r ,..4 .~ :v ",7.\:7L“Z’(:,'A{‘3. fl':\\\\ s- 41m”, , vii} ,.?_ ”(VIII “1’." OVERDUE FINES: 25¢ per day per item RETURNING LIBRARY MTERI!‘J.S: Place in book return to remove charge from circulation records SPECTROSCOPIC STUDIES OF CHROMATIUM FLAVOCYTOCHROME 9552 By Mark R. Ondrias A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1980 ABSTRACT SPECTROSCOPIC STUDIES OF CHROMATIUM FLAVOCYTOCHROME 2552 By Mark R. Ondrias Various spectroscopic techniques were employed to examine Chromatium flavocytochrome 3552 in an effort to elucidate the structural and functional aspects of this multicenter electron transport protein. The primary tech- niques used in this study were absorption, fluorescence, magnetic circular dicroic (MCD), electron paramagnetic resonance (EPR) and resonance Raman spectroscopies. Absorption and fluorescence spectrosc0pies yielded information concerning the binding of exogenous ligands to heme and flavin prothestic groups of the protein and showed a difference in the binding mechanisms between sulfur con- taining ligands (S20; and 30;) and the cyanide ion. The relative midpoint reduction potential of the hemes in 9552 was determined to be who mV higher than the flavin by monitoring the absorption spectrum of the protein during reductive titrations. Mark Ondrias The two hemes in 3552 were demonstrated to exist as magnetically isolated centers by their MCD and EPR spectra. Their protein environments, however, were found to be quite different. The EPR spectra of the protein indicate that one heme is maintained in a single, invariant protein environ- ment, whereas the other can experience at least two dif- ferent environments. Both techniques showed that at physiological pH only one of the hemes binds carbon mon- oxide. Only a small amount of flavin semiquinone was detected with EPR during the course of a reductive titra- tion of 9552, indicating that the major pathway for flavin oxidation and reduction is a concerted two electron process. Resonance Raman spectra with both Soret and visible excitation have been obtained for Chromatium flavocytochrome 9552 and its isolated diheme subunit under varying condi- tions of pH and inhibitor binding. The spectra are gen- erally consistent with previously established classifica- tion schemes for porphyrin ring vibrations. The presence of covalently bound flavin in the protein was apparent in the fluorescent background it produced and the ease with which heme photoreduction took place. No flavin modes were present in the Raman spectra nor was any evidence of direct heme—flavin interaction found using this technique; however, a systematic perturbation of heme Blg vibrational frequencies was found in the oxidized holoprotein. The heme vibrational frequencies of 9552 are compared to those Mark Ondrias of the diheme peptide and of other getype cytochromes. The results of these investigations are discussed in the context of the mechanism of intramolecular electron transfer in 9552. They are consistent with an interpre— tation that involves pH—dependent changes in heme axial ligation and treats the hemes and flavin as isolated chromophores communicating via protein-mediated inter- actions. To my parents ii ACKNOWLEDGMENTS I would like to acknowledge the contributions of Pro- fessors George Leroi and Gerald Babcock to my graduate edu- cation. They each, in their own way, provided the support. advice and direction necessary for this undertaking. The efforts of Peri-Anne Warstler and Kathryn Nyland in the preparation of this manuscript are greatly appreciat- ed. Eric Findsen's aid with the growth of the Chromatium cultures and £552 preparation is gratefully acknowledged. I would also like to express my gratitude to all my compatriots who made the basement of the Chemistry Build- ing a warm and human place. So to, Bill B., Chris Y., Dave R., Dan T., Katie N., Eric F., Dan 0. Sunil K., Pat C., Tom C., Tom P. and all the others, I extend a heart- felt thanks. I couldn't have done it without you. Finally, a special thanks to Bea Botch who, more than anyone else, endured with me the trials and tribulations associated with my research. The weight seems much lighter when shared by two. iii Chapter LIST OF LIST OF CHAPTER CHAPTER A. B. CHAPTER A B. C D E. CHAPTER II. TABLE OF CONTENTS TABLES. . . . . . . a . FIGURES . 1 INTRODUCTION . 2 MATERIALS AND METHODS. Chromatium Growth and Protein Preparation . Experimental. 3 ABSORPTION AND FLUORESCENCE SPECTROSCOPY OF FLAVOCYTO- CHROME 9552. . . . . . . . . . . . Theory of Heme Absorption Results Binding Studies . . . . . . . . . . . Reductive Titration Fluorescence Results. A MAGNETIC TECHNIQUES APPLIED TO FLAVOCYTOCHROME c -552' Magnetic Circular Dichroism . . . . . A. MCD Theory. B. MCD Results . . . . . . . . . . . Electron Paramagnetic Resonance Spectroscopy. . . . . . . . . . . . . A. Theory. iv Page vi vii 15 15 20 2A 25 32 39 I49 60 67 68 68 72 8O 80 Chapter Page B. MCD Results . . . . . . . . . . . . . . 85 C. Reductive Titration . . . . . . . . . . 92 D. Flavin Semiquinone. . . . . . . . . . . 97 E. Exogenous Ligand Binding. . . . . . . . 100 CHAPTER 5 RESONANCE RAMAN SPECTROSCOPY OF FLAVOCYTOCHROME £552 . . . . . . . . . A. Raman Theory. . . . . . . . . . . . . . . . 107 106 B. Raman Results and Discussion. . . . . . . . 118 CHAPTER 6 CONCLUSION . . . . . . . . . . . . . . . INS A. Heme/Heme Interactions. . . . . . . . . . . 1&7 B. Heme/Flavin Interactions. . . . . . . . . . 1A8 C. Flavin Environment. . . . . . . . . . . . . 150 D. Heme Environments . . . . . . . . . . . . . 151 E. Exogenous Ligand Binding. . . . . . . . . . 152 F. Protein Mediated Communication Between Redox Centers . . . . . . . . . . . 153 APPENDIX. . . . . . . . . . . . . . . . . . . . . . 159 REFERENCES. . . . . . . . . . . . . . . . . . . . . 163 Tables Al LIST OF TABLES Extinction Coefficients of 695 nm Absorption Bands. . . . Cyanide Binding to Flavocyto— chrome 9552 . Ligand Field Parameters for Various Hemes g . . . . . . . . . . . . Raman Modes for Flavocytochrome 9552 Obtained with AA1.6 nm Excitation. High Frequency Raman Modes for Flavocytochrome 9552 Species Obtained with 51U.5 nm Excitation Calculated Values of AE From a 3552 Reductive Titration. vi Page 38 U2 91 123 127 162 Figure LIST OF FIGURES Page The structure of heme g, . . . . . . . . . 3 The structure of the flavin moiety and its various reduction products . . . . . . . . . . . . . . . . . 5 Postulated electron transport chain for Chromatium non-cyclic photosynthesis. Components are arranged vertically in order of decreasing reduction potential. Abbreviations: Bchl., bacterial chlorophyll; 9555, cytochrome 3555; NADP, nicotinamide adenine dinucleo- tide phosphate . . . . . . . . . . . . . . 10 Gel electrophoresis of flavotyco- chrome 9552 preparations. Gels are non-denaturing type in 5% acrylanide-N,N'-methylene-bis- acrylanide. Electrophoresis was conducted with a constant 2 mA current and gels were stained with Coumassie blue. Protein: Soret ratios for the various vii Figure Page samples were A (1.13), B (1.00), C (0.76), D (1.50), E (0.69) and F (1.2U) . . . . . . . . . . . . . . . . . l9 Schematic diagram of laser Raman experimental arrangement . . . . . . . . . 22 Spatial and nodal characteristics of the lowest unfilled (eg) and highest filled (alu, a2u) por— phyrin orbitals (from Reference uO). . . . . . . . . . . . . . . . . . . . 27 Relative energy levels of por- phyrin (-‘-) and iron (———) or- bitals for ferric porphyrin complexes (from Reference A2). . . . . . . 31 Absorption spectra of N6 uM oxidized (~-—) and reduced (———) flavo- cytochrome 3552 in 0.1 M Tris buffer, pH 7.5. Insert: m6 uM oxidized 3552 in 0.1 M MES buffer, pH 6.1 before (———) and after (---) addition of 2 mM Na28203. . . . . . . . . . . . . . . . . . 33 Transition axes giving rise to the NH5O nm and N360 nm flavin absorption bands . . . . . . . . . . . . . 35 viii Figure 10 ll 12 13 14 Absorption spectra of oxidized (---) and reduced (———) diheme peptide of flavocytochrome 9552 in 0.1 M Tris, pH 7.5. . . . . Absorption spectra of reduced 3552 under Ar (———) in 0.1 M Tris, pH 7.5, reduced 3552 under 6 psi of CO (...) in 0.1 M Tris, pH 7.5 and reduced 3552 under 6 psi of co in 0.1 M CAPS, pH 10.0 Absorption spectra of ’\’30 HM 9552 in 0.1 M MES, pH 6.1 (———), 5 minutes after the addition of 5 mM CN' (~---), 25 minutes after the CN- addition (---), and 180 minutes after the CN- addition (—--) Absorption spectra of m30 HM 9552 in 0.1 M MES pH 6.1 (--->, 15 minutes after the addition of 5 mM 320;. . . Absorption spectra obtained from a reductive titration of flavocyto— chrome 3552 in 0.1 M Tris, pH 7.5 under an Ar atmosphere ix Page 37 HO HA A6 52 Figure 15 16 17 '18 19 The anaerobic titrator used in the reductive titrations of 9552 Standardization plot for di- thionite reductant used in reductive titrations . . . A plot of AA”75 vs. AA552 during a reductive titration of flavocytochrome 3552 ( ) and horse heart cytochrome c (x), (type II obtained from Sigma Chemical), both in 0.1 M Tris, pH 7.5 . The extent of heme (x) and flavin (0) reduction as a function of electrons per molecule added during a reductive titration of 9552. Open squares and tri- angles denote the theoretical curves for two two-electron couples with AB = A2 mV and 32 mV, respec— tively.. The extent of heme reduction as a function of electrons per molecule added during a re- ductive titration of thiosulfate bound 9552 in 0.1 M MES, pH 6.1 under an Ar atmosphere . . . . X Page 53 5A 56 58 59 Figure 20 21 22 Page Left panel: Fluorescence emission intensities of: (a) 0.1 “M ribo- flavin (———J; (b) 2 “M glucose oxi- dase (---) and (c) 6 “M flavocyto- chrome 9552 (-----) all in 0.1 M Tris buffer, pH 7.5. Right panel: F1uorescence.emission intensities of: (a) 6 uM in 0.1 M MES, pH 6.1 9552 one-half hour after addition of 2 mM KCN (———>; (b) 6 pM 9552 in 0.1 M MES pH 6.1 one-half hour after addition of 2 mM Na28203 (—o--). The excita- tion wavelength was AA2 nm for both sets of spectra. Concentrated solu- tions of KCN and Na2S2O3 were prepared in 0.1 M MES and adjusted to pH 6.1 prior to addition to the protein sample . . . . . . . . . . . . . I . . . . 62 Fluorescence excitation spectrum of flavocytochrome 3552 in 0.1 M Tris, pH 7.5 with emission intensity monitored at 523 nm... . . . . . . . . . . 6A A schematic representation of the physical processes giving rise to MCD A, B, and g_terms (lower) and xi Figure 22 23 2A 25 Page the respective bandshapes of these terms (upper). In the upper panel dotted lines denote the absorption of left and right circularly polarized light while the solid trace is their difference (from Reference 59) . . . . . . 71 The MCD spectra of the reduced, oxidized, and reduced +C0 forms of 9552 in 0.1 M Tris, pH 7.5 in the Soret region are pictured above the absorption spectra of those species. . . . . . . . . . . . . . . . . . 73 The MCD spectra of reduced and oxidized £552 in 0.1 M Tris, pH 7.5 in the visible region are pictured above the absorption spectra of those species . . . . . . . . . 75 The MCD spectra obtained in the Soret region from a reductive titra- tion of 3552 under 6 psi of CO in 0.1 M Tris, pH 7.5. Traces of the peak from left to right (or through from right to left) are: L———), 0% reduced (---), 10% re- duced, (°°°°), 25% reduced, (-——), xii Figure Page 25 35% reduced, (-——), A5% reduced, (°°°'), 100% reduced protein . . . . . . . 78 26 The effect of successive symmetry reductions upon the energy levels of the iron d-orbitals . . . . . . . . . . . 81 27 An EPR spectrum of 100 uM oxidized 9552 in 0.1 M Tris, pH 7.5 obtained at 6.2°K with 2 mw of 9.132 GHz radiation and 10 G modulation. . . . . . . 87 28 Microwave power saturation curve for the gz = 2.88 (A), gz 3.02 (X), gy = 2.35 ( ) and g y 2.25 (o) resonances of 9552 under the same conditions as Figure 27 . . . . . . . 88 29 .Aplot of rhombicity vs. tetro- gonality for the hemes in flavo- cytochrome 3552 ( ) and various monoheme g_proteins (o). . . . . . . . . . 90 30 The EPR/Absorption anaerobic titrator . . . . . . . . . . . . . . . . . 93 31 The decay of the g = 3.02 reson- ance Ki- electrons per molecule added during the course of a reductive titration of m100 pM 3552 in 0.1 M Tris, pH 7.5 with xiii Figure 31 32 33 3A 35 36 37 Page the same instrumental parameters as Figure 27 . . . . . . . . . . . . . . . 95 Flavin semiquinone signal obtained from 100 M 9552 in 0.1 M Tris, pH 7.5 at 1A3°K after the introduction of 3 electrons per molecule. Microwave power was .5 mW at 9.122 GHz with 5 G modulation . . . . . . . . . . . . . . 99 EPR spectra obtained during a reductive titration of mlOO uM 9552 under 6 psi of carbon mon- oxide. Temperature and instrumental parameters are the same as Figure 27 . . . . . . . . . . . . . . . . . . . . 101 EPR spectra of m100 uM 3552 with flavinzheme ratio ml.0 in 0.1 M Tris, pH 7.5, after the addition of 5 mM CN‘. . . . . . . . . . . . . . . . 10A Raman scattering processes . . . . . . . . 109 Tensor symmetries for the reson- ance Raman active vibrational groups of hemes g. . . . . . . . . . . . . 116 Resonance Raman spectra of flavo- cytochrome 9552 obtained with AA1.6 nm excitation. The power was xiv Figure 37 38 Page 10 mw and the 3552 concentration was 75 uM in 0.1 M Tris, pH 7.5. . . . . . 120 Resonance Raman spectra obtained with 51A.5 nm excitation of (a) 70 uM ferrous flavocytochrome 3552'diheme peptide in 0.1 M Tris pH 7.5 with 350 mW of laser power; (b) 100 pM ferrous flavocytochrome 9552 in 0.1 M CAPS, pH 10.0 with 180 mW of laser power; (c) 80 pM ferrous flavocytochrome in 0.1 M Tris, 9—552 pH 7.5 with 250 mW of laser power; (d) 100 uM ferrous flavocytochrome 9552 in 0.1 M MES, pH 6.05 with . 250 mW of laser power; (e) 200 uM ferrous horse heart cytochrome g.in 0.1 M Tris, pH 7.5 with laser power equal to 250 mw. Frequency positions of the principal bands are given in Table 5. The fluorescence background of the diheme peptide spectrum arises from a small amount of residual flavin peptide which could not be separated from the sample. . . . . . . . . . . . . . 126 XV Figure 39 A0 Page Resonance Raman spectra obtained with 51A.5 nm excitation of (a) 70 uM ferric £552 diheme peptide in 0.1 M Tris, pH 7.5 with 250 mw of laser power; (b) 100 uM ferric flavocytochrome 3552 in 0.1 M MES, pH 6.05 2 mM Na28203 with 200 mW of laser power; (c) 80 uM ferric flavocytochrome £552 in 0.1 M Tris, pH 7.5 with 95 mw of laser power; (d) 200 uM ferric horse heart cytochrome g_in 0.1 M Tris, pH 7.5 with 250 mw of laser power. Frequency positions of the principal bands are given in Table 5 . . . . . . . . . . . . . . 130 Resonance Raman spectrum of 80 uM ' ferric flavocytochrome 3552 in 0.1 M Tris, pH 7.5 obtained with 315 mw of 51A.5 nm laser light incident upon the sample. Insert: The position and intensity dependence of the oxidation state marker band in ferric 3552 as a function of 51A.5 nm intensity upon the sample . . . . . . . . . . . . . . . . . . 133 xvi CHAPTER 1 INTRODUCTION Biological systems depend upon the production of high- energy chemical intermediates such as adenosine triphos— phate (ATP) as a source of stored energy that is avail- able for cellular anabolic processes. The elaborate mechan— isms evolved for ATP synthesis range from the photosynthetic systems of plants and some bacteria (1) to the oxidative respiration of higher animals (2). However, virtually all of these processes are ultimately dependent upon the energy generated by the transport of electrons down a potential gradient in a chain of electron transport proteins. ATP production is coupled to this electron transport (3), and conserves the potential energy lost by the electrons in a form convenient to the organism. Electron transport proteins, therefore, are fundamental to the life processes of organisms and stand at the center of a living cell's ability to construct a complex biological order from the entropic chaos of the inorganic world. Protein electron carriers have been isolated from a large number of cellular systems, but the precise nature of the mechanisms by which these enzymes function remains an unsolved problem of fundamental importance. In order to participate in electron transfer, a protein must possess an active center that can undergo reversible oxidation and reduction at a potential within the biologically viable range of —500 to +800 mV. Two such active centers in electron transport proteins are heme and flavin moieities. Perhaps the most extensively characterized of any of the electron transport proteins are the high potential cytochromes 3. These proteins contain heme g as the active redox center and are ubiquitously distributed among both eukaryotic and prokaryotic organisms (A). The heme redox center itself consists of an iron ion chelated within a porphyrin macrocycle which is trapped within the protein polypeptide matrix. Variations in peripheral substituents to the porphyrin ring system lead to differentiation between types of heme centers designated as heme a, b, 3, etc. (5). The structure of heme g is shown in Figure 1. In cyto- chromes g the heme is bound to the polypeptide matrix of the protein by two covalent thioether linkages between the heme 9 vinyl substituents and the sulfhydryl group of cysteine amino acid residues of the matrix. Additionally, other protein amino acid residues (most commonly histidine, methionine or lysine) form coordinate bonds with the heme iron, placing it in a five or six-coordinate octahedral ligand field. The iron ion serves as the vehicle for storage and release of electrons by alternating its valence CH3 R \ 1““4 A; \ H / c c\ /ca-—— cm I N l 0C3 c I c xCH3 c’ \ g / \c N---F'e- --N l HOOC -CH2-CH2-C\ C/ : c f, C‘R I N l /C_C/ \C—g H \ l H c_c I \ CIH2 CH3 (aHg coon CYTOCHROME c R=-$—CHCH3 'f CYSTEINYI. Figure l. The structure of heme c. state between +2 and +3.. The exact mechanism of electron transport to and from the iron in heme proteins remains a subject of controversy (see, for example, Moore and Williams, 1976 (6)) and evidence exists for multiple path- ways for electron transfer within mitochondrial cyto- chrome g_(7). Extensive amino acid sequence studies (8) suggest that several dissimilar subclasses of cytochromes g exist, while high resolution NMR (9) and crystallographic studies (10-12) indicate that there is a large degree of protein structural homology within given subclasses. Flavin redox centers are found in a wide variety of electron transport proteins. They are crucial to the func— tioning of photosynthetic systems in both plants and bac- teria and play a primary role in eukaryotic oxidative metabolism (13). The flavin moiety itself is a modified isoalloxazine ring system that is stable in a number of oxidation states (See Figure 2). Flavins can exist in stable one electron reduced states (called semi—quinones) in either neutral or anionic forms (1A) or in a fully reduced state formed by the addition of two electrons. Thus flavins can participate in electron transfer with either one or two electron redox centers. This makes them obvious choices as crossover points in an electron transfer chain that requires a change from concerted 2 electron transfers early in the chain to sequential one electron transfers at its terminus (15). Such a crossover is R . 2):}:3 HF!” reduction oxldaflon he") 9 (-o‘) V ‘36? {23:3 A reduction oxidation (4- 0") ( - o ') Figure 2. The structure of the flavin moiety and its various reduction products. critical to respiratory electron transport in mammals (l5). Flavins generally serve as non-covalently bound cofactors in flavoproteins, although several instances of covalent protein-flavin binding have been found (17). Considerations of the mechanisms of flavoprotein electron transport have largely been limited to elucidation of the oxidation/reduction kinetics of the isolated flavin chromo- phore (15). However, recent studies (18) have focussed upon the kinetics of the intramolecular flow of electrons in flavocytochromes 62 which contains both heme and flavin centers. Electron transport in biology is not limited to small single redox center proteins. Numerous multicenter electron transport proteins exist and are necessary for the proper function of a variety of biological systems. The importance of this class of proteins is underscored by the fact that often there are unusually severe kinetic or thermodynamic barriers to rapid and efficient progress in the reactions they catalyze. Examples include the role of nitrogenase in nitrogen fixation, cytochrome oxidase in mitochondrial oxygen reduction and the manganese-containing protein involved in photosynthetic water oxidation. Re- cently these proteins have attracted considerable interest both because of the complex nature of the reactions in which they are involved and because of the likely occurr- ence of electronic and protein mediated interaction between the various intramolecular redox centers. It is my intention in this thesis to examine in some detail the spectroscopic prOperties of an electron transport protein that contains both heme g and flavin redox centers. The enzyme chosen for this study is Chromatium flavocyto- chrome 3552. It is a soluble component in the photo- synthetic system of Chromatium vinosum, a purple sulfur bacterium. Flavocytochrome 9552 was first reported by Newton and Kamen (1955) and subsequently purified and characterized (19). Further studies established the molecular weight (~72,000), oxidation-reduction potential (E; = 10 mV) and isoelectric point (5.1) of the protein (20). The multicomponent nature of 3552 is evidenced by the fact that it contains two heme and one flavin moieties per molecule. The flavin prosthetic group has been identi- fied as a substituted 8-a-f1avin adenine dinucleatide (FAD) linked to the protein via a thiohemiacetal bond to a protein cysteine moiety (21) while the hemes in 9552 have been demonstrated to be the mesoheme typical of ggtype cyto- chromes (20) and thus are bound to cysteine protein resi- dues through thioether linkages. The structure of the protein moiety of 9552 has been the subject of several investigations. Kennel and Kamen (22) found that treat- ment with mild denaturants such as 8M urea dissociated 9552 into two subunits. One subunit contains the flavin prothestic group in a single polypeptide having a molecular weight of m50,000 daltons, the other contains both hemes in a single polypeptide weighing ~21,000 daltons (23). Attempts to reconstitute the protein from its subunits by Fukumori SE 31. (23) were unsuccessful. The subunit structure and prosthetic groups of 3552 are representative of a broader class of enzymes found in bacteria utilizing non-cyclic photophosphorylation as an energy source. Soluble proteins have been isolated from Alcaligenes Eutrophus (2A), a hydrogen metabolizing bac- terium, the green sulfur bacterium Chlorobium thiosulfato- philum (20), and Pseudomonas putida (25), that contain both flavin and heme redox centers. The flavocytochrome derived from Chlorobium, in particular, appears to be analogous to £552 in both structure and function. It is dissociable into flavin and heme containing subunits, although only one heme g is present in the heme subunits (25). The biological function of flavocytochrome 3552 has been the subject of numerous investigations that monitored light induced oxidations within whole cells (26,27), sub- cellular chromatophores (28) and ig_viyg preparations (29). These investigations have indicated that_c_:_552 mediates the noncyclic transfer of electrons from soluble 2-) to the membrane- reduced sulfur compounds (8:, S20“ bound constituents of the photosynthetic centers. It is postulated to function as the initial electron transport intermediate between the reduced sulfur substrates and the other chain components. In this role, its function is analogous to that of the water-oxidizing manganese-con- taining complexes of green plants. The less demanding requirements of sulfide oxidation (E; = -230 mV), however, apparently allow the organism to employ a more conventional set of biological redox centers. A generalized scheme for electron transfer in Chromatium is shown in Figure (3). Flavocytochrome 3552 has been shown to exhibit an 12.2322 sulfide-cytochrome 3 reductase activity and also catalyzes the reduction of elementary sulfur to sulfide with reduced benzylviologen as the electron source (23). Neither flavin nor heme subunits of 3552 display any sulfide-cytochrome 3 reductase activity. Thus 3552 may function as a multi— center electron transport vehicle intermediate in complexity between the relatively well characterized monocenter electron transport proteins such as mammalian cytochrome c and the highly complex, membrane-bound multi-protein electron transport chains found in higher animals. The characteriza- tion of intramolecular electron flow in flavocytochrome 9552 would lend valuable insight into the general problem of electron transport in more complex systems. While a multiplicity of biochemically oriented investi- gations of 3552 have been undertaken, there have been only a limited number of studies focussing on the physical and spectroscopic properties of the protein. The optical ab- sorption spectra of the purified holoprotein (20) and its 10 .opmsamona ooHuooHoSCap ocfismcm opHEmchOOH: .maopnn< .Hmap tampon coauoSBmp wcammopoop mo gouge CH zHHmOaumm> Umwcmppm ohm mucoCOQEoo .mfimonpczmouoza oaaomolco: Ezfipmsopno pom Cacao upoomcmmu coppooflo pmpmHSumom ._....._m..m.. >533 / .m opsmfim 11 diheme peptide (30) have been characterized. The oxidized and reduced spectra of both holo— and ape-protein largely resemble those of cytochrome g except for a pronounced flavin absorption in the holoprotein. However, 3552 dis- plays an ability to bind exogenous ligands that cytochrome c lacks. The protein binds CO (19) to its heme moieties and CN‘, 30;, and 320; (30) to its flavin moiety. The optical rotation properties of £552 absorption have been investigated by Bartsch 33 a1. (20) and Vorkink (30). Derivative shaped bands with centers at the Soret absorption maxima were found in ferric, ferro and ferro-CO forms of the protein. These were interpreted to indicate heme-heme or heme-flavin interaction. In particular, the increased magnitude and decreased bandwidth of the Soret transition in ferro-CO c552 circular dichroism spectra led to the speculation that the C0 molecule was intercalated between the closely spaced hemes. On the other hand, optical rotatory dispersion spectra of ferro-CO 3552 obtained by Yong and King (31) were interpreted by them as involving CO binding to only one of the two interacting hemes. Some ambiguity concerning the extent of heme-heme interaction is raised by the observations of Moss 33 a1. (32) that the Massbauer spectra of ferric and ferrous 3552 show no indication of heme-heme magnetic coupling, although a dramatic alteration of the Mfissbauer spectral behavior occurs upon CO-binding. The absence of any discernable heme-heme magnetic coupling 12 was further substantiated by an investigation of 3552's mag- netic susceptibility and EPR properties of 9552 by Strekas (33). The magnetic susceptibility of 3552 was found to be indica- tive of low spin heme g. The protein exhibits two sets of EPR resonances arising from uncoupled low spin (S = 1/2) hemes 3, showing no indication of spin-spin interactions. One set of EPR resonances displays a marked pH dependence, whereas the other set is insensitive to environmental pH. Moreover, CN' binding to the protein was found to convert the pH-labile resonances exclusively to their low pH form. This was interpreted as an indication of heme-flavin inter- action. Finally, several investigations into the energetics of 3552 oxidation and reduction have been made. Vorkink (30) expanded on earlier investigations (20) of the enzyme's mid-point reduction potential and determined the potentials of both the flavin and heme prosthetic groups to be approxi- mately 0 mV. This potential is anomalously low relative to that of other heme 3 containing proteins and much higher than the E; of -l87 mV measured for the flavin containing peptide of 9552 produced by peptic digestion of the pro- tein (3A). Stopped flow techniques were used to monitor the in_yivg kinetics of 3552 oxidation and reduction by a variety of donors and acceptors (30). Although 3552 does auto-oxidize in aerobic solutions, its rate of oxidation by ferricyanide is about seven orders of magnitude faster 13 than with molecular oxygen. Rates of reduction of 3552 by sulfur-containing compounds (8:, 820:) were measured and are 1-2 orders of magnitude slower than rates measured using non-biological reductants of comparable E;. This strongly implies the existence of a multiplicity of oxida- tion—reduction pathways in the protein. The value of E; varies widely with pH but is largely invariant to changes in ionic strength (29), indicating the probable uptake of a proton to keep the net charge of the cytochrome constant upon reduction. The body of research done to date concerning flavo- cytochrome £552 is sufficient to expose the complexity of (fluaproteinn It provides a foundation for the further application of physical techniques to the elucidation of the intramolecular structure and function of this protein and other multicenter enzymes of its general class. This thesis will pursue that application by extending the use of some of the methods already mentioned and bringing new techniques to bear on this problem. Absorption, magnetic circular dicroic (MCD), fluorescence, electron paramag- netic resonance (EPR) and resonance Raman spectroscopies were utilized during the course of this study. Absorption spectroscopy was useful in the characterization of the protein's purity and binding behavior. The magnetic tech- niques (MCD, EPR) were employed to define the extent and nature of the heme magnetic environment in the protein and 1A thus served as a complement to the heme vibrational in- formation obtained via resonance Raman scattering. Flavin fluorescence proved to be a sensitive indicator of the intramolecular interactions of that chromophore. Applied to a molecule as complex as 9552 each of these techniques alone provides information that is too specific to admit an unambiguous interpretation. However, when multiple techniques are employed a synergistic effect is realized and a consistent picture of the structure-function relationships within 3552 can be obtained. CHAPTER 2 MATERIALS AND METHODS A. Chromatium Growth and Protein Preparation Flavocytochrome g 2 was isolated from heterotrophically 55 grown Chromatium vinosum st. D obtained from the American Type Culture Collection (culture E-l7899). The nutrient medium of Cusanovich (35) was used to grow the photo- synthetic bacteria in 10 8 carboys. The yield of wet cells was 6-8 g/liter and displayed little dependence on the incident light intensity on the carboys. Light levels between 100 and 800 ft. candles produced nearly equivalent yields. Na2S was used as the source of reduced sulfur substrate for the organism. Concentrations of S= exceed- ing mo.150 g/R led to an accumulation of elemental sulfur particles within the bacterial cells (a distinguishing characteristic of purple sulfur bacteria) and a subsequent depression of cellular growth rate evidenced by decreased yields and a pronounced change in culture color from deep purple to pink. Maintenance of the Chromatium culture was accomplished by continuous growth of the organism in l 1 bottles using an autotrophic medium of Cusanovich (35). 15 l6 Heterotrophic growth was initiated by introducing a ml% innoculum of late log phase autotrophs into the anaero- bic heterotrophic medium. Heterotrophs reached stationary phase in 3-10 days (depending on light level and size of innoculum) and were subsequently harvested by centrifuga- tion. The cells were resuspended in two volumes of buffer (.lM Tris pH 7.5) and broken by l to 2 passes through an Aminco 50 m1 French Pressure Cell at 16,000 psi. Purification of_c_552 from the crude cellular extract generally followed a procedure detailed elsewhere (30) and is briefly summarized as follows: 1. Crude cellular extract was centrifuged on a Beck- man ultracentrifuge at 90,000 g for m5 hrs to remove particulate matter. (The pellets from centrifuga- tion could be resuspended and recentrifuged result- ing in a m15% increase in protein yield.) 2. The supernatant from above was next chromatographed on a DEAE cellulose column. High potential iron- sulfur protein was eluted with .05 M NaCl in .02 M Tris, cytochrome 2' with .10 M NaCl in buffer and cytochrome 3552 with .15-.18 M NaCl in buffer. 3. The 3552 fraction could be subsequently purified by gel filtration carried out on Sephadex G—150 in 0.02 M Tris, 0.05 M NaCl, pH 7.5 an ammonium sulfate precipitation (a5 95% saturation) filtration 17 or a cembination of both filtration and precipitation. A. The yield of purified protein from the above pro- cedure was 8-1211moles per kilogram wet weight of cells. Criteria for £552 purity have been established by Barstch et 31. (20) using the.protein's optical properties. Ratios of Au75/A525 = 1.25 (flavinzheme) and A280/AA10 = 0.58 (protein: Soret) indicate purified 3552. All samples used in this study had Au75/A525 : 1.15 and A280/AA10 50.75 unless otherwise noted. Two other measures of protein integrity were also used. The ability of c552 to bind CN- and so; (accompanied by a decrease in Au75/A525) was found to be a sensitive criterion of protein integrity. Assays of the protein by gel electrophoresis were also conducted by Eric Findsen. The effect of increased protein:Soret ratio on the number of subunits seen with gel electro- phoresis is demonstrated in Figure (14), from which it is apparent that the increased.protein:Soret ratio reflects the presence of several additional protein subunits. By far the most mutable of the parameters mentioned above was the flavin:heme ratio. Some preps yielded protein with a flavinzheme ratio as low as 0.95. The AA75/A525 value could be "restored" by treatment with mild oxidants such as ferricyanide but this resulted in a loss of the proteins CN- binding ability. Minor deviations in either 18 .A:N.HV m use Am©.ov m .Aom.Hv a .Ams.ov o .Aoo.fiv m .AmH.Hv < mama mmflaewm macspms on» new mOAump popom "cfiopopm .msan mammmezoo Qua: pocfimum who: mHow pcm psoALSo <5 m pcmpmcoo m spas wouozpcoo mm: mfimopocQOLuoon .ocficmampomumfiplocoamnuoEI.z.znopficwfizpow am CH camp wcHLSDmcoclcoc ohm mamo .mCOHumgwqopd mmmm oEopzooumoo>mHm no mfimoponaopuooaw How .: mpsmfim 19 : opswfim 20 flavin:heme or protein:Soret ratios had no apparent effect on the spectral or chemical binding properties of the protein. The diheme peptide of 3552 was prepared by overnight incubation of the holoprotein in either 0.10 M CAPS buffer at pH = 11.0 or an 8 M Urea solution (22) and purified using gel chromatography (Sephadex G-100) to eliminate the dis- sociated flavin-containing subunits. Urea treatment resulted in significantly higher yields and was the method of choice. B- Experimental Resonance Raman spectra were obtained with two similar spectrometers utilizing different laser light sources. Investigations at Aex = AA1.6 nm were conducted on an instru- ment previously described (36) utilizing an RCA LD 2186 helium-cadmium laser. Spectra with 51A.5 nm excita- tion were obtained with a Spectra-Physics 16A argon-ion laser and a spectrometer consisting of a Jarrel-Ash Model 25-100 double Czerny-Turner monochromator equipped with a thermoelectrically cooled RCA C3103A photomultiplier tube. Baird-Atomic spike filters were employed to help eliminate laser-plasma lines. Since resonance Raman requires intro- ducing relatively high powers of radiation into an absorp- tion band of the sample, local heating of the sample can 21 be a problem, particularly when biological molecules are involved. In order to circumvent this problem, the protein samples were cooled during Raman spectroscopy by passing cold dry nitrogen gas through a copper housing which held either 5 mm or 10 mm optical cuvettes. The sample tempera- ture was controlled at 5°C i 2° by regulation of.the cool- ing gas flow rate. Bottom illumination geometry was em- ployed in both spectrometers. A photon counting detection mode was used with the helium-cadmium instrument while the argon ion instrument employed direct current detection. In both cases slit widths which provided a spectral band— 1 were used. A schematic diagram detailing pass 1.8 cm- the above instrumental arrangement is given in Figure (5 ). Optical spectra were obtained with Cary 17 or McPherson EU—707D recording spectrophotometers. Protein fluores- cence was monitored with a Perkin Elmer MPF—2 dual-scanning fluorimeter. A Varian EA spectrometer equipped with an Oxford Instruments ESR-9 liquid helium cryostat was utilized to obtain EPR spectra at low (T i 50°K) temperature. For EPR measurements at -1A5°C a Varian Cryostat was used. MCD spectra were taken with a computer-interfaced spec- trometer previously described (37). Reductive titrations were accomplished utilizing anaerobic titrators described later in the text. Samples for titration were gently but extensively degassed by exposure to alternate cycles of vacuum and either argon or carbon monoxide gas. Fully 22 .pcoEmwcmppm Hmpcoefisoaxm :mEmm sommH mo Empwmfip ofimemnom .m opswfim 3362:0232 _ < «3.30 _ cancum .31 “m”. _ .35.... ._0n_ Flmllu _ 3.30 AV 3:02.00 0 23 reduced samples were obtained by the introduction of excess sodium dithionite into sealed cuvettes under an argon gas atmosphere. CHAPTER 3 ABSORPTION AND FLUORESCENCE SPECTROSCOPY OF FLAVOCYTOCHROME £552 The optical absorption spectrum of c552 is dominated by the heme prosthetic groups of the protein. Thus, know- ledge of the phenomena contributing to heme absorption is a necessary prerequisite for the appreciation of the protein's absorption characteristics. The absorption spectra of por- phyrins in general (and hemes in particular) have long been the focus of extensive inquiry both theoretical and experi- mental. In this chapter a brief explanation of the theo- retical interpretation of heme optical properties will be made. It will be followed by an examination of those properties as they pertain to 9552. Absorption spectroscopy is useful in the characterization of the mechanism of exo- genous ligand binding to heme and flavin centers in the molecule and can be employed for the determination of rela- tive reduction potentials of the hemes and flavin. The application of fluorescence spectroscopy to £552 yields additional information concerning the protein environment of the flavin chromophore. 2A A. 25 Theory of Heme Absorption The absorption spectra of porphyrins and porphyrin metal complexes are, in general, characterized by two strong transitions (38). These are the intense Soret (or B) transition in the near uv spectral region of the spectrum and a less intense visible (or Q) absorption in the 500 to 600 nm region. The Q transition is usually accompanied by a vibronic sideband and may or may not be split depending on the symmetry of the specific porphyrin. As a first approximation, the absorption spectra of por- phyrins may be treated as arising from a conjugated 16- member cyclic polyene. The effects of peripheral sub- stituents and the central metal ion (if present) are then viewed as perturbations upon the polyene spectra. The earliest attempt to explain porphyrin spectra that met with qualitative success was the free electron model advanced by Simpson (39). This model assumes that the n-electrons of porphine are essentially free particles on a ring composed of the 16 lattice points (atoms) in the conjugated pathway of the molecule. The l8anelectrons of the molecule are then paired in orbitals of increasing angular momentum, filling the system to the R = :A level. Transitions then occur from that level to the R = :5 orbitals resulting in transitions with A2 = :1 or :9. The former transitions are allowed by angular momentum 26 selection rules whereas the latter are not. Moreover, Hund's rule predicts that the A1 = :9 transition will be lowest in energy. Thus, this model achieves qualitative agreement with both the relative positions and intensities of the porphyrin visible and Soret transitions. However, it is too simplistic to be useful in appreciating the more esoteric aspects of porphyrin spectra, such as coupling between n-n* transitions or the interaction of metal orbitals with the n system. The formulation of a model based on the configuration interaction (CI) of cyclic polyene molecular orbitals (MOS) by Gouterman (A0) provided a more detailed picture of porphyrin spectra. The model considers only the two highest filled and the two lowest unfilled cyclic polyene orbitals and is frequently referred to as the four orbital model. The MO procedure when applied to cyclic polyenes results in the prediction that the lowest unfilled states are a degenerate orbital of e symmetry, with the highest 8 filled states being orbitals of alu and a2u symmetry. In the cyclic polyene the highest filled states are energet- ically degenerate. This is not the case for hemes, but the generality of the argument still obtains. The energetic relationships and the nodal characteristics of the four orbitals involved in the Gouterman model are shown in Figure 6. Simple MO theory would ascribe the observed Q and B transitions indiVidually to eg + a2u and 8g + alu 27 Figure 6. Spatial and nodal characteristics of the lowest unfilled (eg) and highest filled (alu, a2u) porphyrin orbitals (from Reference A0). 28 transitions. However, at this level the Q and B transi- tions are degenerate. Since both transitions possess the same symmetry (Eu’ in the DAh symmetry group) configura— tion interaction between them is allowed and the observed transitions must be mixtures of such configurations. Configuration interaction occurs because the states a2ueg and a are solutions to the unperturbed eigen- lueg problem for one-electron orbitals, Heff(a2ueg) = E(a ) 2ueg whereas the full Hamiltonian is, A A=A ' H Heff + H Electron repulsion terms, e2/rij, provide the principal contributions to H' and cause states of the same symmetry to be mixed and driven apart in energy. The new states (i.e., eigenvectors of H instead of fieff) are now, 0 = - .— Bx /2/2 (a2uegx aluegy) 0 _ By - /2/2 (a2uegy + aluegx) O - QX "' )/_2-/2 (a2uegx + aluegy) Q0 = /§/2 (a e — a e ) y 2u gy lu gx 29 and are referred to as 50-50 mixtures of the configurations for obvious reasons. Further mixing of states is required between the above states in order to obtain a model con- sistent with experimental data. The reason for this is that the dipole intensity of the Q0 states (in the above model) would be zero. This is easily demonstrated as follows: ,Let Rly (aluegXIYIwo> R2y = now the dipole strength, q2, of a transition is computed as the square of its transition moment. Thus for Q; )2 = 2 1 q ((a2uegy - aluengylwo> 2(Rly - R2y o for B : .Y I. 2 - 2 2 q - <(a2uegy + aluegx)|y|wo> 1 with the same results for Bi, Q: dipole strengths. It can be shown that in rigorous DAh symmetry R1 = R2, and q2(Q; y) = 0. The observed dipole intensity of the Q 9 state in porphyrin spectra derives from a coupling of the Q0 and B0 transitions. If it is assumed that no 30 mixing occurs between states of different polarization, then the application of perturbation theory yields the Q states of an arbitrary porphyrin as, Q N £3 ‘<30 + >’ CD 0 Where A is a coupling parameter dependent on the form of the perturbation operator used. The perturbation operator can be a simple electronic coupling (A05529i31pg from a splitting in the orbital degeneracy of the alu and a2u states or one involving vibronic coupling introduced by the breakdown of the Born-Oppenheimer approximation (Al). The former is a strong function of the electronic effects of porphyrin substituents upon the molecule's w-system, whereas the latter arises from the parametric dependence of the electronic energy of the molecule upon its vibra- tional motions. This will be discussed in some detail in Chapter 5. The situation found in heme g is complicated by the pos- sibility of interaction between iron d orbitals and the porphyrin system. Inclusion of iron d orbitals in extended Hackel calculations (A2) results in the energy levels depicted in Figure 7. It can be seen that the ef- fect of strong field axial ligands (such as CN’) is to 31 .Am: monopmmmm Eogmv moonoEoo gflpmzdpoo mayhem pom mamufinso ¢IIIV COLH pcm AI.IV gfimznapoa mo mam>oa magmco m>HumHom .s opsmfim “Each. IoEbon. 20:58 a. 2 a r <0 <0 mu lo...- |.|-...-.- - -..|.li - - .. -- |.ltsm_o I...I...- MHHIII I II: 1530 0 III II 1....- ||.|I InIII-lfl-h l.|.. .J / liaisons m I //I \ \. \ \ \ Absenvoml 0.0—I .IID-fl . X. 3 1/ U / a I In // b ..I.u.- .A II III/ I10.m.l \II X... a , .. l|3~32e M II Il------ll II------ I IIIII Eon. I |--.---l- ----- III lod- - assuage 32 drive dz2 and dx2-y2 up in energy, forcing the iron into a low spin (S = 1/2) configuration. The charge transfer transitions a + d 2, d 2 2 would occur at energies z x -y lu’ a2u comparable to the porphyrin alu’ a2u + eg transitions; however, mixing of the states involved in these transitions is symmetry-forbidden. The only other unfilled iron or- bital in ferric low-spin hemes (d1r = d dyz) is of e xz’ g symmetry, and since alu’ a2u and dTr are nearly isoenergetic the transition would occur in the near IR spectral region. In ferrous low-spin hemes, the dTr orbitals are filled and the possibility of this transition is eliminated. Thus, in either the ferric or the ferrous case the effect of the iron d orbitals on the heme visible spectrum is small. If the iron is complexed to weak-field ligands the de- creased splitting between dx2-y2’ dz? and d1T orbitals forces the iron into a high—spin (S = 5/2) configuration. The dTr orbitals are now unfilled and charge transfer transitions are expected in the visible region of the spectrum. These transitions have been observed in a variety of high-spin metalloporphyrin complexes (A3). B. Results The heme moieties in flavocytochrome 3552 display optical absorption spectra (shown in Figures 8,12) consistent with their assignment as low—spin heme g. The oxidized protein has a Soret maximum of A10 nm and a 33 .mOmmmmz :5 N no soapfippm AIIIV pmumm cam Alllv whomon H.m mo .gommzn mm: 2 H.o :H mmmm cmuaofixo z: m8 uphomCH .m.> ma .pommsn maps 2 H.o :H mmmm oEog200pm00>mHm Alllv Umodpog cam AIIIV Umnapfixo S: we no wppooan coaudpomn< .m ogswfim A65 2 00¢ con _ .c _ mum .34 APE: nbv OOn nun Ono _ _ L _ nJAYnN. 0.. “nq I llama .. . .- .....-xo «who is 3A poorly-resolved visible band that peaks at 525 nm. The addition of another electron to the iron d-orbital system results in a shift in the Soret maxima to A16 nm in the ferrous protein. The visible transitiOns are intensified and distinct QOO (at 552 nm) and Q01 (at 523 nm) components become apparent. These effects are primarily due to the decrease of central ion charge and the consequent lowering of the a + eg energy gap and to the decrease in lu’a2u the spin and orbital angular momentum of the iron d- electrons, resulting in decreased d1T - porphyrin magnetic interaction and better QOO’QOl resolution. Such behavior is typical of all ggtype cytochromes (AA). There are, however, some noteworthy deviations from typical heme g behavior evidenced by 3552. Its covalently bound flavin moiety is apparent in the oxidized protein as pronounced shoulders on the low and high energy sides of the Soret peak. Flavin absorption arises from the long and short axis n + n* transition dipoles of the isoallox- azine portion of the molecule shown in Figure 9. This shoulder bleaches completely upon flavin reduction and is sensitive to CN', SD; and 820; binding to oxidized £552. Additional structure is also apparent in the 600-700 nm region of the absorption spectrum of oxidized £552 where weak bands at N650 nm and N695 nm are present in 3552. The 695 nm transition is observed in horse heart cytochrome E and other small molecular weight mono-heme 9 proteins (A5) and could arise from a charge transfer from porphyrin 35 Figure 9. Transition axes giving rise to the NA50 nm and N360 nm flavin absorption bands. 36 n orbitals to the unfilled iron dTr or from iron d1r orbitals to distal protein sulfur orbitals. Experimental studies . (A6) have assigned this transition to the charge transfer interaction between the heme iron and the sulfhydryl group of a methi nineoamino acid residue acting as a heme axial ligand. The intensity of the 695 nm band in 3552 is relatively invariant to pH changes (See Table 1) indicat- ing that methionine remains coordinated to the heme over the pH range of 6.0 - 10.3. The 650 nm band has no analog in horse heart cytochrome 3 spectra. The most likely assignment of this band is to a high-spin heme charge— transfer band. The existence of a d + n* charge transfer interaction in high spin hemes is well documented (A3); they typically produce absorption bands in the near-infrared with extinction coefficients of 0.050 mM-1 cm'l. The extinction coefficients of the 3552 N650 nm band is variable but is always less thanN0.0015 mM"l cm-l. 'Thus, 2-3% high spin heme in the sample could account for the observed effect. The optical absorption of the diheme peptide of 3552 lacks the flavin effects mentioned above and is analogous to that of small molecular weight monoheme proteins (See Figure 10). The near-infrared spectrum of ferric heme peptide offers no evidence of a methionine-iron charge- transfer band, indicating that neither heme has methionine as an axial ligand in the apoprotein. 37 550 500 450 x(nm) 400 L5 '6 Absorbcnce in Figure 10. Absorption spectra of oxidized (---) and re- duced (-—) diheme peptide of flavocytochrome 3552 in 0.1 M Tris, pH 7.5. 38 Table 1. Extinction Coefficients of 695 nm Absorption Bands. 6 (mM"1 cm-l) Horse heart cytochrome 3, pH 7.5 .215 (Ref. A5) .200 (This study) Flavocytochrome c552, pH 7.5 .210 " pH 6.05 .205 " pH 10.3 .220 " pH 6.05 + SmM CN- At = 5 min .155 At = 25 min .070 At = 180 min <.030 Calculation of 695 extinction coefficients was based on the values of Q band extinction coefficients of Vorkink (30). 39 C. Binding Studies FlaVOCYtochrome £552 binds a variety of exogenous ligands. The heme prosthetic groups of £552 bind CO while its flavin moiety binds CN', 803 and 8203 These binding characteristics are well documented (30,A7). It is my intention in this section to confirm such binding behavior as a necessary prelude to investigation of the effects of binding on the EPR and fluorescence spectra of 3552' and to examine the behavior of the near IR 9552 transitions with respect to binding. Reduced c 2 binds CO under a wide range of pH condi- -55 tions. The effects of CO binding at pH 7.5 and 10.0 are shown in Figure 11. The intensification of Soret absorp- tion and a decrease in the Q00 band oscillator strength resulting from C0 binding are indicative of decreased coupling between heme B and Q electronic states and is much more pronounced at high pH. Binding decreases the extinction coefficient for the Q00 transition by N15% at pH 7.5 and by N50% at pH 10.0. An obvious explanation for the smaller effect at low pH is that only one heme is accessible to the CO whereas at high pH both hemes are able to bind CO. This binding scheme is substantiated by the MCD and EPR behavior of the c -CO complex and will be -552 elaborated upon later. The effects of CN', 80;, and S20; binding to 3552 are most obvious in the bleaching of flavin absorption at A0 E‘- H H u.- I u I Z J ‘ <1 1 00 0: .1 o : .7 8 . /' ' A I r q / 0 V I / o... “J~/ I ‘0 ex _ / O... ‘ ' / 0.0 I L 600 500 400 X(nm) Figure 11. Absorption spectra of reduced 3552 under Ar 0———) in 0.1 M Tris, pH 7.5, reduced £552 under 6 psi of CO (~--) in 0.1 M Tris, pH 7.5 and reduced 2552 under 6 psi of CO in 0.1 M CAPS, pH 10.0. A1 A75 nm. All three of these exogenous ligands bind most efficiently at low pH «LJ.M MES pH = 6.0 - 6.2 was used as a buffer system for all binding experiments) and result in a lowering of Au75/A525 to 0.85. The spectral changes occasioned by CN' binding are shown in Figure 12. A cal- culation of the dissociation constant for the binding of CN' from the data shown in Table 2 yields a value of “1.5 x 10'5 M-1 which correlates well with previously measured values of Kd (30). Such bleaching marks the protein's flavin moiety as the site of binding. Indeed, extensive studies with both free and bound flavins by Massey et a1. (A8) have established that formation of a direct adduct between flavin and so; proceeds through binding at the flavin (N5) position. No such adduct formation was found using CN', however. Since thiosulfate can act as a natural electron source for Chromatium (35), its binding to the flavin group of the proteinjgivitro indicates that it is the likely point of protein:substrateinteraction in_vivo. In vivo cytochrome g_reductase activity in the presence CN- and so; = has b en d monstrated for . of s e e 9552 (23) can be postulated to serve as substrate analogs on the basis of their effects on flavin absorption. However, 3 flavin and remain bound unless removed via dialysis, CN— while 820; and SO apparently bind reversibly to the displays irreversible behavior. Subsequent to cyanide binding, Au75/A525 returns to its initial value and cannot A2 Table 2. Cyanide Binding to Flavocytochrome - —1 A.3 11M 0 1.23 O " no uM .965 72.8 1.5 x 10'5 " 100 uM .899 90.9 1.0 x 10‘5 " 150 uM .893 92.6 1.2 x 10'5 " 250 uM .890 93.A 1.8 x 10'5 " 500 uM .876 100 " 1000 uM .876 100 A3 Figure 12. Absorption spectra of N30 0M 3552 in 0.1 M MES, pH 6.1 ( ), 5 minutes after the addition of 5 mM CN- (---—), 25 minutes after the CN- addition (—--), and 180 minutes after the CN- addition (---). ABSORBANCE AA .02- 600 650 700 >.(nm) Figure 12 A5 Figure 13. Absorption spectra of N30 uM 9552 in 0.1 M MES pH 6.1 (———), 15 minutes after the addition of 5 mM 820% (---). Absorption A6 I I I 600 6 50 700 >- (nm) Figure 13 A7 be re-bleached by further addition of CN’. Moreover, this was accompanied by a significant alteration of the heme Q-band absorption, which decreased in intensity and shifted to N528 nm. Such behavior usually occurred within 1/2 hour of the initial CN' addition, but its time de- pendence was variable. A distinction between the effects of CN- and 820; binding to 9552 can also be seen in behavior of the near IR (750-600 nm) absorption bands. (See Figures 12,13.) On the basis of its extinction coefficient, the 695 nm band in 3552 can be postulated to arise from a single methionine-heme interaction in the protein (See Table 1). Neither N5 nor any of the sulfur-containing flavin ligands has any effect on the intensity of either the 650 or 695 bands. Upon cyanide binding the Spectrum displays a rapid broad increase in absorption in the 600-750 nm region ac- companied by a bleaching of the 650 band and a time de- pendent decrease in intensity both of the broad 600-750 nm background and of the 695 band that parallels the previously discussed time dependent increase in absorption at A75 nm. Apparently, the initial effects of CN’ upon the flavin are reversed by other cyanide-protein interactions. The alteration of the visible heme absorption spectrum implicates at least one of the hemes as the site of this interaction. The decrease in intensity of the 695 nm band is indicative of a disruption of methionine-heme axial A8 ligation. CN- readily serves as a strong field ligand for hemes g_(A9) and could be expected to assume this role with The fact that the hemes in 9552 are already 9552' 6-coordinate would require a ligand displacement reaction in order to bind CN' to the hemes. Whether methionine- heme ligation is disrupted by such a direct replacement or as a result of a protein conformational change induced by CN’ attack on the non-methionine ligated heme is unclear. The kinetic barriers encountered in the replacement of an axial amino acid residue with CN' might produce the ob- served long time dependence of cyanide effects. The pro- tein conformational changes associated with axial ligand displacement could also induce a significant reduction in the ability of the flavin group to bind CN‘ and result in the restoration of absorption intensity at A75 nm. Evi- dence is found in the EPR spectra of 9552 to substantiate this scenario and will be discussed in a later chapter. Since the sulfur containing flavin ligands do not exhibit any substantial interactions with the hemes in 9552, they can be postulated to react with the protein in a single-step flavin-adduct formation. Despite the fact that the high spin heme resulting from "damaged" protein makes up a very minor component of our samples and must be considered of parenthetical interest, some insight into the heme protein environment can be gained by examination of its behavior with respect to A9 exogenous ligands. If the 650 nm band results from high spin heme charge transfer bands, the addition of a strong field heme ligand like CN- should produce the observed 'bleaching. What is Surprising is the lack of a similar effect with a 5 mM N3 addition, since azide is also a strong field heme ligand, although not as strong as CN'. The complete lack of effect by N; would appear to indicate that the high spin band in 3552 is not the result of a simple unfolding of the protein upon denaturation. Such a situation would result in exposure of the hemes to the solvent and provide equal accessibility to cyanide and and azide. The high spin heme in "damaged" 9552 is ob— viously still in an environment that can discriminate between ligands. D. Reductive Titration Monitoring the absorption spectrum of 3552 during the course of a reductive titration of the protein provides a simple means of determining the relative reduction po- tentials of the protein's two types of chromophores. Ap- plication of the Nernst Equation to a system of two active redox couples results in: 2.303 RT nF EO(Flavin) + log [FlJOX/[Fl]red = E [H] 2.303 RT OX “—517— 109 m (3'1) = EO(Heme) + 50 where n', n = number of e_ transferred; F = Faraday's constant; E0 = midpoint potential of flavin or heme; [Fl]ox, [Fl]red = concentrations of oxidized and reduced flavin; and [Hjox’ [ered = concentrations of oxidized and reduced heme at 25°C this reduces to 1/n AE = .059 log [F130x 0 [H1 l/n'/£F111/“EHJ§§“' (3.2) red red where AEO = EO(Heme) - EO(Flavin) Thus, determination of the relative concentrations of reduced and oxidized flavin and heme species allows for calculation of the relative potentials of the two redox couples involved. The above derivation contains the implicit assumption that the hemes in 9552 behave as one two-electron redox couple having an average midpoint po- tential of Eo(heme) rather than acting as two independent one-electron couples. Analysis of the data from reductive 51 titrations indicates that such an assumption is justified. Calculation of ABC from Equation (3.2) resulted in a rela- tively constant value throughout the course of the titra- tion whereas solving the Nernst equation using three inde- pendent redox couples (one flavin and two hemes) produced widely and systematically varying values of ABC at different points in the titration. (See Appendix.) Figure 1A shows the spectra collected from a single reductive titration of 9552. The titrations were performed in a sealed anaerobic titrator shown in Figure 15. Samples were extensively degassed by alternate exposure to vacuum and Ar gas. A sodium dithionite solution was introduced via an air-tight syringe in measured aliquots. The concentra— tion of the sodium dithionite solution was determined by performing a reductive titration upon a solution with a known concentration of 1umiflavin-3-acetate (obtained from Dr. Graham Palmer, Department of Biochemistry, Rice University), a two-electron acceptor having an absorption maximum at AA6rm1 with As (oxidized—reduced) of 10.8 mM"1 cm'l. A representa- tive plot of dithionite standardization against 1umiflavin- 3-acetate is shown in Figure 16. Optical spectra were re- corded and the procedure repeated until reduction was com- plete. Flavinzheme ratios of the samples used in different titrations varied from 1.25 to 1.12 with no apparent effect on the quantification of the titrations (1,2,, all titrations consumed between A.03 and A.18 electrons per molecule). 52 .oLNQQmoEpw p< cm Nope: m.» mm .meB E H.o :H mmmm oEopzoouzoo>mHu mo coapmppfip o>apo§pom m soap pmcfimuno mpuomam coapgpomn< .:H ossmfim EB? soc ohm man com Re , I On- , < .n< I 00. («g 20=<~== ._Huozpms :a com: unspospop opficoanufip pom pOHQ :oapmNfippmpcmpm .mH osswfim 2:255 t oE:_o> _ _ _ _ 4-III .. 9.0 V I q . a I 10N.O m. . n . a I 1 Omd « I I I 1 Ovd 55 Determination of relative concentrations from titra- tion data required some deconvolution of the spectral data. . Relative concentrations of oxidized and reduced hemes can be obtained directly from AA552 measurements; however, flavin concentrations can be determined from AA“75 only after heme effects at that wavelength have been accounted for. A plot of AA552 vs AAu75 (Figure 17) clearly shows non-linear be- havior. This is due to the fact that initially the hemes provide the only active redox couple leading to a slope of AAu75(heme)/AA552(heme) No.25. A reductive titration of horse heart cytochrome g_yielded this value of AAu75/AA550 throughout the course of the entire titration (see Figure 17). As the flavin begins to become reduced the slope changes to [Aeu75(heme)+Aeu75(flavin)J. ACflavin/Acheme/A8552(heme)>0.25 where ACi = change in concentration of species 1 By subtracting the effect of the heme component at A75 nm an accurate determination of AC(f1avin) can be made. Pre- vious attempts (30) to determine AB; between flavin and heme did not recognize the dual contribution to AA75 and conse- quently derived erroneous values for AEE. Determination O o of AEm by the above method yielded a value of EHeme - EFlavin = 37 i 5 mV. The reduction of heme and flavin 56 .m.s mo .aase s H.o ca shoe .AasoHEoso meHm Eopm pocfimpno HH 093v .33 m NEOL300p>o ammo: capo: new Alv mmmlo. mEopzoouzoo>mam no coapmpuau m>Hpospmp m wgfiLSU mmm<< .mM msa<< mo poad < .NH mesmfim nn¢<< ooe. com. com. 8.. _ _ _ _ .- X I I Ill, 00... u _. Sm<< nlin2uu. 57 redox centers in the protein as a function of electrons/ molecule added is plotted in Figure 18 along with the theoretical lines for two two—electron redox couples with as; = 32 mV and as; = A2 mV. The binding of a thiosulfate ion to the oxidized flavin moiety affects its redox capabilities. This is to be expected since the coupling of the binding and reduction reactions of the flavin will result in stabilization of the oxidized form of that center. This lowers its ap- parent midpoint potential. Figure 19 attests to the fact that thiosulfate binding lowers the flavin potential to the extent that hemes in 9552 become fully reduced after the introduction of only two equivalents of electrons into the molecule. When protein with a flavin heme ratio of less than one was used, the titrations described above yielded a bi- phasic behavior in flavin reduction. Quantification of the reduction remained atN Ae'/molecule but the flavin reduction occurred in a stepwise manner. Two components, one N20 mV and the other NAO mV lower than the hemes potentials, were present. A mixture of high flavin:heme (1.25) and low flavin:heme (.85) forms of the protein could be expected to produce the observed behavior. In any case, it is apparent that a low flavin:heme ratio in the protein does not arise from partial flavin reduction as the mole- cule still requires A e' to become fully reduced, nor 58 .zao>fipooamop .>E mm was >8 m: n m< Spas mmaazoo coppomHolozu 03» new mo>p30 HmOHumgomnp one muocop mofiwcmfipu new mopmsvm Como .mmmm mo coapmpufip o>Huospop m wcfipsc poops manooaofi pod mcogpomao mo coapocsm m mm soaposcop on cfi>mHm pew Axv @805 we ucopxo one .mfi mpsmfim m_:oo_o_>_\-o 0.». c.» on 0.. _ _ _ 0000 _mm 09000 no as a... e a I o c .8 on m c we I a o «s cc o< one I w 4X0 om o 4 q mm d x II nuOG d.x x0 nzw e 4 . a O 4 x s .. e a .. . o. oo. :32“. o 6661 x 59 . mml .mpozomoEum p< cm popes H.m mm .mm: 2 H.o :H m o meson mumMHSmOHnu mo coapmspfip o>Hpospop m mgfipsp , poppm manooHoE god msoppooao mo coapocsm m we coHpospop 050: we pcmpxo one .mH ogsafim m_:um_o_>_\-m on Io...- _ A x x x IIIIAXN % x x I oe mm D- x n on w. x I m. x U x [on x x x I00. 60 does it result from the binding of exogenous ligands to the oxidized flavin since this would serve to increase- AE; rather than diminish it. Further, there exists a cor- relation between the bleaching of flavin absorption and an alteration of its redox capabilities. The flavin moiety in the molecule apparently exists in an oxidized state irrespective of the absorption properties; however, its environment within the protein matrix is subject to changes which dictate its spectral and redox behavior. The mutability of the flavin's environment in 9552 is even more graphically illustrated by the fluorescence properties of the protein. E. Fluorescence Results The fluorescence properties of 3552 are totally domin- ated by the protein's flavin moiety and arise from the same isoalloxazine n + n* transitions seen in the absorp- tion spectrum of the molecule. The position and shape of 3552 fluorescence emission strongly resemble those of free riboflavin. At room temperature, both display a featureless emission band with a maximum at 525 i 3 nm. However, the fluorescence of £552 is strongly quenched relative to both free riboflavin and glucose oxidase, a flavoprotein containing no heme groups. Figure 20 shows the fluorescence emission spectra of the oxidized form of 61 .oHCEMm CHNCOCC on ou COHpHpUm ou CoHCQ H.@ ma 0» pmpmdnpm pCm mm: 2 H.o CH pogmampa NCNB mONmmmz pCm 20m Co mCOHpsHom ComepCmoCoo .wppooem mo mpom Cuon Com EC ms: mm: prCmHo>m3 COHCMCHoxo one .AI..IV memmmwz 2E m C0 CoHqupm Lopes CCOC CHMCIoCo H.m mm was s a.o ea mmmm as e gee mg-.-.-e m.s.ea .mase a a.o ea amen a: s has MAI-Iv H.m mm mm: 2 H.o CH mmmm E: m Any MA V zox SE N no CoHpprm poems Cson CHmnloCo H.w ma .mm: 2 H.o CH mmmm z: m Amv "no moHpHmCmpCH COHm ImHEo moCoommCosHm "HoCma quHm .m.> ma .Commsn mHCB 2 H.o CH HHm AI.I.IV mmmm oEopCooumoo>mHm z: m on oCm AIIIV ommpon mmoosHm z: m ADV m¢lllv CH>mHConHC z: H.o Amv "mo moHuHmCopCH COHmmHEm moCoomoCosHm "HonC whoa .om ossmae )‘ex: 442nm 62 GOO 500 600 $00 (D J) AJJSNBINI BONBSBHOO'H AHVHJJBBV Mnm) Figure 20 63 those three molecules with actinic excitation at AA2 nm. The position of the maxima (at NAA5 and 355 nm) in the excitation spectrum (shown in Figure 21) clearly demon- strates the participation of the flavin n + n* transitions in the fluorescence process. After heme reabsorption is taken into account, the fluorescence quantum efficiency of 9552 is approximately 1% that of free riboflavin and 33% that of glucose oxidase. This intensity is relatively invariant over the pH range 6.5 - 9.0, decreasing slightly to a minimum at pH N7.5, but increasing rapidly as the high pH limit (pH N10.0) for hemezflavin subunit binding stability is reached. The extent of this quenching is quite striking in that free flavins are strongly fluores- cent, some having quantum efficiencies as high as 58% (50). The quenching of flavin fluorescence upon binding to peptides is well established. Flavin interactions with aromatic amino acid residues in both flavoproteins and model complexes have been shown to result in flavin fluorescence quantum yields as low as 3% (51). The quench- ing exhibited in 3552 results in a quantum efficiency of N0.7% that of riboflavin in the same solvent, or N0.20% absolute efficiency (based on the Kotaki gt_al. value of 26% for riboflavin quantum efficiency). This anomalously low fluorescence level can be attributed to two effects. The first, and more important of the two, is the quench- ing due to protein tyrosine residues. Sequencing and CD 6A Excitation spectrum _. Of 9552 _1 Aem’ 523nm )- t: — _ (I) Z I LIJ F- ; __ __ I I I 1 II 300 400 ' 500 Mnm) Figure 21. Fluorescence excitation spectrum of flavocyto- chrome 3552 in 0.1 M Tris, pH 7.5 with emission intensity monitored at 523 nm. 65 studies on 9552 digestive peptide fragments containing the covalently bound flavin by Kenney 33 a1. (52) have estab- lished that the flavin is intimately associated with at least one and probably two tyrosine residues of the pro- tein. The circular dichroism spectra of these peptide fragments were indicative of direct tyrosin-flavin inter- action resembling the parallel stacking arrangement found in flavodoxin (53). They found that the fluorescence efficiency of the flavopeptides prepared by trypsin-chymo- trypsin digestion to be 1% that of free riboflavin, only marginally larger than that of the holoprotein. Heme reabsorption is the second factor which influences the apparent fluorescence yield and produces a decrease of ap- proximately 15% in apparent flavin fluorescence at 525 nm (based on a 5 mm pathlength in a 6 uM solution of 9552). These two considerations appear to be sufficient to explain the quenching of flavin fluorescence in 9552 and obviate the necessity of evoking direct heme-flavin energy transfer of the F6rster type (5“)- The binding of exogenous ligands to 3552 has a pro- nounced effect on the protein's fluorescence properties. The initial results of 0N” and 520; binding upon the visible absorption spectrum of 9552 are quite similar. Their effect on flavin fluorescence, however, is markedly different (See Figure 20). Shortly after cyanide binding the fluorescence yield of £552 dramatically increases as 66 Au75/A525 is restored to its original value, whereas S20; binding results in nearly complete quenching of flavin emission for an indefinite period of time. Clearly, two different effects are experienced by the flavin for the two different ligands. The CN', in time, .disrupts the delocalization of the flavin excited state in a manner that diminishes the effect of radiationless transfer (presumably to protein tyrosine residues). The thiosulfate binding, on the other hand, contributes to the efficiency of such processes. One possible explanation for these effects is a difference in the mechanism of ligand-pro- tein interaction between CN- and 820;. Thiosulfate behavior is probably due to direct adduct formation, as discussed earlier, which would be expected to increase quenching. .The increase in flavin fluorescence subsequent to cyanide binding cannot be attributed to adduct formation and must reflect more pervasive protein-cyanide interactions. Manifestations of these interactions are seen in the ef- fects of CN- upon the absorption and EPR spectra of the hemes in 3552 which indicate that CN- significantly alters the protein environment of those redox centers. The fluor- escence spectra of cyanide treated 9552 implies that these environmental changes extend to the flavin moiety as well. CHAPTER A MAGNETIC TECHNIQUES APPLIED TO FLAVOCYTOCHROME 9552 The application of MCD and EPR spectroscopies to flavo- cytochrome 3552 provides probes of the heme magnetic environments in the protein. This is desirable for a number of reasons. Specific magnetic interactions between the two heme groups in 3552 can be determined via MCD spectroscopy. MCD is a far more sensitive indicator of hemezheme interaction than simple CD since the latter tech- nique relies on the global effects of the molecule's inter— action with light whereas MCD is not directly influenced by the protein matrix. EPR spectroscopy allows discrimina- tion between the two heme groups and thus serves as an effective complement to the absorption and resonance Raman investigations which yield information concerning the average properties of two hemes. Moreover, the data ob- tained with absorption, MCD and EPR spectroscopies are progressively more specific to the d-orbital electronic system of the heme iron. Absorption spectroscopy probes primarily the w electronic system of the heme porphyrin. The spin-orbit interactions between iron d- and porphyrin 67 68 W— electrons are evident as perturbations upon the general properties of the porphyrin absorption spectrum. These interactions are directly manifest in the MCD spectra of hemes and give rise to the character and intensities of the transitions observed. EPR spectra are characteristic of the energy level spacing of the iron d-orbitals which is dictated by the combined ligand field effects of the porphyrin macrocycle and the local protein environment of the heme. I. Magnetic Circular Dichroism A. MCD Theory The qualitative sensitivity of hemeprotein optical spectra to the spin state of the heme iron has been dis- cussed in Chapter 3. MCD spectroscopy provides a rela- tively straightforward means of assessing the degree to which such iron spin state influences exist and of deter- mining the extent of interaction between the two heme centers in 9552. The theory for the origin of magnetically induced dichroism in hemes, however, is still in its forma- tive stages (55,56). In this application to 3552 MCD is used primarily as an analytical tool and, as such, only a brief description of the basis for such effects is given here. An MCD spectrum is simply a plot of the differential absorption of left circularly polarized and right circularly 69 polarized light by a molecule under the influence of a Zeeman field (measured in Tesla). It is conventionally plotted as Ae/Tesla (where A8 = 8R - eL) XE wavelength. There are three types of MCD effects arising from differ- ent origins. These are designated Faraday A, B, or B terms and each gives rise to a characteristic band shape in MCD spectra. A terms arise from transitions to orbitally degenerate excited states whose degeneracy has been split by an external magnetic field. Heme w-w* transitions conform to this situation and, in the absence of spin- orbit coupling, would be expected to yield A terms in their MCD spectra. Their shape resembles the first derivative of the absorption band. Faraday B terms originate 122 transitions from Zeeman split ground states. Thus, they occur in paramagnetic materials and exhibit a difference in absorption intensities for left and right circularly polarized light due to the fact that there is a Boltzmann population distribution in the split ground states. This population difference is exponentially proportional to l/T and imparts a temperature dependence to B term intensity. B terms are thus expected in all hemes transitions that include the iron d-orbitals. Moreover, Byterm behavior has been found to dominate the MCD of w-w* transitions in a variety of heme proteins (57,58) and is indicative of iron spin-porphyrin orbit coupling. The bandshape of C-terms resembles that of the absorption peak. Faraday B terms 70 .Amm ooCoCoCmm EOCCV ooCoComme CHme mH oomCu CHHom on oHHCz uCaHH poNHCmHoo mHCmHsoCHo pCMHC pCm pmmH mo CoHpQComnw on» ouOCop moCHH coupon HoCmQ some: one CH .ACNCQSV mECop omon mo monCmUCmp o>HuooommC on pCm ACN3OHV mECop o pCm m .< Go: on omHC wCH>Hm mommNOOCQ HNonmCQ oCu mo COHpmuComoCCoC oHmeoCom < .mm ossta 71 mm QHSMHm + JI--..- I 22m .. acacia Il- PH xn-II..." 22m 02:8 mEcma. m Elma < 72 arise from a perturbative mixing of transitions by the applied magnetic field. They occur, to varying degrees, in all compounds but are generally weak in gytype hemes (59). B terms also resemble the bandshape of the absorp- tion spectrum. A schematic representation of the processes giving rise to Faraday A, B, and B terms is shown in Figure 22. B. MCD Results The MCD spectra in the Soret region of oxidized, reduced and reduced +CO (at pH 7.95) 3552 are shown in Figure 23. The holoprotein displays MCD intensities and bandshapes that closely parallel those of other low spin mono-heme 3 proteins (60). The Soret MCD in oxidized £552 is composed of mixed A and Bfterms which produce a derivative shaped .MCD curve with a zero crossover at A10 nm. Previous studies (57) have established an empirical correlation between the intensity of the low energy trough and the percentage of the molecular population in a low spin con— figuration. The values obtained for 9552 indicate that it is exclusively low spin heme g_(1;§;, no thermal equilib- rium exists between low-spin and high-spin configurations). There is no indication of any magnetic coupling between the heme groups of oxidized 3552 seen in its MCD spectrum. Such coupling could be expected to result in a mixing of i do 0 I _ Reduced _ __ Oxidized -I20r .._._ Reduced + CO ‘ e/H (M‘cm-TY' 460— _ ABS I00 e (mM"-cm") 50 350 460 790 Mnm) Figure 23. The MCD spectra of the reduced, oxidized, and reduced +CO forms of 9552 in 0.1 M Tris, pH 7.5 in the Soret region are pictured above the absorption spectra of those species. 7A states by magnetic dipole coupling or electron spin-spin interactions. The former effect would lead to a dramatic increase in Beterm intensities introducing a severe asym- metry to the derivative MCD Soret bandshape whereas the latter would produce a quantum mechanical admixture of spins resulting in a departure from low spin behavior. The com- plete absence of any observable deviation from low spin mono-heme g behavior in the MCD spectra of oxidized 3552 clearly rules out strong coupling such as would occur for a bridging ligand between hemes, and makes even weak direct heme-heme interaction unlikely. Upon reduction the MCD spectrum exhibits the characteristics of an S = 0 heme system. Heme spin-porphyrin orbit coupling is no longer a factor and the spectrum is composed of a mixture of A and B terms peaking at A17 nm. The MCD behavior of £552 in the visible region of '3 absorption spectrum is 2552 qualitatively different from that in the Soret region while remaining entirely consistent with behavior exhibited by mono-heme 9 proteins (See Figure 2A). In contrast to the Soret region, A terms form the dominant contribution to the visible MCD. The zero crossovers of these terms coincide with QOO band (at 552 nm) and the various vibronic components of the Q01. These are expected from porphyrin n-n* transitions that are not spin-orbit coupled to the paramagnetic iron d-orbital system and are stronger for the reduced protein than the oxidized, reflecting the 75 200 MCD I60 —' fl — l20 — - -460-— I —- -200 — -— Reduced .— --- Oxidized ABS 500 550 x(nn0 Figure 2A. The MCD spectra of reduced and oxidized 3552 in 0.1 M Tris, pH 7.5 in the visible region are pictured above the absorption spectra of those species. 76 increase in Q00 absorption upon reduction. While the Q01 band of oxidized 9552 display a greater absorption in- tensity than the Q00 band, the MCD is much weaker. This arises from the fact that vibrational components of dif- ferent symmetry can have A terms of opposite sign and will tend to cancel each other. A reductive titration of c _5 (performed in the same manner as the absorption titration 52 under an Ar atmosphere described earlier) resulted in a smooth transition between oxidized and reduced MCD bandshapes and provided no evi- dence of bi-phasic behavior with respect to reduction in either the Soret or visible regions. The hemes in 9552 then, apparently either contribute equally to the observed MCD spectra or titrate with approximately the same redox potential or both. The binding of CO to reduced 9552 produces a peak sharpening in the Soret MCD spectrum and a shift of the peak to A12 nm reflects the changes seen in its optical spectrum. However, the spectrum now clearly exhibits two components. It is composed of the "typical" combination of A_and B terms seen with the unbound reduced protein with a broad peak and trough at A17 nm and A29 nm respectively, super- imposed upon a sharper A term with a zero crossover of AlA nm. The multicomponent nature of this spectrum strongly implies that CO binding occurs to only one of the two hemes in 3552 producing the sharp Afterm in the MCD spectrum 77 with the unbound heme producing the remaining MCD features. Moreover, the Soret MCD displays bi-phasic behavior during the course of a reductive titration under a CO atmosphere. This is shown in Figure 25. The growth of the sharp A term preceeds the gradual transition of the features aris- ing from the "unbound" heme. This is to be expected since under a CO atmosphere, the reduction of the "bound" heme: heme g(Fe3+) + e— z heme 9(Fe2+) is coupled to the chemical binding of CO 2+). heme B(Fe2+) + CO I heme B(Fe CO The effect of this coupling on the heme midpoint potential is O _ O .059 1 EM+co ‘ EM-co + r1 1°3 (1 + KaECOJ) where Ka is the CO affinity constant for £552 previously measured (30) to be 7 x 103 M'l. The solubility of co in cold buffer is 1.A mM. Thus 0 _ O EM+CO - EM_CO + 0.058/1 and a 58 mV difference in potential should exist between the binding and nonbinding hemes. This is sufficient to insure that nearly all of the electrons initially introduced KN) I I l l I l l I I I l I I MCD . Reductive Titration . under CO __ L . F — E .. 9 2 ,., :1: ° , d \ . .e — u .' -—6o _ -so 5 I — .I . —|OO l L I I I III I I I I I I I 360 “mm 50 Figure 25. The MCD spectra obtained in the Soret region from a reductive titration of 3552 under 6 psi of C0 in 0.1 M Tris, pH 7.5. Traces of the peak from left to right (or trough from right to left) are: (-——), 0% reduced (---), 10% reduced, (°---),25% reduced, ( ), 35% reduced, (---), A5% reduced,(----),100% reduced protein. 79 will produce the CO-bound species. These results clearly show that CO binding to £552 occurs at only one heme and does not result in CO intercalation between hemes as has been suggested as an explanation for the molecule's CD spectra (20). A CO reductive titration performed at pH 11.1 showed no multicomponent features in the Soret MCD nor was any bi-phasic behavior with respect to heme reduction potential found. This suggests that at high pH a conformational change in 9552 occurs that allows CO access to both of the hemes. The MCD spectra of oxidized 9552 heme peptide and CN-- bound 3552 at pH 6.0 were also investigated. Neither was found to deviate substantially from the spectrum of the unbound oxidized holoprotein. Thus, the alteration or removal of the flavin prosthetic group has little or no effect on the MCD properties of the hemes in 9552. Flavins themselves exhibit quite weak MCD spectra. Because of the limited symmetry (C8) of the isoalloxazine moiety these spectra would be expected to be composed of Beterms. Re— cent studies (61) have determined that two weak (As/Tesla 10) positive MCD transitions occur at N370 nm and NA60 nm in FAD. The former transition would be completely obscured by the intense Soret heme MCD in 2552' Some evidence for the A60 nm transition can be found in a comparison of the oxidized heme peptide and holoprotein MCD spectra, but it is too weak to quantify with confidence. 80 II. Electron Paramagnetic Resonance Spectroscopy A. Theory EPR spectroscopy has been extensively applied to both high Spin (S = 5/2) ferric hemoproteins (62). Although other valence states of iron are paramagnetic, the spectra of these states are extremely broad (presumably due to in- creased spin-orbit relaxation) and generally undetectable. Only a brief explanation of heme EPR specific to low spin heme B_will be given here. The model employed here was originated by Griffith (63). More detailed theoretical explanations can be found in the work of Kotani (6A) or Weissbluth (65). All low spin hemes display three distinct g-values which can be analyzed assuming that the cubic (octahedral) splitting, A, is sufficient to force all five iron d electrons exclusively into the t28 orbitals, g = dyz, n = dzx’ and c = dxy (See Figure 26). It is more con- venient to deal with a single electron hole than five electrons, thus, a wavefunction for one of the six pos- sible low spin configurations is: Ia'n2c2> = |€+> (u.1) where i denotes unpaired electron spin and paired orbitals 81 .mHmquCOIp COCH 0C» mo mHm>oH meoCo on Com: mCoHpoSCNC zpumeshm o>Hmmooosm mo Commwm oCB me 1.9 6589:”. .0800th 6.63060 co. moi CV36 --..- #1 Xaa. / . / xn .~> .>x < \w p .0 “/3. C023 . /// fi \\\ fi / o \\\I spit o .9. / $8 .- _ / x \ \ GOO \\ \ _ \ NC [/0/1 \u \\ «#115me m .wm opszm 82 (having no net orbital angular momentum) are supressed in the ket expression. Similar expressions are obtained for the other five configurations. Since gx # gy # gz for low spin hemes a reduction of field symmetry from octahedral to rhombic (D2) is anticipat- ed. The energy level diagram for successive levels of d- orbital symmetry reduction is shown in Figure 26. The origin of the three distinct resonances displayed by low- spin hemes can be traced to this splitting of d-orbital degeneracy and subsequent spin-orbit interactions between these orbitals, as follows: Spin-orbit coupling will mix the orbitals resulting in three sets of Kramer's doublets corresponding to m values of il/2, 13/2, and i5/2. These 3 are linear combinations of d-orbitals having nonvanishing spin-orbit matrix elements among themselves that are eigen- functions of the total Hamiltonian including the coupling between spin (S) and orbital (I) angular momentum. The spin-orbit operator 7‘5 is nonzero between (5+,n+,;') and (5', n',c+), thus the resulting wavefunctions are: Ag+ + iBn+ + 0; €- V II (A.2) -Ag' + iBn' + Cg+ '6 V II for the lowest lying (ms= 1/2) Kramer's doublet. Moreover, the eigenvalues of the orbital interaction with an external 83 magnetic field can be determined from the matrices of magnetic interaction operators AZ + 2sz 1x + 23x’ and 2y + 2sy using 9+ and w- as the basis functions. This yields energy separation of: )2 - 02] ()_() ()_ AB 2 - E+z — E_z - 2BHz[(Al — B1 1 where TD ll Bohr magnetron; and Hz applied magnetic field. betweenluf> and I¢7> and electron paramagnetic resonance will be observed at: Thus 2|(A - B)2 - 02| 09 N ll Similarly, 2|(A - C)2 - B2| (1.3) 2|(B + 0)2 - A2| 09 N II 8A These relationships plus the normalization requirement A2 + 82 + 02 = 1 allows for determination of A, B, and C from experimental g values. A, B, and C can be directly translated into the rela- tive energies of the g, n, and c orbitals by determining the eigenvalues of the combined spin-orbit and crystal field Hamiltonian, +-> HT = "*8‘3 + V(crystal field) 1 for the free ferric l- spin-orbit parameter = A35 cm— ion V = potential of the rhombic field. + Solving Hw+ = Em one obtains . 1 l _ Meg - E) - 1B .2 1 + 7 10 - 0 i 1 _ A§A+iB(en-E)--2-lC-0 (A.A) A i _ A-2-+iB‘2-A+C(EC-E)-O which can easily be solved for the energy differences 85 between 5, n, and C given the values of A, B and C. These energy differences are conveniently expressed in multiples of A, the spin-orbit coupling constant for the system. .In the free ion 1 = A35 cm"1 and is probably lower in complexes due to a ligand induced decrease in spin density at the metal (66)- Two parameters useful for the quantification of the asymmetric ligand field experienced by the iron d-orbitals in hemes g are the tetragonality (A/l) and rhombicity (V/A) of the field. The tetragonality reflects the con- tribution of the large axial field component and thus depends largely on the charge donation ability of the z- ligands. The rhombicity, on the other hand, is a measure of the overall geometric distortion of the complex and is sensitive to distinctions between ligands in the x-y plane of the complex. A systematic classification of iron porphyrins based on their ligand field parameters was originated by Peisach and co-workers (67) and later ex- tended to a variety of hemes 9 (68,69). B. EPR Results The EPR spectrum of oxidized is consistent with c -552 that expected of low spin heme B, but is complicated by the fact that at least three separate heme ligand fields are evident in the holoprotein. A previous study of the 86 protein's EPR spectra by Strekas (33) revealed three sets of resonances whose respective low field components (gz) occur at g = 3.35, g = 3.00 and g = 2.89. He interpreted the pH dependence of these signals as a pH dependent interconversion of the signals associated with gZ = 3.35 and g2 = 2.89 hemes with the gZ = 2.89 form being favored at high pH, while noting that gz = 3.00 signal displayed no pH dependence. Cyanide binding was found by Strekas to reduce the intensity of the resonances associated with the gz = 2.89 heme and he suggested that this implied heme/ flavin interaction. The EPR spectra of 9552 obtained in our laboratory generally confirm those obtained by Strekas. Figure 27 shows the EPR spectrum of ferric 9552 at pH 7.5 and 70K. Three distinct sets of EPR resonances are apparent, one with gz = 2.89, g = 2.35 and a very weak gx = 1.55, a y second with gZ = 3.02, gy = 2.25 and a very weak gx = 1.36, and the third with gZ = 3.35 and both gx and gy too broad to detect. With these g- values it is possible to assign the heme axial ligands for each set of reson- ances. This can be accomplished by correlating either the gZ values or the ligand field parameters of 9552 to those of other gftype hemes with known axial ligands. Actually, these two methods are largely similar since both the value of gz and the tetragonal field are directly proportional to the electron donating ability of the axial ligand. The 87 .COHpmHSUOE 0 OH Cam COHpchmp wa NMH.m mo 35 N CCHB xom.m as eoeaeneo m.s mo .mHse : H.o cH 2mm .5. .595 «one 2100. . ON.N.O onNuO eosaeaxo z: oeH no esaocoon mam ca Io exec .nm NCCmHm 88 9. X‘102 A-ZBB _. 0-235 9‘0 o - 2.25 BIDP- x Z£>- . o GAD- :n 2: "3 5£)- C: 0 +— 55- x n 4&)- 3.0 -— ° 21)-- x A l£)—- 0 a I i l I!) 2&) 3M) . I (Microwave Power) ’2 Figure 28. Microwave power saturation curve for the 9:2 = 2.88 (A), 82 = 3.02 (x), gy = 2.35 (D) and gy = 2.25 (o) resonances of'g552 under the same conditions as Figure 27. 89 correlation of 9552 gZ-values and ligand field parameters with those of other hemes g is displayed in Figure 29. Note that a general correlation between tetragonality, gz values and the electron donating power of axial ligands exists for the species shewn in the order amine > imidazole > imidazoleO > methionine. The gz value of the pH stable heme falls within the range expected of either bis-imif dazoleOO or imidazoleo/methionine ligation. The low value of its tetragonal field, however, makes the latter axial ligation scheme more likely. The gZ—values of the high and low pH forms of the pH—labile heme are those expected from amine/imidazoleO and bis-imidazoleoo respectively, although the rhombicity of the bis-imidazole form is somewhat higher than that of heme g_models (69). This is not unexpected. In fact, in the limit where all six nitrogens coordinated to the heme iron (four from the porphyrin and one from each histidine imidazole) donate electrons equally, a purely rhombic field (V/A = .69) should result. In the absence of the solvent effects found in the bis-imidazole heme 2 models such a situation may obtain for the pH-labile heme in 9552. Thus, it can be seen that the EPR spectrum of £552 is consistent with an axial ligation scheme of involving one pH-stable heme with methionine/histidine ligands and one pH-labile heme favoring lysine/histidine ligands at low pH and histidine/ histidine at high pH. Table 3 summarizes the g-values, 90 .on mCHopoCQ .m oEmCOC-mE nSOHCa-S 0C0 A00 mmmm mEopCoouzoo>mHm CH meson 0C0 Com mpHHmCowogump .m> mpHoHCEOCC mo pOHQ < .mm opsmHm Q): >:_ocooo.:m._. 0.0 0.0 0.? 0.m 0.N _ _ _ R H In 20.: 0:6 020: I ON. A0 N0 In .Iinm .02-Eu omnm .02.. EU uua 0:30 0:206 Cu w low. M. D. .656. o (M ./0 To». 38> N1 0.020. In mama “”030 0951 . W oE.\oE. errata-0n. 0 /Oc a m 30 can... 100 I A0).-£01.00; O 0.50. In Nanak In 30. w .30 00.61 bu 1 Om. 91 Table 3.‘ Ligand Field Parameters for Various Hemes g. Species gx gy gz A/X V/D Ligands Ref. 9552 heme 1 3.02 2.25 1.“ 2.8“ .57 Met/His° heme 2 2.89 2.35. 1.6 3.10 .68 His°/His° 3-35 ---- --- Lys/His° horse heart cytochrome 3 pH 7.0 3.06 2.25 1.3 2.56 .58 Met/His° 68 pH 11.0 3.37 2.10 --- 6.2 .25 Met/Lys 69 pH 2.5 2.90 2.“ l 5 2.9 .80 His°/His° 69 CM-Met 65,80 3.UO 2.08 --- 6.4 .27 His°/Lys 69 bis-imidazole heme g 2.92 2.30 1.5 3.12 .62 Im°/Im° 68 Yeast iso-l (pH 1n) cytochrome c 2.71 2.26 1.8 3.22 .67 His°/His° 69 Euglena cytochrome c 3.20 2.05 1.4 “.5 .35 Met/His” 69 CM-Met = carboxymethylated. Met-methionine, His°-neutra1 histidine, His'-deprotonated histidine, Lys-lysine, Im°-neutra1 imidazole. 92 ligand field parameters and axial ligand assignments for the hemes shown in Figure 29. C. Reductive Titration The ability of EPR spectroscopy to distinguish between the two hemes in 9552 allows the determination of.their relative redox potentials. Since ferrous heme is dia- magnetic (S = 0), it is possible to monitor the extent of reduction of a given heme by measuring the decay of its EPR signal intensity. Thus, by plotting the decrease in signal at g = 3.02 vs the "average" extent of heme reduc- tion obtained via absorption spectroscopy, comparison of the relative redox potentials of the two hemes in 3552 could be made. Simultaneous determination of the absorption and EPR spectra of degassed 3552 under an Ar atmosphere during a reductive titration was accomplished by modifying the anaerobic titrator shown in Chapter 3 to include a side— arm/value system for the removal of an EPR sample (See Figure 30). An EPR sample tube could be attached to the sidearm and evacuated using the two-way stopcock (#1). The titrator was turned so that the sidearm nipple was covered with sample and the desired amount (m0.3 ml) of sample was drawn into the sidearm which was calibrated to determine sample volume to the nearest0.051m1. The titrator was then turned so that the nipple was no longer covered 93 EPR SAMPLE TUBE T0 VACUUM ABSORPTION C UVETTE Figure 30. The EPR/Absorption anaerobic titrator. 9” with sample and the sample now in the sidearm was trans- ferred to the EPR sample cell by opening both stopcocks l and 2. The now-filled EPR sample tube was then anaerobic- ally removed (by closing both stopcocks) and frozen in liquid nitrogen. Stopcock 2 could now be removed from the sample tube, more reductant added, and the process repeated for the next point. A sodium dithionite solution whose concentration was previously determined as in the titrations described in Chapter 3 was used as the reductant in these titrations. . A comparison (Shown in Figure 31) of the decay of g = 3.02 vs extent of total heme reduction reveals that the methionine/histidine ligated heme in 2552 has approxi— mately the same reduction potential as the protein's pH- labile heme. This precludes the possibility of sequential reduction of the two hemes and, in light of the previous potentiometric titrations (30), establishes both hemes as having E; ~ 0 mV. This is an anomalously low potential, particularly for the heme with methionine/histidine liga- tion.©orse heart cytochrome g_which also has methionine/ histidine ligands has E; ; 265 mV)- The nature of heme axial ligands has been postulated to be the dominant effect upon heme redox potential (70). All other factors being equal, the methionine/histidine ligated heme would be expected to have a much higher redox potential than either form of the pH-labile heme due to 95 a .5m onsmfim mm mpmumEmgmd HmpcoESppmcH mean 0:» Spas m.w ma .mfine 2 duo CH mmmm z: ooaé mo COHQMApHp 0>Hpozomp 0 mo 009500 on» mafipso cooom cacooaos pom mcoppooam amw mocmcommp mo.m u w 050 00 amooo one .Hm opsmfim 0.30.0.2 \-o 0.¢ 0m 0.N 0.. _ 5 _ & gm 0 oo 0 20:8... mam. a a. o 4 0.. I. ococSEEom :32... . o o « 1 0m All 0 . « 0 4 Q . 4 o.m I o « low % O O ‘ M H% Oml 0 << 100 D. nu o do 0.. o 4 I! 0 v I. 00 4 d l 00 w < «a o 4 4 Illv... O < 44 a a q «a 2 22.2.: 60.60. 2.6.... o loo. 226...: 60.60. 0E0... a .8». .. Nu . 8.6:: mam: 9:01 < 96 the greater n-acceptor power of methionine over either histidine or lysine. Investigations of well characterized heme g analogs suggest that thioether/imidazole ligation is responsible for an 160 mV shift in iron redox potential relative to bis-imidazole ligation, independent of environ- ment (71). The same trend is apparent in small heme proteins. Monoheme 9 proteins with methionine/histidine ligands have redox potentials of between -60 and +400 mV whereas those with histidine/histidine ligands are generally much lower (-200 to -500 mV) in potential (72). This is obviously not the case for 3552. Both hemes exhibit approximately the same redox potential despite the disparity in their axial ligands. There are other exceptions to the general trend in the dependence of heme potential upon axial ligand configuration, most notably spinach cytochrome g which has lysine and histidine as axial ligands but possesses a midpoint potential of +H20 mV (73). This indicates that factors other than axial ligation figure strongly in the determination of heme redox potential. One such influence is the hydrophobicity of the medium surrounding the heme. Theoretical models (7h) suggest that small differences in the volume and dielectric constant of the immediate heme environment can result in changes of hundreds of millivolts in the apparent potential of the heme. As the heme en- vironment becomes more hydrophobic, the ferrous heme is more stabilized relative to the more highly charged ferric 97 state and the midpoint potential of the heme would be expected to rise. Experimentally, the potential exhibited by hemes in a homologous series of bacterial monoheme 9 proteins has been linked to the degree of hydrophobicity of the heme protein environment by Pettigrew 22.2l- (75). This effect is independent of axial ligand effects. The difference in potential between the pH stable heme in 3552 and horse heart cytochrome 9 indicates that the former may be in a distinctly more hydrophilic environ- ment than the latter. Moreover, in order for the pH- stable (methionine/histidine) and pH-labile (histidine/ histidine or lysine/histidine) hemes of £552 to have the same potential they likely exist in different protein en- vironments, with the environment of the pH-labile heme being more hydrophobic. D. Flavin Semiquinone Flavins can exhibit a number of oxidation states during the course of a reductive titration depending upon whether the reduction proceeds through a one- or two-electron step. Addition of one electron to the flavin results in the crea- tion of a paramagnetic semiquinone free radical (76). The unpaired electron in the radical is extensively de- localized and appears in an EPR spectrum as a resonance at g a 2.00. Several flaVOproteins, most notably amino acid oxidase (77) and flavocytochrome b2 (78), have been 98 shown to exhibit semiquinone behavior upon partial reduc- tion. The simultaneous addition of two electrons to the flavin results in a diamagnetic species that is EPR silent. Samples of 9552 obtained during a reductive titration under an Ar atmosphere were examined for the existence of flavin semiquinone. EPR spectra were obtained at N13O°K since EPR signals from free radical systems saturate too easily to be observed at liquid helium temperatures. Record- ing a spectrum at the higher temperature has the added advantage of increasing the dipolar broadening of the heme systems to the point where the heme resonances are un- detectable and no longer complicate the spectrum. A weak signal at g = 2.01 (See Figure 32) was observed in par- tially reduced 9552. This signal increased in intensity as the extent of reduction reached 3e-/molecule and thus is consistent with the relative redox potentials between flavin and heme obtained via absorption spectroscopy. However, the magnitude of the signal was quite small and could account for no more than 5% of the total flavin content of the sample. This indicates that the major pathway for flavin reduction does not involve the semi- quinone form of this species and implicates a concerted two-electron transfer as the vehicle for flavin oxidation and reduction in the protein. The EPR spectrum of the flavin free radical can also be used to determine the ionization state of the semiquinone species formed. 99 g=200 .l dH L w FI-——a>- Figure 32. Flavin semiquinone signal obtained from mlOO uM £552 in 0.1 M Tris, pH 7.5 at 143°K after the introduction of 3 electrons per molecule. Micro- wave power was .5 mw at 9.122 GHz withESG modu- lation. 100 According to Palmer gt al. (78) the bandwidth of the neutral radical is 190 while the anionic semiquinone has a band- width of 15G. The bandwidth of the flavocytochrome £552 radical is 1ui2 G and thus it can be assumed to be anionic. It is interesting to note that during the reduction of flavocytochrome 92’ which contains only one heme per flavin, an initial burst of fully reduced flavin is followed by the accumulation of up to 50% of the flavin as semi-- quinone (18). This has been interpreted as resulting from the rapid distribution of one electron from reduced flavin to the heme followed by a slower addition of a third electron to the flavin semiquinone. For flavocytochrome 9552 it is possible that the two hemes in the protein act cooperatively to allow the simultaneous transfer of two electrons from the reduced flavin, thus obviating the necessity of the flavin semiquinone in the reduction mechan- ism of that molecule. E. Exogenous Ligand Binding EPR spectroscopy also allows for the unambiguous assignment of the CO binding site in 3552. At neutral pH absorption and MCD spectroscopy indicate that only one heme in 9552 binds CO. Since CO-binding serves to increase the apparent heme redox potential by m58 mV, an EPR reductive titration performed under a CO atmos- phere can easily identify the CO-binding heme. Figure 33 101 EPR Rahdwemehn vim umerO 0:215 9:215 ' Y .3135 93'2” 0%!“ ED 20% I! H —->- Figure 33. EPR spectra obtained during a reductive titra- tion of ~100 uM 9552 under 6 psi of carbon monoxide. Temperature and instrumental parameters are the same as Figure 27. 102 illustrates the effect of CO-binding on the £552 EPR spectrum. Clearly, the pH-labile heme now titrates with a higher potential than the pH—stable heme as the reson- ance at gz = 2.89 is almost completely removed by the addition of 0.8 e‘/molecule. CO binding in thus g5'52 necessarily involves a ligand displacement at the pH- labile heme and further substantiates the mutability of its protein environment. The pH-stable heme, on the other hand, does not undergo ligand displacement (i;g;,'bind CO) until the protein has begun to denature at high pH (pH = 11.0). As indicated in the absorption studies discussed earlier, the effects of CN- are more pervasive than simple flavin-CN- adduct formation. Strekas interpreted the reduction of the gz = 2.89 resonance in the 3552 EPR spectrum that results from the addition of CN- to the sample as arising from an alteration of the pH-labile heme's environment. However, he found no effect upon 3 ference in mechanism between S20= and ON. binding. 3 In an attempt to elucidate the effects of CN' upon . S20 binding to the protein. This implies a clear dif- the hemes in 9552, CN was added to a sample of low flavin:heme 2552 at neutral pH. Under these conditions, no effect on the absorbance at N75 nm was noted. None- theless, the EPR spectrum taken from a sample frozen 1/2 hour after CN' addition displayed the bleaching of the 103 resonances at g = 2.89 and 2.35 seen by Strekas. For a sample frozen 1.2 hours after CN' addition, the EPR spec- trum begins to display alteration of the g = 3.35 and the g = 3.02, 2.25 resonances as well. Figure 3D displays the EPR spectra of a sample that had been treated withESmM cyanide. Initially cyanide affects only by the pH-labile heme. It is apparent that histidine/histidine ligation at the pH-labile heme is significantly disrupted independent of flavin CN' adduct formation. The incorporation of CN' as a heme axial ligand results in a broad gz z 3.h5 resonance accompanied by a marked decrease in rhombicity in myoglobin derivatives (79). Some alteration of the g = 3.5 region is seen in the 3552 CN- spectra, but it is too broad to quantify. On a longer time scale disruption of the methione/histidine axial ligation of the pH—stable heme is evidenced by a decrease in intensity of both its gZ = 3.02 and gy = 2.25 components. This parallels the observed decrease in the 695 nm optical absorption band as a function of CN' binding. CN- binding to the heme moiety in horse heart cyto- chrome g has been characterized by Dyer gt_al. (80) who found that CN' bound in a ligand displacement reaction to the native protein and a variety of cytochrome g_deriva- tives. Their mechanism postulated a protein conformation step which opened the heme crevice as the rate determin- ing step. Activation energies as large as 17 kcal and 1014 g=30l I SmM CN— Af= lOmin. 9:225 dxll dH 973.0! 5mM CN' 9_ -2. 25 A1=L0hr. g= 2.35| Figure 3A. EPR spectra of mlOO uM 3552 with flavin:heme ratio ml.0 in 0.1 M Tris, pH 7.5, after the addition of 5 mM CN‘. 105 reaction rate constants as slow as 6.0 x 10'2 sec"1 were found for the cytochrome c derivatives. Those findings correlate well with the observed behavior of 9552 upon exposure to CN'. The cyanide apparently attacks the pH- labile heme particularly when it has two histidines as axial ligands. This is another indication of the mutability of the environment about this heme. The effects of ON“ on the pH-stable heme could arise from two sources. If its protein environment presents CN- with too large an activa- tion barrier to allow reaction at all, then the disruption of methionine-heme interaction would have to occur via a protein conformational change induced by attack upon the pH-labile heme. Alternatively, the activation barrier might simply be high enough to force the ligand displace- ment to proceed very slowly. The data obtained to date do not allow us to differentiate between those possibilities. The changes in heme environments resulting from cyanide- heme interactions may explain the observation by Vorkink (30) that CN' binding to 3552 resulted in a loss of the derivative shape in the heme Soret CD spectrum whereas SC; and $20; binding had no such effect. This was inter- preted by Vorkink as evidence of heme-flavin interaction, but more likely arises from the direct alteration of the heme environment by the cyanide. CHAPTER 5 RESONANCE RAMAN SPECTROSCOPY OF FLAVOCYTOCHROME 3552 Resonance Raman spectroscopy has been shown to be a powerful tool for the elucidation of the structure and function of biological molecules, in particular those proteins which contain a heme moiety (“1). The informa- tion obtained in a resonant scattering experiment is specific to the vibrations of the heme active site, and thus can be used to characterize both radical changes in iron redox and spin states as well as the more subtle perturbations due to alterations of the protein environ- ment surrounding the active site. For example, the vibra- tions of porphyrin ring substituents can be observed directly (81,82), insight into the planarity of the porphyrin ring can be obtained (83), and porphyrin metal- axial ligand properties can be determined independent of the magnetic state of the metal (8h). The technique should be particularly sensitive to vibrational manifesta- tions of chromophore interactions. In the experiments described in this chapter no evidence of direct heme/ flavin or heme/heme interactions through either the heme 106 107 axial positions or periphery was found. The resonance Raman s ectra of c p —552 direct, protein-mediated heme/flavin interactions. can be interpreted in terms of in- A. Raman Theory The total intensity of radiation scattered into a solid angle of Mn due to a Raman transition in a molecular system is (85): Total _ 7 5 u l 2 IScattered ' (2 " /a)IoVs 020°(aoo)GFI (5'1) 3 where IO and V0 the incident radiation intensity and frequency respectively and vs represents the frequency of the scattered radiation (i;§;, Vs = v0 i VGF, where the minus sign refers to Stokes and the plus sign anti-Stokes scattering) (apo)GF is the po component of the polariz- ability tensor connecting the initial and final molecular eigenstates defined by the relationship: Px 0Lxx axy O‘xz EX Py = ayx dyy adyz Ey P d a a E 108 and can be evaluated via a second order perturbation ap— proach originally developed.by Kramers and Heisenburg by analogy with classical dispersion theory. 2L_ + E hv A ‘2 V II OIF" 1 (5.2) GE-hvo+iFE thE+hvo+iFE where up, “o are the electric-dipole moment operators in the directions 0 and p (e.g., u = £e(r ) ) and (r ) is the --- O K K p K 0 th p'th component of the K electron's position vector. P E is the natural half-width of the state IE>. The summation runs over all intermediate states, lE>, exclusive of IG> and |F>. A schematic representation of the relative energy levels of the states involved in a resonance Raman experiment is shown in Figure (35). Far from the resonance hng - hvo >> O and both terms in Equation (2) contribute to the scattered intensity. As resonance is approached the energy denominator of one vibronic manifold becomes much smaller than the rest and it dominates the summation over states: The remainder of this section will confine itself to the resonance case. The evaluation of the interaction between the electric- dipole operator and the eigenstates of the molecular system in question lies at the heart of Raman theory. The symmetry and intensity of these interactions are directly manifested 109 E2 E / ' * GE) ‘— hpo h"s ? two hvs K _“‘_F F—‘+'_ G Normal ~ Resonance Ramon Roman Figure 35. Raman scattering processes 110 in the polarizability tensor, and hence the scattered radiation. Several methodologies exist for such evalua- tion. One of the most useful from a spectroscopic stand- point is the vibronic coupling model first advanced by Albrecht (86) and subsequently expanded upon and applied to porphyrin systems by others (87,88,89). It is summarized as follows: The terms |G>, lF>, and |E> in Equation (5.2) represent the wavefunctions associated with the total (vibrational and electronic) Hamiltonian of the system. The adiabatic Born-Oppenheimer approximation is employed in which the vibronic states are constructed as products of pure vibrational states, |n(R€)>, and pure electronic states, |g(R€,r)>. Thus 30> = |g(R€,r)> n (Rg)> (5,3) where RE and r are the vibrational normal coordinates and the electronic coordinates of the molecule, respectively, and: n(Qg) = fi¢g ¢g = harmonic oscillator i i i wavefunctions Substituting the above Born-Oppenheimer states into Equation (5.2) one obtaine- 111 ) = (5.”) l gm+gn c (“no 2 z e V hvgm,gn-hvo+irev for the resonance case. Here age = fg(Q€’r)uU 9(Qgr)dr and represents the electronic transition dipole between the electronic states |g> and |e>. In order to perform the remaining integration over nuclear coordinates, the parametic dependence of o,p on nuclear coordinates must be removed. To that end, it is necessary to expand the matrix elements 0 and p in a Taylor series about the equilibrium nuclear coordinates Q =0 _ ' _ 30 _ , 0(Qg) - om.g — 0) + (ggg) g Qg + ... - o + 0 Q5 + Truncating the series after the second term and substitut- ing into Equation (5.“) yields ( ) d = Z 2 pg gm+gn e V E + g pég J (5.5) 112 The fourth term in this expression is small relative to the others and may be neglected. Assuming a vibronic coupling of states, oée and oée can be evaluated via perturbation theory. This is referred to as the Herzberg- Teller expansion of the state |e> = Z 2 o ' E. (5.6) SE 6 Sfe Ee - ES 020) 1 deg The extent to which the vibronic coupling operator gg’: 5 serves to mix the other molecular eigenstates, |s>, with |e> depends upon the relative magnitudes of the coupling matrix element, ,and the energy separation between the states. The expression for the polarizability tensor now becomes 0 < m> ) - z 2 pg? 83 av VI V a -— po gm+gn c e hvgm,ev'hvo+irev H ( hE es hv (hvgm,ev e s¢e v e,s 'hvo+irev) x [ ogepsg + pgeosgleElvxvlm)J (5'7) E _ 3H where hes — /Ee - ES 113 The first term in Equation (5.7) involves no vibronic coupling between electronic manifolds and is known as the Albrecht A-term. Since the vibrational integrals in A-term scatter- ing are simple overlap integrals it is sometimes referred to as Franck-Condon (F-C) scattering. Far from resonance, closure may be applied to the F-C vibrational integrals: - hv + if a gm,Br 0|}-J E -l 3 hBQE(hVB,Q)(hvgm,Br-hvo+irBr)J X E x [ongQg prgQg] (5.8) with an analogous expression for visible (Q) resonance. Since pr, 083 >> OQg’ pQg, F-C scattering dominates the R.R. spectrum obtained with Soret excitation and only Alg modes are enhanced. On the other hand, excitation in resonance with the heme Q transition leads to scattering _ E _ via the H T terms since ng OgQ << ng GBg hBQ for vibra tional modes that efficiently couple the B and Q states. In order for th to be non-zero, the symmetry of the vibra- tional normal coordinate must be contained in the direct product of the coupled electronic states. Both B and Q transitions possess Eu symmetry and Eu x Eu = Alg + Blg + B2g + A2g. Thus, modes of these symmetries are the only 115 ones enhanced by H-T scattering. Modes of A symmetry 18 have been shown to be ineffective in vibronic mixing for the cyclic polyene model (91). Heme vibrational modes of the allowed symmetries correspond to in-plane deformations of the porphyrin macrocycle. The relationship between polarizations of the exciting and scattered radiation is expressed as the depolarization ratio for the band in question. It is defined as ”2 = Ii/III intensity of scattered radiation with polarization H '__ ll 1 to that of the incident radiation. H II intensity of scattered radiation with polarization [I to that of the incident radiation. This ratio is a function of the symmetry of the polariz- ability tensor of the scattering state. Symmetry patterns for vibrations in the Duh symmetry group have been de- termined (92) and are given in Figure (36). The depolarization ratio can be redefined in terms of invariants of the preceding tensors p, = 3g8 + SEA/logo + ugs for 90° scattering geometry. The tensor invariants are: 116 IO ono Aug: 0| 0 Azg=-I 00 000 000 moo ouo 8'9- o-I o 329:.- I 00 000 000 Figure 36. Tensor symmetries for the resonance Raman active vibrational groups of hemes g. 117 Tr(SOSO+) the isotropic invariant gO Tr(SSSS+) the symmetric invariant g8 the antisymmetric invariant ga = -Tr(AA+) where l 823 = §'°ij(axx + axy + O‘zz) s l 813 = 2° 3/M) modes of Azg symmetry in Duh hemes. Spectra of flavocytochrome 3552 have been obtained in both of the above- mentioned scattering regimes with “41.6 nm and 51A.5 nm excitation. Fluorescence is a much stronger process than resonance Raman scattering. Thus, flavin fluorescence, despite the fact that it is extensively quenched, posed a significant obstacle to these resonance Raman studies and manifests itself as rising baseline in the spectra pre- sented here. 119 The high and low frequency regions of resonance Raman spectra obtained with 441.6 nm excitation are shown in Figure (37). Scattering from the polarized heme modes of the ferric and ferrous forms of the protein are typical of other getype cytochromes investigated at this wavelength (94). The low signal-to-noise ratio displayed by most the F-C active modes in c is to be expected since 441.6 nm -552 excitation lies in the pre-resonance enhancement region of the protein's absorption spectrum. Resonance enhance- ment via simple F-C scattering is proportional to the ratio of the electronic matrix elements and the resonance denomin- ator of Equation (5.7) (95) although this is recognized as only a first approximation (96). 2 2 _ 2 R - e /(hvgm,ev - hvo) + T (5.9) where 82 = extinction coefficient of the resonant absorp- tion band. thm,ev’ hvo are the energies of absorption peak and exciting light, respectively, and P is the natural absorption linewidth. For h = 24390 9552 cm“1 and F = 1200 cm-1. Thus, the enhancement at 441.6 ng,ev nm is approximately 1/3 that at 410.0 nm. This situation is compounded by the fact that for hvo 3 24390 cm"1 the vibrational overlap integrals arising from F-C scattering should exhibit a destructive interference between scattering 120 Figure 37. Resonance Raman spectra of flavocytochrome 9552 obtained with 441.6 nm excitation. The power was 10 mw and the £552 concentration was 75 uM in 0.1 M Tris, pH 7.5. INTENSITY ABITRARY 121 i l l l l l l I 000 I IOO l200 l300 1400 ISOO l600 .. 4:2 4?5 748 see ' - 690 ' 75: 4:3 1 l l l l l l 300 400 500 600 y 700 800 900 Av (cm") Figure 37 122 from the vibronic components of the Soret band of £552 (97) further diminishing their intensity. The spectrum of reduced 3552 displays a selective enhancement of the 690 cm'1 and 1360 cm"1 modes over the other Alg vibra- tions. This behavior has been observed in other heme proteins (98). Vibrational modes active in F-C scattering gain intensity from the simple overlap integrals of Equation. (5.8), i;g;, , and, therefore, their intensity is dictated by the distortion in their equilibrium posi- tions between excited and ground electronic states. The selective enhancement of 690 cm-1 and 1360 cm"1 modes indicates that the excited state of the heme Soret band experiences a porphyrin macrocycle expansion in the direc- tion of those modes' normal coordinates. These coordinates, particularly for the 1360 cm"1 band, have been shown to originate primarily from in-phase CaN symmetric stretch- ing in metallooctaethyl porphyrins (99). For oxidized 1 3552 the relative enhancement of both the 696 cm” and 1360 cm.1 modes is diminished, presumably reflecting a smaller distortion of CaN stretch in the Soret excited state due to the increased central ion charge. Even at the extremely low laser power used (~10 mW) a small amount of heme photoreduction is evident as a low energy shoulder on the 1370 cm'1 band of the oxidized protein. This band has been used as an indication of heme redox state (100) and photoreduction (lOlgyS) although some ambiguity exists 123 Table 4. Raman Modes for Flavocytochrome 3552 Obtained With 441.6 nm Excitation. Ferric Ferrous 412 (w) . 413 (w) 496 (m) 696 (w) 690 (s) 748 (m) 751 (m) 1186 (w) 1229 (w) 1370 (S) 1361 (s) 1504 (w) 1492 (w) 1586 (m) ' 1593 (m) 1637 (m) -1 A11 frequencies in units of cm w - weak m = medium 5 - strong. 124 in interpretations based on its position since the mode is also sensitive to the basicity of heme axial ligands (81). No evidence of resonance enhancement of flavin vibrational modes is observed. Resonance Raman spectra of reduced 3552 and its diheme subunit, obtained with 514.5 nm excitation and shown in Figure (38), display the variety of mode symmetries ex- pected for heme visible band resonance scattering. The resonance Raman spectrum of reduced horse heart cyto- chrome g is included for ease of comparison. Vibrational symmetries have been assigned (See Table 5) on the basis of the depolarization ratios obtained from the reduced protein under the assumption of D4 heme symmetry. Both h the holoprotein and heme peptide spectra are quite similar to those of horse heart cytochrome g_and other small molecu- lar weight monoheme 3 proteins (100,102) the only substantial difference being the flavin fluorescence background. The anomalously polarized (ap) band at 1586 cm"1 and the de- l, which are sensitive to polarized (dp) band at 1621 cm- heme spin state (83), appear at frequencies consistent with low spin heme g. This confirms the assignment made from previous magnetic studies (33). The position of the 1 polarized oxidation state marker at 1363 cm' offers no evidence of anomalous heme axial ligation such as that seen with the P-450 cytochromes (103). Figure (39) displays spectra of the ferric forms of 3552, Figure 38. 125 Resonance Raman spectra obtained with 514.5 nm excitation of (a) 70 uM ferrous flavocytochrome 9552 diheme peptide in 0.1 M Tris pH 7.5 with 350 mw of laser power; (b) 100 uM ferrous flavo- cytochrome 9552 in 0.1 M CAPS, pH 10.0 with 180 m8 of laser power; (c) 80 uM ferrous flavo- cytochrome 2552 in 0.1 M Tris, pH 7.5 with 250 mw of laser power; (d) 100 uM ferrous flavocyto- chrome 3552 in 0.1 M MES, pH 6.05 with 250 mw of laser power; (e) 200 uM ferrous horse heart cytochrome g in 0.1 M Tris, pH 7.5 with laser power equal to 250 mw. Frequency positions of the principal bands are given in Table 5. The fluorescence background of the diheme peptide spectrum arises from a small amount of residual flavin peptide which could not be separated from the sample. Arbitrary Intensity 126 a) home peptide pH 7.5 b) £55sz no.0 c)g,552,pH7.5 d) 9552 ,pH 6.05 e) cyt. 9 ,pH 7.5 J I l l l l l l l l l 1600 I400 ”1200 A7 (cm") Figure 38 127 as. amaa as. amaa as. mmaa an. amaa as. mmaa .oa.av ame Asa mmaa as. amaa as. mmaa Ana mmaa as. mmaa .om.v amm as. emma as. mmma as. omma .a. amma .sv mmma .oa.. mam an>a aama Ame. mama Ame. aama aaev oama Ame. mama .mm.aa mma .av mama .aa mama Ame. mama as. mama am. mama .mm.v maa as. moaa Ana mama .a. moaa as. aoaa an. aoaa .ma.v mam lea mama .aa mama ... aama .ea aama .aa mama .am.a mam .av aama .aea mama Ame. aama .aa mama .aev aama .aa.ma ama as. amaa as. amaa as. amaa. as. mmaa was. amaa amm.v mam .o.oa ma. .mo.a maa .m.m ma. .m.m ma. am.» ma. manossmm mmmm .amo mmmm .amo mmmm amo oeaaamm tam: m .aao oeoz msopgom npaz UoCaMpno momooam mmmm oEogsoOono>mHm pom mono: mmmwmmwwmwwuwmmmmmwm .m manme .xwoz mao> u 3> .xmoz u 3 «EaapoE n E awsonum u m .wCOppm mpo> u m> . so no mums: ca mam moaocosaoam amcoapmana>a 128 a- Ana mmaa nun: am. amaa axe. mmaae am. amaa mma Asa maaa Asa maaa asv maaa as. maaa aaa oaaa mmm am. mama as. aama aea aama .ea aama aev amma..ea amma mam awe. mama .aa aama aea mama Asa mama .aa mama mma as. mama as. eaaa .ea mama as. mama as. mama mam .50 maaa as. aoaa Asa aoaa as. moaa Asa maaa mam am. aama Ame mama ama aama aea aama am. mama mam amea mama .ma aama am. aama ama aama am. mama ama amea maaa awe. maaa awe. aaaa amea maaa amea mmaa mam .o.oa mo. amo.a mav am.a ma. .m.a may .m.a maa maaosEam mama mmmm .aao mmmm .mao ooaaaoa oEom m .0m0 oooz . mmmm.pmo MMMMMM .ooseaoeoo .m oaoma Figure 39. 129 Resonance Raman spectra obtained with 514.5 nm excitation of (a) 70 uM ferric 3552 diheme peptide in 0.1 M Tris, pH 7.5 with 250 mw of laser power; (b) 100 uM ferric flavocytochrome £552 in 0.1 M MES, pH 6.05 2 mM Na28203 with 200 mw of laser power; (c) 80 uM ferric flavocytochrome 3552 in 0.1 M Tris, pH 7.5 with 95 mw of laser power; (d) 200 uM ferric horse heart cytochrome c in 0.1 M Tris, pH 7.5 with 250 mw of laser power. Frequency positions of the principal bands are given in Table 5. Intensity Arbitrary 130 a) home peptide p+|25 8’ Si4:5:52' $205 pFISIXS |600 I400 I200 IOOO Av (Cm-4) Figure 39 131 its diheme peptide and horse heart cytochrome c, in the high frequency region, obtained with 514.5 nm excitation. As with the ferrous spectra, the positions and intensities of the Raman bands of the holo- and apo-protein are generally typical of monohemes 3. Two major departures from this "typical" behavior are evident. The first is a small but consistent deviation in the positions of the high fre- quency modes with B symmetry, as may be noted from 1g Table (5 ). These bands are characterized by depolariza- tion ratios equal to 3/4 and occur at 1635, 1562, 1412, and 1250 cm“1 in horse heart cytochrome c. In ferric 3552 these modes, and only these modes, show wavenumber shifts relative to horse heart cytochrome g. The shifts range from +9 cm.1 (for the 1571 cm"1 band) to -5 cm'1 (for the 1407 cm-l'band). The observation that the positions of only a specific symmetry class of resonance Raman active heme vibrations are anomalous is suggestive of some specific perturbation of the heme environment in 3552. Moreover, this perturbation of the B1g modes is not ob- served in the ferrous form of the protein (See Table 5). The second case of anomalous behavior lies in the rela- tive ease with which the heme in the holoprotein is photo- reduced in the laser beam. The small amount of photo- reduction observed with Soret excitation increases dramatically with the higher laser powers used for visible excitation. Figure (40) shows a spectrum of 132 .maqum one son: mpamcoch e: m.:am mo coapocaa a ma mmmm oaaaoa 2a pawn poxams oumum coaumpaxo ocu no cocoocoaop muamcouca cam coauamoa one appomcH .oaaemm 0:0 com: 0:00a02m unwaa momma Es m.:am mo 38 mam Spa; pagampno m.a ma .maLB z a.o Ca mmmm oannoopmoo>oam oaaaom :1 cm mo Esmpooam :mEmm monocomom .o: maswmm 133 E {3 ‘1- ? i 2 .2 3 .a' E 8 Sf‘ :36 g: aEQ AJJSNBINI AHVHJJBBV ezu— J CLII— stal— men-- 99::— soon— sass.—~ 989l— 42> 099I——-¢: 7fi“fin AllSNElNI AHVHIIBHV I400 I300 I200 Av (cm") |500 I600 Figure 40 134 resting 9552 at 315 mw incident laser power, plus an inset displaying spectra in the oxidation state marker band region as a function of incident laser power. It is apparent that at high laser intensity this band shifts from 1370 cm-1 to 1356 cm-1, a position indicative of ferrous iron (100). At intermediate powers a double peak is clearly visible. Neither horse heart cytochrome g_nor the heme peptide displays this behavior; indeed, spectra of the ferric forms of these proteins were routinely obtained with 200 mw of laser power. The binding of thiosulfate and cyanide to the oxidized protein resulted in an initial lowering of flavin fluores- cence. However, CN' binding was found to be short-lived. Subsequent to cyanide binding, the ratio A475/A525 returned to its initial value and flavin fluorescence increased greatly. Thiosulfate, on the other hand, remained bound and continued to quench flavin fluorescence. Relative fluorescence levels of these species are contrasted in Figure (21). The instability of the 95 precluded observation of its resonance Raman spectrum. 52-CN' complex The spectrum obtained with 514.5 nm excitation of 3552'3203 is included in Figure 39. It differs from the unbound holoprotein only in the substantial increase in l l anomal- intensity exhibited by the 1315 cm- ’and 1589 cm- ously polarized modes. The pH dependence of the holoprotein Raman spectra 135 was also investigated. Fluorescence levels increased markedly at both the high pH (N10.0) and low pH (m6.0) limits of 9552 stability. This resulted in a deteriora- tion of Raman spectral quality and a spectrum of the ferric protein at high pH could not be obtained. As with other cytochromes, 9552 band positions display some pH dependence. In particular the band which appears at 1544 cm"1 and 1571 cm-1 in the ferrous and ferric holoprotein, respectively (at pH 7.5) was sensitive to pH changes. Figure 38 exhibits the pH dependence of scattering from ferrous £552 and attests to the dramatic rise in background fluores- cence as the high pH limit of hemezflavin subunit binding stability is reached. A summary of band positions, inten- sities and depolarization ratios obtained for the various forms of flavocytochrome 9552 with 514.5 nm excitation is given in Table 5. The general features of the resonance Raman spectra of flavocytochrome 3552 presented here conform well to the classification methodology developed for monoheme proteins by Spiro and Strekas (100). Upon closer examination, however, several distinguishing aspects of the 3552 spectra become apparent. These can be divided into two categories: general effects involving the flavin moiety of the protein and specific perturbations of heme vibra- tional modes. The former have the more obvious impact on the spectra, whereas the interpretation of the latter 136 provides insight into the multicomponent nature of flavo— cytochrome 9552. Flavin Effects The most salient of the flavin effects is the broad fluorescence background it produces in the resonance Raman spectra of 2552. This is unusual not because of its existence, but rather because it is weak enough to permit the observation of resonance Raman scattering. The factors leading to the quenching of flavin fluorescence have been discussed in detail in Chapter 3 and need not be reiterated here. However, the presence of the proteinksflavin moiety has other effects upon the heme resonance Raman spectra of 3552. An indication of heme/flavin communication is apparent in the heme photoreduction at high incident laser power. The shift in the "oxidation state marker" frequency from 1370 cm'1 to 1360 cm'1 has been used as an indication of increased electron density located on the heme iron (104). Such a shift occurs in flavocytochrome 3552 spectra at moderate laser powers, but it is absent in the heme peptide and horse heart cytochrome 9 spectra even at high power. Photoreduction has been observed in other heme proteins, most notably cytochrome oxidase, and was attributed to a flavin contamination which was postulated to be the initial site of photoreduction (101). In the case of 3552 137 the flavin is an integral part of the functional protein. The flavin excited state is extensively delocalized by interactions with protein aromatic residues (as indicated by fluorescence quenching) and thus could be expected to provide an efficient pathway for heme photoreduction. Under aerobic conditions the photoeffects described here are most evident for the 1372 cm'1 (dp) band. At the highest laser power used, two of the B heme modes (at 1 1571 and 1407 cm-1) begin to decrease 1: frequency; however, the systematic lowering of heme vibrational frequencies known to result from chemical reduction is largely absent despite the fact that the oxidation state marker appears at 1356 cm.1 in the "photoreduced" protein (7 cm"1 lower in wavenumber than in chemically reduced 3552). In par- ticular, the 1640 cm'1 (dp) mode which undergoes large changes in both position and intensity depending on the iron redox state clearly retains its oxidized character. This indicates that the lability of the "oxidation state marker" is predicated, at least in part, upon factors independent of the formal iron redox state. Anomalous oxidation state marker behavior has also been observed in carbon monoxy and oxyhemoglobin and myoglobin resonance Raman spectra (81). The resonance Raman spectra obtained with Soret excita- tion are noteworthy for the absence of any bands attributable to flavin scattering despite the fact that the 441.6 nm 138 exciting radiation is in resonance with the flavin absorp- tion at 450 nm in 3552. This absorption arises from an in-plane flavin n+n* transition (105). Thus, resonance enhancement of in-plane isoalloxazine vibrational modes would be expected to result in several peaks in the 1000- 1 region of the resonance Raman spectrum. These 1600 cm" modes have recently been observed in resonance Raman spectra of protein-bound FAD (106)and in CARS spectra of FAD and glucose oxidase (107) The absence of flavin bands in our spectra can be readily explained by the fact that at 441.6 nm heme scattering dominates the spectrum because of its greater extinction, leaving the flavin modes unobservable at the laser power used. Resonance enhancement via simple Franck-Condon scattering is proportional to the quantity R given in Equation (5.9). For the following conditions: v = 22,645 cm"1 (441.5 nm), vo(flavin) = 22,422 cm-1 (446 nm), vo(heme) = 24,390 cm"1 (410 nm), r (flavin) l 1 -1 , P (heme) N 1200 cm-1, em(heme) = 125 mM- cm and em(flavin) = 10 mM-1 cm'l, the ratio R(heme)/R(flavin) m 1500 om’ is calculated to be more than one hundred. Thus any flavin bands would be at least an order of magnitude less intense than the 9552 heme bands, and undetectable with a conven- tional spectrometer. 139 Heme Vibrational Bands All the features found in the resonance Raman spectra of £552 can be readily interpreted as monoheme scattering coupled with the flavin effects discussed above. The ap- plicability of previous classification schemes (100 to the heme scattering of 3552 is obvious and confirms their assignment as low-spin heme g. This is particularly striking in the spectra of reduced £552 and its heme peptide under visible excitation. For both the holo- and apo- 1 protein all bands are within :2 cm- of their horse heart cytochrome g values, except the depolarized band at ~1545 cm'1 which displays a pH—dependent position. Investigations of Desulfovibrio vulgaris cytochrome 33 by Kitagawa 23 al. 002) indicate that this band is sensitive to the nature of heme axial ligands. In fact, they used its position to monitor a pH-dependent heme ligand change: a shift in l to 1536 cm-1 frequency from 1541 cm- was interpreted as resulting from the replacement of histidine by lysine in the protein at high pH. The existence of an analogous pH dependence in the heme axial ligation of ferricg552 has been indicated in a previous EPR study of the protein by Strekas (33). The situation is somewhat more complicated in 3552 than in cyto- chrome 33 because only one of the protein's two hemes dis- plays pH-dependent behavior. The EPR spectra of ferric 140 3552 obtained by Strekas (33) and reproduced in our labora- tory can be interpreted to arise from a heme axial ligation scheme involving one pH—stable heme with methionine/histi- dine ligands and one pH-labile heme favoring lysine/histi- 'dine ligands at low pH and histidine ligands at high pH. The pH-dependent behavior of the 1545 cm'1 Raman band in ferrous indicates that a similar axial liga- 9552 tion scheme obtains for the reduced protein. If this is the case, the resonance Raman spectra of ferrous 9552 at neutral pH should, in principle, exhibit three peaks for the ligand-sensitive band at 1545 cm'l: one at 1547 cm'l, another at 1541 cm‘1.and a third at 1536 cm’l, correspond- ing to methionine/histidine (as in ferrous horse heart cytochrome c), histidine/histidine, and lysine/histidine ligation, respectively. In practice, a single asymmetric peak appears in this region: at 1542 cm'"1 in the holo- protein at pH 6.05, 1544 cm-1 at pH 7.50 and 1548 cm"1 at pH 10.0, suggesting that the pH—dependent ligand shift is also operative for the reduced protein. For the ferrous heme peptide at pH 7.5 the band is observed at 1541 cm’l, indicating a preference for the low pH ligand (lysine) in the apoprotein. The spectra of ferric 9552 holo- and apo-proteins, while retaining the general characteristics of monoheme 3 spectra, display a larger deviation from horse heart cytochrome g behavior than do the ferrous spectra. Two 141 effects are most likely responsible for this behavior: ligand effects seen in both 9552 and the heme peptide, and a general perturbation of heme modes of Blg symmetry observed only in the oxidized holoprotein. The ligand effects which were limited to a single mode in reduced 9552 are more widespread in the ferric spectra.‘ The ligand effects seen here can be interpreted as reflecting a change in the heme iron electronegativity as. a function of axial ligand electron donating capabilities. A recent study by Kitagawa, gt al. (108) utilizing metallo- porphyrins with a variety of central metal ions indicates that the position of several high frequency Raman bands and the Q00 optical absorption maximum are directly cor- related to the electronegativity of the central metal ion. Increased metal electronegativity allows for better con- Jugation of the metal rpZ orbital with the porphyrin a2u orbital, shifting the Q00 transition to higher energy and giving rise to a stronger n-bonding system in the porphyrin macrocycle. Stronger n bonding results in higher frequencies for the affected vibrations. These results can be extrapolated to heme proteins with the realization that while the central metal ion does not change, its apparent electronegativity is a direct result of the electron donor power of the axial ligands. Thus, the frequencies of the ligand sensitive Raman bands should 142 increase in the order of methionine to histidine to lysine based the electron donating capabilities of those ligands. This expectation has been shown to be true for several high frequency heme 3 Raman bands. Kitagawa 31; al. (102) have 1 found that the bands at 1635, 1562, and 1372 cm' are pH (ligand) dependent in ferric horse heart cytochrome g, changing to 1641, 1568 and 1375 cm-1 upon replacement of the axial methionine by lysine at high pH. The Raman data presented here confirm that this situation also applies to flavocytochrome 3552. The heme peptide frequencies parallel the high pH values of horse heart cytochrome g (iifia’ lysine/histidine ligands), indicating the effect of the pH labile heme. For the ferric holoprotein, however, the positions of all of the high frequency Blg modes are shifted (relative to horse heart cytochrome g). Particularly evident is the 9 cm-1 change in the position of the 1562 cm-1 band. The band positions of the ferric heme peptide B1 modes are, with the exception of the 1642 cm”1 band, 8 intermediate between the values for horse heart cytochrome g and halo-cytochrome 9552. On the basis of axial ligation it would be expected that the situation would be reversed; the holoprotein with its mixture of lysine and histidine ligands would have band positions closer to horse heart cytochrome 9 than to the heme peptide. The expected situation holds for the low pH holoprotein, but is not the case at pH 7.5. Thus, relative wavenumber shifts in the 143 spectra cannot be explained as arising solely from axial ligand changes; the changes in frequency of the Blg modes must alSo be diagnostic of some other protein influence. Normal coordinate calculations (99) have indicated 1 that B modes in general and the ~1565 cm- mode in par- 18 ticular involve out-of-phase stretching of atoms at the porphyrin periphery (either CB - C8 or Cm-H stretches), whereas the 1372 cm'1 mode is closely associated with C-N symmetric stretching. In fact, studies of metalloetio- porphyrins by Spaulding gt_§l. (83) have stressed the im- portance of contributions from the core expansion of the inner 16-membered ring to both the oxidation state (A13) and spin state (A2g) marker bands. The depolarized modes 1 and 1250 cm-1 in ferric hemes appearing at ~1565 cm- have been shown to be particularly sensitive to substi- tuent effects. The former shifts from 1547 cm"1 in ferrous cytochrome g to 1538 cm-1 in protoheme reconstituted ferrous cytochrome 25, indicating the presence of the two peripheral vinyl groups in the protoheme. A similar effect was noted in ferric heme g by Kitagawa 31; al. (109), who observed 1 that the 1564 cm- mode in a bis-imidazole iron-proto- porphyrin complex shifted to 1555 cm"1 in the heme §_bis- imidazole complex, reflecting the contribution of the heme 1 mode has been g peripheral carbonyl group. The 1250 cm- shown to be sensitive to deuteration of the methine hydrogens in ferrous mesoporphyrin IX dimethyl ester 144 complexes 010). These observations indicate that the depolarized modes of the heme are more sensitive to peri- pheral influences on the porphyrin than are the high fre- quency A2g or Alg modes. In flavocytochrome 3552 these Blg modes all experience frequency shifts (with respect to horse heart cytochrome g) in the oxidized protein that are absent in the reduced protein. The effect is ob- scured by the axial ligand dependence of the 1640 cm"1 and 1571 cm‘1 bands, but is clearly independent of it since the 1407 cm"1 band shows no axial ligand effect in reduced or oxidized 9552 or in any of the ferric bacterial cytochromes g_studied~by Kitagawa gt g}, (102. The binding of S to the low pH form of the oxidized 203 protein produces no appreciable change in the heme Raman frequencies, implying that there is very little perturba- tion of the local heme vibrational environment upon sub- strate binding. Thus, in its low pH form, the heme en- vironment is already in a conformation amenable to sub- strate binding. However, the intensity of Herzberg-Teller active heme modes is dependent upon the extent to which they couple the porphyrin Q and B states. The large increase in the relative intensities of the two anomalously polarized 1 bands (at ml315 cm- and m1589 cm'l) in going from c -552 at pH 7.5 to the thiosulfate-bound protein may be indica- tive of an alteration in the electronic environment of the heme. 145 It is apparent that the redox state of the protein (and by implication the flavin) has a noticeable effect on the peripheral environment of at least one of the heme moieties. The effect is small. No extra bands occur in the heme spectra, nor are there any dramatic changes in electron density at the heme iron (as evidenced by the position of the 1370 cm”1 band). This suggests that there is no direct heme/flavin interaction through either electronic resonance or the axial heme positions, and if heme/flavin interac- tion occurs it does so via a protein-mediated heme/flavin communication through the heme periphery. The heme vibra- tional mode frequencies are consistent with an interpre- tation which considers both heme axial ligation and an indirect protein mediated heme/flavin interaction. CHAPTER 6 CONCLUSION The results obtained from the application of a variety of spectroscopic techniques to flavocytochrome 3552 have been described in the preceding chapters. Individually these techniques provide insight into a number of specific properties of the protein, but their true utility lies in a synthesis of these specific results which leads to the elucidation of general relationships between the protein's structural and functional aspects. These relationships can be divided into three general categories: 1) chromophore : chromophore interactions; 2) chromophore : protein interactions; 3) chromophore interactions with exogenous ligands This chapter serves to summarize the nature and extent of these interactions in an effort to arrive at a consistent description of flavocytochrome £552 and the general class of multicomponent enzymes it represents. 146 147 A. Heme/Heme Interactions Communication between redox centers in a multicomponent enzyme such as 3552 is a necessary prerequisite for the proper function of the protein in electron transport. Perhaps the most straight forward mechanism for such com- munication lies in the direct coupling of the action centers by electronic, magnetic or chemical means. No evidence for direct heme/heme interaction in flavocytochrome 3552 can be found with any of the spectroscopic techniques used in this study. The EPR and absorption spectra of the protein are completely consistent with isolated low spin heme g. MCD and EPR spectroscopies would be particularly sensitive to heme/heme magnetic interactions within 3552, yet both yield spectra of 3552 that are typical of isolated hemes. No magnetic coupling such as spin-spin or dipole-dipole interactions are evident using either of these techniques. These findings preclude the possibility of heme:heme stack- ing or even the sharing of a common axial ligand (Saga: histidine) between the two hemes, as either of these would lead to a significant distortion of heme magnetic properties such as that found in bacterial gjcytochromes 011) or mito- chondrial cytochrome g_oxidase (112. Heme/heme interaction in 3552 was originally postulated as an explanation for the protein's derivative shaped Soret CD spectrum. Exciton coupling does lead to derivative shaped bands, however, it is not a necessary condition 148 for such CD behavior. Derivative Soret CD spectra have been observed in monoheme proteins such as cytochrome g from Candida lerusei.CLr3)and horse heart cytochrome g, where no possibility of heme/heme interactions exist. The global properties of the protein molecule must be considered in an interpretation of heme g_CD spectra (11”) Myer (115) interpreted the negative peak of horse heart ferricytochrome g as resulting from heme-protein inter- actions which reflected the conformational integrity of the heme crevice. Considering the lack of corraborative evidence for heme/heme magnetic interaction, the CD characteristics of the Soret region of 9552 must be as- cribed to a similar heme-protein interaction. The dis- ruption of this interaction by CN' is not surprising con- sidering the extensive effect CN- has on the environment of the pH-labile heme. B. Heme/Flavin Interactions The interaction between the dissimilar redox centers in 3552, while more in evidence than heme/heme interactions, is indirect in nature. The most obvious manifestation of heme/flavin communication is found in the ease with which the heme moieties of ferric 9552 are photoreduced during the course of resonance Raman spectroscopy. This photo- reduction is dependent upon the flavin group (it is absent 149 in the diheme peptide of 2) and is indicative of the 955 existence of a pathway for electron transport between flavin and heme. Heme vibrational modes of Blg symmetry display a dependence upon the oxidation state of the pro- tein (and by implication the flavin). This can be inter- preted as resulting from a perturbation of the peripheral environment of at least one of the hemes. The resonance Raman data cited above establish that some form of indirect contact between flavin and heme in 9552 exists,however there is no datum available at this time that requires direct heme/flavin coupling for its explanation. Flavin-fluorescence is significantly quenched in 9552, but the extensive quenching exhibited by heme- free, flavin-containing, peptide fragnmnts of the protein leaves any possible further effects due to flavin-heme Fdrster coupling undetectable. The alteration of the protein's EPR spectrum upon CN' binding is most easily explained as resulting from a direct CN‘ attack upon the pH-labile heme rather than a flavin/heme interaction. The absence of evidence for direct coupling between chromophores requires that the mechanism of electron flow in 9552 actively involve the polypeptide matrix. Thus, the protein environment of the redox centers in £552 be- comes critical to their proper function. A partial defini- tion of chromophore environments within 3552 is possible from the spectroscopic data and is summarized below. 150 C. Flavin Environment The flavin in 3552 displays no magnetic or Raman spectra and, therefore, is less "visible" than the protein's heme groups. Nonetheless, it is possible to infer a number of characteristics of its protein environment from absorption and fluorescence data. The ability of the flavin to form adducts with exogenous ligands dictates that the portion of the group containing the N-5 position is accessible from the external solution. However, the nearly complete quenching of flavin fluorescence and the CD spectra of the flavin—containing peptides argues that it is closely associated with aromatic (tyrosine) residues in the protein matrix. This is corroborated by the increased reduction potential (E0 = 0 mV) of the flavin in 3552 relative to free flavins (E0 = -200 mV), which indicates a highly hydrophobic flavin environment. The hydrophobicity of the flavin environment could also inhibit the formation of the charged semiquinone species and favor a concerted, two electron transfer at the flavin. A flavin configuration that exposes only the central edge of the isoalloxazine ring system to the solution while maintaining the rest of the flavin in a crevice of hydrophobic residues is consistent with the above observations. 151 D. Heme Environments The two hemes in g 2 exist in distinctly different 55 protein environments. Optical and MCD data confirm that both hemes are low-spin six-coordinate heme 9. However, the EPR spectra of 9552 clearly indicate that the axial ligands of two hemes are different. One heme has a pH- invariant axial ligation of methionine and histidine whereas the other possesses either histidine/histidine or lysine/histidine ligands, favoring the former at high pH and the latter at low pH. Resonance Raman spectra of the protein are consistent with this picture and further sug- gest that at least one of the hemes (presumably the pH- labile one) experiences a perturbation of its peripheral environment due to the redox state of the flavin moiety. The reduction potentials of the two hemes are approximately equal, a result unexpected in light of their differing axial ligands. This strongly implies that the degree of hydro- phobicity of the two heme environments is markedly dis- similar or that a high degree of positive cooperativity exists between them. The spectra of low pH (6.05) 9552 and 3552-820; are nearly identical, suggesting that the protein configuration inducing low pH (histidine/lysine) form of the pH-labile heme is more amenable to substrate binding by the flavin. 152 E. Exogenous Ligand Binding The behavior of the hemes in 9552 toward exogenous ligands serves to emphasize the differences in their pro- tein environments. The pH-labile heme is quite accessible to CO and CN‘. EPR data unequivocally establishes it as the CO binding site in the protein at physiological pH and also indicate that its axial ligation is greatly dis- rupted by CN'. Binding of both CO and CN' can be expected to proceed via a ligand displacement reaction. This indicates that the native protein heme ligands are rela- tively loosely held and their displacement represents only a small kinetic barrier to the binding reaction. The loosely— held pH—labile heme is also the probable source of the small high spin heme signal observed in both absorption and EPR spectra of 3552. The high spin signal can be postulated to arise from some protein conformation where neither lysine nor histidine occupies the 6th ligand position of the pH- labile heme. The observation that N} does not quench this high spin signal whereas CN- does indicates that the heme is not exposed to the solvent, but occupies some internal, largely hydrophobic pocket in the protein that can discrim- inate against the azide molecule. The pH-stable heme, on the other hand, displays much less of a propensity to bind exogenous ligands. It binds CO only when the pH of the protein environment is high enough to induce significant 153 changes in the tertiary structure of 2552. CN- binding to the pH-stable heme, as evidenced by the disappearance of the 695 nm absorption band in 3552, proceeds very slowly suggesting that its native axial ligands are tightly-held and/or are not accessible to the solution. Thus, the pH- stable heme apparently exists in a tightly-bound, largely .hydrophilic environment within the protein that allows at most limited access to its axial positions. Thiosulfate, sulfite, and cyanide ions apparently bind to the flavin moiety in 9552 by a simple adduct formation reaction which bleaches the visible absorption of the flavin. The inter- action of CN- with the heme groups in 9552 subsequently leads to a partial restoration of flavin absorption possibly by inducing protein conformational changes that reduce the flavin affinity for cyanide. F. Protein Mediated Communication Between Redox Centers The description of the mechanism of electron transport in multicenter enzymes is indeed a formidable problem with no unique solution. Perhaps the most simplistic concept of multicenter enzymes is to consider the polypeptide portion of the protein as a passive matrix which holds the redox centers in the correct Juxtaposition for efficient coupling between them. This concept is quite useful from a synthetic standpoint, requiring only that the correct positioning of redox centers be mimicked in order to 154 reproduce the functional aspects of the protein. However, this simplified notion finds little applicability to flavo- cytochrome 9552. This multicenter enzyme is obviously dependent upon its polypeptide component in order to func- tion. The protein environment affects the redox potential, binding characteristics and spectral properties of redox centers in the protein. Moreover, there is no apparent coupling of the centers themselves, necessitating a protein- mediated communication between them. Clearly, 3552 is an example of a multicenter enzyme where non-coupled redox centers exist in an active polypeptide matrix. Polypeptide activity could take the form of direct participation of amino acid residues in electron transfer between redox centers or of introduction of conformation changes during the course of oxidation and reduction that bring the previously uncoupled redox centers into direct contact. The second possibility seems unlikely in light of the data obtained during reductive titrations of the protein that show no evidence of chromophore coupling in the MCD, EPR or absorption properties of 9552 at any point in the titration. The direct participation of amino acid residues in electron transport has been postulated as a vehicle for oxidation and reduction in getype cytochromes (116,1171 There is currently no evidence for any axial ligand exchange accompanying electron transfer in cyto— chromes c_(110)suggesting that such reactions proceed via 155 either an outer-sphere charge transfer or a tunneling mechanism. Charge transfer interactions require the ex- tensive electronic interaction of donor and acceptor centers resulting in a splitting of product and reactant potential surfaces 01]). Electron transfer then occurs adiabatically along the lower energy surface. Tunneling processes require only minimal interaction between redox centers (119,118). Thus, a significant barrier to electron transfer exists and transfer occurs via quantum mechanical tunneling. Since tunneling is an intrinsically weak process with an exponen- tial distance dependence, small changes in the orientation and separation of the redox centers can profoundly affect transfer rates. In the absence of direct electronic inter- actions between redox centers, either mechanism would be strongly dependent upon interactions between the poly- peptide chain and the electron transfer prosthetic groups. Electronic and conformational interactions could induce a significant lowering of the tunneling barrier height and a coupling of chromophore and amino acid residue electronic states would be necessary for charge-transfer processes to occur. Both charge-transfer and tunneling have been sug- gested as mechanisms for electron transfer in monoheme g proteins. Tunneling has been implicated in the oxidation kinetics of low potential cytochromes in Rhodopseudomonas (120) and Chromatium (121) and several theoretical treat- ments of the data from these systems exist (122,123). 156 One mechanism for mammalian cytochrome 3 reduction en- visions a series of electron transfer hops from the reduc- tase via a charge—transfer channel of aromatic amino acid residues to the porphyrin edge of the heme center in the.~ protein C117) . The application of such an electron transfer schemes to flavocytochrome £552 is intriguing. Without crystallo- graphic data, the relative positions of the redox centers and amino acid residues in 9552 is uncertain. This makes definitive assignment of electron transfer pathways and mechanisms in the molecule impossible. Nonetheless, the spectroscopic data gathered from the protein provides many indications that such pathways may well exist. Hemes and flavins are well suited to function as electron transfer centers. Both have rather unstable electrons in their reduced states and extended molecular orbital fl-systems which are highly polarizable. Moreover, their reduction potentials in 3552 are nearly equivalent and electron transfer (via either charge-transfer or tunneling) is most rapid between species where donor and acceptor potentials are closely matched 0J7). Thus, the hydrophobic protein environment of the pH-labile heme in 2552 serves to modify its redox potential in a manner that would make it more amenable to electron transfer from the flavin. Aromatic residues (specifically two tyrosines) in 3552 also sig- nificantly modify the excited state of the flavin moiety, 157 delocalizing it and presumably providing a pathway for radiationless energy transfer. It has been demonstrated that flavins can form charge transfer complexes with tyrosine, tryptophan and phenols under inorganic conditions (124,125), and in view of their intimate contact with protein tyrosine residues, it is reasonable to assume that such an interaction occurs in 9552. The observation that photo-excitation of the flavin in 3552 results in the photoreduction of at least one of the protein's hemes indicates that the delocalized flavin excited state is communicated to the heme(s) in question. The pH-labile heme is the most likely candidate for inter— action with the flavin. That its hydrophobic environment contains the aromatic amino acid residues intimately associated with the flavin is a matter of speculation. However, the variability of its environment makes it the obvious choice as the source of the peripheral perturba- tion evident in the heme resonance Raman spectra. It is instructive to note that studies of the oxidation potentials of mesotetraphenyl-porphyrins by Giravdeau gt_g1. (125) have suggested the existence of two sites for electron transfer in metalloporphyrins - the pyrrolic nitrogens (which are sensitive to iron-axial ligand electron density) in oxidation and the peripheral fl-electron system in reduction. If these results extrapolate to £552, the alteration of the heme peripheral environment in the oxidized protein may be 158 indicative of an electron transfer pathway via peripheral aromatic residues that is altered upon reduction of the protein. The electron transfer pathway between hemes in 9552 is even more obscure but presumably would entail a protein mediated communication between the loosely held axial position of the pH-labile heme and the fl-system of the pH-stable heme. The picture that emerges for flavocytochrome 3552 is I one in which electron transfer is initiated by formation of a flavin-substrate adduct. The disparity in redox poten- tial between S= and the protein's flavin moiety results in reduction of the flavin. The excess electrons in the iso- alloxazine n-system are delocalized through tyrosine amino acid residues near the flavin and are transferred to the hemes which act as a two-electron acceptor at a slightly higher potential than the flavin. The transfer can be postulated to occur via the periphery of the pH-labile heme. Reduced substrate could again bind to the flavin, filling the system to its 4 electron capacity. The pH—stable heme presumably could then communicate with a higher potential cytochrome (possibly 3555) in either the organism's light-driven photosynthetic chain or its ATP-coupled dark reactions. APPENDIX APPENDIX If the three redox centers in 3552 are assumed to behave as independent redox couples then the following system of equations must be solved: E = E13, + 4925-9— log [FJOX/[Fjred (A1) E = so + '059 lo [H(l)] /[H(1)1 (A2) H(l) I g ox red E = E° + egig-io [H(2] /[H(2)l (A3) H(2) l g ox red where Efi(1); Efi(2) are the midpoint potentials of the individual hemes. Dividing Equations (2) and (3) by 2 and subtracting from Equation (1) yields: 0 = (E° _ -o) + .059 log [Fjox[H(l)]red[H(2)]red F H ‘2 [FJred[H(l)]ofo(2)]ox 0 _ O 0 159 160 Thus, AK 5 Bo _ Eo = .059 [FJOX[H(1)]Ped[H(2)]red H F 2 10% [F] [H(lnoxmzfiox (Au) red In order to evaluate AE from Equation (4) the concentra- tions of the oxidized and reduced species of each of the two hemes must be determined from the average value ob- tained from a reductive titration. The extent to which each heme contributes to the amount of heme reduction measured is governed by the relative midpoint potentials of the two hemes. Three different cases need to be con- sidered: (1) If Efi(1) = Efi(2), then . 2 _ .059 [Fjox [HJred “‘3 ‘ 2 log [P] {le red ox Where [ered’ [Hjox = average concentration of reduced and oxidized hemes. 0 O (2) If EH(1) > EH(2) then Equation (4) holds and [H(l)]red’ [H(2)]red can be determined from the relative midpoint potentials of heme (l) and heme (2). (3) 161 O 0 If EH(1) >> EH(2)’ then heme (l) is completely reduced before either heme (2) or the flavin begin ' to accept electrons and only a two center equilib- AB All three variation titration the data. rium need be considered. Thus, 1/2, = Eo _ E0 = -059 1o I:FJOX [H(2)]red H(2) F [F1888'[H(2)]ox of the above situations lead to a systematic of the "constant" B over the course of the (See Table 1) and thus are not consistent with 162 >.mm m.mm o.H: o.m: m.mm m.w: m.am m.mm m.a: 0.0m m.mm m.mm m.ma o.mm m.am o.mm a.mw a.om m.mm m.am o.m: o.ma z.a~ m.ma >.mm m.w:, a.wm o.m m.ww m.o o.Hm H.©: 0.0: 0.0 n.0m m.:m o.mm :.mm 0.0 m.m: >8 ooa + >5 om + AmvmmuAavmm magsoo tom poospom .poospom oEom amammuaavm amammuaaam we mosom eaemam a a ommaoea Lou A>E :mv m< .coaumapae o>apadpom mmmw m anm m< no monaw> poumaaoawo .H< oanme REFERENCES 10. 11. 12. 13. REFERENCES Govindjee and GovindJee, R., Scientific American, 231, 68-82 (1976). White, A., Flandler, P., and Smith, B., "Principles of Biochemistry", McGraw-Hill, New York (1973). Boyer, P. R., Chance, B., Ernster, L., Mitchell, P. Racker, E., and Slater, E. C., Ann. Rev. of Biochem., 35. 955-1026 (1977). Dickerson, R. E., and Timkovich, R. in "The Enzymes", Vol. XI (Boyer, P. D., ed), 397-550 (1975). Marks, G. S., "Heme and Chlorophyll", VanNostrand, London, London (1969). Moore, G. R., and Williams, R. J. P., Coord. Chem. Rev., is. 125—197 (1976). Creutz, C., and Sutin, N., Proceed. Natl. Acad. Sci., USA, 19, 1701—1703 (1973). Ambler, R. P., Meyer, T. E., and Kamen, M. D., Pro- ceed. Natl. Acad. Sci., USA, 73, 472-475 (1976). Cookson, D. J., Moore, G. R., Pitt, R. C., Williams, R. J. P., Campbell, 1. D., Ambler, R. P. Buschi, M. and LeGall, Jr., Eur. J. Biochem., 83, 261-275 (1978). Takano, T., Trus, B. L., Mandel, N. Mardel, G., Kallan, U. B., Swanson, R., and Dickenson, R. F., J. Biol. Chem., 252, 776-785 (1977). Salemme, F. R., Freer, S. T., Nguyen, H. X., Alden, R. A., and Kraut, J., J. Biol. Chem., 248, 3910—3921 (1973). Timkovich, R., and Dickerson, R. E., J. Biol. Chem., 251, 4033-4046 (1976). Mayhew, S. G. and Ludwig, M. L. in "The Enzymes", Vol. XII, (Boyer, P.D. ed) 57-120, Academic Press, New York (1975). 163 l4. l5. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 164 Palmer, G., Muller, F., and Massey, V., in "Flavins and Flavoproteins", 3rd Intntl. Symp. (Kamin, M. ed.) pp. 123-140 (1970), Elsevier. Palmer, G., and Massey, V. in "Biol. Oxidations" (Singer, T. P. ed), pp. 263-299 (1968) John Wiley & Sons. Lehninger, A. L., in "Biochemistry" Worth Publishers, New York (1970). Singer, T. P., Edmondson, D. E., and Kenney, W. C. in "Flavins and Flavoprotein" (Singer, T. P. ed) (1976) Elsevier. Capeillere-Blandin, C. Brey, R. C., Iwasubo, M., and Labeyrie, F., Eur. J. Biochem. 54, 549-566 (1975). Bastsch, R. G. and Kamen, M. D., J. Biol. Chem., 235, 825-831 (1960). Bartsch, R. G., Meyer, T. E., and Robinson, A. B. in "Structure and Function of Cytochromes" (Okunuki, K., Kamen, M. D. and Sekuzv, I. eds.) pp. 443-451 Univ. of Tokyo Press (1968). Walker, W. H., Kenney, W. C., Edmondson, D. F., and Singer, T. P., Eur. J. Biochem., 48, 449-453 (1974). Kennel, J. R. and Kamen, M. D., Biochim. Biophys. Acta, 234, 458-467 (1971). Fukumori, Y., and Yamanaka, T., J. Biochem. (Tokyo), g5, 1405-1414 (1975). Probst, 1., Wolf, B., and Schlegel, H., Biochim. Biophys. Acta. 576, 471-478 (1979). Hopper, D. J., and Taylor, D. G., Biochem J., 167, 155-162 (1977). “’ Siebert, M. and DeVault, D., Biochim. Biophys. Acta, 205, 220-231 (1970). VanGrondelle, R., Duysens L., Vanderwel, J. and VanderWel, H., Biochim. Biophys. Acta, 461, 188-201 (1977). Romijn, J. C. and Amesz, J., Biochim. Biophys. Acta, 461. 327-338 (1977). 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 165 Case, 0., and Parson, W., Biochim. Biophys. Acta, 292, 677-684 (1973). Vorkink, W., Ph.D. Thesis, University of Arizona. Yong, F. C. and King, T. E., J. Biol. Chem., 245, 1331-1338 (1970)- Moss, T. H., Bearden, A. J., Bartsch, R. G., and Cusanovich, M. A., Biochemistry, 1, 1583-1590 (1968). Streéas, T. C., Biochim. Biophys. Acta, 446, 179-191 (197 ). Edmondson, D. E. and Singer, T. P., J. Biol. Chem., 248, 8144-8149 (1973). Cusanuvich, M. A., Ph.D. Thesis, University of California, San Diego, CA. Salmeen, 1., Rimai, L. and Babcock, G. T., Biochemistry, 16, 800-806 (1970). Sutherland, J. C., Vickery, L. E., and Klien, M. P., Rev. Sci. Instrum., 45, 1089-1094 (1974). Gouterman, M. in "The Porphyrins", Vol. III, (Dolphin, D. ed), pp. 1-165 (1978), Academic Press. Simpson, W. T., J. Chem. Phys., 11, 1218-1221 (1949). Gouterman, M., J. Chem. Phys. 30, 1139-1161 (1959). Felton, R. H. and Yu, N. T. in "The Porphyrins", Vol. III (Dolphin, P. ed.), pp. 347-393 (1978), Academic Press, New York. Zerner, M., Gouterman, M., and Kobayushi, H., Theoret. Chim. Acta (Berlin), 6, 363-400 (1966). Smith, D. W., and Williams, R. J. P., in "Structure and Bonding", Vol. 7, pp. 1-45, (1970), Springer Verlay, Berlin. Adar, F., in "The Porphyrins", Vol. III, (Dolphin, D. ed), pp. 167-209 (1978), Academic Press, New York. Drew, H. R., Dickenson, R. E., J. Biol. Chem., 253, 8420-8427 (1978). Lambeth, D., Campbell, K., Zand, R., and Palmer, G., J. Biol. Chem., 248, 8130-8136 (1973). 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 166 Meyer, T. E., and Bartsch, R. G. in "Flavins and Flavo- proteins", (Singer, T. P. ed.), pp. 312-317 (1976) Elsevier. Muller, F. and Massey, V., Jour. Biolg. Chem., 44, pp. 4007-4012 (1969). Beetlestune, J., and George, P., Biochemistry, 3, 707-714 (1964). Kotaki, A., and Yagi, K., J. Biochem., 66, 509-56 (1970)- Chance, B., Erecinska, M., Lee, C. P., Oshino, R., Ohnishi, T. and Prig, M., in "Flavins and Flavopro- teins" (Kamin, H. ed) pp. 669-680 (1971) Elsevier. Kenney, W. C., Edmondson, D., Seng, R., and Singer, T. P., Biochem. Biophys. Res. Communs., 52, 434-439 (1973). ‘— Edmondson, D. and Toller, G., Biochemistry, $6, 113- 122 (1971). F5rster, Th., Disc. Faraday Soc., 21, 7-17 (1959). Hatano, M. and Nozawa, T., Advs. in Biophys., 11, 95-149 (1978). Stephens, P. J., Ann. Rev. Phys. Chem., 26, 201-232 (1974). Vickery, L., Nozawa, T., and Saver, K., J. Amer. Chem. §22m. 98. 343-350 (1976). Briat, B., Berger, D. and Leliboux, M., J. Chem. Phys. 21. 5606-5607 (1972). Vickery, L., Mtds. in Enzymology, 4;, 64-82, (1977). Vickery, L': Nazawa, T. and Sauer, K., J. Amer. Chem. £22-. 2Q. 351-359 (1976). Templeton, D. M., Hollebone, B. R., and Tsai, C. 8., submitted to Biochemistry (1980). Palmer, G., to be published in "The Porphyrins", Vol. IV, (Dolphin, D., ed.), Academic Press, NY (1980). Griffith, J. 3., Nature, 180, 30-31 (1957). 64. 55. 66. 67. 68. 69. 70. 71. 72. 73- 74. 75. 76. 77. 78. 79. 167 Hotgni, M., Prog. Theoret. Phys. Sgppl., ll, 4-13‘ 19 l . Weisbluth, M., in "Hemoglobin: Cooperativity and Electronic Processes", Springer-Verlay (1973). Mun, S. K., Chang, J. C., and Das, T. P., J. Amer. Chem. Soc., 101,5562-5569 (1979). Peisach, J., Blumberg, W. E., and Adler, A., 4m. N. Y. Acad. Sci., 206, 310-327 (1973). Taylor, C. P. S., Biochemices et Biophysica Acta., 191, 137-149 (197777 Brautigan, D. L., Feinberg, B. A., Hoffman, B. M., Margoliash, E., Peisach, J., and Blumberg, W. E., J. Biological Chem., 252, 574-582 (1977). Moore, G. R., and Williams, R. J. P., FEBS Lett., 19. 229-232 (1977). Mashiko, T., Marchon, J. C., Musser, P. T., Reed, C. A., Kastner, M. E., and Scheidt, W. R., J. Amer. Chem. Soc., 101, 3653-3654 (1979). Lemberg, R. and Barrett, J. in "The Cytochromes", Academic Press, New York (1975). Siedow, D. W., Vickery, L. E. and Palmer, G., submitted to J. Biol. Chem. (1980). Kassner, R. J., J. Amer. Chem. Soc., _6, 2674-2677 (1973). Pettigrew, G. W., Bartsch, R. G., Meyer, T. E. and Kamen, M. D., Biochemica et Biophysica Acta, 503, 509-523 (1978). Palmer, G. and Massey, V., in "Biological Oxidations" (Singer, T. P. ed), pp. 263-299 (1968), Wiley. Wellner, D., and Meister, A., J. Biol. Chem., 236, 2357-2362 (1961). Palmer, G., Muller, F., and Massey, V. in "Flavins and Flavoproteins", (Kamin, H. ed.), pp. 123-140 (1970). Hori, H., Biochemica et Biophysica Acta, 251, 227- 235 (1971). 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 168 Dyer, C., Schubert, A., Timkovich, R. and Feinberg, B., Biochemica et Biophysica Acta, 579, 253-268 (1979). Salmeen, 1., Rimai, L., Gill, D., Yamamoto, T., Palmer, G., Hartzell, C. R., and Beinert, H., Biochem. Bio- phys. Res. Communs., 32, 1100-1107 (1973). Lutz, M., Biochimica et Bigphysica Acta, 460, 408- 430 (1977). Spaulding, L. D., Chang, C. C., Yu, N. T. and Felton, R. H., J. Amer. Chem. Soc., 21, 2517-2524 (1975). Burke, J. M., Kincaid, J. R., and Spiro, T. G., J. Amer. Chem. Soc., 100, 6077-6083 (1978). Tang, J., and Albrecht, A. C., in "Raman Spectroscopy", Vol. 2 (Szymanski, H. A., ed), pp. 33-69, (1970) Plenum Press. Albrecht, A. C., J. Chem. Phys., 33, 156-171 (1960). Nafie, L. A., Pezolet, M. and Peticolas, W. L., Chem. Phys. Lett., 39, 563-568 (1973). . Friedman, J. M. and Hochstrasser, R. M., J. Amer. Chem. Soc., 36, 4043-4048 (1976). Shelnutt, J. A., O'Shea, D. C., Yu, N. T., Cheng, L. D. and Felton, R. H., J. Chem. Phys., 64, 1156- 1165 (1976). Clark, R. J. H., and Stewart, B., Structure and Bond- ing. 39. 1-80 (1979). Perrin, M. H., Gouterman, M., and Perrin, C. L., 3. Chem. Phys., 36, 4137-4150 (1969). McClain, w. M., J. Chem. Phys. 32, 2789-2795 (1971). Collins, D. W., Fitcher, D. B., and Lewis, A., 3. Chem. Phys., 62, 5714-5719 (1973). Yamamoto, T., Ph.D. Thesis, University of Michigan (1974). Albrecht, A. C. and Hutley, M. C., J. Chem. Phys., _6, 4438-4443 (1971). Champion, P. M., and Albrecht, A. C., J. Chem. Phys. 1;, 1110-1121 (1979). 169 97. Friedman, J. M. and Hochstrasser, R. M., Chemical Physics, 1, h57-“67 (1973). 98. Remba, R. D., Champion, P. M., Fitchen, D. B., Chiang, R., and Hager, L. P., Biochemistry, 18, 2280—2290 (1979). . 99. Kitagawa, T., Abe, M., and Ogushi, H., J. Chem. Phys. £3. “516-u523 (1978). 100. Spiro, T. G. and Strekas, T. C., J. Amer. Chem. Soc., 2g. 338-3u5 (197u). lOl. Adar, F., and Yonetani, T., Biochimica et Biophysica Acta, 502, 80-86 (1978). 102. Kitagawa, T., Ozaki, Y., Teraoka, J., Kyogoku, Y. and Yamanaka, T., Biochimica et Biophysica Acta, 59“, loo-118 (1977). 103. Ozaki, Y., Kitagawa, T., Kyogoku, Y., Shimada, T., Iizuka, T., and Ishimura, Y., J. Biochem. (Tokyo), 89, 1uu7-1u51 (1978). 10H. Kitagawa, T., Kyogoku, Y., Iizuka, T., and Saito, M. 1., J. Amer. Chem. Soc., 98, 5169—5173 (1976). 105. Sun, M., Moore, T. A., and Song, P. S., J. Amer. Chem. Soc., 9_, 1730-1740 (1972). 106. Kitagawa, T., Nishiru, Y. Kyogoku, Y., Yamano, T., Ohishi, w., Takai-Suzuki, A., and Gahi, K., Biochem— istry, 18, 18ou-1808 (1979). “ ”"— 107. Dutta, P., Nestor, J., and Spiro, T., Proceed. Natl. Acad. Sci. (USA), 15, Alas-ulug (1977). 108. Kitagawa, T., Ozaki, Y., and Kyogoku, Y., Advs. Bio- Phys., 11, 153-196 (1978). 109. Kitagawa, T., Kyogoku, Y., and Orii, Y., Arch. Biochem. Biophys., 181, 228-235 (1977). 110. Spiro, T. G., and Burke, J. M., J. Amer. Chem. Soc., 28, 5u82-5u88 (1976). 111. Maltempo, M. M., Moss, T. H., and Casanovich, M. A., Biochimica et Biophysica Acta, 3&2, 290-305 (197“). 112. Babcock, G. T., Vickery, L. E. and Palmer, G., J. Biol. Chem., 251, 7907-7919 (1976). 113. 11“. 115. 116. 117. 118. 119. 120. 121. 122. 123. 12“. 125. 126. 170 Hamaguchi, K., Ikada, K., and Narita, N., in "Struc- ture and Function of Cytochromes", (Okunuki, K., Kamen, M. D., and Sekuzu, I., eds), pp. 328-33h, Univ. Park Press, Baltimore, MD (1968). Hsu, M.-C., and Woody, R. W., J. Amer. Chem. Soc., 21, 3515-3525 (1971). Meyer, Y. P., Biochemistry, 1, 765-772 (1968). Salemme, F. R., Kraut, J., and Kamen, M. D., 1. Biol. Chem., 2H6, 7701-7716 (1973). Moore, G. R. and Williams, R. J. P., Coord. Chem. Rev., 18, 125-197 (1976). Salemme, F. R., Ann. Rev. Biochem., 88) 299-329 (1977). Blankenship, R. E., and Passor, w. W., to be published in "Topics in Photosynthesis", Vol. III, (Barber, J., ed), Elsevier, Amsterdam. Kihara, J. F., and McCray, J. A., Biochimica et Bio- physica Acta, 292, 297-309 (1973). Devaut “D., Parks, J., and Chang, B., Nature, 215, 692-695 (1967). Itopfield, J. J., Biophys. J., 18, 311-321 (1977). Jortner, J., J. Chem. Phys., 88, N860-M867 (1976). Harbury, H. A. and Foley, K. A., Proceed. Natl. Acad. Sci. (USA), 88, 662-666 (1958). Fleishman, D. E., and Tollen, G., Biochimica et Biophysica Acta, 28, 255-270 (19657? Giravdeau, A., Callot, H. J., and Gross, M., Inorganic Chemistry, 18, 201-206 (1979).