deuterated buffers obtained in the region of the cytochrome a EPR signal (g=3.0) are shown in Figure 6.10. The spectra are centered at the free proton Larmour frequency of 9.57, 9.58, and 9.69 MHz, respectively. The proton hyperfine couplings of these heme Q proton ENDOR lines are 1.16, 0.92, 0.76, 0.72, and 0.16 MHz from 1 t0 4, respectively. The proton ENDOR lines in Figure 6.10 are narrow in width, especially the two larger couplings, suggesting well—defined proton sites around the cytochrome a Fe nuclei. To study the ENDOR spectra of CuA2*, two regions of the. EPR spectrum corresponding to g (g =2.03) and g (g =2.17) were selected. Setting the magnetic field at different 218 Figure 6.9: Electron paramagnetic resonance (EPR) spectra of cytochrome oxidase. Oxidase samples are: A) native in H20; B) redox-cycled in 020. Experimental conditions: temperature 4.2 K, frequency 9.46686 0H2; modulation field 3600 Gauss (centerfield). Intensity EPR 219 A” Native Oxidase 320 r v “.2 A— f h ‘H (v I l iB. Redox Cycled I D20 gx=1.45 I I J] g=2 .02_TL + 1995 2834 3684 4723 Magnetic Field (Gauss) 220 Figure 6.10: ENDOR spectra of cytochrome 33+ at the electronic 9 value of g = 3.03. ENDOR conditions are: temperature 4.2 K, frequency 7-12 MHz, field modulation 2248, 2251, and 2275 G, respectively. The spectra are centered at teh free proton frequency V 9.58 MHz. H 2 The oxidase samples are: A) resting oxidase in H20; B) resting oxidase in D20; C) and redox-cycled oxidase in 020. ENDOR first derivative amplitude -—--> 221 ENDOR Spectra of Cytochrome a3+ at g-3.03 A. Native Oxidase/320 B. Native Oxidase/D20 C. Redox Cycled/Dzo —«.. I f i I *T 7 8 9 IO II I2 Frequency, MHz ——>- 222 g—values allows for selection in the orientation (Eachus and Ohms, 1985; O’Malley and Babcock, 1986). The ENDOR spectra of resting and redox—cycled oxidase in protonated and deuterated buffers in the g$=2.02 region of the CUA EPR signal shows 4 pairs of peaks (lines a through d in Figure 6.11) centered at the free proton frequency ( Va of 14.15 MHz), with proton hyperfine couplings of 1.70, 1.30, 0.80, and 0.40. Upon use of deuterated buffers, the lineshape, relative intensities, and peak positions remain unaltered. These lines are relatively narrow, again suggesting that the protons giving rise to these ENDOR transitions are from protons in well—defined sites relative to CuA2+ and unable to undergo H/D exchange. The corresponding ENDOR spectra recorded at g =2.1717 of the CuA EPR signal are shown in Figure 6.12. Seven pairs of lines are clearly observed with hyperfine couplings of 3.24, 2.54, 1.84, 1.44, 1.14, 0.70, and 0.54 from a to g, respectively. In contrast to the similarities noted for the spectra depicted in Figure 6.11, a similar interpretation of the observed g lines appear to be complicated by changes in relative intensities. Thus peaks with the smallest coupling are observed to decrease in relative intensity (i.e., line g); lines d and e have become narrower, while line g broadens. This observation might be an indication of underlying lines arising from other proton sites. The large couplings, those of lines a and b in Figure 6.12, appears 223 Figure 6.11: Matrix ENDOR spectra of CuA2+ in cytochrome oxidase at the electronic 9 value of 9 =2.02 (perpendi- cular region). ENDOR experimental conditions are: temperature 2.2 K, frequency 9.4395 GHz, amplitude 10.7- 15.7 MHz, magnetic field 3104.1 G. The oxidase samples are: A) resting oxidase in H20 ; B) resting oxidase in 020; C) and redox-cycled oxidase in 020. 224 Matrix OR Spectra of (If CuA + at g=2.03 T ‘3' 14.15 MHz 7'? $ A. Native Oxidase/320 é I I a Q) I c I ‘0 ed .é’ % 3. Native Oxidase/D20 C 0) > '6': C .2 g Q) '0 ch U) .2 9.. C. Redox Cycled/D20 O: O C) 2 LIJ I24 I31 I4.l I5.) F requency, MHz ——-> 225 Figure 6.12: Matrix ENDOR spectra of CuA2+ in cytochrome oxidase at the electronic g value of g = 2.1717 (parallel region). Same oxidase samples and experimental conditions as in Figure 6.11. ENDOR first derivative amplitude ——-* 226 ' Matrix ENDOR.Spectra of f CuA + at g=2.1717 «I6 V3 = 13.22 MHz — .— _— .A. Native Oxidase/320 - 0Q *- 0— a... B. Native Oxidase/D20 C. Redox Cycled/DZO l T T l ' I I2 l2.2 I32 I42 I52 Frequency, MHz ——-> 227 very broad in all the spectra shown in this Figure, sugges- ting loosely define proton sites nearby the CuA atom. However, upon H/D-exchange, these two lines occur at the same positions and does not display any significant decrease in relative intensity. III. Discussion: A. Enhancement of Cytochrome a Formyl H—Bonding Sensitive Modes:Slow H/D Exchange Process. Our results in the resonance Raman difference spectra of resting (fully oxidized) cytochrome oxidase agree with those of Copeland and Spiro (1986), and support the proposed H-bond structure of cytochrome a in cytochrome oxidase (Callahan and Babcock, 1983). The deuterium shift obtained in the cytochrome a carbonyl frequency bands at 1650 cm"1 (Av c=o :-3.85 cm'l) and at the 1619 cm"1 RR envelope (Avczo =—4.31 cm'l), suggest that the cytochrome a formyl- protein H—bond is fairly strong. This contrasts with recent H-bonding studies on heme g model compounds in which a lack of an isotope effect ( A < 1 cm'l) on the isolated vc=o stretching frequency was reported (Centeno and Babcock, in preparation; see Chapter III). The occurrence of protein exchangeable sites, as studied by the deuterium isotope effect, have been also documented in other structurally related H-bonded heme proteins, such as carbonyl-horseradish Q peroxidase (Smith et.al., 1983) oxymyoglobin and oxy— cobalthemoglobin (Kitagawa et al., 1982). In the latter case, a ~2.5 cm-1 in the Vc=o stretching frequency upon 228 protein deuteration was detected, while in the latter case a 2-5 cm-1 upshift in the Co-Oz stretching frequency was obtained by RDS techniques. The sensitivity of the cytochrome a carbonyl group stretching frequency to redox changes at the heme a iron in the presence deuterated buffers was recently studied (Copeland and Spiro 1986; Argade et al., 1986). Both groups interpreted their results as evidence against the involve- ment of the cytochrome a formyl hydrogen bonding structure in the oxidase proton pumping activity, since no frequency shift in the cho line of ferrous cytochrome a was detec— ted, after presumably turning the enzyme over in the pre- sence of deuterated solvents with solid N8232O4 as the electron donor. However, the above results should be inter- preted with caution since it is known that the presence of S204: in solution will quickly react with all the oxygen available (Lambeth and Palmer, 1973). Hence, the oxidizing for cytochrome oxidase, 02, will be depleted and turnover of the enzyme in the presence of dithionite will be sharply curtailed. In the work presented here, we have followed changes in the cytochrome a C=O stretching frequency for oxidase samples which have been reduced with TMPD and ascorbate (a known poor oxygen scavenger) as the electron sources. Our results indicate a significant decrease in the cytochrome a2+ C=O stretching when the ascorbate/TMPD couple was used as the reductant. A simple interpretation of these results indicates a need for substrate molecules 229 during the reduction of cytochrome a and cytochrome as to trigger conformational changes at the cyt a (-CHO...H— protein) site. Important insights in the structure and solvent accessibility of the cytochrome a heme site have emerged from these experiments. A comparison of the oxidase/D20 spectra allows us to distinguish between fast H/D-exchange sensitive modes detected at 1186 cm'l, 1306 cm‘l, 1330 cm‘l, and 1625 cmrl; and slow (hydrogen bond related) H/D exchange modes at 1619 cm-1, 1608 cm'l, 1398 cm-l,”1228 cm‘l, 1246 cm'l, and 661 cm'l. The data in Tables 6.1 and 6.2, on the classification of slow- and fast— exchange modes, indicate that the frequency position at which the latter modes occur are consistent with a substantial contribution to their normal mode of motion from the peripheral vinyl groups at position 4 (see Chapter IV), while the former set of fre— quencies appears to be associated with the cytochrome a formyl H-bonding modes (Callahan and Babcock, 1983; Choi et a1., 1983; Babcock and Callahan, 1983; Centeno et al., in preparation). The occurrence of "relatively" faster and slower hydrogen-exchange classes in structural proteins has been well documented by Englander and coworkers (1972 and 1984) and recently by Segawa and Kume (1986) in their studies with myoglobin and lysozyme proteins, respectively. These studies indicated that, for structurally related H— bonded amides, H—exchange occurrs relatively slowly, while side-chain hydrogens exchanged rapidly. H—bonding in these 230 TABLE 6.1: Tentative Raman Frequency Assignments for Cytochrome a Formyl Modes in Cytochrome Q Oxidase ASSIGNMENTS: OXIDASE DERIVATIVES Resting Mixed-Valence Begased E129 D20 1.1.2.9 12.29 .1129 12.2.9 Vazo 1650 1646.2 (”1619)3 1608 (”1619)8 1608 (1612)b 1608 I1612Ib 1608 V. VCb~Cb —— —- 1586 1583 1586 1583 6H-co 1398 1399.2 1395 1398 1396 1398 V13 Ca-Cb -— —- 1247 1249 1247 1250 -- —- 1228 1231.3 1229 1232 vCb—CHo 2 2 1236c 1249c 1236c 1249 —— —- 1131 1132 1131 1132 60:0 652 647 660(1 656 660 655 V. 6ob-S -~ -- 342 341 342 341 9 Cb-S —- -- 265 262 270 267 a) The carbonyl stretching frequency is expected at ~1612 "buried" cm'l (Callahan and Babcock, cm'l. b) Frequency observed with 441.6 nm excitation. deuteration, this line increases in intensity and shifts 1983); however, within the strong envelope centered at ~1619 deuteration the Vc=o is downshifted and detected cm'l; upon at 1608 Upon down to 1609 cm—1 (1608 cm—1 as detected by the RDS) in the presence of ascorbate as reductant. c) Frequencies observed in the Raman difference spectra. The group assignment is consistent with results from heme a model compounds. d) A 5 cm'1 frequency shift in the motion was calculated from the RDS. éIC=OI in—plane wag TABLE 6.2: Summary of Raman Frequency Assignments for Cytochrome a Vinyl Modes in Cytochrome c Oxidase ASSIGNMENTS: OXIDASE DERIVATIVES:(a) Easting Mixed-Valence 3692999 H29. 12.2.9 5.2.9. 9.2.9.. £29 12.3.9. VC=C ? ? 1625 1625 1625 1625 6(2CH2) 1330 1335 1331 1336 1331 1336 5ICH=I 1306? 1310? 1305 1310 1305 1310 VCb-Ca-CB 1174 1184 1173 1183 1173 1183 VaoV(Cb-S) 1157 1167 ( 6011:) or 982(w) 979(w) 981 978 981 978 V 4 5 E u 820b 820 914c 915 Pyr. fold. —— —— 438 430 439 431 a) All the vinyl modes listed in this Table were observed to exchange quickly. b) This line is seen to decrease in intensity in the spectrum of the deuterated protein. 0) The enhancement of the 914 cm—1 resonance Raman mode is observed only in the spectrum of the deuterated protein. 232 structural proteins was found to be the major factor in modifying these hydrogen exchange rates. Interestingly, the observation of slow— and relatively fast— exchangeable protein-hydrogens on intact membraneous cytochrome oxidase was previously reported by Capaldi (1973) in an IR study of intact mitochondria. While 41% of the peptide hydrogens were quickly exchanged, 40% were found to be slowly exchanged, taking up to 72 hours for the exchange to level off (Capaldi, 1973). Capaldi accounted for his observations suggesting that protons involved in the slow-exchange process were those stabilized by H—bonds. In addition to the frequency differences detected in the 1619 cm-1 envelope and 1612 cm'1 lines in Figures 6.6 and 6.7, other spectral differences in the RDS of the redox— cycled protein are also noted in the mid-frequency region for modes at 1184 cm‘l, ~1228-1236 cm‘l, 1246 cm‘l, 1306 cm-1, and 1330 cm-1. The difference pattern observed in the RDS of Figure 6.7A with negative peaks at 1216 cm“1 and 1249 cm-1 and a positive peak at 1236 cm‘l, was previously reported by Argade et al. (1986) where negative peaks at 1223 cm‘1 and 1249 cm-1 and a positive at 1236 cm‘1 were observed. To explain the apparent Splitting in the 1236 our1 line to 1223 cm-1 and 1249 cm'l, Rousseau and coworkers (1986) invoked a Fermi resonance mechanism which was suggested to occur due to a downshift of a weak mode located above the 1236 cm'1 line upon protein deuteration. As is shown on Figure 6.7A, we have detected frequency upshifts 233 for Raman modes at 1228 cm‘1 and 1246 cm-1 to 1230 cm*1 and 1250 cm'l, respectively. We also observed the emerging of a new vibration at 1216 cm“1 which we assign here as due to a D20 vibration. To disentangle the origin of these frequency perturbations and the direction of shifting in the spectrum of cytochrome a2*, we have conducted the following experi— ments: 1) hydrogen bonding experiments with benzaldehyde compounds bearing {—CH0) and (-CD0) groups; 2) hydrogen bonding studies on heme a (N-MeIm)z model compounds, as well as formyl—isotopic substitution on model compounds (Centeno et al., in preparation); 3) selective pH—modification studies on resting, reduced, and mixed-valence protein (Callahan et al., 1987). Isotopic studies on benzaldehyde compounds have assigned the ring-formyl stretching frequency ( VCb-cno ) at 1206 cm-1 and at 1215 cm'1 for ¢—CHO and ¢-CDO, respectively. This mode is observed to upshift to 1212 cm-1 and 1223 cm‘l, respectively, upon use of H—donors to the benzaldehyde C=O group. Our investigation on the spec- troscopy of formyl~isotopically substituted heme a (N—MeIm): and Cu2+ porphyrin a, suggested substantial contribution from a formyl—related mode, presumably Vcb-cao, to a Raman peak centered at 1238 cm'1 shifting up to ~1248 cm-1 upon H— bonding formation. We find no evidence of new vibrations at 1218—1223 cm-1 or Fermi resonance mode coupling in these hydrogen bonded heme a model compounds in the Raman region from 1200—1300 cm'l. In addition to the work with the model compounds, we have recently observed that upon disruption of 234 the cytochrome a H-bonding interaction, the 1246 cm'1 and 1228 cm"1 lines were downshifted to 1243 cm—1 and 1226 cm'l, respectively (Callahan et al., 1987). The RDS of the pH modified oxidase (RDS : pH7.4 - pH5.35; Figure 5.2 Chapter V), displays positive peaks at 1249 cm-1 and 1226 cm-1 and a negative peak at 1236 cm‘l, indicating the sensitivity of the 1236 cm'1 and 1246 cm'1 lines to H-bonding effects at the cytochrome a formyl group. Based upon these observa- tions with the heme a model compounds and the enzyme data presented here, we suggest that the VCb-CHO of cytochrome a lies below the strong mode at 1246 cm'l, presumably at ”1236 cm“1 which is the frequency observed for the isolated heme a models. Upon protein deuteration this stretching motion shifts up under the 1249 cm'1 (~+8-10 cm"1 upshift as detected with the isolated model compound). B. Low—Frequency Region: Protein induced alterations in the low-frequency modes of cytochrome a2+ were detected for modes at 439 cm'l, 661 cm-1, 820 cm-1, and 981 cm'l. The downshift in the 439 cm‘1 line to 432 cm“1 induced by deuterated solvent was original- ly reported by Argade et al. (1986), and interpreted as due to internal formyl-proton deuteration in cytochrome a2*, presumably catalyzed by an unusual protein environment. Downshifted frequency for this 439 cm‘1 mode has also been observed in the Soret excitation RR spectra of intact mito— chondria in deuterated buffers, as well as in isolated oxidase studied with excitation in the visible bands 235 (Centeno and Babcock, unpublished results). This mode appears to be insensitive to enzymatic cycles of reduction and reoxidation (see Figure 6.8A), and.does not seem to require prolonged sample incubation in D20. Its Raman intensity was observed to be largely diminished upon pH- modification of reduced and mixed-valence oxidase (Callahan et al., 1987; see also Figure 5.3, Chapter V). Its RR enhancement with visible excitation support its origin as due to a cytochrome a vibrational mode. Interestingly, Raman studies of benzaldehyde and metal substituted heme a model compounds reveal a low—frequency mode at 441 cm“1 (see Table 3.1, Chapter III) which decreases markedly in intensity upon H—bond formation to the isolated C=O group in heme a, but shifts towards lower frequencies upon formyl—proton deutera— tion in benzaldehyde. This mode was assigned in the model compounds as due to a pyrrole folding mode containing a substantial contribution from the peripheral formyl group. A low-frequency mode at 661-cm'1 is observed to shift down to 656 cm"1 upon use of deuterated buffers as detected by the RDS (see Figure 6.8B). Raman spectra of benzaldehyde and heme model compounds display a mode at 649'cm-l and 638 cm-1, respectively, assigned to the 6c=o in plane motion. H-bonding formation in these model compounds shifts the 6c:o to 641 cm“1 and 635 cm"1 (Centeno et al., in prepara- tion), respectively. Accordingly, the sensitivity of this mode in the isolated enzyme to H/D exchange supports its formyl character. 236 C. ENDOR Results: The ENDOR results reported appears to confirm the occurrence of exchangeable sites nearby the Cu;2+ site, however, our data can not confirm the proposed involvement of this metal center in the redox-linked proton pumping activity of cytochrome oxidase (Chan et al., 1979; Blair et al., 1986). The g ENDOR spectra of resting and redox—cycled oxidase in protonated and deuterated buffers showed identi— cal features. This suggests that these proton ENDOR signals might arise from protons at the nitrogen ligands of CUA2+ (Stevens et al., 1982). On the other hand, the ENDOR spectra obtained on the g region, show a complicated peak intensi- ty pattern that could result from unresolved proton sites with slightly different hyperfine couplings. If the decrease in relative intensity of the small coupling lines in the g proton ENDOR spectra is due to accessibility of H/D exchange sites near the CUA atom, then we will expect that upon repeated cycles of reduction and reoxidation, the site will be fully deprotonated (Blair et al., 1986), and hence its ENDOR signal should either vanish or reappear. Nevertheless, the ENDOR spectra of the resting and turnover enzyme in D20 are completely identical. Alternatively, it has been pointed out that the presence of D20 will generally broaden the proton ENDOR lines, presumably due to its influence on proton relaxation time properties (Eachaus and Olms, 1985; 237 Baker et al., 1986), which in turn will result in intensity differences. IV. Conclusions: The original proposal of Callahan and Babcock (1983) of the involvement of the cytochrome a formyl group on a H- bonding interaction with a nearby protein residue was based upon the red-shift on the cytochrome a a~band and the downshifted Vc:o frequency relative to its low-spin heme a model compounds. It was later recognized by the same workers that the strength of the cytochrome a (-HC=O...H-donor) interaction increases when cytochrome a was in the reduced state (Babcock and Callahan, 1983). These observations triggered the formulation of a redox-linked proton pump model (Babcock and Callahan, 1983) that predicts the occurrence of exchangeable protein protons and the enhanced incorporation of deuterium atoms at the H-bonding site as a function of enzyme catalytic redox-cycles. In the present study, resonance Raman evidence has been presented supporting the mechanistic role of the cytochrome a formyl group in the oxidase proton pump mechanism. We based our conclusions on the frequency shifts in the cyto— chrome a carbonyl sensitive modes in resting, mixed-valence and reduced oxidase derivatives, upon H/D exchange, and on the sensitivity of these modes to redox cycles, especially modes at ”1619 cm-1, 1396 cm'l, 1228—1246 cm'l, and 661 cm*1 in the protonated enzyme. 238 The vibrational differences between the cytochrome a formyl and vinyl modes presented here allow us to dis- tinguish the accessibility of fast and slow-protein labile sites near the periphery of the cytochrome a heme chromo— phore. This observation might be a direct indication of non— structurally related protein exchangeable sites, as well as H—bonded, and hence structurally related, labile sites at the cytochrome a protein vicinity. Our present ENDOR data on resting and redox—cycled cytochrome oxidase have been interpreted as the occurrence of H/D exchangeable sites in the immediate protein environ— ment around CuA2+ metal center. However, the likelihood of this CuA2+center being the oxidase proton pumping site, remains unproven. It appear that other factors, such as temperature and proton relaxation rates in D20 (Eachaus and Olms, 1985) might be significantly in influencing the dyna- mics of proton relaxation times as to obscure possible isotopic effects. Nevertheless, this work provide evidence that tightly bound protons (or matrix protons) are largely unlikely to exchanged with deuterium atoms from the solvent. CHAPTER VII Summary and Future Work I. Summary: The research described in this thesis deals with the application of resonance Raman spectroscopy to study struc- tural and environmental aspects of cytochrome a in cyto— chrome oxidase. The H-bond structure at the formyl group of and by monitoring vibrational perturbations in the spectrum of cyt a2+ introduced by remOval of this interaction and/or in vivo deuterium incorporation. The relationship between the cyt a hydrogen bonded structure and the oxidase redox— linked proton pumping activity was investigated by using oxidase samples in which the redox state of the cyt a heme iron and the peripheral formyl group vibrations were simul— taneously studied. The implications of the CuA2+ center in the oxidase proton pumping activity was investigated by using ENDOR spectroscopy. In an attempt to uncover formyl (—CHO) modes contri— buting to the resonance Raman spectrum of cyt a, we recorded (Chapter III) the RR spectra of (bis-imidazole) heme a and Cu2+ porphyrin a model compounds after formyl isotopic substitution and H-bonding conditions. The results reveal strong similarities between the models and the intact chro- mophore,in terms of the identification of formyl vibrational 239 240 modes, but differ in that the H—bond in the isolated chromo— phore is a weak interaction, while in cyt a the H-bond is observed as a rather strong one. The visible excitation spectra (605 nm excitation) of reduced cytochrome oxidase reported in Chapter IV produced essentially cyt a2+ vibrational modes. The influence of dissolving oxidase in buffered D20 on cyt a2+ formyl-related modes and the enhancement of symmetry-forbidden Eu-type vibrations was interpreted as due to the strongly electron- withdrawing character of the H—bonded C=O group. When these data are compared with previous results from the alkaline pH denaturation studies (Callahan and Babcock, 1983) a good correlation is found between those cyt a modes sensitive to H/D exchange and pH denaturation effects, thus supporting the H—bond structure of cyt a. Acidic denaturation of reduced and mixed-valence cyto- chrome oxidase produced structural changes in both chromo— phores (Chapter V). The action of low pH on the cyt as site transforms this site from a high to a low-spin and exposes its peripheral carbonyl group to the aqueous environment. In cyt a, on the other hand, acidic pH resulted in the removal of its H—bond structure and hence, in the spectral separation of those RR modes that arise from this interaction. The structural relationship between the cyt a formyl H— bonded structure and the oxidase redox—linked proton pump activity was studied under conditions in which the heme a 241 iron redox-state and formyl vibrational changes were both monitored in the presence of buffered D20. The occurrence of relatively "fast" and "slow" H/D-exchange sensitive modes in the RR spectra of cyt a were investigated and explained on the basis of conformational changes at both the formyl and vinyl substituents of cyt a. Spectroscopic studies of the CuA2* site of cytochrome oxidase were described in Chapter VI. ENDOR Spectroscopy indicates the accessibility of protein exchangeable sites near the CUA ligand coordination sphere; however, the identity and involvement of these protein exchangeable sites near CuA2+ in the oxidase proton pump remains unproven. II. Future Work A. Spectroscopy of Cytochrome Oxidase and Heme a Models The work presented here on the spectroscopy of the heme a(N—MeIm)2 model compounds has provided evidence for the identification of formyl (~CHO) vibrations that are sensi- tive to H-bonding effects in the intact cyt a chromophore. In addition, structural differences in heme a formyl and vinyl contributions to the vibrational spectrum were also noticeable (see Chapter III). To understand fully the individual influences of the formyl and vinyl groups on the vibrational properties of heme a, infrared studies in the C—H stretch and low frequency regions should be pursued further. Since IR spectroscopy is able to monitor all mole- cular bonds in the heme a macrocycle, including conjugated 242 and saturated substituents, a comparative study of the C-H stretch region (about 2800 cm'1 to 3000 cm‘l) of heme a and protoporphyrin IX should be useful in establishing the conformation of the -CH=CH2 group. Similar studies have already been reported in other metalloporphyrins. For instance, in chlorophylls a and b the intensities and frequency shifts of the C-H vibrational stretches were used to monitor the state of aggregation of these materials (Chapados and Leblanc, 1983; Chapados, 1985). The extent to which the H-bond structure of cyt a increases the ring molecular asymmetry should be investigaf ted further. Resonance Raman excitation profiles and detailed measurement of dispersion in the depolarization ratio (1,211/1”) of various Raman lines are suggested as two suitable experimental tools to gain insights in heme—protein interaction effects. Since the resonance Raman intensity depends on the extent of vibronic coupling, an analysis of the excitation profile (i.e., the Raman intensity as a function of the wavelength of the exciting line) of intact cytochrome oxidase and isolated heme a models should provide information on the properties of heme a in the ground and excited electronic states, and hence on the individual influences of the formyl and vinyl groups. We should continue our efforts toward the understanding of factors that govern the coupling between electron trans— fer and proton translocation in cytochrome oxidase. In this thesis evidence has been presented that supports the cyt a 243 H-bonded structure (Chapters IV and V), and its mechanistic implications in the redox-linked proton pump activity were considered (Chapter VI).. On the other hand, subunit III has been recently suggested as being an integral part (Pentilla, 1983; Sarti et al., 1985; Prochaska and Reynolds, 1986) or to serve as a regulatory control (Thompson et al., 1985) to the redox-linked proton pump activity. It has been shown that removal of this polypeptide causes a slight decrease in the electron transfer activity of the oxidase complex, but partially or completely abolishes the ability of the oxidase to catalyze redox-linked proton pumping. A resonance Raman study of cytochrome oxidase depleted of subunit III is therefore suggested as an alternative to obtain information on the H*-pumping activity of cytochrome oxidase. Based upon our results and the proposed role of subunit III (Pentila, 1983; Thompson et al., 1985) it is predicted that in subunit III depleted oxidase, the H-bond interaction in cyt a will be a weaker interaction. Finally, structural comparisons between cytochrome oxidase from other eukaryotic and procaryotic species can now be used to extend the cyt a2+ hydrogen bond assignments made in this work. For instance, our recent comparison of the resonance Raman spectra of isolated maize oxidase and mammalian protein, shown in Figure 7.1, indicates that the heme a environment of maize oxidase is altered, suggesting that this structural perturbation is associated with the 244 RELATIVE INTENSITY FREQUENCY SHIFT (CM-1) Figure 7.1: Comparison of the resonance Raman spectra of cytochrome oxidase isolated from maize (corn seedlings) and bovine beef heart mitochondria. Traces (a) and (b) correspond to the RR spectrum of reduced and cyanide inhibited maize oxidase, respectively. Traces (c) and (d) correspond to the RR spectrum of reduced and mixed- valence mammalian oxidase. Excitation wavelength is at 441.6 nm. Oxidase concentration: maize 9—14/uM; bovine 30—35/“M. 245 formyl group of the heme a porphyrin macrocycle (Dutch et al., 1986). This is evident by the absence of resonance enhancement of the C=O stretching frequency at 1612 cm‘l. Since the formyl group has been proposed to play a role in the proton pumping activity of cytochrome oxidase (Babcock and Callahan, 1983), further studies on the maize oxidase should be of considerable interest. B. Hydrogen Bonding To measure the magnitude of the H—bond interaction in heme a model compounds and in cyt a we have used a modified Badger-Bauer (1937) relation involving the observed wave- number shift on C=O stretching frequency upon H-bond forma— tion. This approximation normally neglects three important aspects of H—bond interactions: first, it ignores the directionality of the oxygen lone pair electrons; it assumes that the kinematic coupling between the C=O normal coordinate and other donor vibrations is negligible; and finally, it ignores the force constant of the O...H bond, allowing one to treat the H—bond as a diatomic molecule. An assessment of this latter approximation should become available by investigating the resonance Raman spectra of cytochrome a and its heme a model compounds in the region from 80~400 cm-l, where vibrations coresponding to the stretching frequency of the O...H—O bond are expected to appear (Ginn & Wood, 1965; Spinner, 1983). 246 Insights into the dynamics of kinematic coupling between the C=O and other H-bond stretching vibrations can be studied by using modern calculations such as ab-initio SCF-MO as applied to hydrogen systems (Ditchfield et al., 1971; Chean & Krimm, 1986). For small systems these calculations have proven to be useful in determining the effective force resulting from interaction between the C=O and H—bonded stretching vibrations. References Cited: Aasa, R.; Albracht, S.P.J.; Falk, K.E.; Lanne, B.; & Vanngard, T. (1976) Biochim. Biophys. Acta 422, 260—272. Abe, M.; Kitagawa, T.; & Kyoguku, Y. (1978) J. Chem. Phys. 9%. 4526-4534. Abe, M.; Kitagawa, T.; and Kyuguku, Y. (1978) J. Chem. Phys. 92. 880-890. Adman, E.T.; Skenkump, R.E.; Sieker, L.E.; & Jensen, L.H. (1978) J. Mol. Biol. (123, 35. Anderson, L.A.; Loehr, T.M.; Change, C.K.; & Mauk, G.M. (1985) J. Am. Chem. Soc. 101 ,182-191. nan-u».- Antonini, E.; Brunori, M.; Colosimo, A.; Greenwood, C.; & Wilson, M.T. (1977) Proc. Natl. Acad. Sci. USA 14, 3128-3132. Argade, P.V.; Ching, Y.C.; Sassaroli, M.; & Rousseau, D.L. (1986) J. Biol. Chem. 261(3), 5969-5973. Arnett, E.M.; Mitchell, E.J.; & Murty, T.S.S.R. (1974) J. Am. Chem. Soc. 96, 3875-3891. Azzi, A. (1980) Biochim. Biophys. Acta 694, 231-252. Babcock, G.T. (1987) in "Biological Applications of Raman Scattering” (Spiro, T.G., ed.) in press. Babcock, G.T.; Callahan, P.M.; McMahon, J.J.; Ondrias, M.R.; & Salmeen, I. (1982) in Electron Transport and Oxygen Utilization, Elsevier (North Holland Inc.) Chien Ho (ed). Babcock, G.T.; Callahan, P.M.; Ondrias, M.R.; & Salmeen, I. (1981) Biochemistry 20, 959-966. Babcock, G.T. & Callahan, P.M. (1983) Biochemistry 22, 2314-2319. Babcock, G.T.; Jean, J.M.; Johnston, L.M.; Palmer, G.; & Woodruff, W.H. (1984) J. Am. Chem. Soc. 196, 8305-8306. Babcock, G.T.; Jean, J.M.; Johnston, L.M.; Woodruff, W.F.; & Palmer, G. (1985) J. Inorg. Biochem. 28, 2337—2346. Babcock, G.T.; & Salmeen, I. (1979) Biochemistry 18, 2493- 2498. 247 248 Babcock, G.T.; Ondrias, M.R.; Gobeli, D.A.; VanSteelandt, J.; & Leroi, G.E. (1979) FEBS Lett 108 (1), 147-151. Babcock, G.T.; VanSteelandt, J.; Palmer,G.; Vickery, L.E.; & Salmeen, I. (1979) in Cytochrome Oxidase (King, T.E.,et. al., eds.) Elsevier/North—Holland Biomedical Press, pp. 105- 115. Babcock, G.T.; Vickery, L.E.; & Palmer, G. (1976) J. Biol. Chem. 251, 7907-7919. .um-n-u. Babcock, G.T.; Vickery, L.E.; & Palmer, G. (1978) J. Biol. nu . nu" Babcock, G.T.; Widger, W.R.; Cramer, W.A.; Oertling, W.A.; &» Metz, J.G. (1985) Biochemistry 23, 3638-3645. Baker, G.J.; Knowles, P.F.; Pandeya, K.B.; & Rayner, J.B. (1986) Biochem. J. 222, 609-612. Baker, G.M.; Masato, N.; & Palmer, G. (1987) J. Biol. Chem. 222, 595-604. Beinert, H.; Griffiths, D.E.; Wharton, D.C.; & Sands, R.H. (1962) J. Biol. Chem. 222, 2337-2346. Beinert, H.; Hansen, R.E.; & Hartzell, C.R. (1976) Biochim. Biophys. Acta 322, 339. Beinert, H.; Shaw, R.W.; Hanse, R.E.; & Hartzell, C.R. (1980) Biochim. Biophys. Acta 59;, 458. «mm-cu Bickar, C.; Bonaventura, J.; & Bonaventura, C. (1982) Biochemistry 21, 2661—2666. ...uo—n Bickar, D.; Lehninger, A.; Brunori, M.; Bonaventura, J.; & Bonaventura, C. (1985) J. Inorg. Biochem. 22, 365-372. Blair, B.F.; Gelles, J.; & Chan, S.I. (1986) Biophys. J. £9. 713-733. Bocian, D.F.; Lemley, A.T.; Peterson, N.O.; Brudvig, G.W.; & Chan S.I. (1979) Biochemistry 22, 34396-4402. Bocian, D.F.; Masthay, M.A.; & Birge, R.R. (1986) Chem. Phys. Letts. l2§(5,6), 467-472. Boelens, R.; Rademaker, H.; Weber, R.; & Van Gelder, D.F. (1984) Biochim. Biophys. Acta 222, 196-209. Boettcher, J.W. & Drago, R.S. (1974) J. Chem. Phys. 22, 429. 249 Boobyer, G.J. & Orville-Thomas, W.J. (1966) Spectrochim. Acta 22, 147. Bovil, A.J.; McConnell, A.A.; Nimmo, J.A.; & Smith, W.E. (1986) J. Phys. Chem. 29, 569-575. Brudvig, G.W.; Blair, D.F.; & Chan, S.I. (1984) J. Biol. Chemo g-§~g.’ 11,001-11,0090 Brudvig, G.W.; Stevens, T.H.; Morse, R.H.; & Chan, S.I. (1981) Biochemistry 22, 3912-3921. Brunori, M.; Bickar, D.; Bonaventura, J.; & Bonaventura, C. (1985) J. Biol. Chem. 260(12L, 7165-7167. gum-«outta mlmum Brunori, M.; Colosimo, A.; Wilson, M.T.; & Sarti, P. (1983) Juan-....- ----- Brunori, M.; Sarti, P.; Antonini, G.; & Malatesta, F. (1986) Bioelectrochem. and Bioenerg. 22, 159-165. Brunori, M. & Wilson, T. (1982) Trends Biochem. Sci. 2, 295. Buse, G.; Meinecke, L.; & Bruch, B. (1985) J. Inorg. Biochem. 23, 149—153. nun—mu Callahan, P.M. (1983) Dissertation, Michigan State University, East Lansing, Michigan, 48824. Callahan, P.M. & Babcock, G.T. (1983) Biochemistry 22, 452- 461. Callahan, P.M. & Babcock, G.T. (1981) Biochemistry'2g, 952-958. Callahan, P.M.; Centeno, J.A.; Ingle, R.; & Babcock, G.T. (1987) Biochemistry (to be submitted). Capaldi, R.A.; (1982) Biochim. Biophys. Acta 222, 291— 306. Capaldi, R.A. (1973) Biochim. Biophys. Acta 292, 237-241. Casey, R.A.; Chappell, J.B.; & Azzi, A. (1979) Biochem. J. igg, 149-156. Centeno, J.A.; Callahan, P.M.; and Babcock, G.T. (1987) in preparation. Centeno, J.A.; Bender, C.; & Babcock, G.T. (1987) in preparation. 250 Chan, S.I.; Bocian, D.F.; Brudvig, G.W.; Morse, R.H.; Stevens, T.H. (1979) Cytochrome Oxidase. ( King, T.H.; Orii, Y.; Chance, B.; & Okuniki, K. eds.) PP. 177-188 Elsevier, Amsterdam. Chance, B.; Saronic, C.; & Leigh, J.S., Jr. (1975) Proc. Natl. Acad. Sci. U.S.A. 22, 1635-1640. Chance, B.; Saronic, C.; & Leigh, J.S., Jr. (1975) J. Biol. Chem. 229, 9226-9237. Chance, B.; Kumar, C.; Powers, L.; & Chin, Y.C. (1983) Chancellor, T.; Quill, M.; Bergbreiter, E.; & Newcomb, M. (1978) J. Org. Chem. §§, 1245-1246. Chapados, C. & Leblanc, R.M. (1983) Biophysical Chem. 11, 211-244. Chapados, C. (1985) Biophysical Chem. 22, 227-242. Cheam, T.C. & Krimm, S. (1986) J. Mol. Struct. 222, 175-189. Ching, Y.C.; Argade, P.V.; & Rousseau, D.L. (1985) Biochemistry 22, 4938-4946. Choi, S.; Lee, J.J.; Wei, Y.H.; & Spiro, T.G. (1983) J. Choi, S. & Spiro, T.G. (1982) J. Am. Chem. Soc. 292, 3683- 3692. Choi, S.; Spiro, T.G.; Langry, K.C.; & Smith, K.M. (1982-a) Choi, S.; Spiro, T.G.; Langry, K.C.; Smith, K.M.; Budd, D.L.; & La Mar, G.N. (1982-b) J. Am. Chem. Soc. 293, 4345- 4551. Choi, S.; Spiro, T.G.; Langry, K.C.; & Smith, K.M.; (1982) J. Am. Chem. Soc. 292, 4337—4344. Clark, R.J. & Stewart, 9.9 . Reinhammer, (1983) Cline, J.; Hoffman, B.M. G.M.; B.G. Clore, Malmstrom, ibid (1980) Biochem. Andreasson, (1980) B. (1979) in "Structure and Bonding" B.; Jensen, P.; Venters, R.; & J. Biol. Chem. 222, 5124-5128. L.E.; Karlsson, B.; Aasa, R.; & 185, 155-167. uu «soc-anus-.. 251 Clore, G.M. & Chance, E.M. (1979) Biochem. J. 177, 613- 621. WNW Clore, G.M. & Chance, E.M. (1978) Biochem. J. 173, 811- 820. MW“ Colman, D.M.; Freeman, H.C.; Cuss, J.M.; Murata, M; Norri, V.A.; Ramshaw, A.M.; & Venkatapa, P. (1978) Nature, 272- 319. Copeland, R.A.; Naqui, A.; Chance, B.; & Spiro, T.G. (1985) FEBS Lett. 222, 375-379. Copeland, R.A. & Spiro, G.T. (1986) FEBS Letts. 22122222 239-243. Denis, M. (1981) Biochim. Biophys. Acta 223, 30-40. Denis, M. & Clore, M. (1981) Plant Physiol. 22, 229-235. Denis, M.; Gayda, J.P.; Bertrand, P.; & Richaud, P. (1984) Physiol. Veg 22(51J 589-595. Denis, M. & Richaud, P. (1982) Biochem. J. 06, 379-385. - "...-um Denis, M. (1986) Biochimie. §§, 459-470. De Vrij, W.; Azzi, A.; & Konings, W.N. (1983) Eur. J. Biochem. 222, 97-103. Dockter, M.E.; Steineman, A.; &Schatz, G. (1978) J. Biol. Chem. 222, 311-317. Dutch, R.; Ingle, R.; Centeno, J.A.; Peiffer, P. ; Babcock, G.T.; and Ferguson-Miller, S. (1986) in Plant Mitochondria (Ed. Beechy, B. and Moose, A.) Plenum Publishing Co., New York. Eachus, R.S. & Olm, M. (1985) Science 229, 268-274. Einarsdottir, O. & Caughey, W.S. (1984) Biochim. Biophys. Res. Commum. 222 , 836-842. Einarsdottir, O. & Caughey, W.S. (1985) Fed. Proc. 12, 1780 (Abst.) Emsley, J. (1984) in Structure and Bonding, pp. 147-191, Springer-Verlag Berlin, Heidelberg. Englander, S.W.; Downer, N.W.; & Teitelbaulm, H. (1972) Ann. Revs. Biochem. 32, 903—924. Felton, R.H.; & Yu, N.T. (1978) in Porphyrins,,2, 347-393. 252 Findsen, E.; Centeno, J.A.; Babcock, G.T.; & Ondrias, M.R. (1987) J. Amer. Chem. 800., in press. . Frey, T.G.; Kuhn, L.A.; Leigh, J.S., Jr.; Costedo, M.J.; & Chan, S.P. (1985) J. Inorg. Biochem. 22, 155-162. uuuuuu ... Fuhrop, J—H. & Smith, K.M. (1975) in "Laboratory Methods in Porphyrin and Metalloporphyrin Research” Elsevier Scientific Publishing Company. Fuller, S.D.; Capaldi, R.A.; & Henderson, R. (1979) J. Mol. Biol. 134, 305-327. "anon-«1...... Fuller, S.D.; Capaldi, R.A.; & Henderson, R. (1982) Biochemistry 22, 2525-2529. Fukumori, Y.; Nakayawa,, H.; & Yamanaka, T. (1985) J. Biochem. 98, 493-499. ooooooooooo Gelles, J.; Blair, D.F.; & Chan, S.I. (1986) submitted. Georgevich, G.; Darley-Usmar, V.M.; Malatesta, F.; & Capaldi, R.A. (1983) Biochemistry 22, 1311-1322. Gibson, Q.H. & Greenwood, C. (1963) Biochem. J. 22, 546- 554. Ginn, S.C. & Wood, J.L. Proc. Chem. Soc. 222 (1965). Glasoe, P.D. & Long, F.A. (1960) J. Phys. Chem. 22, 188— 190. Gorren, A.C.F.; Dekker, H.; & Wever, R. (1985) Biochim. Biophys. Acta 809, 90—96. ..l'llllltMl Gouterman, M. (1959) J. Chem. Phys. 29, 1139—1161. Gouterman, M. (1961) J. Mol. Spectrosc. g, 138. Greenwood, C. & Gibson, Q.H. (1967) J. Biol. Chem. 222, 1782-1787. Greenwood, C.; Hill, B.; Barber, D.; Englinton, D.C.; & Thomson, A.J. (1983) Biochem. J. 222, 303-316. Halaka, F. (1981) Dissertation, Michigan State University. Halmann, M. & Pinchas, S. (1958) J. Chem. Soc. 1703. Hartzell, C.R. & Beinert, H. (1976) Biochim. Biophys. Acta 362, 318—338. ltolmwflm Honami, K. & Oshima, T. (1984) Biochemistry 22, 454-460. 253 Hurts, G., Kraft, K. Schultz, R.; & Kreilick, R. (1982) J. Mag. Res. 22, 159. Jones, G.D. & Wilson, M.T. (1984) J. Inorg. Biochem. 21, 159-168. “M" Kamoun, M. & Mirone, A. (1980) Chem. Phys. Letts. 25(2) 287- 289. ” Karabatsos, G.J. (1960) J. Org. Chem. 25(3), 315-318. nnnnnn no Karlson, B.; Aasa, R.; Vanngard, T.; Malmstrom, B.G. (1981) FEBS Lett 222, 186-188. Keifer, W. (1973) Appli. Spectrosc. 27, 253. ........ Keifer, W.; Schmid, W.J.; & Topp, J.a. (1975) Appl. Spectrosc. 434-436. Keifer, W. (1977) Adv. Infrared and Raman Spectrosc. 2, 142. Keilin, D. & Hartree, E.F. (1939) Proc. Roy. Soc. London ”...“..- Kent, T.A.; Munk, E.; Dunham, W.R.; Filter, W.F.; Findling, K.L.; Yoshiba, T.; & Fee, J.A. (1982) J. Biol. Chem. 252, 12,489—12,492. dunno-...“ Kincaid, J.; Stein, P.; & Spiro, T.G. (1979) Proc. Natl. ...uuu ua Kitagawa, T.; Ogoshi, H.; Watanabe, E.; Yoshiba, Z-I (1975) J. Phys. Chem. 72(24), 2629—2635. ccccccccc Kitagawa, T.; Abe, M.; Kyoguku, Y.; Ogashi, H.; Watanabe, E.; & Yoshibe, Z. (1976) J. Phys. Chem. (11) 22, 1181—1186. Kitagawa, T.; Kyoguku, Y.; & Orii, Y. (1977) Arch. Biochem. Biophys. 182, 228—235. ...-cun- Kitagawa, T.; Ondrias, M.R.; Rousseau, D.L.; Ikeda-Saito, M.; & Yonetani, T. (1982) Nature 222, 869-871. Krishnan, R.S. & Shankar, R.K. (1981) J. Raman Spectrosc. .19., 1‘7 . Kumar, C.; Naqui, A.; & Chance, B. (1984) J. Biol. Chem. 259(4), 2073—2076 numuouunmu-uo-«c Kumar, C.; Naqui, A.; & Chance, B. (1984) J. Biol. Chem. 259(9), 11,668-11,671. .. ouautanu-w—nuuu-ou a 254 Lambeth, D.O. & Palmer, G. (1973) J. Biol. Chem. 248L}7l, 6095-6103. --------------- “a Lanne, J. & Keifer, W. (1980) J. Chem. Phys. 12, 5305 Lemberg, M.R. (1969) Physiolo. Rev. 22, 48. Lorosh, J.; Haase, W.; & Huong, D.V. (1986) J. Inorg. Biochem. 22, 53-63. Lorusso, M.; Capuano, F.; Boffoli, D.; Stefanelli, R.; & Papa, S. (1979) Biochem. J. 182, 133—147. m. ooo m ------ Ludvig, B. & Schatz, G. (1980) Proc. Natl. Acad. Sci. USA 22, 196-200. Malmstrom, B.G. (1985) Biochim. Biophys. Acta 222, 1-12. Malmstrom, B.G. (1986) Chem. Scripta Martin, C.T.; Scholes, C.P.; & Chan, S.I. (1985) J. Biol. .4 .woou 04—.“- -------------- McGovern—Moroney, P.; Scholes, T.A.; & Hinkle, P.C. (1984) Biochemistry 22, 4991-4997. Mincey, T. & Traylor, T.G. (1978) Bioinorg. Chem. 2, 409- 420. Mitchell, P. (1979) Eur. J. Biochem. 2Q, 1—20. Mitchell, P. & Moyle, J. (1983) FEBS Lett 151, 167-178. Musci, G.; Desideri, A.; Murpurgo, L.; & Tosi, L. '(1985) J. Inorg. Biochem. 22, 93-102. Nalecz, K.A.; Bolli, R.; & Azzi, A. (1985) Biochim. Biophys. Res. Commun. 222, 822-828. Naqui, A.; Chance, B.; & Cadenas, E. (1986) Ann. Rev. Biochem. 22, 137-166. Nicholls, P. (1978) Biochem. J. _222, 1147-1150. Nicholls, P. & Chanady, G.A. (1981) Biochim. Biophys. Acta Qfifl, 256-265. Novak, A. (1974) in Structure and Bonding vol. 22. 177-216. O’Malley, P. & Babcock, G.T. (1986) J. Am. Chem. Soc. 222, 3995-4001. Ogura, T.; Sone, N.; Tagawa, K.; & Kitagawa, T. (1984) Biochemistry 22, 2826-2831. 255 Ramaswamy, K.; Pichai, R.; & Chanadesikan, S.G. (1967) J. Mol. Spectrosc. 22,.416-424. Reichardt, J.K.V. and Gibson, Q.H. (1983) J. Biol. Chem. 258, 1504-1507. Reinhard, B.; Nalecz, K.A.; and Azzi, A. (1985) Arch. Biochem. BiOphys. 222 (1), 102-116. Richaud, P. and Denis, M. (1984) Arch. Biochem. Biophys. 222, 8-16. Rousseau, D.L.; Friedman, J.M.; and Willams, P.E. (1979) in "Topics in Current Physics" (A. Weber, editor) Spring-Verlag Berlin, Heidelberg, 1;. 203-252. Rousseau, D.L. (1981) J. Raman Spectrosc. 22, 94-99. Rousseau, D.L.; Ondrias, M.R.; LaMar, G.N.; Kong, S.B.; & Smith, K. M. (1983) J. Biol. Chem. 222(3), 1740-1746. Sarti, P.; Jones, M.G.; Antonini, G.; Malatesta, F.; Colosimo, A.; & Wilson, M.T. (1985) Proc. Natl. Acad. Sci. U.S.A. 22, 4876-4880. Scheiner, S. & Hillenbrand, E.A. (1985) Proc. Natl. Acad. Scholes, C.P.; Janakiraman, R.; Taylor, H.; and King, T.E. (1984) Biophys. J. .491. 1027-1030. Segawa and Kumer (1986) Seiter, C.H.A. and Angelos, S.G. (1980) Proc. Natl. Acad. Sci. USA 21, 1806-1808. Shaw, R.W.; Hansen, R.E.; Beinert, H. (1978) J. Biol. Chem. 222, 6637-6640. Shelnutt, J.A.; Rousseau, D.L.; Dethmers, J.K.; and Margoliash, E. (1981) Biochemistry 22, 6485-6497. Sievers, G.; Osterlund, K. & Ellfolk, N. (1979) Biochem. Biophys. Acta 581. 1-14. "...—um”: Sigel, E. and Carafoli, E. (1978) Eur. J. Biochem. 22, 119. Singh, T.R. & Wodd, J.L. (1969) J. Chem. Phys. 22(8), 3572- 3576. Smith, M.L.; Ohlsson, P—I.; and Paul, G.K. (1983) FEBS Letts. 222(2), 303—305. 256 Ohnishi, T.; Lo Brutto, R.; Salerno, J.G.; Bruckner, R.C.;& W- Okunuki, K.; Hagihora, B.; Sekuzu, I.; & Horio, J. (1958) Proceedings of the International Symposium on Enzyme Chemistry, Tokyo and Kyoto, p. 264, Maruzen, Tokyo. Ondrias, M.R. (1980) Ph.D. Dissertation Michigan State University, East Lansing, MI. 48824 Orii, Y. (1984) J. Biol. Chem. ggg, 7187-7190. Orii, Y. & King, T.E. (1976) J. Biol. Chem. 22;, 7487- 7493. Ozaki, T.; Katagawa, T.; & Ogoshi, H. (1979) Inorg. Chem. 22, 1772-1776. Palmer, G.; Babcock, G.T.; & Vickery, L.E. (1976) Proc. Natl. Acad. Sci. USA 7., 2206. w Papa, S.; LoruSso, M.; Capitanio, N.; & De Nitto, E. (1983) FEBS Lett 222, 7-14. Penttila, T. (1983) Eur. J. Biochem. 222, 355. Powers, L.; Blumberg, W.E.; Chance, B.; Barlow, C.H.; Leigh, T.S., Jr.; Smith, J.; Yonetani, T.; Vik, S.; & Peisach, J. (1979) Biochim. Biophys. Acta 546, 520-538. Powers, L. & Chance, B. (1985) J. Inorg. Biochem. 22, 207-217. Powers, L.; Chance, B.; Ching, Y.; & Angiolillo, P. (1981) Biophys. J. 22, 465-498. Prochaska, L.J.; Bisson, R.; Capaldi, R.A.; Steffens, G.C.M.; and Buse, G. (1981) Biochim. Biophys. Acta 222, 361-373. Prochaska, L.J. & Reynolds, K.A. (1986) Biochemistry_2§, 781-787. Proteau, G.; Wrigglesworth, J.M.; and Nicholls, P. (1983) Biochem. J. 210, 199-205. “M... Puttner, I.; Carafoli, E.; and Malatesta, F. (1985) J. Biol. Chem. 262, 3719-3723. anon-o. Raman, C.V. (1928) Nature London 222, 501. 257 Soliocz, M.; Carafoli, E.; and Ludwig, B. (1982) J. Biol. Chem. 2&1. 1579-1582. Sone, N. and Hinckle, P. (1982) J. Biol. Chem. 222, 12,600-12,604. Sone, N.; Naqui, A.; Kumar, C.; and Chance, B. (1984) Biochem. J. 222, 809-813. Sone, N. and Kosako, T. (1986) The EMBO Journal 2221, 1515- 1519. Sone, N. and Yanagita, Y. (1982) Biochim. Biophys. Acta 222, 216-226. Spinner, E. (1983) J. Am. Chem. Soc. 222 (4), 756-761. Spiro, T.G. & Strekas, T.C. (1972) Proc. Natl. Acad. Sci. USA 69(9) 2622-2626. anon-con Spiro, T.G. & Burke, J.M. (1976) J. Am. Chem. Soc. 22, 5482- 5489. Spiro, T.G. (1983) in "Iron Porphyrins" (Lever, A.B.D. and Gray, H.B., eds.) Addison-Wesley Publishing Company, part II pp. 91-157. Stevens, T.H.; Martin, C.T.; Wang, H.; Brudvig, W.; Scholes, C.P.; and Chan, S.I. (1982) J. Biol. Chem. 222 (20), 12,106-12,1130 Suarez, M.D.; Revzin, A.; Narlock, R.; Kempner, E.S.; Thompson, D.A.; Ferguson-Miller, S. (1984) J. Biol. Chem. ggg, 13,791-13,799. Takemori, S. & King, T.E. (1965) J. Biol. Chem. 222, 504- 513. Thelen, M.; O’Shea, A.S.; and Azzi, A. (1985) Biochem. J. .221, 163—167. Thompson, D.A. and Ferguson—Miller, S. (1983) Biochemistry 22, 3178. Thompson, D.A.; Gregory, L.; and Ferguson-Miller, S. (1985) J. Inorg. Biochem. 22, 357-364. r MAID! Thomson, A.J.; Greenwood, C.; Gadsby, P.M.A.; Peterson, J.; Englinton, D.G.; Hill, B.G.; and Nicholls, P. (1985) J. Inorg. Biochem. 22, 180—197. 258 Thomson, 24u1.; Greenwood, (3.; Peterson, J.; and. Barrett, C.P. (1986) J. Inorg. Biochem. 28, 195-205. "...-u Tsubaki, M.; Nagai, K.; & Kitagawa, T. (1980) Biochemistry 12, 379-385. Tweedle, PLIR; Wilson, IQHJ.; Garcia—Iniquez, I“; Babcock, G.T.; and Palmer, G. (1978) J. Biol. Chem. 252, 8065- 8071. Tzagoloff, A.; Marino, G.; and Sebald, W. (1979) Ann. Rev. ..... Valance, W.G. & Strekas, T.C. (1982) J. Phys. Chem. 26‘, 1804-1808. Vanderkoii, G. and Stotz, L. (1965) J. Biol. Chem. 232; 3418—3424. Van Gelder, B.F. (1966) Biochim. Biophys. Acta 18, 36. Van Gelder, E.F. and Beinert, H. (1969) Biochim. Biophys. Acta 222, 1. Vanneste, W.H. (1967) Biochemistry 5, 838-848. VanSteelandt—Frendrup, J.; Salmeen, I.; and Babcock, G.T. Vinogradov, S.N. and. Linnel, R.H. (1971) in "Hydrogen Bonding," Van Nostrand Reinhold Company, New York, Chapter 6, pp. 147—175. Ward, B.; Callahan, P.M.; Young, R.; Babcock, G.T.; and Chang, C.K. (1983) J.Amer. Chem. Soc. 105 634-636. . mun Wharton, D.C. and Tzagoloff, A. (1964) J. Biol. Chem. 232, 2036. "nan—u Wikstrom, M.K.F. (1977) Nature 222, 271-273. Wikstrom, M.K.F. (1984) Nature 302, 558—560. . ...-"nun- Wikstrom, M.K.F. and Krab, K. (1979) Biochim. Biophys. Acta 542, 177—222. nan-oug- Wikstrom, M.K.F.; Krabu K.; and. Saraste, M. (1982) in "Cytochrome c Oxidase: A Synthesis," Academic Press, London. Wikstrom, M.K.F. and Saari, H.T. (1977) Biochim. Biophys. Acta ggg, 347. 259 Willems, D. & Bocian, D.L. (1984) J. Am. Chem. Soc. 1 6, 880-890. “"“ Willems, D. & Bocian, D.L. (1985) J. Phys. Chem. 24, 5662. Willems, D. & Bocian, D.L. (1985) J. Phys. Chem. 89, 234. H-v-o-w Wilson, E.B.; Decius, J.G., Jr.; & Cross, R.C. (1955) in "Molecular Vibration" McGraw Hill, New York, N.Y. p.336. Wilson, M.T.; Peterson, J.; Antonini, E.; Brunori, M.; Colosimo , A. ; and Wymann , J . ( 1981 ) Proc . Natl . Acad. Sci. USA lg. 7115. Winter, D.E.; Bruyninckx, W.J.; Foulke, F.G.; Grinich, N.P.; and Mason, H.S. (1980) J. Biol. Chem. 2525, 11,408- 11,414. Woodruff, W.H.; Kessler, R.J.; Ferris, N.S.; Dallinger, B.F.; Carter, K.; Antalis, T.M.; 8c Palmer, G. (1982) Adv. Chem. Ser. 222, 625. Woodruff, W.H.; Norton, K.A.; Swanson, B.L.; and Fry, H.A. (1984) Proc. Natl. Acad. Sci. USA 21, 1263-1267. Yamanaka, T.; Fujii, K.; and Kamita, Y. (1979) J. Biochem. §§. 821-824. Yonetani, T. (1960) J. Biol. Chem. 235, 845-852. Young, L.J. and Palmer, G. (1986) J. Biol. Chem. 261(28), 13,031-13,033. Yoshiba, T. anui Fee, J.A. (1984) J) Biol. Chem. 59, 1031-1046. Yoshida, T.; Lorence, R.M.; Choc, M.G.; Tarr, G.E.; Findling, K.L.; and Fee, J.A. (1984) J. Biol. Chem. 25”, 112-123. Zwarich R.; Smolarek, J.; and Goodman, L. (1971) J. M01. 9 Spec. 2“, 336—337. ._ ' r H A -L. {2“ H.;:_rl-_r___- _ ,, v STATE UNIVE Y LIBRARIES MICHIGAN RSIT 1293 030616860