UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Interaction of cytochrome b₅ and cytochrome c Eltis, Lindsay David 1989

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1989_A1 E47.pdf [ 7.14MB ]
Metadata
JSON: 831-1.0098188.json
JSON-LD: 831-1.0098188-ld.json
RDF/XML (Pretty): 831-1.0098188-rdf.xml
RDF/JSON: 831-1.0098188-rdf.json
Turtle: 831-1.0098188-turtle.txt
N-Triples: 831-1.0098188-rdf-ntriples.txt
Original Record: 831-1.0098188-source.json
Full Text
831-1.0098188-fulltext.txt
Citation
831-1.0098188.ris

Full Text

INTERACTION OF CYTOCHROME b5 AND CYTOCHROME c by L I N D S A Y D A V I D ELTIS B . S c , The University of Toronto, 1984 A THESIS SUBMITTED IN P A R T I A L F U L F I L L M E N T OF T H E REQUIREMENTS FOR T H E D E G R E E OF DOCTOR OF PHILOSOPHY in T H E F A C U L T Y OF G R A D U A T E STUDIES T H E D E P A R T M E N T OF BIOCHEMISTRY We accept this thesis as conforming to the required standard T H E UNIVERSITY OF BRITISH C O L U M B I A Apri l 1989 © Lindsay David Eltis, 1989 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of iQ&t+eftl) / The University of British Columbia Vancouver, Canada Date DE-6 (2/88) A B S T R A C T The interaction and kinetics of electron transfer between cytochrome b5 and cytochrome c, two well characterised soluble electron transfer proteins, have been investigated by three techniques. First, fluorescence quenching experiments were done with cytochrome b5 and porphyrin cytochrome c, a fluorescent analogue of cytochrome c. These quenching studies yielded association constants and structural information for the cytochrome ^cytochrome c complex. Second, an analysis of the rate of flavin semiquinone reduction of the cytochromes within the cytochrome ^cytochrome c complex yielded information about the structure and electrostatics of die complex. Third, the second order rate constant for ferrocytochrome b5 reduction of ferricytochrome c was determined under a variety of solution conditions by stopped-flow spectroscopy. Particular effort was directed at evaluating the role of the cytochrome bs heme propionates in the interaction and electron transfer between cytochrome b5 and c through performing each of diese three studies with a derivative of cytochrome bs in which die heme propionates had been esterified (referred to as D M E -cytochrome b5). The fluorescence quenching studies on the interaction of cytochrome b5 and porphyrin cytochrome c and die kinetics of flavin semiquinone reduction of the proteins within the cytochrome ^•cytochrome c complex provided evidence that esterification of the cytochrome b5 heme propionates, influences the docking geometry of the two proteins. These findings support the predictions of electrostatic calculations [Mauk, M . R . , Mauk, A . G . , Weber, P .C . & Matthew, J.B. (1986) Biochemistry, 25, 7085] in two respects. First, esterification of the cytochrome bs heme propionates detectably increases the separation between die two heme groups in die cytochrome ^cytochrome c complex. Second, die cytochrome c heme group is not as sterically hindered in the DME-cytochiome ^cytochrome c complex as in die cytochrome ^cytochrome c complex. The stopped-flow studies demonstrate that the bimolecular rate constant for ferrocytochrome b5 reduction of ferricytochrome c is determined in part by the rate of association of the two proteins. i i This rate of association is strongly influenced by electrostatic forces. The principal effect of esterification of the cytochrome b5 heme propionates on the second order rate of electron transfer between cytochromes bs and c is to influence complex formation between these two proteins. The stopped-flow studies further suggest that the reduction potentials of native and DME-cytochromes b5 are not significantly different when these proteins are bound to cytochrome c. The nature of electron transfer protein-protein interactions and protein-protein electron transfer is discussed with respect to these findings. i i i T A B L E O F C O N T E N T S page A B S T R A C T ii T A B L E OF C O N T E N T S iv LIST OF T A B L E S vi LIST OF FIGURES vii A B B R E V I A T I O N S ix A C K N O W L E D G E M E N T S x INTRODUCTION 1 Overview 1 The Theory of Protein-Protein Electron Transfer Studies 2 Experimental Studies of Protein-Protein Electron Transfer Reactions 7 Protein-Protein Complexes 7 Electron Transfer Kinetics 16 Cytochromes 24 Cytochrome b5 26 Cytochrome c 30 Cytochrome ^Cytochrome c Complex 34 Objectives • 42 E X P E R I M E N T A L PROCEDURES 44 General Procedures 44 Interaction of Porphyrin Cytochrome c with Native and DME-Cytochrome b5 45 Preparation of protein 45 iv Binding measurements 45 Energy Transfer measurements 46 Polarization measurements 47 Kinetics of Flavin Semiquinone Reduction of Cytochrome b5 and Cytochrome c 47 Ferrocytochrome b5 Reduction of Ferricytochrome c 49 RESULTS 55 Interaction of Porphyrin Cytochrome c with Native and DME-Cytochrome b5 55 Kinetics of Flavin Semiquinone Reduction of Cytochrome b5 and Cytochrome c 65 Ferrocytochrome b5 Reduction of Ferricytochrome c 73 DISCUSSION 80 Interaction of Porphyrin Cytochrome c with Native and DME-Cytochrome b5 80 Kinetics of Flavin Semiquinone Reduction of Cytochrome b5 and Cytochrome c 89 Reduction of cytochrome c by LfH- 89 Reduction of cytochrome c by F M N H " 92 Reduction of native and DME-cytochrome b5 by LfH- 93 Reduction of native and DME-cytochrome b5 by F M N H - 95 Ferrocytochrome b5 Reduction of Ferricytochrome c 98 C O N C L U S I O N 110 R E F E R E N C E S 112 A P P E N D I X A 122 A P P E N D I X B 124 v LIST O F T A B L E S page I. Salt linkages of the model cytochrome ^cytochrome c complexes. 38 II. The pH dependence of the extinction coefficient of porphyrin cytochrome c at 505 nm. 55 III. Parameters used in fitting the ionic strength dependence of the association constant of the cytochrome b5 porphyrin cytochrome c complex to the Van Leeuwen equation. 62 IV. Second order rate constants for flavin semiquinone reduction of cytochrome c. 68 V . Second order rate constants for LfH- reduction of cytochrome b5. 69 V I . Second order rate constants for F M N H - reduction of cytochrome b5. 72 VII. Solvent exposure of cytochrome heme groups in the two predicted docking geometries of the cytochrome ^cytochrome c complex 73 VIII. Parameters used in fitting the ionic strength dependence of the second order rate constants of ferrocytochrome b5 reduction of cytochrome c to the Van Leeuwen equation. 76 IX . Calculated monopoles and dipoles of the cytochromes used in this study 79 X . The reported extinction coefficients of porphyrin cytochrome c at 505 nm. 81 X I . Properties of lumiflavin and F M N . 91 XII . Thermodynamic parameters of several second order protein-protein electron transfer reactions. 108 vi L I S T O F F I G U R E S page 1. ProtohemelX 25 2. Structure of cytochrome b5 28 3. Electronic absorption spectrum of cytochrome b5 29 4. Structure of cytochrome c 32 5. Electronic absorption spectrum of cytochrome c 33 6. Structure of Salemme (edge on) cytochrome ^cytochrome c complex 36 7. Structure of edge off cytochrome ^cytochrome c complex 37 8. Stopped-flow spectrophotometer mounted in a glove box 50 9. Simulated difference spectrum of cytochrome b5 and cytochrome c 52 10. Emission spectrum of porphyrin cytochrome c 56 11. The p H dependence of the fluorescence intensity of porphyrin cytochrome c 57 12. Fluorescence titration curve of porphyrin cytochrome c with cytochrome b5 59 13. The ionic strength dependence of the K ,^ of cytochrome b5 and porphyrin cytochrome c 60 14. The pH dependence of of cytochrome b5 and porphyrin cytochrome c 61 15. The ionic strength dependence of Q m a x 63 16. The p H dependence of Q m a x 64 17. Reduction of cytochrome c by lumiflavin 65 18. Reduction of cytochrome c by F M N H - 67 19. Reduction of cytochrome b5 by LfH- 69 20. Reduction of cytochrome b5 by F M N H - 71 21. Ferrocytochrome bs reduction of ferricytochrome c 74 22. The ionic strength dependence of ferrocytochrome b5 reduction of ferricytochrome c 75 23. The pH dependence of ferrocytochrome b5 reduction of ferricytochrome c 11 24. Eyring plot of ferrocytochrome b5 reduction of ferricytochrome c 78 vii 25. Structure of flavins and their oxidation states A B B R E V I A T I O N S C D circular dichroism Co(dipic) 2 ammonium bis(dipiciolinato) cobaltate(III) DME-cytochrome b5 dimethyl esterified heme substituted cytochrome b5 E D T A ethylenediaminotetraacetic acid EPR electron paramagnetic resonance E X A F S extended x-ray absorption fine structure spectroscopy F M N flavin mononucleotide F M N H - flavin mononucleotide semiquinone FTIR fourier transform infrared spectroscopy L H N M R proton nuclear magnetic resonance spectroscopy L f lumiflavin (7,8,10 - trimethylisoalloxazine) LfH- lumiflavin semiquinone M C D Magnetic circular dichroism Throughout this dissertation, amino acid residues are referred to by the three letter abbreviations recommended by the I U P A C - I U B Commission on Biochemical Nomenclature (1972). A number following a three letter abbreviation refers to the position of that residue in the amino acid sequence of the protein being discussed. The amino terminal amino acid residue is numbered "1". For example, "horse heart cytochrome c His-18" refers to the histidine residue at position 18 of horse heart cytochrome c. ix ACKNOWLEDGEMENTS I will never be able to express my gratitude to all the people whose friendship and inspiration have contributed to bringing this project to fruition. I am particularly thankful to my supervisor, Grant Mauk, for his patience and guidance in developing my abilities as a scientist. I am also very grateful for the friendship, advice and assistance of my lab colleagues: Bhavini Sishta, Marcia Mauk, Ed Chang, Alf Gartner, Linda Pearce, Paul Barker, Steve Rafferty, and especially Tony Lim. Gordon Louie, Albert Berghuis and Mike Murphy provided valuable assistance with the IRIS computer. Finally, I should like to thank my family, David, Suzan, Christopher and Jonathan, for encouraging and supporting the various projects I have undertaken during my life. x INTRODUCTION Overview Electron transfer between metallo- or flavoproteins is central to many metabolic processes, including respiration, photosynthesis, and the catalytic cycles of many enzymes. These reactions can involve soluble or membrane bound proteins. In either case, the structural factors that are thought to modulate the rate of electron transfer between the proteins are the same. Electron transfer involving a soluble protein is facilitated by the formation of an intramolecular complex of the two reacting proteins prior to the electron transfer event (Scott et al, 1985). Much effort has been directed at analyzing the structural and functional properties of these precursor complexes. Characterizing the interaction of electron transfer proteins is not only essential to understanding biological electron transfer reactions, but also provides insight into protein-protein interactions in general. The reaction between cytochrome b5 and cytochrome c serves as an excellent model system for the study of protein-protein electron transfer mechanisms for several reasons. The three dimensional structures of these proteins have been determined in both oxidation states by x-ray crystallography (Mathews et al, 1979; Dickerson et al, 1981a,b) and lH N M R (Veitch et al., 1988; Williams et al., 1985a; Moore et al., 1985). Partly based on this information, Salemme (1976) proposed a model for a 1:1 complex that might form between these proteins in solution using computer graphics. This model stimulated several studies aimed at characterizing the nature of the complex (Ng et al, 1977; Stoneheurner et al, 1979; Mauk et al, 1982; Eley & Moore, 1983; Rodgers et al, 1988; Holloway & Mansch, 1988), which have led in turn to further computer graphics modelling (Mauk et al, 1986) and molecular dynamic calculations (Wendoloski et al., 1987). To characterize further the complex that forms between cytochrome b5 and cytochrome c and to address some of the questions that have arisen from recent modelling studies, the interaction between these proteins has been studied in the current work by three new approaches. 1 Spectrofluorometry has been used to measure die binding constant between cytochrome b5 and porphyrin cytochrome c, an analogue of cytochrome c. Spectrofluoromety also provided structural information on die complex. Additional information about the cytochrome ^cytochrome c complex was obtained through an analysis of the rates of reduction of the heme proteins within this complex by flavin semiquinones. Finally, stopped flow spectrophotometry was used to measure the bimolecular reduction of ferricytochrome c by ferrocytochrome b5 under a range of solution conditions. Particular attention has been directed at the role of the heme propionates of cytochrome bs in complex formation through studies widi a derivative of cytochrome b5 in which the heme propionates are esterified widi methyl groups (DME-cytochrome b5). The Theory of Protein-Protein Electron Transfer Reactions Although the theory for electron transfer is more thoroughly developed in nonbiological systems, many of the conceptual insights provided by this work are applicable to protein-protein electron transfer (Devault, 1984; Marcus & Sutin, 1985). The transfer of an electron from one protein to another can be represented by the following scheme (Scott et al., 1985). In diis A" + B < - ^ - > A ' : B — A : B " < - ^ - > A + B" representation, die two proteins, A and B, first come togedier to form a precursor complex, A" :B, with an association constant K j . Within this complex, an electron is transferred from A to B with a rate k e t . The complex dien dissociates at a rate that is assumed to be rapid. The overall measurable rate is given by equation 1. For die convenience of discussion, die reaction can be k = K . k , , [1] envisaged to involve these three distinct processes, each of which is influenced by different factors. 2 A n expression for the rate of electron transfer can be derived by considering diat the rate is the product of the frequency with which the system achieves a suitable nuclear configuration (Franck-Condon factor) and the probability that an electron wil l move from one prosthetic group to the other (PAB) once this configuration has been achieved (Devault, 1984). The potential energy of the reactants and the surrounding medium is a function of all the nuclear coordinates of the system and can be represented by an n-dimensional surface. The potential energy of the products can be represented in a similar manner. The Franck-Condon principle ensures that electron transfer can only occur when the free energies of the reactants and products are equal. For a particular nuclear configuration, the rate of electron transfer is determined by the extent of electronic coupling between the donor and acceptor orbitals as measured by the matrix element, H A B . In equation 2, h 2ir P A B = — | H A B | 2 ( 4 ^ A k B T ) - 1 / 2 [2] h represents Planck's constant, k B represents Boltzmann's constant and A represents the reorganizational energy of the donor and acceptor sites (vide infra). In biological electron transfer systems, electronic coupling between die donor and acceptor sites is often very poor, hence die probability of electron transfer is small, and the reaction is referred to as being non-adiabatic. In such cases a tunnelling mechanism is invoked. Several expressions for the electronic coupling matrix element have been presented (Hopfield, 1974; Jortner, 1976) in which diis value depends exponentially on the distance separating the donor and acceptor orbitals. In a subsequent treatment, Larsson (1983) showed that the height of the tunnelling barrier is significantly reduced by the presence of saturated or conjugated TT systems between the donor and acceptor centres, rather than water. Such intervening material is predicted by Larsson to facilitate faster electron transfer over greater distances. These theoretical considerations dictate that a complex formed between two electron transfer proteins can optimize the rate of electron transfer between the two proteins by minimizing the distance between the prosdietic groups and that any significant separation is occupied by amino acid residues. 3 Marcus derived an expression for electron transfer rates using a statistical mechanical approximation of the nuclear motions as reviewed by Marcus and Sutin (1985). A similar equation may be derived from a simplified one dimensional model (Devault, 1984). Treating the vibrations of reactants and products as harmonic oscillators, the activation energy of the reaction (E*) can be expressed as a function of the reorganizational energy (A) and the driving force ( A E ) . (A - A E ) 2 E * = [3] 4A Averaging over a Boltzmann distribution of energies, the rate of electron transfer is -(A - A E ) 2 27T [ ] k et = — | H A B I 2 (4T Ak B T)" 1 / 2 e 4 A k B T [4] h The Marcus relationship implies that proteins can regulate electron transfer rates by modulating the reorganization energy and die reduction potential of the prosthetic group. Another equation (5) that arises from Marcus theory relates the rate of electron transfer between two species (k) to the rate of self exchange of these two species (kjj , k 2 2 ) and the equilibrium constant of the reaction (Kgq). k = (k ] 1 k 2 2 K e q / ) ' /2 [5] l o g / = (log K ^ / r ^ l o g t k , , k 2 2 / Z 2 ) ] [5a] In this relationship, Z is the collision frequency. In cases where the equilibrium constant approaches unity, / , the collection of constants given in equation 5a, is assigned a value of one. This relationship holds providing that the electron transfer reactions are adiabatic or uniformly nonadiabatic ( P 2 ^ = P A A P B B ) -Alternative treatments of biological electron transfer incorporate nuclear tunnelling to account for temperature independent rates, as first observed in the photo-induced cytochrome oxidation in Chromatium vinosum at low temperatures (DeVault & Chance, 1966). Hopfield (1974) developed a 4 semiclassical solution to this problem through an analogy with radiationless energy transfer (Forster, 1948). In the Hopfield treatment, the vibrational energy states of the reactants and products are quantized with a spacing equal to hw between these levels, but these states are populated according to a classical probability distribution. A full quantum mechanical treatment of vibration coupled to changes in electronic state as it applies to biological systems was presented by Jortner (1976). Both the full quantum mechanical treatment and the semiclassical equation reproduce the classical Arrhenius thermally activated behaviour at high temperatures. Ironically, the observations by Devault and Chance that prompted these theoretical developments have since been interpreted as involving the oxidation of two cytochromes (Bixon & Jortner, 1986). With an increasing number of three dimensional protein structures being determined and the development of molecular graphics capabilities, several detailed hypodietical models for the interaction of electron transfer proteins have been reported. Computer graphics has been used to develop structural models of some of the relatively stable complexes that are known to form between electron transfer proteins including cytochrome c and cytochrome b5 (Salemme, 1976), cytochrome c peroxidase and cytochrome c (Poulos & Kraut, 1980), cytochrome c and flavodoxin (Simondsen et al., 1982), cytochrome b5 and methemoglobin (Poulos & Mauk, 1983), and tetra-heme cytochrome c 3 with both flavodoxin (Stewart et al., 1988) and ferredoxin (Cambrillau et al., 1988). More quantitative approaches to the modelling of protein-protein interactions have made use of electrostatics, Brownian dynamics and molecular mechanics calculations. For example, calculations of association constants based on modified Tanford-Kirkwood theory (Tanford & Kirkwood, 1957; Shire et al., 1974) reproduce the ionic strength dependence of the association constants of flavodoxin and cytochrome c, suggesting that electrostatic forces constitute the principal contribution to the free energy of association of electron transfer proteins (Matthew et al., 1983). However, in the case of the complex formed between cytochromes b5 and c, Tanford-Kirkwood dieory does not account for all of the energy of association (Mauk et al., 1986), implicating the importance of odier forces, such as hydrophobicity, in stabilizing these complexes. The considerable entropic effect of the release of 5 solvent molecules at the interface of the interacting proteins (Chothia & Janin, 1975) overcomes the loss of individual molecular degrees of freedom. The predominant role of electrostatic forces predicts a positive entropy and a negligible enthalpy of association (Ross & Subramanian, 1981), as observed in the case of cytochrome b5 and cytochrome c (Mauk et al., 1982). Brownian dynamics calculations have been used to simulate the association of horse heart cytochrome c and yeast cytochrome c peroxidase (Northrup et al., 1988). The monopoles, dipoles and quadrupoles of these proteins were calculated by assigning a partial charge to each non-hydrogen atom in the structure (Northrup et al., 1986) and the Brownian motions of these models were stochastically simulated. The usefulness of these calculations is attested to by their successful prediction of the behaviour of chemically modified cytochrome c. Brownian Dynamics simulations have suggested several characteristics of the cytochrome c peroxidase-cytochrome c complex (Northrup et al., 1988). Of particular note is the conclusion that the proteins can assume several electrostatically stabilized docking geometries that fulfill the requirements for electron transfer, and that no single complex appears to be energetically favoured. Indeed, several docking geometries may be sampled during one encounter. In addition, although dipole-dipole interactions may be important for steering the molecules towards each other from a distance, surface electrostatic effects dominate at close range. The electrostatic attraction between these proteins ensures that their association rates are close to the diffusion controlled limit, despite the fact that dieir active sites constitute a low percentage of the surface area of the protein. Notably, this study did not (1) explicitly include water molecules, (2) consider the rearrangement of residue side chains during the docking process, or (3) account for non-electrostatic forces that might contribute to complex formation. Molecular dynamics simulations of the cytochrome ^cytochrome c complex on a picosecond time scale (Wendoloski et al., 1987) support the notion of flexibility in a given docking geometry once a particular complex is formed. This work further suggests that rearrangement of amino acid side chains within a complex can lead to docking geometries that are more favourable for electron transfer (vide infra). Each of these three approaches (electrostatics, Brownian dynamics and 6 molecular dynamics) addresses different, yet complementary, aspects of electron transfer protein interaction. Ultimately, it wil l be of interest to see these three approaches combined in a single algorithm, that also accounts for solvent effects, to simulate the entire encounter, recognition and realignment process. From this discussion, it should be clear that the structures of electron transfer proteins can influence the reactions in which these proteins are involved in many ways. In most of the simple cases studied, electrostatic forces modulate the rate and stability of precursor complex formation. The rate of electron transfer widiin this complex depends on the diermodynamic driving force of the reaction, the reorganization energies of the prosthetic groups and the separation and relative orientation of, and the nature of the intervening medium between the donor and acceptor sites. Experimental Studies of Protein-Protein Electron Transfer Reactions Electron transfer in biological systems has been studied by many techniques to assess the various mechanistic features discussed above. These studies have been directed towards (1) determination of the association constants for complex formation between soluble proteins, (2) elucidation of the structures of these complexes, (3) measurement of the rate constants of electron transfer between proteins, (4) evaluation of the magnitude of die parameters that may effect these rates (vide supra), and (5) determination of the structural features of the proteins that influence each of these characteristics. (i) Protein-protein complexes The complexes that form between electron transfer proteins have been characterised by a number of methods. These -studies have involved determining the association constants of the complexes, evaluating the structural features of the partner proteins that contribute to complex stabilization, elucidating the structure of the complex, and determining how die complex influences 7 the electron transfer between the proteins. Association constants have been measured by several techniques including gel filtration, analytical ultracentrifugation, electronic absorption spectroscopy, spectrofluorometry, *H N M R , and, recently, microcalorimetric and pH-stat titrations. Physical measurements of electron transfer protein-protein complexes were first reported for cytochrome c peroxidase and cytochrome c using ultracentrifugation and gel filtration (Nicholls & Mochan, 1971). Although these initial studies did not furnish an association constant for the cytochrome c peroxidase-cytochrome c complex, these values have since been evaluated by these two techniques (Kang et al., 1977; Dowe et al., 1984). The dissociation constant of the cytochrome ^cytochrome c complex has also been estimated by both of diese techniques (Stoneheurner et al., 1979). Gel filtration has been used to determine the dissociation constants of cytochrome c and cytochrome c oxidase (Osheroff et al., 1980) and cytochrome c and sulphite oxidase (Speck et al., 1981). Spectrofluorometric studies involve the use of a fluorescent derivative of one of the partner proteins. In native heme proteins, the porphyrin fluorescence is quenched by the paraniagnetically coupled iron atom. Fluorescent derivatives of hemoproteins are generated by removing die iron atom and sometimes replacing it with another metal such as tin or zinc (Vanderkooi et al., 1976). Quenching of the modified fluorescent protein is observed when another, native heme protein binds to form a protein-protein complex. This approach was used to measure the binding constants between porphyrin cytochrome c peroxidase and cytochrome c (Leonard & Yonetani, 1974; Vitello & Erman, 1987), zinc substituted cytochrome c and cytochrome b2 (Thomas et al., 1983a), and porphyrin cytochrome c and cytochrome c peroxidase (Kornblatt & English, 1986). The dissociation constant of cytochrome c to porphyrin cytochrome c peroxidase was found to be strongly ionic strength dependent (Vitello & Erman, 1987), varying from 0.4 (±0 .1 ) fiM at I = 0.01 M to 180 (±140) nM at I = 0.1 M (pH 6.0, 25 °C). The usefulness of fluorescent analogues in the measurement of protein-protein association constants is attested to by how well these data compare with determinations performed on the native proteins by electronic absorption spectroscopy (Erman & Vitello, 1980) [for instance, 8 K j = 0.2 ( ± 0 . 1 ) fj.M for native cytochrome c peroxidase and cytochrome c (I = 0.01 M , pH 6.0, 25°C)]. In these studies, die use of die modified cytochrome is based on the assumption that the derivative has similar hydrodynamic and electrostatic properties to the native protein. In the case of a porphyrin derivative, this assumption is imperfect because the removal of the iron may affect the tertiary structure and the thermal stability of the protein. These effects might be expected to be more pronounced in cytochrome c than in cytochrome c peroxidase given the smaller size and die greater heme exposure of the former. Nevertheless, viscosity measurements, circular dichroism spectra and tryptophan fluorescence studies indicate that the structure of porphyrin cytochrome c is very similar to that of die native protein (Fisher et al., 1973). Iron reconstituted porphyrin cytochrome c has the same reduction potential and acts as a substrate for cytochrome b2 with the same affinity and velocity constants as native cytochrome c, indicating that removal of the iron does not irreversibly denature cytochrome c (Vanderkooi et al., 1980). This study also reports diat in steady state kinetics widi cytochrome b2, porphyrin cytochrome c has an inhibition constant which is approximately the same as the Michaelis-Menton binding constant for ferric cytochrome c. A substantial body of literature has accumulated with the implicit assumption diat die fluorescent analogues of cytochromes are interchangeable with the native proteins (for example, McLendon & Miller, 1985; Cheung et al., 1986; Koloczek et al., 1987; Liang et al., 1988). The association constants for complex formation by four pairs of hemoprotein complexes have been determined by monitoring die small perturbation in the electronic absorption spectrum that arises when the proteins form a complex. These four pairs of hemoproteins are: cytochrome c peroxidase and cytochrome c (Erman & Vitello, 1980); cytochromes b5 and c (Mauk et al., 1982, 1986); cytochrome b5 and methemoglobin (Mauk & Mauk, 1982); and cytochrome c and cytochrome c oxidase (Michel & Bosshard, 1984; Michel et al., 1989). In each of these cases, the maximum change in absorption caused by protein-protein association is no more than 2% of the total absorbance of the sample, limiting the determination of association constants by this technique to conditions under which complex formation is highly favoured. 9 Most N M R studies have measured the off rates of protein-protein complexes and determined which protons are affected by complex formation. However, *H N M R studies have detected a 1:1 complex between cytochromes b5 and c with an association constant of 10 3 M " 1 (Eley & Moore, 1983) as discussed below. The high protein concentrations necessary for N M R work raise the ionic strength of the solution to the point where comparisons with association constants measured by other techniques is difficult. Two additional techniques have been used to measure the association constant of a 1:1 complex formed between cytochrome c 3 and ferredoxin, microcalorimetry and pH-stat titrations (Guerlesquin et al., 1987). These techniques show considerable promise, because the thermodynamic parameters and proton release associated with complex formation are measured, but they have yet to be applied to other electron transfer protein-protein complexes. Each of these techniques monitors a different characteristic of the complex. N M R measurements are based on the broadening of proton resonances that are associated with interacting residues. Absorption and fluorescence spectroscopies rely on an electronic interaction between the prosthetic groups. Ultracentrifugation and gel fdtration detect differences in the hydrodynamic properties of the free and bound proteins. These latter two techniques might detect complexes that do not result in a perturbation of one of the characteristics upon which the first three tecluiiques are dependent. More generally, differences in the association constants evaluated by these techniques might reflect differences in the populations of docking geometries that each technique detects. N M R shows die most promise in discerning different docking geometries when complexes involve proteins whose N M R resonances have been completely assigned. It wil l be interesting to see, in these cases, whether more than one docking geometry can be detected as a function of solution conditions. Information about die precise docking geometry of electron transfer protein-protein complexes comes from an even greater range of techniques. Methods diat have been used include N M R , spectrofluorometry, C D , M C D , chemical modification, chemical crosslinking studies, and variable reactivity with small electron transfer reagents. The lU N M R studies reported for the cytochrome 10 ^cytochrome c complex are discussed below. N M R has also been used to characterize the complex that forms between cytochrome c and flavodoxin from two species of bacteria (Hazzard & Tollin, 1985; Tollin et al, 1987) and cytochrome c peroxidase and cytochrome c (Satterlee et al, 1987). Hemoproteins are particularly amenable to this type of investigation as die paramagnetic ferric centre induces large chemical shifts in the heme resonances that make these protons easy to identify and monitor. While the high concentrations of protein required for N M R studies can create complicating effects, such as high solution viscosity and protein-protein dimerization (Dixon, 1989), the results of these ! H N M R studies are found to be in agreement with predictions of the computer graphics models. Chemical modification studies have been developed to evaluate the role of individual amino acid residues in stabilising a protein-protein complex. Two basic strategies have evolved based on this approach, both of which have been used to map the electron transfer domain of cytochrome c widi its various electron transfer partners (Margoliash & Bosshard, 1983). The first of these techniques involves preparing derivatives of the protein of interest, singly modified at a variety of surface residues, and testing the reactivities of these modified proteins with a partner electron transfer protein. Trifluoroacetylation and trifluoromethylphenylcarbamoylation of cytochrome c lysine residues were used to map die interaction domain of this protein with cytochromes bs and cl and cytochrome oxidase (Ng et al, 1977; Smith et al, 1980). A similar approach using 4-carboxyl-2,5-dinitrophenylated lysine modified cytochromes c (Brautigan et al, 1978a, b) has been used to map the interaction domain of this protein with cytochrome c oxidase (Ferguson-Miller et al, 1978), ubiquinol-cytochrome c reductase (Speck et al, 1979), cytochrome c{ (Konig et al, 1980), sulphite oxidase (Speck et al, 1981), and yeast cytochrome c peroxidase (Kang et al, 1978). Similar work has been undertaken with cytochrome b5 and NADHxytochrome b5 reductase (Loverde & Strittmatter, 1968; Dailey & Strittmatter, 1979), though the work with cytochrome b5 is less satisfying as individually modified proteins were not purified. These chemical modification studies have been used not only to evaluate the role of individual residues in stabilizing protein-protein complexes, but also 11 to implicate the role of higher order electrostatic forces (e.g., dipole moments) in protein-protein interactions (Koppenol & Margoliash, 1982). In each of these studies, die interaction of the proteins is monitored by measuring a kinetic parameter, however, studies in which die dissociation constants were measured directly yielded similar results (Osheroff et al., 1980; Speck et al., 1981). Unfortunately, in some cases the modifying group is bulky, and it is not clear whether the altered behaviour of a modified protein can be attributed to electrostatic or steric effects. The advent of site directed mutagenesis has facilitated the introduction of more conservative changes, as has been reported for the cytochrome ^cytochrome c complex (vide infra). A second chemical modification strategy has been to compare die susceptibility to modification of residues on the protein free in solution with that of residues on the protein bound to an election transfer partner. The premise of this strategy is that residues involved in protein-protein interaction should be protected from modification. This strategy of differential chemical modification has made use of the methylation (Pettigrew, 1978) and acetylation (Reider & Bosshard, 1980) of lysine residues of cytochrome c, and has implicated the same lysine residues in the interaction domain of cytochrome c with cytochrome oxidase, cytochrome reductase, cytochrome c, and cytochrome c peroxidase as the individually modified cytochromes c did. Chemical modification studies of tiiis type have been instrumental in establishing diat cytochrome c has one interaction surface when it interacts with its physiological electron transfer partners. This surface includes most, i f not all of the exposed heme edge of the cytochrome as well as a number of hydrophobic residues (Margoliash & Bossshard, 1983). Chemical crosslinking has been used to study the structure and dynamics of the complexes formed by several pairs of electron transfer proteins including cytochrome c and cytochrome c peroxidase (Bison & Capaldi, 1981; Waldmeyer & Bosshard, 1985; Moench et al., 1987), cytochromes b5 and c (Mauk & Mauk, 1989), cytochrome c and plastocyanin (Geren et al., 1983; Peerey & Kostic, 1989) and cytochrome b5 and NADPHxytochrome b5 reductase (Hacked & Strittmatter, 1984). In these studies, the partner proteins were covalently coupled with zero length crosslinking reagents, 12 such as l-emyl-3-[3-(dimediylamino)propyl]carrxxliimide hydrochloride, under conditions that promote precursor complex formation. While it is theoretically possible to identify the sites of covalent attachment between the proteins by peptide mapping, the covalent complexes do • not behave as conventional substrates to some proteases and the peptides containing die crosslink often have anomalous elution profiles on the chromatographic columns used to separate them (Mauk & Mauk, 1989). This has made definitive identification of the sites of covalent attachment quite a challenging task. Apart from the usefulness of these studies in characterizing the precursor complexes, crosslinking experiments can provide complexes of electron transfer proteins in fixed, defined geometries that can be useful for evaluating the parameters that influence electron transfer within the complex. Some characterization of covalent cytochrome c peroxidasecytochrome c complexes (Erman et al., 1987; Hazzard et al., 1988) and a covalent cytochrome fo5NADH:cytochrome b5 reductase complex (Hackett & Strittmatter, 1984) have been reported. The lack of electron transfer within four covalent cytochrome cplastocyanin complexes (Peerey & Kostic, 1989) suggests that certain degrees of freedom are necessary for protein-protein electron transfer. While these studies generally ignore the multiplicity of complexes that are formed by chemical crosslinking, as the technique develops it wil l be interesting to measure the reduction potentials of the proteins in these complexes, as has been done in cytochrome cplastocyanin complexes (Geren et al., 1983), particularly as the solution conditions diat promote noncovalent precursor complex formation (low ionic strength) are not conducive to electrochemical measurements. Spectrofluorometry can provide structural information about the precursor complex if the mechanism of quenching of a modified hemoprotein by a native hemoprotein is interpreted in terms of Forster energy transfer theory (Forster, 1948). According to the precepts of this theory, maximal quenching is dependent on the separation and the relative orientation of the donor and the acceptor. This treatment has permitted the estimation of the distances between die prosthetic groups of cytochrome c and porphyrin cytochrome c peroxidase [15 A](Leonard & Yonetani, 1974), zinc and tin cytochrome c and cytochrome c oxidase [36 A](Vanderkooi et al., 1977), and zinc cytochrome c and 13 cytochrome bs [18 A](McLendon et al., 1985). Three different fluorescent derivatives of cytochrome c were used to estimate the distance between the heme of this protein and the heme of cytochrome b2 [18 A](Vanderkooi et al., 1980). In another study, proteolytic fragments of cytochrome b2 were used to localise the interaction domain of this protein with cytochrome c (Thomas et al., 1983b). Using two genetically engineered cytochromes b5 singly modified at surface residues with acrylodan, Stayton et al. (1988) have presented evidence that cytochromes c and P-450 interact with cytochrome b5 differently than myoglobin does. However, it is not clear how the fluorescent probe in this study affects the properties of cytochrome b5. Interestingly, in one study of the quenching of zinc cytochrome c phosphorescence by cytochrome c peroxidase at least part of the quenching was attributed to a conformational change in cytochrome c that occurs upon binding to the peroxidase (Koloczek era/., 1987). Another experimental approach diat has provided useful information about the properties of precursor complexes involves the kinetic analysis of the reduction of die individual protein components of the complexes by exogenous, photolytically generated flavin semiquinones. One appeal of this approach is that alteration of the flavin substituents can be used to generate reductants of different charges, sizes and reduction potentials (Figure 25, Table XI) . Numerous kinetic studies have been reported concerning the reactions of a variety of flavins with various small, soluble electron transfer proteins (Tollin et al., 1986). These studies have established that the mechanism of electron transfer from a semiquinone to a cytochrome involves the transfer of an electron from the N-5 dimethyl benzene ring portion and the C-4a position of the flavin to die exposed heme edge of the protein (Meyer et al., 1984; Pryzsiecki et al., 1985). Using flavins widi different characteristics, it possible to use flavin semiquinone reduction kinetics to probe the environment around the heme edge. A particularly striking example of this is the ionic strength dependence of the rate constant of reduction of Rhodospirillum rubrum cytochrome c 2 by F M N semiquinone (Meyer et al., 1984), indicating that the heme edge of this cytochrome has a positive electrostatic surface, even though the net charge of the protein is zero. A n extension of the use of flavin semiquinones to probe the 14 active sites of electron tranfer proteins involves performing the reduction kinetics of an electron transfer protein bodi in the presence and absence of a partner protein. Changes in the kinetics observed in the presence of an interacting protein can be ascribed to the electrostatic and steric influence of the partner protein on the heme edge. In studies of the flavodoxincytochrome c complex (Hazzard et al., 1986), a decrease in the rate constant of reduction of ferricytochrome c by the electrostatically neutral lumiflavin semiquinone from 7.1 (±0 .5 ) x 10 7 M ' V 1 to 2.9 (±0 .3 ) x 10 7 M ' V 1 in the presence of flavodoxin was interpreted as arising from steric constraints and reduced accessiblity of die cytochrome c heme edge in the presence of flavodoxin. A similar effect on the rate constant of reduction of ferricytochrome c by riboflavin semiquinone was observed [5.4 (±0 .3 ) x 10 7 M ' V 1 in the absence of flavodoxin and 1.7 (±0 .2 ) x 10 7 M ' V 1 in the presence of flavodoxin]. Reduction kinetics with the negatively charged F M N and C 1 2 F M N semiquinones gave information about die electrostatic potential at the protein-protein interface. Pseudo-first order rate constants for die reduction of bound cytochrome c by these negatively charged semiquinones showed a nonlinear concentration dependence. This non-linearity was interpreted in terms of dynamic motions widiin the complex. Similar studies have provided insight into the complexes that form between cytochrome c and both cytochrome c oxidase (Ahmad et al., 1982) and cytochrome c peroxidase (Hazzard et al., 1987; 1988), and ferredoxin N A D P + reductase and a number of other electron transfer proteins (Bhattacharyya et al., 1986; Bhattacharyya et al., 1987). The idea that the formation of an electron transfer protein-protein complex might modulate the parameters that influence the rate of electron transfer between the two proteins was suggested some years ago (Nicholls, 1974), but there has been surprisingly little investigation into tliis area. The influence of charged groups on the reduction potential (Moore et al., 1986) and the reorganizational energy (Warshel & Churg, 1983; Churg et al., 1983) of a protein indicate that the proximity of an oppositely charged electrostatic field of a partner protein could affect these parameters. There have been attempts to correlate the reduction potential with the overall charge of a protein (Rees, 1985), 15 but clearly, this is but one of many factors and is by no means the most important determinant (Moore & Williams, 1977; Moore et al., 1986). Electrochemical studies demonstrated that the reduction potential of cytochrome c decreases from 290 ( ± 5 ) mV to 250 ( ± 5 ) mV upon binding to cytochrome c oxidase (0.02 M M E S , 0.2 M sucrose, pH 7.0) but is not altered by die presence of cytochrome b5 or cytochrome c peroxidase (Vanderkooi & Erecinska, 1974). These studies were performed at an ionic strength of 0.02 M , which may not be low enough to promote a stable cytochrome ^-cytochrome c complex. Separate studies have shown that the reduction potential of cytochrome c drops 25 mV on binding of the cytochrome to plastocyanin (Geren et al., 1983) and 30 mV on binding to flavodoxin (Hazzard et al., 1986). Each of these studies was limited to a single set of solution conditions (buffer composition, ionic strength, pH and temperature). C D , M C D and phosphorescence measurements have provided some insight into the structure of electron transfer protein-protein complexes. Complexes involving cytochrome c and flavodoxin (Tollin et al., 1987), and cytochrome c and cytochrome c oxidase (Weber et al., 1987) display C D or M C D spectra that differ from the sum of the spectra of the partner proteins, suggesting that complex formation induces changes in the electronic structure of die prosthetic groups. These changes have been interpreted as facilitating electron transfer eidier by decreasing the reorganizational free energy of electron transfer (Weber et al., 1987) or increasing the accessibility of the prosthetic groups (Tollin et al., 1987). Recently, the change in the C D spectrum of cytochrome c in the presence of cytochrome c oxidase has been interpreted as a conformational change involving the movement of phenylalanine-82 away from the cytchrome c heme (Michel et al., 1989a,b). As discussed above, phosphorescence quenching studies of zinc cytochrome c suggest that this protein might undergo a conformational change upon binding cytochrome c peroxidase (Koloczek et al., 1987). The increasing number of calculations based on structural information of protein-protein complexes may stimulate research in this area. The experimental studies reviewed above have supported the theoretical predictions discussed in the previous section in at least two ways. First, a large number of electron transfer protein 16 precursor complexes are stabilized by electrostatic forces. Given the failure of electrostatic calculations to account for all of energy of association of some of these complexes, it is surprising that few experiments investigating the role of uncharged amino acids in stabilizing protein-protein complexes have been undertaken. Second, experimental studies provide evidence that these complexes bring the prosthetic groups of the electron tranfer proteins into close proximity and suggest that the computer graphics generated models are accurate representations of the average docking geometry of these proteins. While it has been suggested that electron transfer proteins can assume several docking geometries in which efficient electron transfer can occur (Maine et al., 1986; Northrup et al., 1988), there is no definitive experimental evidence for this assertion. Finally, although two or three studies suggest that complex formation might influence the electronic and conformational structure of the partner proteins, there have been few documented attempts to evaluate the effect of complex formation on the parameters, such as the reduction potential and the reorganizational energies of die partner proteins, that affect electron transfer. Undoubtedly, part of this deficiency arises from die practical difficulties involved in working at die low ionic strengths required for stable protein-protein complexes to form in most cases. (ii) Electron transfer kinetics Various techniques have been used to evaluate the factors diat affect electron transfer rates, such as the reduction potentials, reorganizational energies and relative positions of the donor and acceptor centres. Experimental studies of protein-protein electron transfer report either first order rates between donor and acceptor centres at fixed separation, or second order rates of electron transfer between soluble proteins. This discussion begins with the first case, which is conceptually easier to envisage. To measure the rate of electron transfer between sites of fixed, known separation, several laboratories have attached ruthenium compounds to specific histidine imidazoles (Mayo et al., 1986). These studies also provide information about the influence of the intervening medium, and in cases 17 where the ruthenium ligands and metal centres can be modified, the driving force, on the rate of electron transfer. These studies appear to corroborate the predictions that the rate of electron transfer depends exponentially on distance and that the rate decreases when the reorganizational energy is increased or the driving force is decreased (Mayo et al., 1986; McLendon, 1988). Alternatively, some of these factors have been evaluated by measuring the rates of electron transfer within hemoglobin hybrids in which the iron atom of one of the subunits has been replaced widi zinc or tin (e.g., McGourty et al., 1983). Electron transfer between the metal centres is initiated by a laser flash to photoproduce an excited triplet state. This excited triplet state can either thermally decay back to the ground state or, in the presence of a suitable acceptor, transfer an electron. The thermodynamic driving forces of these reactions are much greater than physiological electron transfer reactions, a consideration that may also influence the mechanism of electron transfer in some manner diat is difficult to assess. Nevertheless, these studies have provided insight into the distances over which electron transfer can occur and the influence of the intervening medium on die rate of electron transfer. One question that all of these studies raise is why the first order rate constants of electron transfer involving proteins that are putatively designed for electron transfer are so many orders of magnitude smaller than the rate constants of electron transfer measured over comparable distances in non-protein systems (Miller, 1975; Heiler et al., 1987) or in other biological systems, such as the photosynthetic reaction centre (Parson et al., 1978). Measurements of the first order rate of electron transfer between two soluble proteins in a preformed complex have also been reported for cytochromes b5 and c (McLendon & Miller, 1985), cytochrome c peroxidase and cytochrome c (reviewed in McLendon, 1988), and cytochromes b2 and c (McLendon et al., 1987). The attraction of these systems is tiiat diey involve donor and acceptor sites located in physiological environments. One disadvantage is that die precise structures of these complexes is unknown. For each of these systems, the reorganizational energies of the electron transfer reaction may be evaluated through varying the thermodynamic driving forces of die respective reactions and using the Marcus relationship (equation 3). The best example involves the 18 electron transfer reaction between cytochromes b5 and c (McLendon & Miller, 1985), for which the reorganization energy has been estimated to be 0.7 eV (16 kcal/mol). The driving force for the reaction between these proteins was varied by removing the iron of cytochrome c and measuring the rate of electron transfer within the complex formed by cytochrome b5 with native, porphyrin and zinc cytochromes c. The magnitude of the reorganization energy seems reasonable compared to the calculated reorganization energy of cytochrome c [6.4 kcal/mol](Warshel & Churg, 1983; Churg et al., 1983) and considering that the reorganization energy of an electron transfer can be estimated to be the mean of the reorganization energies of the self exchange reactions. While the effects of such a modification on protein conformation and hence the precursor complex geometry might be minimal (vide supra), the effect of such a substitution on the electronic configuration of the porphyrin is more difficult to evaluate. Additional complications arise when considering the direction of electron transfer within the complex and die solution conditions in which the kinetics were performed. In the studies of McLendon and Miller cited above, it is assumed that for the purposes of measuring reorganizational energies, electron transfer rates within a complex are equivalent no matter which direction the electron travels. Recent studies in which a mutation of cytochrome c Phe-82 exerts a greater influence on the rate of electron transfer from the zinc cytochrome c peroxidase excited triplet state to ferricytochrome c than on the rate of electron transfer from ferrocytochrome c to the zinc cytochrome c peroxidase porphyrin cation radical (Liang et al., 1987) indicate that this assumption may not be valid. In light of the observation that the ionic strength of the solution can influence the rate of electron transfer within the complex (Hazzard et al., 1988) and the suggestion that different solution conditions can favour different docking geometries (Mauk et al., 1986), care must be taken to reproduce the solution conditions in each case. Further complexities of protein-protein electron transfer are apparent from a study involving cytochromes b2 and c (McLendon et al., 1987) that is analagous to the study widi cytochromes b5 and c discussed above. The rate of electron transfer within the cytochrome fc2'cytocnrome c 19 complex was found to be independent of the driving force of die reaction. The authors suggested that this observation can be taken as evidence of a conformational gating process controlling the rate of electron transfer. Intriguingly, lU N M R studies have suggested diat the cytochrome domain of cytochrome b2 is quite mobile in the holoenzyme (Labeyrie et al., 1988), though die nature and timescale of this motion is unknown. Further characterisation of this conformational mobility and the structure of the cytochrome fc2cytoc^rome c complex is required before definitive conclusions are drawn. The concept that protein-protein electron transfer might by controlled by a conformational change has been developed theoretically by Hoffman & Ratner (1987). A first order rate constant for the nonphysiological reduction of ferricytochrome c by ferrocytochrome c peroxidase has been reported [3.4 s"1 for the yeast proteins in 0.01 M phosphate, pH 7.0, 24°C](Cheung et al, 1986; Cheung & English, 1987). This reaction was observed to be unimolecular at both high and low ionic strengths, and to exhibit thermodynamic parameters that vary with ionic strength. The authors have suggested that at low ionic strength, it is electron transfer within the complex that is rate limiting and at higher ionic strengths, it is a conformational change in the peroxidase that is rate limiting. While these data are interesting, kinetic data involving cytochrome c peroxidase should be viewed with caution owing to die difficulty in isolating the native, pentacoordinated form of the enzyme uncontaminated with a six coordinate form (Yonetani & Anni , 1987). Whatever die explanation of diese results, it is clear that even in instances where a first order rate is measured between donor and acceptor sites, interpretation of these rates solely in terms of electron transfer theory may be overly simplistic. Measurements of the second order rates of electron transfer between proteins pose several difficulties that make such measurements technically challenging. Many reduced proteins are readily oxidised by molecular oxygen, necessitating the use of anaerobic conditions. Another difficulty presented by intensely coloured heme proteins is the measurement of small changes in optical density against a high background. Despite these problems, the bimolecular rate constants of a number of protein pairs have been measured including horse heart and Candida krusei cytochromes c (Yoshimura 20 et al., 1981), horse heart cytochrome c and Pseudomonas cytochrome c 5 5 1 (Morton et al., 1970) and horse heart cytochrome c with plastocyanin and azurin (Augustin et al., 1983). These measurements were performed on stopped flow spectrophotometers, and all the bimolecular rate constants were on the order of 10 4-10 6 M ' V 1 . Electron transfer between two soluble proteins wi l l involve not only the mechanistic considerations discussed in the case of electron transfer within a protein-protein complex, but also the recognition and association process. There have been few attempts to correlate protein-protein binding data and. rates of electron transfer within protein-protein complexes with measured second order rate constants. This analysis has been reported to a certain extent for the reactions between cytochrome c and cytochrome c peroxidase (Summers & Erman, 1988) and between cytochrome c and plastocyanin (King et al., 1985). Evidence for the rate limiting protein conformational steps that have been invoked to explain the mechanism of protein-protein electron transfer within some complexes (vide supra) has also been seen in bimolecular reactions. The kinetics of electron transfer from reduced azurin to ferricytochrome c 5 S 1 isolated from Pseudomonas aeruginosa are clearly biphasic (Pecht & Rosen, 1973; Wilson et al., 1975) and have been interpreted as involving an equilibrium between two forms of reduced azurin (Wilson et al., 1975) and two forms of ferricytochrome c 5 5 1 (Rosen & Pecht, 1976). The electron transfer reaction of azurin isolated from Alcaligenes faecalis with Pseudomonas cytochrome c 5 5 1 suggests diat there is only one conformational form of Alcaligenes azurin (Rosen et al., 1981). These results were initially interpreted as involving the protonation of histidine residue 35 which is present in Pseudomonas azurin but not Alcaligenes (Rosen et al., 1981). However, the pH dependence of the rate of electron transfer between Pseudomonas azurin and cytochrome c 5 5 1 does not suggest any strong dependence of reactivity on the state of protonation of the protein (Cor ing a/., 1983). In one of the most extensive studies of bimolecular protein-protein electron transfer, the rate constant for the reduction of several c-type cytochromes by flavodoxin semiquinone were determined 21 as a function of ionic strength (Tollin et al., 1984). Although modest success was achieved in fitting the electrostatically corrected rate constants to the equations of Hopfield and Marcus, the scatter in these data illustrates the complexity of the bimolecular electron transfer reaction. In fitting the data to the Marcus relationship, a value of 43 kcal/mol was used for the reorganization energy [A is 4 times A G J at zero driving force (equation 3)]. This is considerably larger than the value of A (11 kcal/mol) used in a similar analysis of the second order rate constants of electron transfer from free F M N to c-type cytochromes (Meyer et al., 1984). The authors equate structural complexity with more extensive structural rearrangements and rationalise the difference in A by the greater structural complexity of flavodoxin as compared to F M N . This interpretation is incompatible with the suggestion that one of the roles of the polypeptide chain of electron transfer proteins is to reduce the reorganisational energy of the electron transfer centre (Churg et al., 1983). This discrepancy aside, considering the potential importance of complex formation and conformational changes in protein-protein electron transfer, it is perhaps fortuitous that any reasonable correlation was observed between the second order rate constants of protein-protein electron transfer and the driving force of the reaction. Wherland and Pecht (1978) measured the second order rate constants of electron transfer between several pairs of electron transfer proteins at constant ionic strength and fitted these rate constants to die Marcus relationship (equation 5), even though it is not clear that protein-protein electron transfer reactions are adiabatic or uniformly nonadiabatic. Although the calculated self exchange rates were remarkably consistent for the series of electron transfer proteins studied in the work of Wherland and Pecht, there is considerably less agreement with self exchange rates calculated from protein-protein electron transfer rate constants measured by other workers, or from electron transfer reactions between proteins and small molecules (Cummins & Gray, 1977). Interestingly, die measured self exchange rate of cytochromes c [6.7 x 10 3 M"'s" ' (I = 0.13 M , pH 7.0, 2 7 ° Q ] and cytochrome cS5l [1.2 x 10 7 M ' V 1 (0.05 M phosphate, pH 7.0, 2 5 ° Q ] (Gupta, 1973; Dixon et al., 1989) agree reasonably well with the values calculated in this study [1.5 x 102 M"'s"' 22 and 4.6 x 10 7 M ' s " 1 respectively (0.05 M phosphate, pH 7.0, 25 °C)]. Related to these kinetic measurements has been the determination of a number of metalloprotein electron transfer self-exchange rates. Generally, diis has been accomplished using ' H N M R , as in the case of cytochrome c (Gupta, 1973; Dixon et al., 1989a), cytochrome c 5 5 1 (Keller et al., 1976; Dixon et al., 1989a), plastocyanin (Armstrong et al., 1985) and cytochrome b5 (Dixon et al., 1989a,b). Studies of the ionic strength dependence of these rate constants (Gupta, 1973; Dixon et al., 1989a,b) and the self-exchange rate constants of chemically modified proteins (Concar et al., 1986) have been useful in establishing the major role of electrostatic forces in protein-protein electron transfer. Attempts to correlate these self-exchange rate constants with distances of electron transfer within the protein-protein complex (Dixon et al., 1989a) and reorganization energies (Dixon et al., 1989b) are interesting, but must be viewed with caution until the mechanism of protein-protein electron transfer has been established. Should this mechanism include a rate limiting step involving complex formation or a change in protein conformation, dien the analysis of protein-protein electron transfer by multiphonon electron transfer theory would not be appropriate. In these and related studies, several formalisms have been used to analyse the ionic strength dependence of the bimolecular rate. The seminal work of Debye and Hiickel (1923) analyses the observed behaviour in terms of monopole-monopole interactions. Several adaptations of the Debye-Hiickel equations to protein electron transfer reactions have been developed (for example, Wherland & Gray, 1976; Stoneheurner et al., 1979). Although these treatments have been used in many situations, they are inadequate to describe macromolecular interactions. Not only do these equation treat proteins as homogeneously charged spheres, which clearly they are not, but above ionic strengths of 0.01 M , higher order electrostatic components contribute to the interactions (Van Leeuwen, 1983). Two approaches that consider the contribution of protein dipole moments are those developed by Koppenol et al. (1978), based on the Brdisted relationship, and Van Leeuwen (1983). Koppenol has revised his equation (Rush & Koppenol, 1988) and it is now virtually identical to the Van Leeuwen approach. An alternative strategy considers the interacting domains of the proteins as 23 two interacting charged plates. An empirical equation based on this approach works reasonably well (Tollin et al., 1984), diough a detailed theoretical development of this model has not been published. These studies indicate that in the simplest cases, biological electron transfer may be reasonably interpreted in terms of electron transfer theory (DeVault, 1984; Marcus & Sutin, 1985). On the whole, there remains a scarcity of data on the reorganization energies and reduction potentials of proteins within a precursor complex. In many instances, electron transfer between soluble proteins seems to involve more than the collision of two proteins followed by an electron transfer event and may be determined by die rate of protein-protein recognition and association, and conformational rearrangements within the complex (Hoffman & Ratner, 1987). Cytochromes The cytochromes are a class of heme containing proteins whose principal biological function is electron transport by virtue of a reversible valency change of dieir heme iron (Leniberg & Barrett, 1973; Pettigrew & Moore, 1987). Although cytochromes were originally discovered in the late nineteenth century by MacMunn (1884), dieir existence was not generally acknowledged until they were rediscovered by Keilin in the 1920's (Keilin, 1966). Since that time, these proteins have been discovered in all living organisms and have been broadly classified according to the structure of their prosthetic groups into a, b, c and d types. Cytochromes have a distinctive red colour and a characteristic four banded electronic absorption spectrum composed of an o band around 550-604 nm, a p band around 520-546 nm, a i or Soret band around 400-450 nm, and a 5 band around 300-350 nm, in addition to the protein absorption bands in the ultraviolet region. The extinction coefficients of these bands were first determined in horse heart cytochrome c (Margoliash & Frohwirt, 1959). The majority of cytochromes studied consist of an iron porphyrin prosthetic group bound to a single polypeptide chain containing no disulphide linkages. The heme consists of an iron atom ligated .to protoporphyrin IX (Figure 1) (Smith, 1975). In some classes of cytochromes, the structure 24 ROOC COOR Figure 1 : The structure of protoheme IX. In the native protoheme, R = H; in dimethylesterified protoheme IX, R = C H 3 . 25 of the heme is modified. For example, the heme of a type cytochromes contains a formyl side chain (Lemberg & Barrett, 1973). Four of the octrahedrally coordinated iron ligands are provided by the pyrrole nitrogens and the odier two axial ligands are derived from amino acid side chains. In the oxidised Fe(III) state, die iron has a single unpaired electron and is paramagnetic. As the iron displaces two protons when it is incorporated into the porphyrin, ferriheme has a formal charge of + 1. In the reduced Fe(II) state, the iron has no unpaired electrons, and the heme has no net charge. (i) Cytochrome b5 As the name implies, cytochrome b5 is a fc-type cytochrome, and as such it has a noncovalently bound heme. The axial ligands of the iron atom are provided by the e-nitrogen atoms of histidine residues 39 and 63. Cytochrome b5 occurs in several forms in vivo (Madiews, 1985). A membrane bound form occurs in the smoodi endoplasmic reticulum of mammalian hepatocytes where it functions in steroid hydroxylation (e.g., Grinstead & Gaylor, 1982), fatty acid desaturation (Strittmatter et al., 1974) and supports the reduction of a variety of cytochromes P-450 (e.g., Peterson & Prough, 1986; Tamburini & Schenkman, 1988). This species is a monomer of about 16,000 daltons consisting of a simgle polypeptide chain diat is folded into two domains. The larger domain consists of the first 86 - 90 amino terminal residues and contains the heme prosthetic group. This domain is hydrophilic and extends into the aqueous environment of the cytoplasm (Tanford, 1980). The second domain is composed of the 37 predominantly hydrophobic, carboxyl terminal residues and is embedded in the lipid bilayer, anchoring die cytochrome to the membrane. A flexible sequence of approximately 7 residues links the two domains (Bendzko & Pfeil, 1983). The complete amino acid sequence of bovine liver microsomal cytochrome b5 was determined (Ozols, 1975) but has recently been corrected at three positions in the hydrophilic heme binding region based on the sequence of the cytochrome bs gene (Cristiano & Steggles, 1989). A soluble form of cytochrome bs occurs in mature erythrocytes where it functions to reduce 26 methemeglobin (Passon & Hultiquist, 1971). This species is apparently derived from a microsomal precursor by cathepsin digestion during erythroid development (Slaughter & Hultquist, 1979). Mature erythrocytic cytochrome b5 has 97 residues. Although the first 96 residues of die porcine and human erythroid proteins are identical to their respective hepatocyte species, a difference in residue 97 suggests that the erythrocyte and hepatocyte cytochromes are derived from two distinct but closely related genes (Abe et al., 1985). Cytochrome b5 has also been isolated from two species of sipunculids where it is believed to function in mememerythrin reduction (Utech & Kurtz, 1985; 1988). A second membrane bound species of cytochrome b5 is found on die inner surface of the outer membrane of mitochondria (Mathews, 1985). The amino acid sequence of die globular heme binding domain of this protein shows 58% sequence homology to the corresponding microsomal heme binding domain (Lederer et al., 1983). It has been suggested that this species of cytochrome bs participates in an electron shuttle between the inner and outer mitochondrial membranes, effecting the aerobic oxidation of exogenous N A D H (Bernardi & Azzone, 1981). In this proposed pathway, cytochrome b5 reacts with intermembrane cytochrome c. Bovine microsomal cytochrome b5 may be isolated in several forms including a lipase solubilized and trypsin solubilized form. The lipase solubilized form comprises residues 5 to 97 of the intact microsomal sequence whilst the trypsin solubilized form is comprised of residues 7 to 88. In the current studies, the trypsin solubilized fragment isolated from bovine liver was used. Currently, the crystal structure of the lipase solubilized form of bovine microsomal ferricytochrome b5 has been solved to a resolution of 2 A (Madiews et al., 1979), but two residues at die amino terminal and ten residues at the carboxyl terminal could not be detected in the electron density maps. Consequently, the crystallographic structure is of a fragment diat approximately corresponds to the smaller, trypsin solubilized species (Figure 2). With resonance assignments for more than 40% of the protons of the tryptic fragment of oxidized and reduced cytochrome b5, two dimensional *H N M R revealed that differences between the crystal and solution structures are limited to small variations in a few side chain orientations (Veitch et al., 1988). Neither technique detected significant conformational 27 Figure 2 : The structure of the tryptic fragment of bovine hepatic ferrocytochrome b5. Tins image was generated with the Raster3D package (D. Bacon, University of Alberta) on a Silicon Graphics IRIS 3130. The heme atoms are shown in red and the aspartate and glutamate side chains, in orange. The atomic coordinates where taken from the Brookhaven Data Bank (Mathews et al., 1979). 28 270 315 360 405 450 495 540 585 630 Wavelength (nm) Figure 3 : The electronic absorption spectrum of the tryptic fragment of bovine microsomal cytochrome b5. The spectra are normalised to an extinction coefficient of 117 mM" 1 at 412.5 nm of the oxidised spectrum (Ozols & Strittmatter, 1964). The oxidised and reduced forms are indicated. Ferrocytochrome bs was produced by the addition of a few grains of sodium dithionite to the solution of ferricytochrome b5. The conditions were 0.02 M sodium phosphate, pH 7.0, 25°C. 29 differences between the reduced and oxidized forms of the protein, diough x-ray crystallography detects the binding of a cation to heme propionate-7 in the reduced form (Argos & Matdiews, 1975). The crystallographic model of the cytochrome b5 indicates that the molecule is cylindrical in shape, measuring 37 A in height and 31 A in diameter. The polypeptide backbone is folded into 6 short a helices and a 5 stranded /? pleated sheet. The heme is buried in a large hydrophobic core such that the propionate groups (numbered 6 and 7) of the heme group are oriented towards the surface of die protein. Propionate-6 extends into the solvent creating a pronounced pimple in the electrostatic surface of the molecule (Argos & Madiew, 1975; Mauk et al., 1986). The second propionate group is not as prominent on the electrostatic surface of cytochrome b5 because it bends back to form two hydrogen bonds with serine 64. J H N M R studies suggest diat in 10% of the protein molecules, the heme is bound in a second orientation (La Mar et al., 1981). This second orientation differs from the first by a 180° rotation about the 0 - 7 - m e s o axis of the heme in the heme binding pocket. Cytochrome b5 is acidic, containing 17 aspartate and glutamate residues, two heme propionates, and 9 basic (lysine and arginine) residues. The excess acidic residues are clustered on the surface of the molecule around the mouth of die heme crevice, giving the molecule a substantial dipole moment. The trypsin solubilized fragment of cytochrome b5, henceforth referred to as cytochrome b5, has a molecular weight of 10,063 daltons and a pi of 5.2 (Reid, 1984). The reduction potential has been measured to be 5.1 (+0.6) mV (pH 7.0, I = 0.1 M , 25°C) versus the saturated hydrogen electrode (Reid et al., 1982). The electronic absorption spectra of ferro- and ferricytochrome b5 are given in Figure 3. (ii) Cytochrome c Cytochrome c is one of the most extensively studied proteins. The amino acid sequences from over 100 different species, including horse (Margoliash et al., 1961), have been determined. Crystal structures of cytochrome c from several eukaryotes, including horse (Dickerson et al., 1971; Bushnell et al., unpublished), bonito (Dickerson et al., 1971), tuna (Takano & Dickerson, 1981a,b), rice (Ochi 30 et al., 1983), and yeast (Louie et al., 1988) have been solved. As with most c class cytochromes studied, the heme of cytochrome c is covalently linked to die peptide chain via two thioether bridges. Horse heart cytochrome c is made up of 104 amino acid residues and has a molecular mass of 12,500 Daltons. The axial ligands of the iron atom are provided by the 6 sulphur atom of a methionine (residue 80) and the e nitrogen atom of a histidine (residue 18). The iron is low spin in both oxidation states of this protein. The three dimensional structure of horse heart cytochrome c has been solved to 2.8 A resolution (Dickerson et al., 1971). The structure of this species of cytochrome c is very similar to that of the tuna protein, which has been solved to 1.5 A and 1.8 A in the reduced and oxidized forms respectively (Takano & Dickerson, 1981a, b) (Figure 4). The solution structure of tuna cytochrome c in both oxidation states has also been studied by *H N M R (Williams et al., 1985a; Moore et al., 1985). Apart from some small differences in three regions of die protein (at the carboxyl terminus, around heme propionate-7 and near thioether-2) diese structures agree remarkably well and the crystal structures accurately describe the solution structures of cytochrome c (Williams et al., 1985b). Two dimensional ' H N M R studies of reduced and oxidised horse heart cytochrome c (Wand et al., 1989; Feng et al., 1989) indicate that the differences between this species of cytochrome and tuna cytochrome c are minor. Bodi ' H N M R and x-ray crystallography studies indicate that there are few conformational differences between ferric and ferrocytochrome c E X A F S studies indicate diat there is very little difference in die immediate coordination environment of the iron atom in the two oxidation states (Korszan et al., 1982). The molecule is a prolate spheroid measuring 30 x 34 x 34 A with the peptide backbone folded into 6 helices that account for 48% of the structure. Residues 1 to 47 are folded on one side of the heme while residues 48 to 91 are on the other. The remaining residues (92 to 104) form a strap that passes across the molecule, behind the heme. The heme is deeply buried in a fairly hydrophobic pocket and is oriented such that one of the heme edges is exposed to the solvent and the heme propionates point towards the bottom of the crevice. 31 Figure 4 : The three dimensional structure of horse heart ferricytochrome c. The image was generated with the Raster3D package (D. Bacon, University of Alberta) on a Silicon Graphics IRIS 3130. The heme atoms are shown in red, the methionine and cysteine side chains are shown in yellow, and the lysine and arginine side chains are shown in blue. The atomic coordinates were generously provided by Prof. Gary D. Brayer. 32 Wavelength (nm) Figure 5 : The electronic absorption spectrum of horse heart cytochrome c. The spectra are normalised to an extinction coefficient of 106.1 mM" 1 at 409 n M of the oxidised spectrum (Frohwirt & Margoliash, 1959). The oxidised and reduced species are indicated. Ferrocytochrome c was produced by adding a few grains of sodium dithionite to the solution of ferricytochrome c. The solution conditions were 0.02 M sodium phosphate, pH 7.2, 25°C. 33 Cytochrome c is a basic molecule with a pi of approximately 10.5. The tuna protein has 18 positively charged residues and 7 negatively charged residues. The excess positively charged residues are clustered around the mouth of die heme crevice, creating a significant charge asymmetry on the surface of the protein. Horse heart cytochrome c has three additional positively charged residues and three additional negatively charged residues. The reduction potential of cytochrome c is 261 mV (pH 7.0, I = 0.01 M , 25°C) versus the saturated hydrogen electrode (Margalit & Schejter, 1973). The electronic absorption spectra of ferri- and ferrocytochrome c are presented in figure 5. Cytochrome c has several roles in vivo which have been reviewed recently (Pettigrew & Moore, 1987). This protein is found in the periplasmic space between die outer and inner mitochrondrial membranes. Its primary role is to shuttle electrons between ubiquinol-cytochrome c oxidoreductase (fccj complex, complex III) and cytochrome c oxidase (complex IV) in eukaryotic respiration. In the periplasmic space, cytochrome c can react with other intermembrane electron transfer proteins such as sulphite oxidase in animals and cytochrome c peroxidase and cytochrome b2 in yeast. Closely related cytochromes occur in other electron transfer pathways, such as in photosynthesis where they shuttle electrons between ubiquinol-cytochrome c2 oxidoreductase and the photochemical reaction centre protein. The Cytochrome /^Cytochrome c Complex The wealth of structural and functional data for cytochromes bs and c makes them excellent choices for use in the study of protein-protein interaction and electron transfer reactions.' In addition, genes encoding for both of these proteins have been expressed in high efficiency (von Bodman et al., 1986; Pielak et al., 1985). This development allows for the production of specifically mutated forms of diese proteins that should be of considerable value in elucidating the contributions of individual amino acid residues to the interaction of and electron transfer between these proteins. Although the physiological significance of the cytochrome ^cytochrome c complex remains a moot 34 point (vide supra), physiologically relevant interactions of cytochrome c with sulphite oxidase (Johnson & Rajagapolan, 1977, 1980) and cytochrome b2 (Briquet et al., 1980), two proteins that contain a cytochrome bs domain, have been documented. Two lines of evidence suggest diat cytochromes b5 and c interact with each other in a manner analogous to that of physiological oxidoreductases. First, the bimolecular rate constant of the reduction of cytochrome c by cytochrome b5 is very similar to that of physiological electron transfer reactions (Strittmatter, 1964). Second, the cluster of positively charged residues around the heme crevice of cytochrome c have been implicated in its interactions with physiological electron transfer partners (Margoliash & Bosshard, 1983). Similarly, die acidic groups clustered around the heme edge of cytochrome b5 are thought to mediate the interaction between this protein and its various physiological partners such as hemoglobin (Mauk & Mauk, 1982; Poulos & Mauk, 1983), NADHrcytochrome b5 reductase (Dailey & Strittmatter, 1979), stearyl CoA reductase and NADPHxytochrome P-450 reductase (Dailey & Strittmatter, 1980). The charge complementarity between cytochrome c and cytochrome bs is thought to stabilise the complex that forms between these proteins in a manner analogous to the physiological examples. Salemme (1976) proposed a model of the complex formed by these two proteins based on the structural coordinates of tuna cytochrome c and bovine microsomal cytochrome b5 (Figure 6). The intermolecular charge and steric interactions between these two proteins were optimized using a least squares fitting process. Aldiough the hemes of the cytochromes were initially constrained to be coplanar, in the final model they lie at an angle of 15° to each other. The complementary charge pairs in the model are listed in Table I. The closest approach of die TT bonded atoms of the respective prosthetic groups is 8.4 A (iron to iron distance of 18 A) . Evidence supporting this model has come from several laboratories. Ferrocytochrome b5 reduction of ferricytochrome c was shown to be strongly ionic strength dependent (Stoneheumer et al., 1979), suggesting that the reaction is mediated by complementary charge interactions. These studies were performed with microsomal preparations containing cytochrome b5 and NADHxytochrome 35 Figure 6 : The structure of the cytochrome ^cytochrome c complex as proposed by Salemme (1976). The heme atoms are shown in red, the side chains of the negatively charged cytochrome b5 and the positively charged residues of cytochrome c that are proposed to form specific salt linkages are shown in orange and blue respectively. This image was generated with the Raster3D package (D. Bacon, University of Alberta) on a Silicon Graphics IRIS 3130. The coordinates for this structure were provided by Dr. James B. Matthew. 36 Figure 7 : The structure of the edge off cytochrome ^cytochrome c complex (Mauk et al., 1986). The image was generated with the Raster3D package (D. Bacon, University of Alberta) on a Silicon Graphics IRIS 3130. The colour scheme is the same as that used in Figure 6. The coordinates for this structure were provided by Dr. James B. Matthew. 37 Table I : Charged groups diat form specific salt linkages in computer graphic models of the cytochrome ^cytochrome c complexes (Salemme, 1976; Mauk et al., 1986). Edge on model Off edge model cytc cyt b$ cyt c cyt fr5 Lys-13 Glu-48 Lys-72 Glu-48 Lys-27 Glu-44 Lys-79 Glu-11 Lys-72 Asp-60 Lys-86 Glu-56 Lys-79 heme propionate-6 Lys-87 Asp-60 Lys-25 GIu-43* * Although not proposed in die original Salemme model, diis salt linkage has been suggested by the work of Ng et al. (1977), Rodgers et al. (1988) and Holloway & Mansch (1988). bs reductase and low enough concentrations of modified cytochromes c so that the reduction of ferricytochrome c by ferrocytochrome b5 was rate limiting. These audiors also detected die formation of a 1:1 complex between cytochromes b5 and c in vitro by gel permeation and ultracentrifugation. A dissociation constant of 20 (0.01 M Tris-HCl, pH 7.5, 25°C) was measured. Derivatives of cytochrome c, modified at individual lysine residues, were used in the same kinetic system to evaluate die role of specific charge pair interactions in stabilizing the cytochrome /^cytochrome c complex. The same lysines of cytochrome c diat Salemme postulated to form complementary charge pairs with carboxyl groups on cytochrome b5, in addition to a fifth one (cytochrome c Lys-25 to Glu-43 of cytochrome bs), were implicated (Ng et al., 1977; Smith et al., 1980). Using difference absorption spectroscopy, Mauk et al. (1982) were able to detect a cytochrome ^cytochrome c complex with a 1:1 stoichiometry. The association constant of this complex was shown to be strongly dependent on ionic strength, decreasing from 4 ( ± 3 ) x 10 6 M " 1 at I = 0.001 M 38 to 8 ( ± 4 ) x 10 M at I = 0.01 M , providing more evidence of the role of complementary charge pairs in the formation of a complex between these proteins. The optimum pH of interaction was 7.5, which is midway between the p i of the two proteins. Further studies using absorption spectroscopy have expanded and refined the view of the precursor complex. To evaluate the role of cytochrome bs heme propionate-6 that in the interaction with cytochrome c, the stability of the complex formed between cytochrome c and DME-cytochrome b5 was determined under a variety of solution conditions (Mauk et al., 1986). As with the complex formed between the native proteins, the stability of the complex formed with DME-cytochrome b5 was found to be strongly ionic strength dependent. However, the association constant for the DME-cytochrome ^cytochrome c complex [3 ( + 1) x 106 M " 1 (I = 0.005 M , pH 8.0, 25 °C)] exhibited a higher optimal pH (8.0) and was actually slightly greater than that of the native proteins [7 ( ± 3 ) x 10 5 M - 1 (I = 0.005 M , pH 7.5, 25°C)]. Electrostatic calculations and computer graphic modelling confirmed that the Salemme model describes the protein-protein orientation that is electrostatically favoured by the native proteins at neutral pH. However, a second docking geometry, referred to as the off-edge complex, was predicted to be favoured at higher pH and by elimination of the charges of the cytochrome b5 heme propionates (Figure 7). The surface residues of cytochromes b5 and c that participate in the proposed salt bridges of diis second geometry are listed in Table I. In the off edge docking geometry, the distance between the iron of cytochrome bs and the iron of cytochrome c is 26 A . Recently, it has become apparent that the residues of cytochrome b5 that were proposed to participate in specific salt bridges were based on incorrect sequence information (vide supra). The original model proposed a salt bridge between Lys-79 of cytochrome c and residue 13 of cytochrome b5. Residue 13 was originally thought to be glutamate but the sequence of the cytochrome bs gene has since indicated that residue 13 is glutamine (Cristiano & Steggles, 1989). Cytochrome b5 Gln-13 can form a hydrogen bond with cytochrome c Lys-79. Alternatively, Glu-11 of cytochrome b5 is close to residue 13 in the primary sequence. Visual examination of this substitution on a molecular graphics terminal indicates that residue 11 is considerably further away than residue 13 from the 39 proposed interface of die two proteins in die off edge docking geometry, and that an interaction with Lys-79 of cytochrome c would only be possible after considerable side chain rearrangement. A more thorough reassessment of alternate docking orientations is in progress. *H N M R measurements also detect the formation of a 1:1 complex between cytochromes b5 and c (Eley & Moore, 1983). The lower association constant that these workers measured (10 3 M" 1 ) can be reconciled with the higher ionic strength used in the N M R experiment. These data also suggest that the pK of a cytochrome b5 carboxyl group decreases upon complex formation with cytochrome c. With the assignment of the heme resonances of cytochrome b5 (McLachlan et al., 1986), this group was identified as heme propionate-6, which is the group that in die crystal structure is showm to extend into the solvent. This finding is consistent with the participation of cytochrome b5 heme propionate-6 in a salt linkage with Lys-79 of cytochrome c. It is also possible that die pKg of the heme propionate is lowered in the complex by a reduction of the negative electrostatic potential of its environment that would be expected to accompany binding to cytochrome c. Subsequent experiments (Hartshorn et al., 1987) suggest that cytochrome b5 heme propionate-6 is free to bind chromium (ethylenediamine)^3 when cytochrome bs is bound to cytochrome c. This result suggests that heme propionate-6 is not involved in a specific salt linkage with a corresponding group on the surface of cytochrome c or that the additon of the chromium compound induces a different protein-protein docking geometry, in which heme propionate-6 is free to bind exogenous cations. The off edge complex is compatible with the availability of cytochrome b5 heme propionate-6 in this maimer. Using site directed mutagenesis, Rodgers and coworkers (1988) converted each of the acidic residues of cytochrome bs that the Salemme models predicts to be involved in a specific salt linkage with cytochrome c to the corresponding amide. Complex formation between these mutants and cytochrome c was monitored under hyperbaric conditions using electronic absorption difference spectroscopy to determine free energies and specific volume changes associated with complex formation. The linear free energy relation between A G and P A V observed in this study suggested that a major contribution to the volume change accompanying complex formation is provided by 40 these salt linkages. The results of these experiments confirm the modelling studies of Mauk et al. (1986). It would be interesting do a similar study to determine the effect of a wider range of mutations on the stability of and volume changes accompanying die formation of the cytochrome fe5cytochrome c complex. Fourier transform infrared spectroscopy (FTIR) measurements indicate that 3 glutamic acid residues of cytochrome bs are involved in salt bridges with cytochrome c (Holloway & Mansch, 1988). Although the original cytochrome ^cytochrome c model complex (Salemme, 1976) implicated two glutamate residues of cytochrome bs in this role, the FTIR finding is" consistent with the involvement of Glu-43 of cytochrome b5 in the formation of a salt linkage with a surface residue of cytochrome c, as proposed by Ng et al. (1977). Chemical modifications implicate these three glutamic acid residues in the formation of salt bridges widi three other physiological electron transfer partners (Dailey & Strittmatter, 1979; 1980). The participation of Asp-60 and heme propionate-6 of cytochrome bs in complex stabilization was not investigated in the FTIR study. Molecular dynamic simulations have been reported for die complex formed by cytochromes b5 and c (Wendoloski et al., 1987). These calculations took die edge on docking geometry of the complex (Salemme, 1976) as the starting orientations of the cytochromes and suggest that within this one docking conformation, the side chains of the protein residues can rearrange, creating a structure more favourable to electron transfer. Rearrangement of amino acid side chains at the protein-protein interface predicted by these calculations resulted in inter-iron distances that are 1.1 to 2.1 A smaller than in the Salemme model. In addition, they predict that phenylalanine 82 of cytochrome c is reoriented so diat it bridges the two hemes in the complex and actually moves back and forth between the heme groups of the two proteins widiin the complex. Despite the increased attention directed toward characterization of die cytochrome ^cytochrome c complex, several critical issues remain to be elucidated. These issues include (1) the exact identity of which residues are involved in salt linkages, (2) the contribution of non-electrostatic forces to the stability and docking geometry of the complex, (3) whether an alternative 41 docking geometry or microheterogeneity of the docking geometries in the cytochrome ^cytochrome c complex can be detected, and (4) the relevance of alternate docking geometries of the complex to electron transfer. Objectives To characterise better the mechanism of interaction and electron transfer between ferrocytochrome bs and ferricytochrome c, several biophysical studies were undertaken. Because the heme propionates of cytochrome b5 create such a prominent feature in the electrostatic surface of the protein, particular effort was directed at evaluating the role of the heme propionates of cytochrome b5 in these processes. The role of these groups has been assessed dirough the use of a derivative of cytochrome b5 in which the native heme was substituted widi a dimethyl esterified heme (DME-cytochrome bs). The interaction between cytochromes b5 and c was studied by fluorescence quenching titrations of porphyrin c, a fluorescent analogue of cytochrome c, with ferricytochrome b5. These measurements yielded stoichiometry, association constants and maximal quenching values. Forster energy transfer theory (Forster, 1948) was used to estimate the relative orientation and separation of the cytochrome b5 heme and the cytochrome c porphyrin within the complex. These measurements were performed under a variety of solution conditions to evaluate the influence of ionic strength and pH on the stability and die conformation of the complex. The complex that forms between cytochromes b5 and c was studied further by measuring the kinetics of reduction of the components of the complex by various flavin semiquinones. This approach has provided information regarding the rate of electron transfer to and within several other complexes tiiat has been interpreted in terms of prosthetic group accessibility, electrostatic environments near the prosthetic groups and dynamic motions within the protein complex (Bhattacharya et al., 1986, 1987; Hazzard et al., 1986, 1987). In the present study, reduction kinetics were performed with die neutral lumiflavin semiquinone and the negatively charged flavin 42 mononucleotide semiquinone. Experiments were repeated at pH 7 and pH 8 with native and D M E -cytochrome b5 in an attempt to detect any difference between the reactivities off edge and the edge on conformations that may be present under different conditions (Mauk et al., 1986). The final set of experiments involved measuring the rate of ferricytochrome c reduction by native and DME-ferrocytochrome b5. These experiments were performed as a function of ionic strength, p H , and temperature to permit correlation of the kinetic results with the restdts of protein-protein association constant measurements under similar solution conditions. The results of these studies provide a more detailed understanding of the mechanism of cytochrome ^cytochrome c recognition and the mechanism of electron transfer from ferrocytochrome b5 to ferricytochrome c. 43 E X P E R I M E N T A L P R O C E D U R E S General Procedures Glass distilled water purified with a Barnstead NANOpure system to a resistivity of 16-17 Mohms was used in all experimental work. A l l pH measurements were performed widi a Radiometer Model P H M 84 meter equipped with a Radiometer Type G K 2321C combination electrode. Except where noted, reagent grade chemicals were used. Phosphate buffers of specified pH and ionic strength were prepared from monobasic and dibasic sodium salts. Most of the spectrophotometric measurements were performed on a Shimadzu UV-250 recording spectrophotometer equipped with an OPI-2 Optional Program interface, a thermostatted cuvette holder, and a thermostatted ( ± 0 . 1 °C) circulating Brinkmann R M 3 water bath. The electronic absorption spectra were collected on a Cary-219 spectrophotometer interfaced to a Zenith ZW-248-82 microcomputer by On Line Instruments Systems, Jefferson, Georgia, U . S . A . (OLIS). Data acquisition and manipulation were performed with OLIS model 4000 Data System Spectrophotometry version 7.05 software. The tryptic fragment of bovine hepatic cytochrome b5 and its dimethylesterheme-substituted derivative were prepared as described previously (Reid & Mauk, 1982; Reid et al., 1984). Horse heart cytochrome c (Type VI) was purchased from Sigma Chemical Co. and purified by anion exchange chromatography to remove deamidated cytochrome (Brautigan et al., 1978c). Protein solutions were concentrated by ultrafiltration using Amicon-52 stirred cells fitted with Y M - 5 membranes, Centripreps and Centricons (Amicon). Protein concentrations were determined using values of £ 4 1 0 = 106.1 m M - 1 and « 412.5 = 117.0 mM" 1 for ferricytochromes c (Margoliash & Frohwirt, 1959) and b5 (Ozols & Strittmatter, 1964) respectively. 44 Interaction of Porphyrin Cytochrome c with Native and DME-Cytochrome b5 Preparation of Protein. Porphyrin cytochrome c was prepared as described previously (Kornblatt & English, 1986) from horse heart cytochrome c, with the assistance of Prof. Jack A . Kornblatt at the University of Concordia, Montreal. Porphyrin cytochrome c was protected from light as much as possible. The extinction coefficient of porphyrin cytochrome c was determined by both the Bradford assay (Bradford, 1976) and by amino acid analysis. The Bradford assay was performed with a kit purchased from Bio-Rad Laboratories, using horse heart cytochrome c as the standard. The second method entailed determining the amino acid composition of a sample of porphyrin c of known absorbance. The amino acid composition of porphyrin cytochrome c was assumed to be identical to that of horse heart cytochrome c The extinction coefficient of the porphyrin derivative was calculated by averaging the values predicted for each residue. Amino acid analyses were performed on a Durrum Amino Acid Analyzer Model D-500 with the assistance of Dr. Robert L. Cutler. The extinction coefficient at 505 nm and the relative fluorescence emission of porphyrin cytochrome c were measured from pH 6.0 to 8.0 (I = 0.01 M , 25°C). Binding measurements. Fluorometry measurements were made widi an S L M SPF-500C fluorometer interfaced to an IBM-PC A T microcomputer. A l l measurements were performed at 25 °C. Fluorescence samples were excited at 505 nm (1 nm bandpass) and die emission was recorded at 624.5 nm (20 nm bandpass). The absorbance of the sample never exceeded 0.1 at the excitation wavelength. Binding measurements were performed over a pH range of 6 to 8 and an ionic strength range of 2 to 50 m M in sodium phosphate buffer. To an appropriately buffered solution of porphyrin cytochrome c (~2 i*M) was added 15 to 25 10 / i L aliquots of a similarly buffered stock cytochrome b5 solution (~50 /xM). The fluorescence of the sample was recorded after each addition of cytochrome bs. To minimise die exposure of the sample to light, the excitation and emission shutters were closed between measurements, and the room lights were shut off while the sample compartment was open. 45 Inner filter correction and quenching values were calculated according to the formulae derived by Hwang and Greer (1979). Binding data were fitted to equation (6) using a nonlinear regression program (Duggleby, 1984). In this equation, B and C represent the concentrations of cytochrome b5 and porphyrin cytochrome c respectively, Q is the degree of quenching observed, Q m a x is the quenching at infinite cytochrome fc5 concentration, and K , is the binding constant. Qmax 1 1 Q = (C + B + ((C + B + — ) 2 - 4BC)*) [6] 2B K , . K , Energy transfer measurements. Maximal quenching values, obtained by addition of saturating amounts of cytochrome b5 to solutions of porphyrin cytochrome c, were determined under the conditions of p H and ionic strength listed above. The theory of Forster (1948) was applied to calculate the distance of separation between the porphyrin groups in the complexes. Based on this theory, the characteristic distance of separation between the donor fluorophore and the acceptor chromophore at which 50% quenching of the fluorescence signal is observed, R 0 , may be calculated 9000 In 10 re2 $ J Ro = P I 128ir 6 « 4 N according to equation 7, where « 2 is the orientation factor, $ is the quantum yield of the donor, J is the overlap integral, n is the refractive index of the intervening medium, and N is Avogadro's number. The quantum yield and K 2 were taken to be 0.01 and 2/3 respectively (Vanderkooi et al., 1980). As the intervening medium can be approximated by a light oi l , n is given a value of 1.4. The overlap integral was evaluated between 560 nm and 760 nm from the electronic spectra of the fd(v)E a(v) J = / dv [8] v4 two proteins using equation 8, in which fd is the fraction of total donor fluorescence occurring at wavenumber v and E a is the corresponding extinction coefficient of the acceptor. With these values, the separation, R, from the efficiency of energy transfer, Q ^ , may be calculated using equation 9. 46 1 - Q n Qn R = R Q ( ) 1 / 6 [9] Polarization measurements. Fluorescence anisotropy measurements were performed on porphyrin cytochrome c in the presence and absence of cytochrome b5. The measurements were performed with an S L M Aminco SPF-500 polarization accessory at an ionic strength of 2 m M , pH 7.0, 25°C. The equations used in analysing these data are given in appendix A . Kinetics of Flavin Semiquinone Reduction of Cytochrome bS and Cytochrome c The kinetics of flavin semiquinone reduction of the components of the 1:1 complex formed by cytochrome c with either cytochrome b5 or DME-cytochrome bs were performed at the University of Arizona with the collaboration of Profs. Michael A . Cusanovich, Gordon Tollin and Dr. James T. Hazzard (Eltis et al., 1988). Tuna cytochrome c (Type XI) was purchased from Sigma Chenucal Co. and purified by anion exchange chromatography (Brautigan et al., 1978d). Lumiflavin was prepared as described by Guzzo and Tollin (1963) and F M N purchased from Sigma Chemical Co. was purified by size exclusion chromatography on a BioRad P2 column (Nagy et al., 1982). Electron transfer was initiated by a 300 ps pulse of light from a nitrogen laser (PRA LN100) pumped dye solution (10 m M C-450 from P R A , Inc. in ethanol) in a 1 cm path length cuvette. The dye solution has an emission maximum at 450 nm which is close to the absorption maximum at 442 nm of lumiflavin and F M N . A quartz-iodide projection lamp focused through a Bausch & Lomb monochromator onto the sample and then onto an R C A 4463 photomultiplier tube (S-20 response) comprised the monitoring beam. A minimum of four decay curves were averaged on a Nicolet 1170 signal averager (Simondsen & Tollin, 1983). The reaction mixtures, containing 50 /xM flavin and 0.5 m M E D T A in phosphate buffer (total ionic strength of 4 mM), were made up in 3 mL cuvettes, sealed widi rubber stoppers (Thomas Scientific), and bubbled for 30 minutes with argon that had passed through an oxygen scrubbing 47 solution (Stone & Hume, 1939) to remove dissolved oxygen. The appropriate cytochromes were then introduced with gas tight syringes, and argon was passed over the top of the solution for an additional five minutes. The photo-generated flavin triplet state in the reaction mixture reacted with the E D T A to produce the flavin semiquinone (Ahmad et al., 1982). Under these conditions, less than 0.1 i t M of semiquinone was produced. Experiments were conducted at 24°C at pH 7.0 and 8.0. The reduction of horse heart ferricytochrome c by LfH- and F M N H - was studied in the absence and presence of ferricytochrome b5 as a decrease in absorption at 565 nm, an isosbestic point in die cytochrome b5 oxidation-reduction difference spectrum. Cytochrome c was preferentially reduced under these conditions, and no evidence of cytochrome b5 reduction prior to or coincident with cytochrome c reduction was observed. The reduction of D M E - and native ferricytochrome b5 by LfH- and F M N H - was studied in the presence and absence of ferrocytochrome c. In experiments involving die reduction of ferricytochrome bs bound to ferrocytochrome c, the cytochrome c was photoreduced by steady-state illumination of the flavin in the reaction mixture prior to the anaerobic addition of ferricytochrome bs, and the reaction was monitored as a decrease in the absorption at 580 nm. The protein concentrations were sufficiently high (> 5 /*M) to ensure pseudo-first order conditions relative to the concentration of semiquinone generated (vide supra) and to ensure that flavin disproportionation did not compete with protein reduction. Kinetic experiments were performed at 6 protein concentrations by adding successive samples (5 - 10 /xL) of stock cytochrome solutions to the reaction mixture with a gas tight syringe. First order rate constants, k o b s , were derived from semilog plots of the kinetic data. These were linear for at least 4 half lives. Second order rate constants were calculated from linear plots of k o b s versus protein concentration. On the basis of multiple determinations, the error in the second order rate constants was estimated to be ± 10%. The solvent accessible surface areas of the hemes of cytochromes b5 and c, unbound and in each of the proposed cytochrome ^cytochrome c complex docking geometries, were calculated with an algoridim that uses the coordinates of the structure, van der Waal's radii of the atoms and a 48 spherical solvent probe with a radius of 1.4 A (Connolly, 1983). Ferrocytochrome b5 Reduction of Ferricytochrome c Rapid mixing experiments were performed with a stopped flow spectrophotometer based on the design of Gibson and Milnes (1964) as marketed by Dionex (Model D-103). Modifications to the reactant delivery system (Scott, 1980) and syringe plungers were described previously (Reid, 1982; 1984). To achieve greater anaerobicity of the reacting solutions, the entire reactant flow system and the photomultiplier tube were mounted in a Vacuum Atmospheres Co. Model HE-243-2 glove box equipped with a HE-63-P pedatrol, a HE-493 Dri-Train and a Sargent Welch Scientific Co. model 1402 vacuum pump. A similar apparatus has recently been described (Leung et al., 1989). Light from an Osram Xenophot H L X 64250 lamp passed through a J-Y Model DH-10 dual grating monochromator residing outside the glove box was delivered to the stopped flow cuvette by a fibre optic bundle. The drive syringes and stopped flow cuvette were bathed with water thermostatted with a Haake Model T-41 water bath and circulated by a Gorman-Rupp Industries M877 pump. The glove box was filled with prepurified argon and maintained at less than 5 ppm oxygen. Although water was added to the cooling bath immediately prior to the experiment and removed immediately afterward, the use of an open water bath for the thermostatting of the drive syringes necessitated frequent regeneration of the catalyst. The remainder of the stopped flow apparatus was designed by OLIS. The lamp was operated at 3 amps by an OLIS XLI50 power supply. The output from an EMI-Gencome model RFI/QL-30F photomultiplier tube was amplified by an OLIS amplifier equipped with a variable (0.05 - 100 ms) time constant. The data acquisition and instrument control signals were collected and manipulated with OLIS model 3920Z Data System Stopped Flow version 6.12 software run on a Zenith Z-100 microcomputer. The apparatus is shown in Figure 8. The dead time of the stopped flow was determined to be approximately 7 ms. 49 Lamp Monochromator Fibre optic cable CH 4 B Signal amplifier Microcomputer Argon Water pump HE-493 Dri-traln Glove box Reactant flow system Water bath Copper tubing Photomultiplier tube Antechamber Butyl rubber glove Figure 8 : Stopped-flow spectrophotometer mounted in a glove box. (A) The fibre optic cable was introduced into the glovebox through a port in the side of the box. Outside of the glove box, the cable was sealed in a stainless steel bellows. (B) The cables for the photomultiplier tube signal, power and the stopped-flow trigger (labelled 1, 2 and 3 respectively) and the argon gas line for the pneumatic flow actuator plunger were introduced through a modified steel plate mounted on the fourth glove port. (C) The copper pipes for the cooling water were introduced through ports in the top of the glove box. 50 The reduction of ferricytochrome c by ferrocytochrome b5 was performed under second order conditions with the reactants at equimolar concentrations (~1.7 A*M) . Ferricytochrome c was prepared by the addition of an excess of ammonium bis(dipiciolinato)cobaltate(III) (Co(dipic) 2), previously prepared as described in Schenkein et al. (1977). Excess Co(dipic) 2 was removed from ferricytochrome c by passing the sample down a 10 x 0.5 cm column of Sephadex G-25, superfine equilibrated with the buffer in which stopped flow experiment was to be performed. The protein solutions were prepared in glass serum bottles that were sealed with cleaned rubber serum stoppers and secured with copper wire. The solutions were degassed for 30 minutes under a stream of prepurified argon that had been passed through a 20 cm column of Chemalog Catalyst R3-11 and a water bubbler. After 30 minutes, 0.2 equivalents of lumiflavin from a 200 i * M stock (~40 itL) was introduced into the cytochrome bs solution with a gas tight syringe. Ferricytochrome b5 was reduced by steady state illumination of tiiis solution for 30 minutes. The needle punctures in the serum caps were then covered with electrical tape, and the serum bottles were transferred into the glove box. The antechamber was evacuated for a minimum of 1 hour after the introduction of an aerobic atmosphere. The oxidation of ferrocytochrome b5 was monitored at 428 nm, where the total absorbance and the change in absorbance of cytochrome c are small. The reduction of ferrocytochrome c was monitored at 416 nm, an isosbestic point in the electronic absorption spectra of ferri- and ferrocytochrome bs (Figure 9). A minimum of 5 decay curves were averaged at each wavelength. The data were transferred to an IBM-PC A T microcomputer where they were fit to a second order rate equation (10) using MINSQ, a least squares fitting program (MicroMath Scientific Software). In AQ - Aoo A t = [10] 1 + kc 0t this equation, A t is the absorbance at time t, AQ and Aoo are the initial and final absorbances of the protein solution respectively, CQ is the initial concentration of each cytochrome species and k is the second order rate constant. At the lower ionic strengths, the initial portion of the reaction 51 Figure 9 : Simulated difference spectrum of cytochrome b5 and cytochrome c. The curve represents the sum of the molar extinction coefficients of an equimolar mixture of ferrocytochrome b5 and ferricytochrome c minus the sum of the extinction coefficients of an equimolar mixture of ferricytochrome bs and ferrocytochrome c. Electronic absorption spectra of the respective species used in this simulation are presented in Figures 3 and 5. The arrows indicate the wavelengths (416 nm and 428 nm) at which ferrocytochrome b5 reduction of ferricytochrome c was monitored. 52 occurred within the dead time of the instrument. In these cases, die initial concentration of die cytochromes (c 0) was determined by normalising the overall change of absorbance ( A A t o t ) to A A t o ( observed for the reaction performed at high ionic strength. The second order rate constants obtained as described above were compared with rate constants obtained by two alternative methods. In the first of these methods, die reactions were performed under pseudo-first order conditions with ferricytochrome c in 5 to 30 fold molar excess over ferrocytochrome b5 (1 /*M). Greater excess of cytochrome c could not be used in these measurements owing to the high background absorbance inherent at such cytochrome concentrations and to the limited time resolution of the stopped flow technique. In die second alternative method, the second order rate constants were measured under conditions of equimolar reactant concentrations but generating ferrocytochrome b5 by anaerobic reduction with sodium dithionite rather than with lumiflavin. Several grains of dithionite were added to a sample (~10 itL) of cytochrome b5. The dithionite was removed by passing the sample over a small Sephadex G-25 column in the glove box. Ferricytochrome bs was collected in a serum bottle containing 12 mL of anaerobic buffer. The concentration of ferricytochrome b5 was determined spectrophotometrically by transferring 1 mL of this solution to an anaerobic cuvette. Both control kinetic experiments were performed at I = 0.5 M , pH 7.0, 25°C. Ferricytochrome c reduction of ferrocytochrome b5 was studied over a pH range of 6.0 to 8.0, an ionic strength range of 0.1 to 1.0 M , and a temperature range of 4°C to 32°C. Thermodynamic activation parameters of the reaction were calculated from an Eyring plot of the temperature dependence data. The ionic strength dependence was analysed by the method of Van Leeuwen (1983). In this analysis, Zj and are the monopole charges of die proteins, and P[ and P 2 are the components of the dipoles of the proteins through their respective exposed heme edges. In cytochrome b5 the heme edge was taken to be the 7 carbon of die methylene bridge between the heme pyrroles with the propionate substituents and in cytochrome c the heme edge was taken to be the methyl carbon of the most solvent exposed heme pyrrole. Rj and R 2 are the radii of die two 53 q 2 ln(k) = ln(koo) - ( Z , ^ + ZP(1 + §R) + PP(1 + §R) 2) f(§) [11] 411« 0 e k B T R Z X P 2 + Z 2 P 1 ZP = [11a] qR P , P 2 P P = [ l i b ] (qR) 2 R = R , + R 2 [11c] l - 2 e 2 $ R 2 f ( § ) = [Hd] 2§R 2 (1 + §R,) § = 0.329(1)* A ' 1 (at 294 K) [1 le] proteins, q is the elementary charge, e 0 is the dielectric permittivity, e is the dielectric constant of water, and k B is Boltzmann's constant. The monopoles and dipoles of the proteins were calculated on a Silicon Graphics IRIS 3130 computer using Brookhaven data bank coordinates of ferric cytochromes b5 and c with a F O R T R A N program (Northrup et al., 1986) generously provided to Prof. Gary D . Brayer by Prof. Scott H . Northrup. This program used a C H A R M M library to assign a charge and mass to each non hydrogen atom in the protein structure. The C H A R M M library contains the charge and mass distribution for a ferrous heme only. Dipoles for the oxidised and reduced cytochromes were calculated from the same structures, but the hemes of the oxidised proteins were assigned an additional positive charge, 75% of which was assigned to the iron and the remaining 25% was distributed amongst the pyrrole nitrogens. The dipole of porphyrin cytochrome c was calculated from a structure generated by removing the ferrous atom from the ferrocytochrome c structure and redistributing the charge density of the iron (0.24), and the mass of two hydrogen atoms, over the pyrrole nitrogens. Finally, the dipole of DME-cytochrome bs was calculated from a structure generated by placing a methyl group on the least sterically hindered oxygen of each of the heme propionates of the cytochrome bs structure. A partial charge was assigned to the various atoms based on the dipoles of the ester bonds. 54 R E S U L T S Interaction of Porphyrin Cytochrome c with Native and DME-Cytochrome b5 The extinction coefficient of porphyrin c at 505 nm was determined to be 13.1 (±0 .2 ) mM" 1 (I = 0.01 M , p H 7.0, 25°C) by the Bradford assay, and 12.9 (±0 .6 ) m M " 1 (I = 0.01 M , pH 7.0, 25°C) by amino acid analysis. The value of 13.1 mM" 1 was used in subsequent analysis. The extinction coefficient at 505 nm was found to vary from 12.6 mM" 1 at pH 6 to 14.0 m M " 1 at pH 8.0 (Table II). At pH 7, the absorption spectrum of porphyrin cytochrome c was found to be essentially invariant with ionic strength. The appropriate value was used to determine the concentration of porphyrin cytochrome c in the analysis of the binding curves determined at various values of pH. The emission spectrum of die porphyrin cytochrome c synthesized and used in this study is shown in Figure 10. The pH dependence of the fluorescence emission spectrum is depicted in Figure 11. The fluorescence intensity of porphyrin cytochrome c is considerably greater at pH 6.0 than at pH 7.0. The fluorescence anistropy of porphyrin cytochrome c was determined to be 0.048, while in Table II : The extinction coefficients of porphyrin cytochrome c at 505 nm at the pH's used in this work. pH « (mM" 1) 6.0 12.6 ± 0 . 6 6.5 12.7 ±0 .5 7.0 13.1 ± 0 . 5 7.5 13.5 ±0 .5 8.0 14.0 ± 0 . 7 55 450 525 600 675 750 Wavelength (nm) Figure 10 : The visible absorption and uncorrected fluorescence emission spectra of porphyrin cytochrome c (1 = 0.01 M sodium phosphate, pH 6.0, 25 °C). The sample was excited at 505 nm. The fluorometer settings of the excitation and emission bandpasses were 1 and 20 nm respectively. 56 540 600 660 720 7 8 0 Wavelength (nm) Figure 11 : The pH dependence of the fluorescence intensity of porphyrin cytochrome c in the absence (panel A) and presence (panel B) of cytochrome b5 (I = 0.002 M , 25°C). In panel A , the concentration of porphyrin cytochrome c was 5.0 i t M . In panel B, the concentrations of porphyrin cytochrome c and cytochrome b5 were 0.74 nM and 3.6 (iM respectively. The emission spectra are not corrected for instrument response or the inner filter effect. 57 the presence of saturating amounts of cytochrome b5 it was 0.084. Figure 12 depicts a fluorescence quenching titration curve of porphyrin cytochrome c by D M E -cytochrome bs. The curve was generated by normalizing the data to a constant concentration of porphyrin cytochrome c (2 nM). In general, the data were not normalized in tiiis way, but were analyzed by equation 5. In all cases, quenching was completely reversed at the end of a titration by the addition of 100 /zL of 3.3 M sodium chloride. The association constant for the formation of the cytochrome 6 5porphyrin cytochrome c was strongly dependent on ionic strength, decreasing from 3.3 (±0 .4 ) x 10 6 M " 1 (I = 0.002 M , pH 7.0, 25°C) to 1.4 (±0 .5 ) x 10 5 M " 1 (I = 0.05 M , pH 7.0, 25°C)(Figure 13). Under the same conditions, the association constant for the formation of the DME-cytochrome ^porphyr in cytochrome c complex [1.2 (±0 .3 ) x 10 6 M " 1 (I = 0.002 M , pH 7.0, 25°C)] was slightly lower than that of native cytochrome b5 and porphyrin cytochrome c. The association constant of the DME-cytochrome & 5porphyrin cytochrome c complex was too small to measure by this technique at ionic strengths of 0.05 M and greater. The curves in Figure 13 represent fits of the data to equation 11. As the least squares fitting routine (MINSQ) failed to converge after 100 iterations when all seven parameters in the equation were allowed to vary simultaneously, two different fits were performed. In die first, the monopoles and dipoles, as calculated from the crystal structures (Table IX) were held constant and in the second, die radii and the dipoles were held constant. The results of both fits are shown in Table III. The calculated association constants at infinite ionic strength were 6000 M " 1 for the cytochrome ^porphyr in cytochrome c complex and 50 M " 1 for the DME-cytochrome ^porphyr in cytochrome c complex. The pH dependence of the association constant for the cytochrome ^porphyr in cytochrome c complex varied with the availability of the cytochrome b5 heme propionates (Figure 14). The association constant of the complex formed between DME-cytochrome b$ and porphyrin cytochrome c was essentially independent of pH (I = 0.01 M , 25 °C). The association constant for the complex formed by native cytochrome b5 and porphyrin cytochrome c at the extreme values of pH studied 58 Figure 12 : Fluorescence titration curve of porphyrin cytochrome c with cytochrome bs. This experiment was performed with cytochrome b5 (I = 0.002 M , pH 7.0, 25 °C). The curve represents a fit of the data to equation 6. The value of Q„ax represents an extrapolated value. 59 0.00 0.08 0.16 0.24 Figure 13 : The ionic strength dependence of the association constant of DME (•) and native (•) cytochrome b5 and porphyrin cytochrome c (pH 7.0, 25 °C). The lines represent fits to die Van Leeuwen equation (11), using the parameters given in Table III. 60 1 I 1 I I 1 1 1 1 1 1 6.0 7.0 8.0 PH Figure 14 : The p H dependence of the association constant of porphyrin cytochrome c and either native ( • ) or DME-cytochrome b5 ( • ) (I = 0.01 M , 25°C). 61 was about half the magnitude that it was over the pH 6.5 to 7.5 range. The maximal quenching, Q m a x , of porphyrin cytochrome c by native cytochrome bs, was greatest at pH 6.0 (0.38 (±0.01)) , and decreases to a constant value at pH 7.0. The Q m a x of porphyrin cytochrome c in the presence of DME-cytochrome b5 was independent of pH at an ionic strength of 0.01 M (Figure 15). In contrast, the Q , ^ of porphyrin cytochrome c in die presence of native cytochrome bs was insensitive to ionic strength at pH 7, while in the presence of D M E -cytochrome bs, the Q m a x of porphyrin cytochrome c increased with ionic strengdi, plateauing at 0.01 M (Figure 16). The value for the overlap integral (equation 8, the units of wavenumber were cm"1) used in calculating the separation of the donor and the acceptor sites from Q m a x was 2.65 x 10" 1 4. The value of the characteristic distance of separation (equation 7) was determined to be 15.3 A . Table III : Parameters used in fitting the ionic strength dependence of the association constants of native and DME-cytochrome b5 widi porphyrin cytochrome c to the Van Leeuwen equation (11). parameter cyt ^porphyr in cyt c DME-cyt ^porphyr in cyt c L n Koo 7.7 (1.5) 8.7(1.6) 2(100) 4(20) R i 8(20) 17(f) 4(100) 17(f) R 2 9 (20) 17(f) 8(25) 17(f) Z l - 9(f) - 150(150) - 7(f) -300 (100) Z2 6.2(0 .6(15) 6.2 (f) .3 (10) P i - 468 (f) - 468 (f) - 391 (f) - 391 (0 p 2 255 (f) 255 (f) 255(f) 255 (f) s.s. .06 .05 2 1.4 S.S. refers to the sum of squares of the fit. The numbers in parandieses are errors; an "f" denotes that the parameter was held constant. 62 Figure 15 : The ionic strength dependence of maximal quenching ( Q m a x ) of porphyrin cytochrome c by DME-cytochrome b5 (pH 7.0, T = 25°C). The quenching of porphyrin cytochrome c by native cytochrome b5 showed no ionic strength dependence. 63 Figure 16 : The pH dependence of maximal quenching ( Q ^ ) of porphyrin cytochrome c by native ( • ) and DME-cytochrome bs ( • ) (I = 0.01 M , 25°C). The lines are drawn for visual aid and have no theoretical significance. 64 Kinetics of Flavin Semiquinone Reduction of Cytochrome bs and Cytochrome c. Pseudo-first order rate constants for ferricytochrome c reduction by LfH- at pH 7 and 8 as a function of protein concentration in the presence and absence of stoichiometric amounts of ferricytochrome b5 are shown in Figure 17. The second order rate constants obtained from these plots are given in Table IV. At pH 7, the rate constant for ferricytochrome c reduction was unaffected by the presence of native ferricytochrome b5, while it was decreased by 35% in the complex formed with DME-ferricytochrome by Increasing the pH to 8 produced two types of kinetic effect. First, the rate constant of ferricytochrome c reduction by L fH ' increased substantially. Second, binding to native and DME-ferricytochrome b5 decreased the rate constant for ferricytochrome c reduction by LfH- by 60% and 24% respectively. The dependencies of the rate constants (k0bs) for horse heart and tuna ferricytochrome c reduction by F M N H - on cytochrome c concentration in the presence and absence of stoichiometric amounts of native or DME-ferricytochrome b5 are shown in Figure 18. The second order rate constants derived from these plots are listed in Table IV. Binding of tuna or horse heart ferricytochrome c to native or DME-ferricytochrome b5 significantly decreases the rate constant of cytochrome c reduction by F M N H - at pH 7 (by 73% and 60% respectively). The variations in observed rate constants for native and DME-ferricytochrome bs reduction by L f H - in the presence and absence of stoichiometric amounts of ferrocytochrome c (pH 7 and 8) are shown in Figure 19. The corresponding second order rate constants are listed in Table V . Under all conditions, the rate constant of reduction of native ferricytochrome b5 is lower than that of DME-cytochrome b5. With an increase in pH from 7 to 8, the rate constant for native ferricytochrome b5 reduction by LfH- is unchanged while the rate constant for DME-ferricytochrome b5 is increased to an extent (~1.7 fold) similar to that observed for LfH- reduction of cytochrome c The rate constants for reduction of both ferricytochromes b5 increase in the presence of stoichiometric amounts of ferrocytochrome c. For native ferricytochrome b5, the increase in the reduction rate constant is more pronounced at pH 8, while for DME-ferricytochrome bs, the increase 65 [Cytochrome c] Figure 17 : Plots of k o b s versus cytochrome c concentration for the reduction of horse cytochrome c by LfH- at pH 7 (A) and 8 (B) in the absence ( • ) and the presence of native ( • ) or DME-cytochrome b5 (•). 66 [Cytochrome c] (pM) Figure 18 : Plots of k o b s versus cytochrome c concentration for the reduction of horse (A) and tuna (B) cytochromes c by F M N H - in the absence ( • ) and the presence of native ( • ) or DME-cytochrome 67 Table IV : Second order rate constants for flavin semiquinone reduction of cytochrome c in the presence and absence of cytochrome b 5 . Reduction by LfH-protein k 2 (x lO ' 7 M ' V 1 ) pH 7 pH 8 horse cytochrome c 6.6 11.0 + cytochrome b 5 7.0 6.7 + DME-cytochrome b 5 4.3 8.4 Reduction by F M N H - at pH 7 k 2 (xlO" 7 M ' ! 8 " 1 ) horse cytochrome c free 14.7 + cytochrome b 5 4.4 + DME-cytochrome b 5 5.6 tuna cytochrome c free 12.5 + cytochrome b5 3.2 + DME-cytochrome b 5 5.3 As noted in the Methods, die standard error in these rate constants is ± 10 % 68 2000 [Cytochrome b 5 ] (pM) Figure 19 : Dependence of the rate constant of cytochrome b5 reduction by LfH- on the concentration of cytochrome b5 at pH 7 (A) and 8 (B). Rates were determined for native cytochrome bs in the absence (O) and presence ( • ) of horse heart cytochrome c and for D M E -cytochrome b5 in the absence ( • ) and presence ( • ) of horse heart cytochrome c. 69 Table V : Second order rate constants for reduction of cytochrome b 5 by LfH- in the presence and absence of ferrocytochrome c. protein k 2 (xlO 7 M-V 1 ) p H 7 p H 8 cytochrome b 5 free 1.3 1.3 + horse cytochrome c 1.5 2.1 DME-cytochrome b 5 free 1.8 3.0 + horse cytochrome c 2.5 4.2 As noted in the Methods, the standard error of these rate constants is ± 10%. on binding to ferrocytochrome c was significant at both values of pH. The kinetic results for the reduction of native and DME-ferricytochrome b5 by F M N H in die presence and absence of tuna and horse ferrocytochromes c are presented in Figure 20. For both forms of free cytochrome b$, reduction by F M N H - was slower than semiquinone disproportionation, so cytochrome reduction proceeded primarily via the fully reduced species, F M N H " . The calculated rate constants for tiiis reduction are given in Table V I . The rate constant of F M N H " reduction of free native ferricytochrome b5 was only slightly smaller than that obtained for die free D M E -cytochrome b5, as was also the case for LfH- reduction (cf. Table V) . Large effects on the reduction kinetics of native and DME-ferricytochrome b5 by reduced F M N occur upon complexation with either horse or tuna cytochrome c. In all cases, reduction of the cytochrome b5 heme within the complex occurs via the semiquinone species, F M N H - , in marked contrast to die free cytochromes b5. Contrary to the data obtained for cytochrome c reduction in the complex (cf. Table IV), plots of k ^ versus complex concentration for cytochrome bs are not linear. These kinetics were fit to 70 800 [Cytochrome b 5 ] (uM) Figure 20 : Variation of the rate constants of cytochrome b5 reduction by F M N H - with the concentration of cytochrome b5 in the presence of horse heart (A) and tuna (B) cytochrome c In panel A , the rates of unbound native cytochrome b5 reduction (O) are shown, and in panel B, the rates for unbound DME-cytochrome b5 (0) are shown. In both panels, the effects of binding the respective cytochromes c to native ( • ) and DME-cytochrome b5 ( • ) are shown. The solid lines represent the theoretical fits to the data described in the Discussion. 71 Table V I : Second order rate constants for the reduction of cytochrome b5 by reduced F M N at pH 7. protein k j t x l O ^ M - V 1 ) k, (s-1) cytochrome b5 free + horse cytochrome c + tuna cytochrome c DME-cytochrome b5 free + horse cytochrome c + tuna cytochrome c 0.14 (fully reduced) 2.7 (semiquinone) 5.9 (semiquinone) 0.22 (fully reduced) 6.4 (semiquinone) 8.1 (semiquinone) The oxidation state of the reductant is indicated in paranthesis. standard error in these rate constants is ± 10%. 460 1040 840 1390 As noted in the Methods, the first order rate limiting processes, the mechanisms of which are presented in the discussion. 72 Table VII : Solvent exposure of cytochrome heme groups in the two predicted docking geometries for the cytochrome ^cytochrome c complex. docking geometry cytochrome c (A 2 ) cytochrome b5 (A 2 ) unbound heme 514 485 unbound proteins 59.0(11.5) 161.0(33.2) edge on (Fig. 6) 63.2(12.3) 166.6(34.4) off edge (Fig. 7) 38.2(7.4) 161.0(33.2) The numbers in paratheses represent die percentage of the heme surface area that is solvent accessible. Ferrocytochrome b5 Reduction of Ferricytochrome c Examples of decay curves (and fits to equation 10) observed at 416 nm and 428 nm upon the mixing of ferrocytochrome b5 and ferricytochrome c are presented in Figure 21. The bimolecular rate constants determined at the two wavelengths agreed within 10%. At I = 0.5 M (pH 7.0, 25°C) the second order rate constant for ferrocytochrome b5 reduction of ferricytochrome c under conditions of equimolar concentration was determined to be 18 ( ± 2 ) x 10 6 M " V when lumiflavin was used to generate ferrocytochrome b5. When dithionite was used to produce ferrocytochrome b5, the second order rate constant for ferrocytochrome b5 reduction of ferricytochrome c was 18.5 (±0 .9 ) x 10 6 M ' V 1 . From a linear plot of k o b s versus the concentration of ferricytochrome c, the mixing of ferrocytochrome bs and ferricytochrome c under pseudo-first order conditions yielded a second order rate constant of 20 ( ± 2 ) x 10 6 M ' V 1 . The reduction of ferricytochrome c by ferrocytochrome b5 showed a marked dependence on ionic strength (Figure 22). The second order rate constant of DME-ferrocytochrome b5 reduction of 73 - 0 . 2 0 I , _, , I 0.0 0.8 1.6 t ime (s) Figure 21 : The change in absorbance at 428 run (A) and 416 nm (B) associated with ferrocytochrome b5 reduction of ferricytochrome c (I = 0.5 M sodium phosphate, pH 7.0, 25°C). Each data set represents an average of 5 decay curves. The smooth lines represent second order fits of the data (equation 10). 74 21 19 C M T 17 15 0.0 0.4 0.8 1.2 ionic s t rength (M) Figure 22 : The ionic strength dependence of the bimolecular rate constant of die reduction of ferricytochrome c by native ( • ) and DME-ferrocytochrome b5 ( • ) (pH 7.0, 25.0°C). The solid lines represent non-linear least squares fits to equation 11. The error in the natural log of the second order rate constants is less than the size of the data symbols. 75 ferricytochrome c was lower than that of native ferrocytochrome b5 reduction of ferricytochrome c except at an ionic strength of 1.0 M (pH 7.0, 25°C). The curves in Figure 22 represent fits of the data to the Van Leeuwen equation (11) using the parameters summarised in Table IX. As with the ionic strength dependence of the association constants of cytochrome b5 and porphyrin cytochrome c, the fits were done in two ways. The results of these fits are presented in Table VIII. At an ionic strength of 0.5 M , 25°C, the second order rate constants for ferricytochrome c reduction by either D M E - or native ferrocytochrome bs were independent of pH (Figure 23). Attempts to measure this pH dependence at 0.1 M ionic strength under conditions of equimolar cytochrome concentrations were unsuccessful because the reaction was too fast to monitor. Table VIII : Parameters used in fitting the ionic strength dependence of die second order rate constant of ferrocytochrome b5 reduction of ferricytochrome c to the Van Leeuwen equation (equation 11). parameter cyt & 5 cyt c DME-cyt & 5 cyt c In koo 13.7 (.1) 13.5 (.4) 14.3 (3) 15.1 (.2) R i 8(5) 17(f) 3(13) 17(f) R 2 9(5) 17(0 15 (10) 17(f) Z i - 10(f) - 280 (50) - 8(f) - 100(20) Z2 7.2(f) •2 (.3) 7.2 (0 1.5 (.5) P i - 343 (f) - 343 (f) - 187 (f) - 187 (f) p 2 246 (f) 246 (f) 246 (f) 246 (f) s.s. .01 .01 .06 .01 The numbers in parantheses are errors; an "f" denotes that the parameter was held constant. S.S. refers to the sum of squares of the fit. 76 CD I O x CM 24 18 -12 -6 0 6.0 7.0 PH 8.0 Figure 23 : The pH dependence of the second order rate constant of ferricytochrome c reduction by native ( • ) and DME-ferrocytochrome bs ( • ) (I = 0.5 M , 25°C) . The solid lines are drawn for visual aid and have no theoretical significance. 77 3.3 3.4 3.5 3.6 1/T ( K ~ 1 x 10 3 ) Figure 24 : Eyring plots of the bimolecular rate constant of ferricytochrome c reduction by native ( • ) and D M E ( • ) ferrocytochrome b5 (I = 0.5 M , pH 7.0). The straight lines represent weighted linear least squares fits to the data. 78 The Eyring plots of the temperature dependence of ferricytochrome c reduction by native and DME-ferricytochrome b5 (I = 0.5 M , pH 7.0) are shown in Figure 24. The thermodynamic activation parameters for ferricytochrome c reduction by native ferrocytochrome bs were determined to be A H t = 7.5 ( ± 0 . 2 ) kcal/mol and A S t = -0.3 (±0 .6 ) e.u.. For ferricytochrome c reduction by D M E -ferrocytochrome b5, values of A H J = 7.9 (±0 .4 ) kcal/mol and A S f = 1 ( ± 1 ) e.u. were calculated under the same solution conditions. The results of the dipole calculations of the cytochromes and their derivatives used in this work are listed in Table IX. Table IX. The calculated dipoles and monopoles of the proteins used in this work. oxidation protein state monopole dipole 8 P cytochrome b5 ferrous - 10 - 501 47 - 343 ferric - 9 - 468 51 - 295 DME-cytochrome bs ferrous - 8 - 412 63 - 187 ferric - 7 - 391 69 - 139 cytochrome c ferrous + 6.2 + 255 24 + 232 ferric + 7.2 + 275 26 + 246 porphyrin cyt c n/a + 6.2 + 255 26 + 229 is the angle between the dipole and the vector from the centre of mass of the protein the heme edge. 79 DISCUSSION The three different techniques that have been used to characterise the interaction between cytochromes bs and c are discussed independendy of one another in the following sections. Some facets of this discussion overlap, such as those sections dealing with the calculations of the electrostatic dipoles of the proteins and the solvent accessibility of the heme groups. These aspects are discussed when they are first referred to. Interaction of Porphyrin Cytochrome c with Native and DME-Cytochrome b5 The absorption spectrum of porphyrin cytochrome c exhibits five maxima in the visible region, including the Soret (Figure 10) which is characteristic of porphyrins in which two of the pyrrole nitrogens are protonated (Falk, 1964). Some discrepancies exist in the literature regarding the extinction coefficient of porphyrin cytochrome c. The values that have been measured at 505 nm and the method of determination are summarised in Table X . The current results agree best with the values determined by Flatmark and Robinson (1968). Based on this result, a value of 13.1 mM" 1 (I = 0.01 M , pH 7.0, 25 °C) was used for the extinction coefficient of porphyrin cytochrome c at 505 nm in the current work. The fluorescence emission spectrum of porphyrin cytochrome c (Figure 10) corresponds closely with that observed by previous workers (Vanderkooi & Erecinska, 1975; Strottman et al., 1984; Kornblatt & Laberge, 1988). The increase in the fluorescence intensity of porphyrin cytochrome c with acidic pH (Figure 11) agrees with the reported pH dependence of the fluorescence intensity of this protein (Kornblatt & Laberge, 1988). These workers have measured an apparent pK of 6.2 associated with the change in intensity and have attributed this increase in fluorescence intensity to the protonation of His-18 of porphyrin cytochrome c. In native cytochrome c, His-18 provides one of the axial ligands of the iron atom. The conclusion that the increase in fluorescence intensity 80 Table X : The extinction coefficient of porphyrin cytochrome c at 505 nm. e (mM*1) method of determination reference 13.0* amino acid analysis Flatmark & Robinson, 1968 7.4 b dye binding (Lowry assay) Vanderkooi et al., 1976 13.1 ± 0 . 2 C dye binding (Bradford assay) current work 12.9 ± 0 . 6 C amino acid analysis current work 1 0.04 M phosphate, pH 6.9; b 0.01 M phosphate, pH 7.2; c I = 0.01 M phosphate, pH 7.0 involves a residue that is in a different environment in porphyrin cytochrome c as compared to its state in the native protein is consistent with the observation that no residues in native cytochrome c titrate between pH 6.0 and 8.0 (Matthew, 1978). Several plausible mechanisms could account for the observed quenching of porphyrin cytochrome c by cytochrome bs, including dipole-dipole energy transfer and electron transfer. In addition, heavy atoms can quench fluorescence by spin orbital coupling and paramagnetic centres can quench fluorescence by promoting intersystem crossing (McGlynn et al., 1964). The authors of several studies involving the interaction of porphyrin or metal substituted cytochromes with native cytochromes conclude diat this quenching is due to dipole-dipole resonance energy transfer (Leonard & Yonetani, 1974; Vanderkooi et al., 1977, 1980; McLendon et al., 1985; Vitello & Erman, 1987). In the most closely related system to the current work, Vanderkooi et al. (1980) found that consistent with a resonance energy transfer mechanism of fluorescence quenching, the fluorescence lifetime of porphyrin cytochrome c is decreased in the presence of cytochrome b2. The heme binding domain of cytochrome b2 is structurally homologous to the tryptic fragment of cytochrome bs (Le & Lederer, 1983). In the present work, the spectral overlap between the emission of porphyrin cytochrome c and the absorption of cytochrome b5 are favourable for energy transfer to occur. Quenching of porphyrin cytochrome c fluorescence by cytochrome bs is inhibited by conditions that are expected to dissociate the complex (high ionic strength). In addition there is no evidence that cytochrome b5 81 was reduced during die course of die experiment, ruling out electron transfer as a quenching mechanism. It is concluded that dipole-dipole (Forster) energy transfer is the mechanism of quenching. Nevertheless, it would be useful to perform fluorescence lifetime measurements of porphyrin cytochrome in the presence and absence of cytochrome bs. Previous reports have established that the complex formed between cytochrome b5 and cytochrome c has a 1:1 stoichiometry (Stoneheurner et al., 1979; Mauk et al., 1982; Eley & Moore, 1983). Two lines of evidence in the current work indicate that this is also true for die cytochrome ^porphyr in cytochrome c complex. First, at the lower ionic stengths, diere is a distinct break point in the titration curves at a 1:1 stoichiometry. Second, the results of die polarization study indicate that the complex has 2.3 times the volume of free porphyrin cytochrome c. Calculating the volume of cytochrome b5 from its crystal structure (Mathews et al., 1979), which indicates that the molecule is a cylinder measuring 37 A in height and 31 A in diameter, estimates the volume of this protein at 2.8 x 10 4 A 3 . The calculated volume of the complex is 4.6 x 10 4 A 3 , which is 2.5 times that of porphryin cytochrome c. The agreement of these values argues that the stoichiometry of the cytochrome ^porphyr in cytochrome c complex is 1:1. While it is true that neither of these proteins nor the complex are perfect spheres, an assumption made in calculating the rotation correlation time from the volume of the protein, the error introduced by this assumption does not affect the conclusion. The Q m a x of porphyrin cytochrome c in the presence of native cytochrome b5 is 0.29 (I = 0.01 M , pH 7.0). Using the values for the constants given in the Results section, calculations according to Forster theory indicate that a Q m a x of 0.29 corresponds to an average separation of 18 A between the donor and acceptor dipoles of the cytochrome ^porphyr in cytochrome c complex. Although the locations of these dipoles have not been unambiguously assigned, they are generally taken to be the centres of the porphyrin groups (Vanderkooi et al., 1977). This result indicates that the separation between the centres of the porphyrin and the heme prosthetic groups in the native cytochrome ^porphyr in cytochrome c complex is 18 A (I = 0.01 M , pH 7.0). This result agrees remarkably well 82 with the intersite separation of the donor and acceptor in the cytochrome & 5 zinc cytochrome c complex which was calculated to be 18 A also (McLendon et al., 1985). The complex formed by DME-cytochrome bs and porphyrin cytochrome c yields a Q m a x of 0.18 (I = 0.01 M , pH 7.0, 25 °C). Using Forster energy transfer theory as described above, this value corresponds to a centre to centre intersite separation of 20 A . This result suggests that the average intersite separation is different in the native cytochrome ^porphyr in cytochrome c complex than in the DME-cytochrome ^porphyr in cytochrome c complex, and therefore that the average docking geometry of the two complexes is different. The largest source of error in Forster energy transfer calculations is introduced by estimating the relative orientation of the donor and acceptor dipoles. The value of K2 (2/3) used in the current study is often used when the relative orientation between the donor and acceptor dipoles is unknown, or when this orientation is randomised by rotational diffusion prior to energy transfer (Lakowicz, 1983). Alternatively, a value of 0.476 may be used when the range of donor-acceptor orientations does not change during die lifetime of the excited state (Steinberg, 1971). The difference in K2 of 0.667 and 0.476 introduces an error of 5% in the calculation of R Q . Modelling studies of the cytochrome c peroxidase-cytochrome c complex suggest that these molecules rotate with respect to one another while they are close to one another (Northrup et al., 1988). However, the current fluorescence polarization measurements suggest that cytochrome b5 and porphyrin cytochrome c do not rotate significantly with respect to one another within the cytochrome ^porphyr in cytochrome c complex. A discussion of the relative orientations of the donor and acceptor dipoles of two hemes suggests that the indeterminancy of K2 would not cause more than 20% error in R c (Vanderkooi et al., 1980). In general, the indeterminancy of K2 has not resulted in serious errors in the calculation of distances (Lakowicz, 1983). Finally, while the absolute values of the separation between the donor and acceptor sites in the cytochrome ^porphyr in cytochrome c might not be accurate, the variation of Q m a x indicates that esterification of cytochrome b5 heme propionates induces a change in the average docking geometry of the cytochrome ^porphyr in cytochrome c complex. 83 The apparent increase in the Q m a x of the native cytochrome i 5 •porphyrin cytochrome c complex below pH 7.0 (Figure 16) can be attributed to the increase in the fluorescence yield of free porphyrin cytochrome c at acidic pH. The lack of a pH dependence of the fluorescence of porphyrin cytochrome c in the presence of cytochrome b5 could be due to either a kinetic or thermodynamic effect. Cytochrome b5 could shield His-18 of porphyrin cytochrome c from die solvent, preventing the protonation of His-18. This mechanism is analogous to the inhibition of cyanide binding of cytochrome c by cytochrome c peroxidase (Hoth & Erman, 1984). Alternatively, cytochrome b5 might shift the pK of His-18. Considering that the heme of cytochrome c is sterically accessible to small, exogeneous reductants (vide infra), it seems likely that the principal effect of cytochrome bs is to shift the pK of His-18. A pH titration of His-18 measured by ! H N M R would be interesting. A similar effect on the pH dependence of the fluorescence of porphyrin cytochrome c has been observed in the presence of cytochrome c oxidase (Kornblatt & Laberge, 1988). This effect is not seen for DME-cytochrome bs and porphyrin cytochrome c, presumably because DME-cytochrome b5 does not shift the pK of porphyrin cytochrome c His-18. This result can be taken as further evidence that the modes of interaction between native or DME-cytochrome b5 and porphyrin cytochrome c are not equivalent. The ionic strength dependence of Q m a x for native cytochrome ^porphyr in cytochrome c complex (pH 7.0, 25°C) (Figure 15) indicates that the average docking geometry of this complex does not vary appreciably under these conditions. On the other hand, the Q m a x of the DME-cytochrome ^porphyr in cytochrome c complex varies between 0.11 and 0.18 over an ionic strength range of 0.002 M to 0.02 M (pH 7.0, 25 °C). These values of Q , ^ correspond to intersite separations of 20 to 22 A . Changes in Q m a x could arise from factors other than changes in the average separation of the porphyrin donor and die heme acceptor dipoles. If electrostatic forces predominate in stabilising a particular docking geometry (vide infra), it is easy to envisage how increasing the ionic strength lowers the energy barrier of the proteins to rotate relative to one another. Rotation in this manner would alter Q m a x through changing /c 2. Nevertheless, the variation of Q m a x with ionic strength 84 indicates that the average docking conformation of the DME-cytochrome i 5-porphyrin cytochrome c complex changes in some manner with ionic strength. This is an interesting observation in light of a recent report in which the first order rate constant of electron transfer within the cytochrome c peroxidase-cytochrome c complex has been shown to be dependent on ionic strengdi (Hazzard et al., 1988). Why the Qmax of the native cytochrome ^porphyr in cytochrome c complex is independent of ionic strength is not clear. The ionic strength dependence of the association constants of the cytochrome ^porphyr in cytochrome c complex (Figure 13) suggests that electrostatic forces play a large role in mediating the interaction of these proteins. The fit of the data to the Van Leeuwen equation (Table III) is much better for the interaction of porphyrin cytochrome c and native cytochrome b5 than porphyrin cytochrome c and DME-cytochrome b5. This could be due in part to the small number of data points collected for the interaction between DME-cytochrome bs and porphyrin cytochrome c. The dipole calculations suggest that esterification of the cytochrome b5 heme propionates results in a substantial change in the electrostatic dipole of the protein (Table IX). For instance, the dipole vector of ferrocytochrome b5 passes through the surface of the molecule between Asn-57 and die heme group. Esterification of the propionates not only diminishes the magnitude of the vector by 20%, but also shifts the vector away from exposed heme edge, towards Glu-49 and Asp-53. These calculations should be viewed as approximations for several reasons. Specifically, the calculations treat the protein as a uniform dielectric, and dipoles are assigned to bonds in residues irrespective of the residue's relative location in the tertiary structure. The uncertainty in the fit of the ionic strength dependence data is too great to determine whether or not the difference in die behaviour of native and DME-cytochrome b5 to porphyrin cytochrome c can be adequately explained by the Van Leeuwen treatment. The pH dependences of the association constants of cytochrome bs and porphyrin cytochrome c (Figure 14) indicate that either none of the groups involved in the interaction titrate over this pH range, or that any changes that do occur compensate for each other. The pH optimum of the 85 association constant of cytochrome bs and porphyrin cytochrome c (7.0) is similar to that reported for the native proteins (Mauk et al., 1982). The lack of pH dependence of the association constant between DME-cytochrome b5 and porphyrin cytochrome c is different dian what is observed for DME-cytochrome b5 and cytochrome c. This discrepancy is discussed below. The current work provides two lines of evidence in addition to die points made in the Introduction, that indicate that porphyrin cytochrome c is a good analogue for native cytochrome c in binding assays. First, neither the magnitude nor the orientation of the dipole of cytochrome c is appreciably perturbed by the removal of the iron atom (Table IX), assuming that at pH 7.0, His-18 is not protonated. Perhaps this is not surprising as the iron atom is close to the centre of mass of cytochrome c and positive charges of ferrous iron are replaced by two protons at neutral pH. These calculations were based on the crystal structure of native cytochrome c (Bushnell et al., in preparation). It is reasonable to expect that some structural rearrangement of the protein would occur upon the removal of the iron atom. This has not been accounted for in diese calculations. Second, the association constants of the cytochrome fe5porphyrin cytochrome c complex measured by fluorescence quenching agree well with the measurements performed by absorption spectroscopy using the native cytochromes (Mauk et al., 1982; 1986). This agreement is not quite as good as that obtained for the association constants of the cytochrome c peroxidase-cytochrome c complex measured by the two methods (Erman & Vitello, 1980; Vitello & Erman, 1987). The agreement between the association constants determined by the fluorescence quenching and absorption spectroscopy indicates that the two techniques detect the same population of docking geometries of the complexes. That porphyrin cytochrome c is a good analogue for native cytochrome c is useful because fluorescence quenching is a more sensitive technique than absorption spectroscopy, allowing the measurements of association constants over a wider range of conditions. Although the association constants of the cytochrome fc5porphryin cytochrome c complex agree within an order of magnitude to those determined by absorption spectroscopy for the native proteins imder similar solution conditions, the association constants measured by the two techniques display different pH 86 optima. Whether this difference represents differences in experimental techniques (the pH studies were done at different ionic strengths), differences between porphyrin cytochrome c and native protein, or experimental error is difficult to say. As discussed in the Introduction, electrostatic calculations of the complexes that form between cytochrome b5 and cytochrome c (Mauk et al., 1986) predict that the overall binding energy of the complex that forms between these proteins is insensitive to pH (over a range of 6.75 to 8.25), but that two docking geometries exist (Figures 6 & 7). For the native proteins, the edge on docking geometry is the more stable of the two orientations below pH 8.0. Above pH 8, the two docking geometries are isoenergetic and there are equimolar amounts of the edge on and off edge orientations. The electrostatic calculations further predict that in the complex formed by D M E -cytochrome b5 and cytochrome c, the two docking geometries are isoenergetic at neutral pH, but as the pH rises, histidine 26 of cytochrome b5 deprotonates, . stabilizing the off edge geometry and increasing the overall association constant of the complex. The current results are consistent with the electrostatic calculations in that esterification of the cytochrome b5 heme propionates induces a change in the average docking geometry of the cytochrome ^cytochrome c complex. The pH dependence of the docking geometries that is predicted by the electrostatic calculations is not observed in the fluorescence quenching experiments. In making these comparisons between the calculations and the experimental observations, it is useful to recall that the binding energies derived from the electrostatic calculations correspond to association constants that are two orders of magnitude smaller than the experimentally observed ones. This suggests that there are forces in addition to electrostatic interactions diat contribute to the stability of the complex formed by cytochrome bs and c. These additional forces may also contribute to stabilising different docking geometries within the complex, and may explain in part why the fluorescence quenching experiments do not reproduce the pH dependence predicted by the calculations. In addition, the calculated difference in electrostatic energy ( A G e ) ) of the predicted docking geometries at pH 7.0 and 8.5 is approximately 1 kcal/mol. This suggests that even under 87 conditions that the calculations predict to favour one docking geometry, this geometry represents a shallow energy minimum in the energy surface of potential conformations and may only exist in solution slightly longer than a myriad of other docking geometries as the cytochromes tumble around each other. This corresponds closely to the view of the interaction of cytochrome c peroxidase and cytochrome c that emerges from Brownian dynamics simulations (Northrup et al., 1988) in which two or three docking geometries are slightly favoured over the many docking geometries that these proteins sample. The fluorescence quenching data indicate that several different docking geometries can exist within the complex formed by cytochrome b5 and porphyrin cytochrome c. It is likely that under any given conditions, a population of docking geometries exist, all differing slightly with respect to intersite separation and orientation. Solution conditions and the electrostatic surface of die proteins determine where this equilibrium lies. If this is taken as a model for the complex that forms between the native cytochromes, the question arises as to which docking geometries correspond to productive electron transfer complexes. In conclusion, cytochrome b5 and porphyrin cytochrome c form a 1:1 complex in solution that has approximately the same stability as previously observed for the complex formed by the native proteins. The stability of the cytochrome ^porphyr in cytochrome c complex is mediated by electrostatic forces. Maximal quenching of porphyrin cytochrome c by cytochrome b5 is consistent with a average docking geometry of the proteins in which the centre to centre separation of the prosthetic groups is 18 A , in good agreement with the distance between die iron atoms in the edge on model of the cytochrome ^cytochrome c complex (Salemme, 1976). Esterification of the cytochrome b5 heme propionates does not appreciably decrease the binding strength of the complex but does induce an altered average docking geometry. This effect of heme propionate esterification on the interaction of cytochrome bs and porphyrin cytochrome c can be taken as a reasonable model the effect that the esterification would have on the interaction of the native proteins. 88 Kinetics of Flavin Semiquinone Reduction of Cytochrome b5 and Cytochrome c The characterisation of a large number of electron transfer proteins and protein-protein complexes by flavin semiquinone reduction kinetics was reviewed in the Introduction. The flavins used in the current study are illustrated in Figure 25. Some of the properties of these two flavins are provided in Table X I . As can be seen from this information, the electrochemical properties of the two flavins are similar while the electrostatic properties are not. Under the conditions employed in this study, the association constant of complexes involving die two cytochromes b5 and cytochrome c is approximately 10 6 M " 1 (Mauk et al., 1982; 1986; this work). In general, differences in the rate constants for semiquinone reduction of one cytochrome in the absence and presence of the second cytochrome can be interpreted in terms of steric and electrostatic changes induced by complex formation. Reduction of Cytochrome c by L fH - : The observation that the rate constant for cytochrome c reduction by LfH- is unaffected by the presence of cytochrome b5 at pH 7, suggests that there is no change in the accessibility of the cytochrome c heme when this protein binds to native cytochrome bs. When DME-cytochrome b5 binds to cytochrome c this heme accessibility is decreased in a kinetically detectable manner, indicating that the stereochemical environments of the cytochrome c heme edge are not equivalent in the two complexes at pH 7. This difference could arise from the increased steric bulk introduced by esterification of the heme propionates or from changes that this modification induces in the geometry of the complex formed by the two proteins. As LfH- is essentially uncharged at pH 7, the observed kinetic difference cannot be attributed to electrostatic effects. The substantial increase in the rate constant of cytochrome c reduction by LfH- at pH 8 has been observed previously by Przysiecki et al. (1985). Similar rate enhancement at pH 8 has been observed for several small electron transfer proteins of varying net electrostatic charge in their reactions with this flavin. This rate enhancement has been attributed to an increase in the 89 Figure 25 : The structure of the three oxidation states of flavin. R is a methyl group ( C H 3 ) and ribityl phosphate for lumiflavin and F M N respectively. The reduction potential of die quinone and the semiquinone (E°) and the pKj's of the semiquinones of these flavins are given in Table X I . 90 Table X I : Some relevant properties of lumiflavin and F M N . Property lumiflavin F M N R (in Fig . 25) methyl ribityl phosphate charge 0 - 2 E° p H 7 - 231 - 238 p H 8 - 259 - 269 pR, 8.27 8.55 disp. rate3 pH 5 6.2 x 10 9 2.6 x 10 9 p H 9 1.1 x l O 9 1.9 x l O 9 a rate of semiquinone disproportionation. Equilibria data were taken from Draper & Ingram (1968) and disproportionation rates were taken from Vaish & Tollin (1970). intrinsic reactivity of the lumiflavin radical upon deprotonation to form an anionic species (Figure 25, Table XI) . The increased reactivity is a manifestation of the lower reduction potential and a change in the electron spin density distribution, both of which are characteristic of die deprotonated semiquinone. Electrostatic effects that result from this deprotonation could also contribute to an increase in the rate constant of reduction, but in previous studies, this effect was observed to be less important (Przysiecki et al., 1985). Interestingly, the rate constant of ferricytochrome c reduction in its complex with native cytochrome b5 was unaffected by the increase in p H . At pH 8, the increase in LfH- reactivity is perhaps offset by an increase in the steric hindrance of the cytochrome c heme. In contrast, the rate constant of cytochrome c reduction in its complex with DME-cytochrome b5 approximately doubles on raising the pH from 7.0 to 8.0, which is similar to the pH induced increase observed for free cytochrome c. Overall, these results are consistent with the suggestion that esterification of the cytochrome b5 heme propionate groups induces a change in the docking geometry for the interaction of this 91 protein with cytochrome c at pH 7. In addition, the current findings are consistent with a reduction of the kinetic accessibility of the partially exposed cytochrome c heme edge in the cytochrome ^-cytochrome c complex at higher pH. The electrostatic environment of the partially exposed native cytochrome b5 edge is highly negatively charged owing to die presence of the two heme propionate groups (cf. Figure 2). This characteristic may inhibit the interaction of the partially anionic lumiflavin semiquinone with the heme of cytochrome c in its complex with native cytochrome b5 at p H 8. Such an electrostatic effect is not expected to contribute to the kinetics of cytochrome c reduction in its complex with DME-cytochrome b5 as the electrostatic environment of the heme edge in this derivative is effectively neutral (Mauk et al., 1986). Analysis of the ionic strength dependence of these reactions is required to clarify this latter point, though the sensitivity of the association constant of the complex to ionic strength precludes an extensive investigation. Reduction of cytochrome c by F M N H - : Use of F M N semiquinone ( F M N H ) as a reductant permits evaluation of the sign and magnitude of the effective electrostatic charge of the protein near the site of electron transfer (Meyer et al., 1984; Hazzard et al., 1987). In addition, use of two species of cytochrome c (horse and tuna) can provide information regarding die effects of amino acid substitution in the region surrounding the cytochrome c heme edge [cf. Table VI in Hazzard et al. (1987)] that are thought to be important in the formation of electron-transfer complexes (Salemme, 1976; Mauk etal., 1986). The minimal differences in the reduction behaviour of the two species of cytochrome c is consistent with previous results (Hazzard et al., 1987). The greater reduction rate constants observed with F M N H - , relative to LfH- , can be attributed to the electrostatic attraction of the negatively charged F M N H - to the highly positively charged surface of cytochrome c surroimding its partially exposed heme edge at the low ionic strengths used in these experiments (Meyer et al., 1984). Other properties of the semiquinones which might affect the kinetics of reduction, such as the reduction potential, are very similar (Table IV). The sensitivity of the rate constants of cytochrome c reduction by die anionic flavin ( F M N H ) 92 to the presence of cytochrome b5 contrasts with the identity of the rate constants of cytochrome c reduction by the neutral flavin (LfH) in the presence or absence of native cytochrome b5 at pH 7. The dependence of kinetic behaviour on the electrostatic properties of die flavin employed probably arises from a differential response of the two flavins to the appreciable neutralisation of the positive electrostatic potential surface of cytochrome c that occurs when it binds to native cytochrome b5. This conclusion correlates with the smaller effect of cytochrome c binding to D M E -cytochrome b5 on the reduction rate constants for F M N H - . That is, it is consistent with the diminished negative electrostatic potential surface of the D M E derivative near the region of die exposed heme edge (Mauk et al., 1986). Furthermore, this interpretation implies that the heme propionate groups are in sufficiently close proximity to the site of F M N H - approach to cytochrome c that they exert a kinetically significant electrostatic influence. Finally, the lack of species dependence of cytochrome c reactivity within the complexes is interesting. It might be worthwhile to extend these experiments to include yeast iso-1-cytochrome c as two factors suggest diat this species of cytochrome may bind to cytochrome b5 somewhat differently. First, this cytochrome has a much larger dipole moment than the horse heart species (results not shown). Second, the bimolecular rate constant of ferricytochrome c reduction by ferrocytochrome b5 is much greater for yeast iso-1-cytochrome c than the horse heart protein (Eltis, Barker & Mauk, unpublished). Reduction of Native and DME-Cytochrome b5 by L f H - : At pH 7, die rate constants for native and DME-cytochrome b5 reduction by LfH- vary directly with the thermodynamic driving forces of the reactions (Meyer et al., 1983). For example, the rate constants for reduction of free cytochrome b5 are 25-45% lower than those obtained for free DME-cytochrome b5, as expected from the finding that heme propionate group esterification increases the reduction potential of this protein by 60 mV (Reid et al., 1982, 1984). Similarly, the rate constants for cytochrome c ( E m 7 = 260 mV vs SHE) reduction are significantly faster than the rate constant of reduction of either form of cytochrome b5. Other, less readily correlated factors, are expected to contribute to die rate constant differences between cytochrome c and cytochrome b5 derivatives observed here. These factors 93 include the extent of heme exposure to solvent [12% in cytochrome c, 33% in cytochrome bs; Table VII][cf. Table I and Figure 1 of Tollin et al. (1986)] and the Franck-Condon activation energies of the two proteins. The solvent exposure of cytochromes b5 and c hemes are larger dian diose calculated by Stellwagen (1978) simply because in the current study, the surface areas were calculated from the edge of the probe sphere rather than from its centre. The differential response to an increase in pH of the rate constants of reduction of the two cytochrome b5 derivatives may be attributed to the sum of two opposing effects. On one hand, the electrostatic effects of partial deprotonation of LfH- to generate the anionic form of the flavin at alkaline pH should decrease the rate constant for the native cytochrome reduction to a much greater extent than in the case of the modified protein as a consequence of the electrostatic effect of the solvent exposed heme propionate groups (Reid et al., 1984). On the other hand, the increased reactivity of the anionic form of the flavin (Przysiecki et al., 1985) wil l tend to increase the rate of reduction of both cytochrome bs derivatives. For the native protein, this second effect appears to be exactly offset by the inhibitory electrostatic influence of the heme propionate groups, while for DME-cytochrome b5 elimination of the electrostatic contribution of the propionate groups permits the increased reactivity of LfH- at alkaline pH to prevail. Clearly, the differential response of die two forms of cytochrome b5 is not related to changes in thermodynamic driving force as the reduction potentials of native and DME-cytochrome b5 are largely unaffected by this change in pH (Reid etal, 1982, 1984). While the mechanistic basis for the increase in the reduction rate constant of cytochrome b5 in the presence of cytochrome c is unknown, it may be related to a decrease in the reorganizational barrier to electron transfer or to an increase in the reduction potentials of both forms of cytochrome b5 that is induced by binding of either protein to cytochrome c. This latter possibility could be related to the increase in reduction potential observed for native and DME-cytochrome bs with increased ionic strength (Reid et al., 1982, 1984), the origin of which is not fully understood. The increase in reactivity of cytochrome b5 associated with protein-protein complex formation is not 94 without precedent. For the reduction of bovine cytochrome c oxidase (Ahmad et al., 1982) and of yeast cytochrome c peroxidase (Hazzard et al., 1987) by free flavin semiquinones, electron transfer is not fast enough to compete with semiquinone disproportionation. In the presence of cytochrome c, however, rapid electron transfer from flavin to either the oxidase or the peroxidase does occur, though the mechanism involves direct reduction of the cytochrome c followed by intracomplex electron transfer. In these cases, it has been argued that binding to cytochrome c establishes an effective electron-transfer pathway through structural changes induced in the ultimate electron acceptor protein. Whether a similar mechanism is responsible for the present observation requires further study. Reduction of Native and DME-Cytochrome b5 by F M N H - : The observation that the rate constant for F M N H - reduction of cytochrome b5 (Table VI) is so much slower than that for the reduction of cytochrome c (Table V) suggests that F M N H " reduction kinetics are significantly affected by strong electrostatic repulsion between the negatively charged F M N H - and bodi free cytochrome bs species, as would be expected from previous electrostatic calculations (Mauk et al., 1986). The small magnitude of the difference between the rate of reduction of native and D M E -cytochromes b5 by F M N H - suggests that although the electrostatic repulsion between F M N H and die protein may be affected by the neutralisation of the heme propionate charge by esterification, other factors such as heme accessibility and the change in the reduction potential also influence the reduction kinetics, resulting in a complex mixture of partially cancelling effects. The second-order rate constants for the reduction of native and DME-cytochrome b5 by Fe(EDTA) 2 ", another small reductant with a net charge of -2, are 112 and 1970 M ' V 1 (I = 0.1 M and pH 7) respectively, (Reid et al., 1984). The difference in the effect of the heme propionate esterification on the reaction kinetics of these two reductants [2-fold for F M N H " versus 10-fold for Fe(EDTA) 2"] suggests a fundamental difference in the mechanism of reduction of cytochrome b5 by these reductants, possibly involving requirement for * -orbital overlap between the flavin and the heme prosthetic group. The non-linearity of the plots of k o b s versus complex concentration for cytochrome b5 in the 95 presence of cytochrome c (Figure 20) suggests that there is a first order rate limiting process in the reaction mechanism. One possible explanation for this is represented by [a] and [b] where cyt fe5(III)-cyt c(II)* represents a form of the complex with which F M N H - preferentially reacts, perhaps as the result of increased accessibility of the cytochrome b5 heme group. Thus, the rate limiting cyt fe5(III)-cyt c(II) ^ - > cyt &5(TiI)-cyt c(U)* [a] cyt &5(HI)-cyt c(IJ.)* + F M N H - - ^ - > cyt fc5(II)-cyt c(II) + F M N [b] process at high concentrations is the first-order isomerisation of the complex. This mechanism in combination with our current results implies that the active form of the complex must have electrostatic properties that are distinct from those of the inactive form and to which reduction by F M N H - (but not L f H ) is sensitive. A second plausible mechanism involves the formation of a kinetically detectable precursor complex between the protein-protein complex and F M N H - as represented by equations [c] and [d]. cyt fc5(ffl)-cyt cQl) + F M N H - k 2 > cyt &5(ffl)-cyt c(II)-FMNH- [c] cyt &5(III)-cyt c(II)-FMNH- —^-> cyt i>5(II)-cyt c(II) + F M N [d] Inasmuch as these two mechanisms are mathematically equivalent (Stricklan et al., 1975), the numerical values for the first- and second-order rate constants are the same. For either mechanism, the second-order rate constant reflects the interaction of the negatively charged F M N H - with the bound cytochrome b$ and is thus sensitive to the effective electrostatic environment near the interaction domain. The large values for the second-order rate constants of semiquinone reduction of bound ferricytochrome b5 compared to those observed with the free proteins indicate that there is a considerable neutralisation of the cytochrome b5 negative charge, and hence a greater reactivity 96 with the reduced F M N , in the complex. The increased rate constant for reduction of D M E -cytochrome b5 in complexes with horse or tuna cytochrome c, relative to native cytochrome b5, is consistent with the results obtained with the free proteins. Whereas no difference between horse and tuna cytochrome c reduction was observed in the complexes with cytochrome b5, there were marked species dependent differences in the cytochrome b5 reduction kinetics. For bodi native and DME-cytochrome bs, the values for the second-order rate constants in the complexes with horse cytochrome c were significantly smaller than those obtained with tuna cytochrome c. This finding suggests that there is either a less negative or a more positive electrostatic potential near the cytochrome b5 reduction site in the complex with tuna cytochrome c than is the case in the complex with horse cytochrome c. The interpretation of the effects on the rate-limiting first-order rate constant is dependent on the mechanism. In one previous case, F M N H - reduction of ferricytochrome c in a 1:1 complex with Clostridium pasteurianum flavodoxin, it was shown unambiguously that the mechanism given by equations [a] and [b] was correct, based on the independence of the first-order rate-limiting process of the identity of the reductant (Hazzard et al., 1986). In the present case, no such comparison was made. However, some discussion concerning the mechanism can be given. According to the mechanism given by equation [a] and [b], a larger value for kj implies that the cytochrome ^cytochrome c complex can explore different docking geometries more readily. One explanation for the larger kj values obtained with complexes involving DME-cytochrome b5 as opposed to native cytochrome b5 is that esterification of the heme propionates makes the cytochrome fe5cytochrome c complex more dynamic (i.e., the barriers to rotation between different docking geometries are lower in the DME-cytochrome ^cytochrome c complex). Similarly, the greater value of k] for tuna cytochrome c complexes would imply that a more dynamic complex is formed with this cytochrome than with horse cytochrome c. Admittedly, the differences in these barriers to rotation are quite small ( A A G t ~0.3 kcal/mol based on the ratio of kj). Although it is more difficult to rationalise the results in terms of the mechanism of equations [c] and [d], such arguments are of course not 97 definitive and further study is required. Calculation of the solvent exposure of the heme groups in the two static docking geometries that have been proposed for the cytochrome ^cytochrome c complex (Salemme, 1976; Mauk et al., 1986) indicates that the exposure of cytochrome b5 heme to solvent is essentially unaffected by cytochrome b5 binding of cytochrome c in either orientation (Table VII). On the other hand, the exposure of the cytochrome c heme is unaffected on binding to cytochrome bs in the edge on docking geometry (Figure 6) while it is reduced ~35% on binding to this protein in the off edge docking geometry (Figure 7). The current results are consistent with these models in that the rate constant for the reduction of cytochrome c by L f H - is unaffected by the binding to native cytochrome bs at pH 7.0 but is reduced by 25% upon binding to DME-cytochrome b5 at pH 8.0. Nevertheless, the value of estimating the effect of complex formation on solvent exposure is clearly limited in light of molecular dynamics calculations of the cytochrome ^cytochrome c complex (Wendoloski et al., 1987) and Brownian Dynamics calculations of the cytochrome c peroxidase-cytochrome c complex (Northrup et al., 1988) which suggest that there may be some flexibility in the solution conformations of electron transfer protein-protein complexes. Ferrocytochrome b5 Reduction of Ferricytochrome c Pseudo-first order conditions for analysis of a second order reaction are satisfied when one of the reactants is in at least a 40 fold excess over the other (Moore & Pearson, 1981). The large extinction coefficients exhibited by cytochromes b5 and c and the rate of ferrocytochrome b5 reduction of ferricytochrome c severely restricted the concentration range over which pseudo-first order kinetics could be performed in the present study. Calculations showed that treating a second order reaction in which one of the reactants is in a five-fold excess over the other as a pseudo-first order process underestimates the second order rate constant by 10%. A similar treatment of a reaction in which one of the reactants is in a ten-fold excess over the other underestimates the 98 rate constant by 5% (data not shown). These errors are within the reproducibility of the second order rate constants measured in this work. The agreement between the second order rate constants of ferrocytochrome b5 reduction of ferricytochrome c obtained under pseudo-first order conditions and the second order rate constants obtained under conditions involving equimolar reactant concentrations argues that the rate constant of ferrocytochrome b5 reduction of ferricytochrome c can be accurately determined by either technique. The agreement between the second order rate constant of ferrocytochrome bs reduction of ferricytochrome c obtained using dithionite reduced cytochrome b5 and that obtained with photolytically reduced cytochrome b5 indicates diat the lumiflavin used to prepare ferrocytochrome b5 does not contribute to die absorbance changes observed in the stopped flow cuvette. Although it is preferrable to keep die reaction mixture as simple as possible, it was difficult to perform the reaction in the absence of lumiflavin because once the reductant was removed from the ferrocytochrome bs, up to 10% of the protein auto-oxidised. In the presence of lumiflavin, any ferrocytochrome bs that oxidised was re-reduced until all of the oxidant, presumably oxygen (Berman et al., 1976) was purged from the protein solution. Strittmatter (1964) first reported the rate of ferrocytochrome bs reduction of ferricytochrome c to be 4.7 x 10 7 M ' V 1 (0.1 M phosphate, pH 7.0, 25°C) in a brief communication that described a novel stopped flow technique. This value is difficult to compare with the current results as the ionic strength used in the earlier study is ambiguous. Furthermore, lipase solubilised cytochrome b5 was probably used in this early report rather than the trypsin solubilised form used here, tiiough this is not stated clearly either. If the buffer was 0.1 M phosphate (pH 7), then the ionic strength was 0.25 M . In this event, the current results [4.7 (±0 .4 ) and 9.5 ( ± 0 . 6 ) x 10 7 M ' V at ionic strengths of 0.3 and 0.2 M respectively] are in good agreement with Strittmatter's measurement. The bimolecular rate constant for ferrocytochrome b5 reduction of ferricytochrome c is extremely large. None of the reported second order rate constants of electron transfer between soluble electron transfer proteins are as great at 0.1 M ionic strength, pH 7.0. Perhaps diis is not unexpected considering that the exclusive biological function of these two proteins is electron 99 transfer. Hemoproteins such as cytochrome c peroxidase and hemoglobin, whose rates of electron transfer with other proteins are considerably slower (Cheung et al . , 1986; Summer & Erman, 1988; Mauk et al., 1984), do not function primarily as electron transfer proteins. It is reasonable to expect that factors such as the reorganisational energy of cytochromes would be optimised for the purpose of electron transfer in a way that could not be achieved in a protein that has evolved to perform other functions. The reorganisation energies of cytochrome c peroxidase and myoglobin are greater than those of cytochromes bs and c (Mayo et al., 1986). Comparison of bimolecular rate constants of electron transfer between soluble electron transfer proteins (Wherland & Pecht, 1978; Augustin et al., 1983; Tollin et al., 1986) reveals that one of the few measured second order rate constants of electron transfer between pairs of electron transfer proteins that approaches diat of ferrocytochrome b5 reduction of ferricytochrome c is that of the oxidation of ferrocytochrome / by plastocyanin [6.0 x 10 7 M ' V 1 (pH 7.0, I = 0.1 M , 25°C)] (Takana et al., 1981). The second order rate constant of ferrocytochrome bs reduction of ferricytochrome c is also much greater than that measured between either of these proteins and a wide variety of small electron transfer reagents (Hodges et al., 1974; Wherland & Gray, 1976; Cummins & Gray, 1977; Mauk et al., 1979; Reid & Mauk, 1982; Reid et al., 1984; Chapman et al., 1984). This observation is discussed below. The unimolecular rate constant of electron transfer from cytochrome b5 to cytochrome c within the complex has been determined by pulse radiolysis to be 1600 (±700) s"1 (t^ = 0.43 ms)[l m M phosphate, pH 7](McLendon & Miller, 1985). At the concentrations of protein used in the current study (1.7 pM), the half life of the bimolecular reaction (I = 0.5 M , pH 7.0) is about 30 ms. The second order rate constant would have to be 1.4 x 10 9 M ' V 1 for the half-life of the bimolecular process to be equal to the half-life of the first order process. If electron transfer between the heme groups within the cytochrome ^cytochrome c complex is comparable at 1 m M phosphate and I = 0.5 M , and ferrocytochrome b5 reduction of cytochrome c involves a protein-protein binding step prior to electron transfer (Scott et al., 1985), then at I = 0.5 M , pH 7 the association constant of ferrocytochrome b5 and ferricytochrome c is M O 4 M " 1 . This value is larger than experimentally 100 determined values [e.g. 10 3 M " 1 measured at I = 0.04 M (Eley & Moore, 1983)] suggesting diat the rate of electron transfer within the protein-protein complex is ionic strength dependent. Neutralising the charge of the heme propionates of cytochrome b5 is expected to influence the bimolecular rate constant of ferrocytochrome b5 reduction of ferricytochrome c in at least two ways: (1) through perturbation of the electrostatic potential surface of the cytochrome b5 and (2) through alteration of the thermodynamic driving force for electron transfer. First, removing two negative charges from the surface of the protein would alter both the monopole and the dipole of die protein. Theoretical calculations indicate that the dipole moment of cytochrome b5 is reduced by this modification (Table IX). According to the model developed for the reaction between cytochrome c peroxidase and cytochrome c (Northrup et al., 1988), the principal effect of diminishing die dipole moment is to decrease the rate at which the proteins come together in solution. A change in the monopole might affect the geometry of the complex. The present studies combined with the binding studies previously reported from this laboratory (Mauk et al., 1986) indicate that a change in die charge distribution on the surface of cytochrome bs stabilises a different population of cytochrome ^cytochrome c docking geometries. Different docking geometries can readily influence the rate of electron transfer between the two proteins through alterations in the separation, orientation and protein structure intervening between the donor and acceptor sites. Second, die reduction potential of DME-cytochrome b5 is 60 mV higher (Reid et al., 1984) than that of native cytochrome bs. As the reduction potential of cytochrome c is 261 mV (Margolit & Schejter, 1973), the effect of increasing the reduction potential of cytochrome bs is to decrease the driving force of the reaction and thereby reduce the rate of electron transfer within the complex. In the current experiments, the second order rate constant for ferrocytochrome b5 reduction of ferricytochrome was decreased by esterification of the heme propionates when the reaction was performed at ionic strengths less than 0.5 M . However, there was no difference in die bimolecular rate constants for ferricytochrome c reduction by native and D M E - ferrocytochrome b5 at ionic strengths of 0.5 M and greater. Several mechanisms could account for these observations. First, 101 the reduction potentials of native and DME-cytochromes bs might be similar when bound to cytochrome c. Alternatively, the bimolecular rate constant for ferrocytochrome b5 reduction of ferricytochrome c might involve a rate determining step such as protein-protein coniplexation or a protein conformational change. In such a mechanism, the rate of electron transfer between the heme groups within the complex would not influence the bimolecular rate of ferrocytochrome b5 reduction of ferricytochrome c and hence factors influencing the rate of electron transfer widiin die complex would not affect the bimolecular rate constant. Measurements of the bimolecular rate constant of ferrocytochrome b5 reduction of mutant yeast ferricytochromes c (Eltis et al., unpublished) indicate that factors such as the driving force and the nature of the protein-protein interface influence the bimolecular rate constant of electron transfer between cytochromes b5 and c. This observation indicates that esterification of the cytochrome b5 heme propionates probably decreases the bimolecular rate constant for ferrocytochrome b5 reduction of ferricytochrome c through perturbation of the interaction of these two proteins and diat die reduction potential of native and DME-cytochromes are similar when bound to cytochrome c Several points regarding the reduction potentials and docking geometry of cytochromes b5 and c are worth considering. The reduction potentials of the cytochromes used in these studies were measured for the proteins free in solution. It is not at all clear that die reduction potential of a cytochrome bound to another protein is the same as the reduction potential of that same cytochrome free in solution. Charged groups and the exclusion of solvent from the heme, factors that change upon complex formation, have a documented effect on the reduction potential of cytochromes (Moore & Williams, 1977; Stellwagen, 1978). The reduction potential of cytochrome c has been determined to be unaffected by the presence of cytochrome b5 (Vanderkooi & Erecinska, 1974), but this study was performed at an ionic strength that is too high to promote stable cytochrome ^cytochrome c complex formation. Electrostatic calculations (Mauk et al., 1986) and the current flavin semiquinone reduction kinetics of cytochrome b5 and cytochrome c argue that in the cytochrome ^cytochrome c complex, the cytochrome bs heme propionates are neutralised by the positive electrostatic field of 102 lysine residues on the surface on cytochrome c. For this reason, the reduction potential of cytochrome bs within the complex might approach the reduction potential of DME-cytochrome bs in solution. The principal effect of the heme propionate esterification on cytochrome b5 behaviour might be to leave a residue on cytochrome c unpaired in the complex, and thus to perturb the reduction potential of cytochrome c in the cytochrome ^cytochrome c complex. It is possible that the difference in reduction potentials of native and DME-cytochromes bs is offset by a difference in another factor that influences electron transfer, such as reorganisational energies. It would be interesting to measure the reduction potentials of the cytochromes within the complex. Alternatively, it would be interesting to investigate the effect of the reduction potential and docking geometry on the rate of electron transfer within a complex by measuring the rate of electron transfer within a covalently crosslinked DME-ferrocytochrome Zj5ferricytochrome c complex of defined structure. With respect to the docking geometries of the cytochrome ^cytochrome c complex, it should be remembered that the models (Figs. 6 & 7) represent shallow minima in the energy surface of possible docking orientations at low ionic strengths (Mauk et al., 1986) and, as is suggested by Brownian dynamic simulations of the cytochrome c peroxidase-cytochrome c complex (Northrup et al., 1988), movement between these geometries could occur very readily, particularly at high ionic strength. Cytochrome b5 heme propionate esterification could influence protein-protein complex formation, and hence the second order rate of ferrocytochrome b5 reduction of ferricytochrome c, by one of two mechanisms. First, reduction of the magnitude of the dipole of cytochrome b5 might decrease the rate of protein-protein association. Second, esterification of cytochrome bs heme propionates might promote the formation of complexes that are not as favourable to electron transfer as complexes formed by the native cytochromes. At this time, it is not possible to distinguish between these two possibilities. In either case, increasing the ionic strength of die protein solution would dampen these effects, consistent with the current observations. Protein conformational rearrangements are not likely to contribute significantly to the rate of 103 ferrocytochrome b5 reduction of ferricytochrome c. Several studies indicate that the structures of the reactants are very similar to those of the products. As discussed in the Introduction, neither cytochromes bs nor c undergoes a change in spin state nor axial ligation with the change of oxidation state. Some conformational changes in the side chains of cytochrome c accompany die change in oxidation state of the protein (Takano & Dickerson, 1980), but the three dimensional structures of ferric and ferrocytochrome bs are reported to be identical to each odier (Mathews et al., 1979). There are several reports of complex formation affecting the environment of die heme groups of the interacting hemoproteins. The M C D and C D spectra of cytochrome c are perturbed upon binding to cytochrome c oxidase (Weber et al., 1987). This spectral perturbation has been interpreted as a conformational change in the heme environment of cytochrome c (Michel et al., 1989b). Measurements of phosphorescence lifetime have suggested that the conformation of zinc cytochrome c changes upon binding to cytochrome c peroxidase (Koloczek et al., 1987). The view that there are no significant protein conformational processes diat affect electron transfer in ferrocytochrome b5 reduction of ferricytochrome c is supported by a preliminary report on ' H N M R studies that indicates that the structures of cytochromes bs and c in the complex are the same as in solution (Veitch et al., 1988). It is not possible to rule out completely the role of conformational changes in ferrocytochrome b5 reduction of ferricytochrome c, particularly as the mechanisms of reactions involving azurin (Corin et al., 1983), cytochrome c peroxidase (Cheung & English, 1987) and cytochrome b2 (McLendon et al., 1987) remain so poorly defined. However, there is no kinetic evidence for a conformational change in the present case. Several authors have analyzed second order protein-protein electron transfer reactions by electron transfer theory as developed by Marcus and others (Wherland & Pecht, 1978; Tollin et al., 1986; Dixon et al., 1989). The Marcus relationship (equation 5) can be used to calculate the bimolecular rate constant for ferrocytochrome bs reduction of ferricytochrome c in a similar maimer. Dixon et al. (1989) have measured the electron transfer self-exchange rate constants of cytochromes b5 and c to be 2600 M ' V 1 (I = 0.12 M , 25 °C) and 5400 M ' V 1 (I = 0.12 M , 27 °C) respectively, 104 and / can be estimated to be 0.43 if the collision frequency is 10 1 0 M _ I s (Reynolds & Luniry, 1966). The second order rate constant for ferrocytochrome b5 reduction of ferricytochrome c, using a thermodynamic driving force of 255 mV, is calculated to be 3.5 x 10 5 MAs~l (equation 5). This value is several orders of magnitude smaller than the observed rate constant. Variation of die collision frequency by two orders of magnitude does not affect / by more than 30%. It is not clear whether the Marcus relationship fails in the present case because the electron transfer reactions are not uniformly nonadiabatic or because electron transfer is not rate limiting. The observed ionic strength dependence of the bimolecular rate constant reflects die large role of electrostatic forces in the association process. Unfortunately, the electron transfer rates and die association constants cannot be directly compared as they cannot be obtained at comparable ionic strengths with the techniques currently available. Nevertheless, the results of die binding studies and the current second order rate studies are consistent in that both the association constants and the second order rate constant of ferrocytochrome bs reduction of ferricytochrome c are strongly dependent on ionic strength and that bimolecular kinetics are observed under conditions where no complex is detected by electronic absorption spectroscopy (Mauk et al., 1982; 1986) or fluorescence quenching titrations. The fits of the data to the Van Leeuwen model (equation 11) appear to be quite good, but it is worth noting two points about the best fit parameters. First, in the fits in which the monopoles of the cytochromes were fixed, the radii of cytochromes b5 and c (Table VIII) are approximately half the size that the crystal data suggest (Mathews et al., 1979; Takano & Dickerson, 1980a,b). Second, in the fits in which the radii were fixed, Z{Z^ for DME-cytochrome bs and cytochrome c is approximately twice that for the native cytochromes (Table Vffl). This suggests that the model needs some refinement. It is interesting to note that Wherland-Gray fits of the association constants for DME-cytochrome b5 and cytochrome c yielded higher Z j / ^ than for native cytochrome b5 and cytochrome c (Mauk et al., 1986). The equivalence of the rate constant for reduction of ferricytochrome c by native and D M E -105 ferrocytochrome b5 at high ionic strength has been discussed above. The rate of ferrocytochrome bs reduction of (trifluoromethyl)phenylcarbamyl-Lys-13 cytochrome c has been measured to be ~37% that of native ferricytochrome c at 0.1 M ionic strength (Ng et al., 1977), but at I = 1.5 M , the two forms of cytochrome c are reduced at the same rate by ferrocytochrome b5 (Stoneheurner et al., 1979). Although the rate constant that these workers report is estimated from steady state kinetics and was not measured directly, the ionic strength dependence that they report is similar to that observed in the present study. It is also intriguing that cytochrome c lysine modifications diat decreased this rate the most at I = 0.1 M (Lys-13, Lys-72) exhibited rate constants that where 35% to 40% that of native cytochrome c (Stoneheurner et al., 1979). In the present study, ferricytochrome c reduction by DME-ferrocytochrome b5 (I = 0.1 M) yielded a rate constant that is 70% smaller than that of ferricytochrome c reduction by native ferrocytochrome b5. While D M E -cytochrome b$ is modified at two positions, these results are consistent with the involvement of cytochrome b5 heme propionate-6 in stabilizing the cytochrome ^cytochrome c complex to die same extent as cytochrome c lysines 13 and 72. This result is clearly consistent with the model of the cytochrome fc5cytochrome c complex proposed by Salemme (1976)(Figure 6, Table I). Steady state kinetic studies that detect differences in reactivity between native cytochrome c and cytochromes c singly modified at various surface lysines with a number of enzymes (Ferguson-Miller et al., 1978; Kang et al., 1978; Speck et al., 1979; Smith et al., 1980; Speck et al., 1981) were performed at ionic strengths below 0.2 M . The second order rate constants of plastocyanin reduction by modified cytochromes c were measured at 0.1 M ionic strength (Augustin et al., 1983). The greatest decrease in the second order rate constant of plastocyanin reduction was 50%, which was observed for the 4-carboxy-2,6-dinitrophenyl-Lys 13 cytochrome c derivative. These previous studies are consistent with the current observations that chemical modifications that alter die surface electrostatic properties of the protein only decrease the bimolecular rate constant of electron transfer below I = 0.5 M and that the magnitude of this decrease is at most 70%. The pH independence of the bimolecular rate constant of ferricytochrome c reduction by either 106 D M E - or native ferrocytochrome b5 is not altogether surprising. Any differences in reactivity that might arise would be due to the titration of a group on one of the proteins that would alter the electrostatics of this protein and hence its capability to interact with the other protein. Electrostatic calculations indicate that His-26 of cytochrome bs, a residue found on die surface of the protein, is the only residue on either cytochromes b5 or c that titrates over the pH range 6.0 to 8.0 (Mauk et al., 1986). However, at the ionic strength of the current study (I = 0.5 M) diere was little difference between the rate constant of ferricytochrome c reduction by D M E - and native cytochromes b5, which differ in charge by -2. This finding suggests that it is unlikely diat a difference of a single charge would affect the rate constant of reduction of ferricytochrome c at I = 0.5 M . It was not possible to repeat the pH dependence at 0.1 M ionic strength as die reaction was too fast to monitor under these conditions. While it is not clear what the reduction potentials of the proteins are in the cytochrome ^cytochrome c complex, it is worth noting diat the reduction potentials of native and DME-cytochromes bs and cytochrome c, free in solution, do not change more than 10 mV between pH 6.0 and 8.0 (Margalit & Schejter, 1973; Reid et al., 1982; 1984). The thermodynamic activation parameters for ferricytochrome c reduction by native and D M E -ferrocytochrome b5 are very similar (I = 0.5 M , pH 7.0). This result suggests that the mechanism of ferrocytochrome b5 reduction of ferricytochrome c is not fundamentally altered by esterification of the cytochrome b5 heme propionates. With the notable exception of self-exchange reactions, the activation enthalpy and entropy of ferrocytochrome b5 reduction of ferricytochrome c fall within the values measured for other bimolecular protein-protein electron transfer reactions (Table XII). While the activation enthalpy of ferrocytochrome b5 reduction of ferricytochrome c is comparable to that of the bimolecular oxidation or reduction of either cytochromes b5 (Reid & Mauk, 1982; Reid et al., 1984; Chapman et al., 1984) or c (Hodges et al., 1974; Cummins & Gray, 1977; Mauk et al., 1979) with a variety of small electron transfer reagents, the activation entropy of the protein-protein reaction is considerably smaller than those of the protein-reagent reactions. The activation entropy for electron transfer between either cytochromes b5 or c and a small reagent varies from -11 e.u. for 107 ferricytochrome c oxidation by r^ntaanmiinepyridineruthenium(III) (I = 0.5 M , pH 5.3, 25°C) (Cummins & Gray, 1977) to -30 e.u. for DME-ferricytochrome b5 reduction by Fe(EDTA)" 2 (I = 0.1 M , pH 7.0, Table XII : Thermodynamic parameters of several second order protein-protein electron transfer reactions. reactants A H 1 (kcal/mol) A S t (e.u.) reference ferrocytochrome/ + cupriplastocyanin l d 10.5 11.0 Wood, 1974 ferrocytochrome/ + cupriplastocyanin2b 8.4 4.9 Niwaera / . , 1980 ferricytochrome / + cuproplastocyanin2b 8.6 3.9 ibid ferrocytochrome/ + cupriplastocyanin3 13.4 21.5 Tanakaer al., 1981 ferrocytochrome c + cupriplastocyanin8 7.6 - 2.4 King etal., 1985 ferrocytochrome c 5 5 1 + cupriazurin 4 0 7.8 - 1.1 Rosen & Pecht, 1976 ferricytochrome c 5 5 1 + cuproazurin 4 0 13.7 18.8 ibid ferrocytochrome c 5 5 1 + cupriazurin 5 d 18.1 31.0 Rosen etal., 1981 ferricytochrome c 5 5 1 + cuproazurin 5 d 15.2 22.5 ibid ferrocytochrome c + ferricytochrome c 8.4 - 11 Barbush & Dixon" ferrocytochrome b5 + ferricytochrome b5 5.5 - 23 Dixon et al., 1989 ferrocytochrome bs + ferricytochrome c e 7.5 - 0.3 this work DME-ferrocytochrome b5 + ferricytochrome c e 7.9 1.0 diis work Solution conditions were: a I = 0.1 M , pH 7.5; b I = 0.2 M , pH 7.0; C I not stated, pH 7.0; d I = 0.1 M , pH 7.0; e I = 0.5 M , pH 7.0. The sources of proteins isolated from non-mammalian tissues were: ^Petroselinum sativum; 2Brassica komatsuna; ^Raphanous sativus; ^Pseudomonas aeruginosa; ^Alcaligenes faecalis. In those instances when the errors in the thermodynamic activation parameters are reported, they are not greater than 1 unit, "unpublished. 108 25°C) (Reid et al., 1984). This result suggests that the principal reason diat the rate of electron transfer between cytochromes bs and c so greatly exceeds the rate of electron transfer between each of these proteins and small reagents is that the activation entropy of die protein-protein electron transfer reaction is so much greater. The large, negative activation entropies of electron transfer between small reagents and proteins has been partly explained in terms of the loss of translational and rotational degrees of freedom accompanying formation of the collisional complex of the reactants (Bennett, 1973). The magnitude of this effect has been estimated to be -13 e.u.. It is possible that in the case of ferrocytochrome b5 reduction of ferricytochrome c, these entropic losses could be compensated by the release of solvent molecules at the protein-protein interface in complex formation (Chothia & Janin, 1975). This interpretation is consistent with complex formation being the rate limiting step in ferrocytochrome b5 reduction of ferricytochrome c. Given the limited range of experimental data for protein-protein electron transfer reactions, it is premature to comment on the generality of this phenomenon. Nevertheless, the variation in the activation parameters of protein-protein electron transfer reactions is remarkable, even amongst similar proteins isolated from different sources (Table XII). It may be that different electron transfer protein pairs have evolved slightly different mechanisms to enhance their reactivities. The present results indicate that the second order rate constant for ferrocytochrome b5 reduction of ferricytochrome c is determined in part by the rate of complex formation which is largely controlled by electrostatic forces. At high ionic strength, die rate of electron transfer between cytochromes b5 and c was unaffected by esterification of the heme propionates of cytochrome bs, suggesting that the reduction potentials of native and DME-cytochromes b5 are similar when bound to cytochrome c. The principal effect of esterification of die cytochrome b5 heme propionates is to influence complex formation with cytochrome c, either by reducing the rate of protein-protein association or promoting the formation of complexes that are not conducive to electron transfer. The results are consistent with the participation of cytochrome b5 heme propionate-6 in an intermolecular salt bridge that stabilises the cytochrome ^cytochrome c complex. 109 C O N C L U S I O N The work outlined in this dissertation has evaluated several aspects of the interaction and kinetics of electron transfer between cytochrome b5 and cytochrome c. Particular emphasis has been given to consideration of the influence of the cytochrome b5 heme propionate groups in these phenomena and on the use of porphyrin cytochrome c as a fluorescent analogue for cytochrome c in such investigations. In addition, the dependence of the mechanism of protein-protein interaction and electron transfer on solution conditions (pH, ionic strength, temperature) has been evaluated for these proteins. The fluorescence quenching studies on the interaction of porphyrin cytochrome c and cytochrome b5 combined with the kinetics of flavin semiquinone reduction of the cytochromes within the cytochrome i 5cytochrome c complex provide evidence that esterification of the cytochrome b5 heme propionate groups influences the docking geometry of these two proteins. These findings extend the conclusions obtained through electrostatic calculations (Mauk et al., 1986) by establishing two principal points. First, the distance of separation between the two heme groups in the cytochrome ^cytochrome c complex is detectably increased when the cytochrome bs heme propionate groups are esterified. Second, die cytochrome c heme group is not as sterically hindered in the complex formed by this protein with DME-cytochrome b5 as in the complex formed between the two native proteins. The stopped-flow studies demonstrate that the bimolecular rate constant for ferrocytochrome bs reduction of ferricytochrome c is determined in part by the rate of association of the two proteins. This rate of association is strongly influenced by electrostatic forces. The principal effect of cytochrome b5 heme propionate esterification on the bimolecular rate constant of electron transfer between cytochromes bs and c is to influence complex formation between diese two proteins. The stopped-flow studies suggest that the reduction potentials of native and DME-cytochromes bs are not significantly different when these proteins are bound to cytochrome c. 110 Much of the literature describing complexes formed between electron transfer proteins treats such complexes as discrete and structurally definable entities. The overall picture that emerges from this work is one in which these protein-protein complexes may exist in a variety of structurally similar orientations. Furthermore, it appears that the relative population of a given complex geometry is influenced by surface modifications on either of the interacting proteins and by solution conditions. It seems likely that a truly rigorous characterization of the multiple equilibria that define protein-protein interactions of this type and of the structures of the complexes involved will require a major investigative effort combining the use of specifically mutated proteins, computational and graphics techniques and a variety of physical techniques. I l l REFERENCES Ahmad, I., Cusanovich, M . A . & Tollin, G . (1982) Biochemistry, 21, 3122 - 3128. Argos, P. & Mathews, F . S. (1975)/. Biol. Chem., 250, 747 - 751. Armstrong, F . A . , Driscoll, P. C . & H i l l , H . A . 0. (1985) FEBSLett., 190, 242 - 248. Augustin, M . A . , Chapman, S. K . , Davies, D . M . , Sykes, A . G . , Speck, S. H . & Margoliash, E . (1983) / . Biol. Chem., 258, 6405 - 6409. Barbush, M . & Dixon, D . W . (1989) Biochem. Biophys. Res. Commun. in press. Bendzko, P. & Pfeil, W. (1983) Biochim. Biophys. Acta, 742, 669 - 676. Bennett, L . E . (1973) Prog. Inorg. Chem., 18, 1 -176. Berman, M . C , Adnams, C. M . , Ivanetich, K . M . & Kench, J. E . (1976) Biochem. J., 157, 237 - 246. Bernardi, P. & Azzone, G . F. (1981) / . Biol. Chem., 256, 7187 - 7192. Bernstein, F . C , Koetzle, T. F . , Williams, G . J . B . , Meyer, E . F . , Bruce, M . D . , Rodgers, J. R., Kennard, O. , Shimanouchi, T. & Tasumi, M . (1977)/. Mol. Biol., 112, 535 - 542. Bhattacharyya, A . K . , Meyer, T. E . & Tollin, G . (1986) Biochemistry, 25, 4655 - 4661. Bhattacharyya, A . K . , Meyer, T. E . , Cusanovich, M . A . & Tollin, G . (1987) Biochemistry, 26, 758-764. Bison, R. & Capaldi, R. A . (1981)/. Biol. Chem., 256, 4362 - 4367. Bixon, M . & Jortner, J . (1986) FEBSLett., 200, 303 - 308. Bradford, M . M . (1976) Anal. Biochem., 72, 248 - 254. Brautigan, D . L . , Ferguson-Miller, S. & Margoliash, E . (1978a)/. Biol. Chem., 253, 130 - 139. Brautigan, D . L . , Ferguson-Miller, S., Tarr, G . E . & Margoliash, E . (1978b) / . Biol. Chem., 253, 140-148. Brautigan, D . L . , Ferguson-Miller, S. & Margoliash, E . (1978c) Meth. of Em.., 53, 128 - 164. Briquet, M . , Purnelle, B . , Beattie, D . S. & Goffeau, A . (1982) Eur. /. Biochem., 127, 329 - 342. Bushnell, G . , Louie, G . V . & Brayer, G . B. (1989) in preparation. Cambrillau, C , Frey, M . , Mosse, J . , Guerlesquin, F. & Bruschi, M . (1988) Proteins, 4, 63 - 70. Chapman, S. K . , Davies, D . M . , Vuik, C . P. J. & Sykes, A . G . (1984)/ . Am. Chem. Soc, 106, 2692-2696. 112 Cheung, E . , Taylor, K . , Kornblatt, J. A . , English, A . M . , McLendon, G . & Miller, J. R. (1986) Proc. Natl. Acad. Sci. U.S.A., 83, 1330 - 1333. Cheung, E . & English, A . M . (1988) Inorg. Chem., 27, 1078 - 1081. Chothia, C . & Janin, J. (1975) Nature, 256, 705 - 708. Churg, A . K., Weiss, R. M . ( Warshel, A . & Takano, T. (1983)/. Phys. Chem., 87, 1683 - 1694. Concar, D . W . , H i l l , H . A . O. , Moore, G . R. , Whitford, D . & Williams, R. J. P. (1986) FEBS Lett., 206, 15 - 19. Connolly, M . L . (1983) Science, 221, 709 - 713. Corin, A . F . , Bersohn, R. & Cole, P. E . (1983) Biochemistry, 22, 2032 - 2038. Cristiano, R. J. & Steggles, A . W. (1989) Nucl. Acids Res., 17, 799 - 80. Cummins, D . & Gray, H . B. (1977)/. Am. Chem. Soc., 99, 5158 - 5167. Dailey, H . A . & Strittmatter, P. (1979)/. Biol. Chem., ISA, 5388 - 5396. Dailey, H . A . & Strittmatter, P. (1980)/. Biol. Chem., 255, 5184 - 5189. Debye, P. & Huckel, E . (1923) Phys. Z., 24, 185 - 206. Devault, D . & Chance, B . (1966) Biophys. J., 6, 825 - 827. Devault, D . (1984) Quantum - Mechanical Tunnelling in Biological Systems, Cambridge University Press, London. Dickerson, R. E . , Takano, T., Eisenberg, D . , Kallai, O. B . , Samson, L . , Cooper, A . & Margoliash, E. (1971)/. Biol. Chem., 246, 1511 - 1535. Dixon, D . W . , Hong, X . & Woehler, S. E . (1989) Biophysical J., in press. Dixon, D . W . , Hong, X . , Woehler, S. E . , Mauk, A . G . & Sishta, B. P. (1989) submitted. Dowe, R. J . , Vitello, L . B . & Erman, J . E . (1984) Arch. Biochem. Biophys., 232, 566 - 573. Draper, R. D . & Ingraham, L . I. (196S) Arch. Biochem. Biophys., 125, 802 - 808. Duggleby, R. G . (1984) Comput. Biol. Med., 14, 447 - 455. Eley, C . G . S. & Moore, G . R. (1983) Biochem. / . , 215, 11-21. Eltis, L . , Mauk, A . G . , Hazzard, J. T., Cusanovich, M . A . & Tollin, G . (1988) Biochemistry, 27, . Erman, J . E . & Vitello, L . B. (1980)/. Biol. Chem., 225, 6224 - 6227. Erman, J . E . , K i m , K . L . , Vitello, L . B . , Moench, S. J . & Satterlee, J. D . (1987) Biochem. Biophys. Acta, 911, 1 - 10. 113 Falk (1964) Porphyrins and Metalloporphyrins, Elsevier/North-Holland Press, The Netherlands. Feng, Y . , Roder, H . , Englander, S. W. , Wand, A . J. & D i Stefano, D . L . , (1989) Biochemistry, 28, 195 - 203. Ferguson-Miller, S., Brautigan, D . L . & Margoliash, E . (1978)/. Biol. Chem., 253, 149 - 159. Fisher, M . T. , White, R. E . & Sligar, S. G . (1986)/. Am. Chem. Soc, 108, 6835 - 6837. Fisher, W . R. , Taniuchi, H . & Anfinsen, C . B . (1973)/. Biol. Chem., 248, 3188 - 3195. Flatmark, T. & Robinson, A . R. (1968) Structure and Function of Cytochromes (Okunuki, K . et al. Eds) University of Tokyo Press, Tokyo, 318 - 327. Forster, T. (1948) Ann. Phys. (Leipzig), 2, 55 - 75. Geren, L . M . , Stoneheurner, J . , Davis, D . M . , & Millet, F . (1983) Biochim. Biophys. Acta, 724, 62-68. Gibson, Q. H . & Milnes, L . (1964) Biochem. / . , 91, 161 -171. Grinstead, G . F . &Gaylor , J. L . (1982)/. Biol. Chem., 1ST, 13937 - 13944. Guerlesquin, F . , Sari, J . -C. & Bruschi, M . (1987) Biochemistry, 26, 7438 - 7443. Gupta, R. K . (1973) Biochim. Biophys. Acta, 292, 291 - 295. Guzzo, A . V . & Tollin, G . (1963) Arch. Biochem. Biophys., 103, 231 - 243. Hacked, C . S. & Strittmatter, P. (1984)/. Biol. Chem., 259, 3275 - 3282. Hartshorn, R. T., Mauk, M . R., Mauk, A . G . & Moore, G. R. (1987) FEBSLett., 213, 391 - 395. Hazzard, J . T. & Tollin, G . (1985) Biochem. Biophys. Res. Commun., 130, 1281 - 1286. Hazzard, J. T. , Cusanovich, M . A . , Tainer, J. A . , Getzoff, E . D . & Tollin, G . (1986) Biochemistry, 25, 3318-3328. Hazzard, J . T . , Poulos, T. L . & Tollin, G . (1987) Biochemistry, 26, 2836 - 2848. Hazzard, J . T. , McLendon, G . , Cusanovich, M . & Tollin, G . (1987) Biochem. Biophys. Res. Commun. in press. Hazzard, J . T., Moench, S. J . , Erman, J. E . , Satterlee, J. D . , & Tollin, G . (1988) Biochemistry, 27, 2002 - 2008. Heiler, D . , Rogalsky, P. & McLendon, G . (1987) / . Am. Chem. Soc., 109, 604 - 606. Hodges, H . L . , Holwerda, R. A . & Gray, H . B . (1974)/. Am. Chem. Soc, 96, 3132 - 3137. Hoffman, B. M . & Ratner, M . A . (1987)/. Am. Chem. Soc, 109, 6237 - 6243. 114 Hollway, P. W. & Mantsch, H . H . (1988) Biochemistry., 27, 7991 - 7993. Hopfield, J. J . (1974) Proc. Nat. Acad. Sci., 71, 3640 - 3644. Hoth, L . R. & Erman, J . E . (1984) Biochim. Biophys. Acta, 788, 151 - 153. Hultquist, D . E . & Passon, P. G . (1971) Nature New Biol., 229, 252-254. Hwang, P. K . & Greer, J. (1979)/. Biol. Chem., 254, 2265-2270. IUPAC-IUB Commission on Biochemical Nomenclature (1971). (1972)/. Biol. Chem., 247, 977 - 983. Johnson, J . L . & Rajagopolan, K . V . (1977) / . Biol. Chem., 252,2017 - 2025. Johnson, J. L . & Rajagopolan, K . V . (1980) CIBA Found. Symp.,12, 119 - 133. Jortner, J. (1976)/. Chem. Phys., 64, 4860 - 4867. Kang, C . H . , Ferguson-Miller, S. & Margoliash, E . (1977)/. Biol. Chem., 252, 919 - 926. Kang, C . H . , Brautigan, D . L . , Osheroff, N . & Margoliash, E . (1978)/. Biol. Chem., 253, 6502 - 6510. Keller, R. M . , Wuthrich, K . & Pecht, I. (1986) FEBSLett., 70, 180 -184. Keller, R. M . & Wuthrich, K . (1980) Biochim. Biophys. Acta, 621, 204 - 217. Keilin, D . (1966) The History of Cell Respiration and Cytochrome, Cambridge University Press, London. King, G . C , Binstead, R. A . & Wright, P. E . (1985) Biochim. Biophys. Acta, 806, 262 - 271. Koloczek, H . , Horie, T. , Yonetani, T., Anni, H . , Maniara, G . & Vanderkooi, J. M . (1987) Biochemistiy, 26,3142 - 3148. Konig, B . W . , Osheroff, N . , Wilms, J . , Muijsers, A . O. , Dekker, H . L . & Margoliash, E. (1980) FEBS Lett., I l l , 395 -398. Koppenol, W . H . & Margoliash, E . (1982)/. Biol. Chem., 257, 4426 - 4437. Kornblatt, J . A . & English, A . (1986) Eur. J. Biochem., 155, 505 - 511. Kornblatt, J . A . & Laberge, M . (1988) Eur. J. Biochem., 175, 475 - 479. Korszun, Z . R. , Moffat, K . , Frank, K . & Cusanovich, M . (1982) Biochemistry, 21, 2253 - 2258. Labeyrie, F . , Beloeil, J . C . & Thomas, M . - A . (1988) Biochim. Biophys. Acta, 953, 134 - 141. Lakowicz, J . R. (1983) Principles of Fluorescence Spectroscopy, Plenum Press, New York. La Mar, G . N . , Burns, P. D . , Jackson, J. T. , Smith, K . M . , Langry, K . C. & Strittmatter, P. (1981) / . Biol. Chem., 256, 6075 - 6079. 115 Larsson, S. (1983)/. Chem. Soc, Faraday Trans., 79, 1375 - 1388. Le, K . H . D . & Lederer, F . (1983) The EMBO Journal, 2, 1909 - 1914. Lederer, F . , Ghrir, R. , Guiard, B . , Cortial, S. & Ito, A . (1983)Eur. J. Biochem., 132, 95 - 102. Lemberg, R. & Barrett, J. (1973) Cytochromes, Academic Press, London. Leonard, J . J . & Yonetani, T. (1974) Biochemistry, 13, 1465 - 1468. Leung, P.-S. K . , Betterton, E . A . & Hoffmann, M . R. (1989)/. Phys. Chem., 93, 430 - 433. Liang, N . , Mauk, A . G . , Pielak, G . J . , Johnson, J. A . , Smith, M . & Hoffman, B . M . (1988) Science, 240,311-313. Louie, G . V . , Hutcheon, W. L . B . & Brayer, G . D . (1988)/. Mol. Biol., 199, 295 - 314. Loverde, A . & Strittmatter, P. (1968)/. Biol. Chem., 243, 5779 - 5787. MacMunn, C. A . (1884)/. Physiol. (London), 5, X X I V . Marcus, R. A . & Sutin, N . (1985) Biochim. Biophys. Acta, 811, 265 - 322. Margalit, A . & Schejter, R. (1973) Eur. J. Biochem., 32, 492 - 499. Margoliash, E . & Frohwirt, N . (1959) Biochem. J., 71, 570 - 572. Margoliash, E . , Smith, E . L . , Kreil , G . &Tuppy, G . (l96l)Nature (London), 192, 1121 - 1127. Margoliash, E . & Bosshard, H . R. (1983) TIBS, 316 - 320. Mathews, F . S., Czerwinski, E . W. & Argos, P. (1979) in The Porphyrins, vol V l l b (ed D. Dolphin), pp 107 - 147, Academic Press, New York. Mathews, F . S. (1985) Prog. Biophys. Molec. Biol., 45, 1 - 56. Matthew, J. B . (1978) PhD Dissertation, Indiana University. Mauk, A . G . , Coyle, C . L . , Bordignon, E . & Gray, H . B. (1979)/. Am. Chem. Soc, 101, 5054 - 5056. Mauk, M . R. , Reid, L . S. & Mauk, A . G. (1982) Biochemistry, 21, 1843 - 1846. Mauk, M . R. , Reid, L . S. & Mauk, A . G . (1984) Biochem. J., 111, 297 - 302. Mauk, M . R. & Mauk, A . G . (1982) Biochemistry, 21, 4730 - 4734. Mauk, M . R. , Mauk, A . G . , Weber, P. C. & Matthew, J . B. (1986) Biochemistry, 25, 7085 - 7091. Mauk, M . R. & Mauk, A . G . (1989) submitted. Mayo, S. L . , Ellis, W. R. , Crutchley, R. J. & Gray, H . B. (1986) Science, 233, 948 - 952. 116 McGlynn, S. P . , Azumi, T. & Kasha, M . (1964)/. Chem. Phys., 40, 507 - 515. McGourty, J . , Blough, N . & Hoffman, B . (1983)/. Am. Chem. Soc., 105, 4470 - 4472. McLachlan, S. J . , La Mar, G . N . & Sletten, E . (1986)/. Am. Chem. Soc., 108, 1285 - 1291. McLendon, G . L . , Winkler, J . R. , Nocera, D . G . , Mauk, M . R. , Mauk, A . G . & Gray, H . B. (1985) / . Am. Chem. Soc., 107, 739 - 740. McLendon, G . L . & Mil ler , J . R. (1985)/. Am. Chem. Soc., 107, 7811 - 7816. McLendon, G . L . , Pardue, K . & Bak, P. (1987)/. Am. Chem, Soc., 107, 7540 - 7541. McLendon, G . L . (1988) Acc. Chem. Res., 21, 160 -167. Meyer, T. E . , Watkins, J. A . , Przyslecki, C . T., Tollin, G . & Cusanovich, M . A . (1984) Biochemistry, 23, 4761 - 4767. Michel, B . & Bosshard, H . R. (1984)/. Biol. Chem., 259, 10085 - 10091. Michel, B . , Mauk, A . G . & Bosshard, H . R. (1989) FEBSLett., 243, 149 - 152. Michel, B . , Proudfoot, A . E . I., Wallace, C . J. A . & Bosshard, H . R. (1989) Biochemistry, 28, 456-462. Miller , J . R. (1975) Science, 189, 221 - 224. Moench, S. J . , Satterlee, J . D . & Erman, J. E . (1987) Biochemistry, 26, 3821 - 3826. Mochan, E . & Nicholls, P. (1971) Biochem. / . , 121, 69 - 82. Moore, G . R. & Williams, R. J. P. (1977) FEBSLett., 79, 229 - 232. Moore, G . R. , Robinson, M . N . , Williams, G . & Williams, R. J. P. (1985)/. Mol. Biol., 183, 429 - 446. Moore, G . R. , Petdgrew, G . W . , & Rogers, N . K. (1986) Proc. Natl. Acad. Sci. U.S.A., 83, 4998 - 4999. Moore, J . W. & Pearson, R. G . (1981) Kinetics and Mechanisms J. Wiley & Sons, New York. Morton, M . R. , Overnell, J . & Harburg, H . A . (1970)/. Biol. Chem., 245, 4653 - 4657. Nagy, J . , Knoll , J . , Kenney, W. C. & Singer, T. P. (1982) in Flavins and Flavoproteins, (eds V . Massey & C . H . Williams) Elsevier North Holland, Inc., New York, 573 - 578. Nicholls, P. (1974) Biochem. Biophys. Acta, 346, 261 - 310. Niwa, S., Ishikawa, H . , Nikai , S. & Takabe, T. (1980)/. Biochem., 88, 1177 - 1183. Ng , S., Smith, M . B . , Smith, H . T. & Millet, F . (1977) Biochemistry, 16, 4975 - 4978. Northrup, S. H . , Reynolds, J. C . L . , Miller, C . M . , Forrest, K . J. & Boles, J. O. (1986) / . Am. Chem. 117 Soc, 108, 8162-8170. Northrup, S. H . , Boles, J . O. & Reynolds, J. C . L . (1988) Science, 1A\, 67 - 70. Ochi, H . , Hata, Y . , Tanaka, N . , Kakudo, M . , Sakuri, T., Aihara, S. & Morita, Y . (1983) J. Mol. Biol., 166, 407 - 418. Osheroff, N . , Brautigan, D . L . & Margoliash, E . (1980)/. Biol. Chem., 255, 8245 - 8251. Ozols, J. & Strittmatter, P. (1964)/. Biol. Chem., 239, 1018 -1023. Ozols, J. (1975) in First International Conference on Solid Phase Methods in Protein Sequence Analysis (Laursen, R. H . , ed) Pierce Chem. Co. , pp. 221 - 228. Parson, W . W . , Schence, C . C , Blankenship, R. E . , Holton, D . , Windsor, M . W . & Shank, C . V . , (1978) Frontiers of Biological Energetics, 1, Academic Press, New York, pp 37 - 44. Pecht, I. & Rosen, P. (1973) Biochem. Biophys. Res. Commun., 50, 853 - 858. Peerey, L . M . & Kostic, N . M . (1989) Biochemistry, 28, 1861 - 1868. Peterson, J . A . & Prough, R. A . (1986) Cytochrome P-450: Structure, Mechanism and Biochemistiy, (Ortiz de Montellano, P. ed) Plenum Press, New York, p. 89. Pettigrew, G . W . (1978) FEBS Lett., 86, 14 - 16. Pettigrew, G . W. & Moore, G . R. (1987) Cytochrome c: Biological Aspects, Springer Verlag, Berlin. Pielak, G . J . , Mauk, A . G . & Smith, M . (1985) Nature, 313, 152 - 154. Poulos, T. L . & Kraut, J . (1980) / . Biol. Chem., 255, 10322 -10330. Poulos, T. L . & Mauk, A . G . (1983)/. Biol. Chem., 258, 7369 - 7373. Przysiecki, C . T. , Tollin, G . , Meyer, T. E . , Staggers, J . E . & Cusanovich, M . A . (1985) Arch. Biochem. Biophys., 238, 334 - 343. Rees, D . C . (1985) Proc. Natl. Acad. Sci., 82, 3082 - 3085. Reid, L . S. & Mauk, A . G . (1982)/. Am. Chem. Soc, 104, 841 - 845. Reid, L . S., Taniguchi, V . T., Gray, H . B. & Mauk, A . G . (1982)/. Am. Chem. Soc, 104, 7516 - 7519. Reid, L . S., Mauk, M . R. & Mauk, A . G . (1984)/. Am. Chem. Soc, 106, 2182 - 2185. Reid, L . S. (1984) Ph. D. Dissertation, University of British Columbia. Reider, & Bosshard, H . (1980) / . Biol. Chem., 255, 4732 - 4739. Reynolds, W. L . & Lumry, R. W. (1966) Mechanisms of Electron Transfer, Ronald Press, New York. Rodgers, K . K . , Pochapsky, T. C. & Sligar, S. G. (1988) Science, 240, 1657 - 1659. 118 Rosen, P. & Pecht, I. (1976) Biochemistry, 15, 775 - 786. Rosen, P . , Segal, M . & Pecht, I. (1981) Eur. J. Biochem., 120, 339 - 344. Ross, P. D . & Subramanian, S. (1981) Biochemistry, 20, 3096 - 3102. Rush, J . D . , Lan, J. & Koppenol, W. H . (1987)/. Am. Chem. Soc., 109, 2679 - 2682. Rush, J . D . & Koppenol, W . H . (1988) Biochim. Biophys. Acta, 936, 187 - 198. Salemme, F . R. (1976)/. Mol. Biol., 102, 563 - 568. Satterlee, J . D . , Moench, S. J . & Erman, J. E . (1987) Biochim. Biophys. Acta, 912, 87 - 97. Schenkein, I., Levy, M . , Franklin, E . C. & Frangione, B . (1977) Arch. Biochem. Biophys. 182, 875-884. Scott, R. A . (1980)/7». D. Dissertation, California Institute of Technology. Scott, R. A . , Mauk, A . G . & Gray, H . B. (1985) / . Chem. Ed., 62, 932 - 937. Shire, S. J . , Hanania, G . H . & Gurd, F . R. N . (1974) Biochemistry, 13, 2967 - 2974. Simondsen, R. P . , Weber, P. C , Salemme, F. R. & Tollin, G . (1982) Biochemistry, 21, 6366 - 6375. Simondsen, R. P. & Tollin, G . (1983) Biochemistry, 22, 3008 - 3016. Slaughter, S. R. , & Hultquist, D . E . (1979)/. Cell Biol., 83, 231 -239. Smith, K . M . (1975) Porphyrins and Metalloporphyrins, Elsevier/North-Holland Press, The Netherlands. Smith, M . B . , Stoneheurner, J . , Ahmed, A . J . , Staudenmayer, N . & Millet, F . (1980) Biochim. Biophys. Acta, 592, 303 -313. Speck, S. H . , Ferguson-Miller, S., Osheroff, N . & Margoliash, E . (1979) Proc. Natl. Acad. Sci. U.S.A., 76, 155 - 159. Speck, S. H . , Koppenol, W. H . , Dethmers, J. K . , Osheroff, N . , Margoliash, E . & Rajagopalan, K. V . (1981)/. Biol. Chem., 256, 7394- 7400. Stayton, P. S., Fisher, M . T. & Sligar, S. G . (1988)/. Biol. Chem., 263, 13544 - 13548. Steinberg, I. Z . (1971) Ann. Rev. Biochem., 40, 83 - 114. Stellwagen, E . (1978) Nature, 275, 73 - 74. Stewart, D . E . , LeGall, J . , Moura, I., Moura, J . J . G . , Peck, H . D . , Xavier, A . V . , Weiner, P. K. & Wampler, J . E . (1988) Biochemistry, 27,2444 - 2450. Stone, H . W . & Hume, D . N . (1939) Anal. Chem. 11, 598 - 602. 119 Stonehuerner, J . , Williams, J . B . & Millet, F . (1979) Biochemistry, 18, 5422 - 5427. Strickland, S., Palmer, G . & Massey, V . (1975)/. Biol. Chem., 250, 4048 - 4052. Strittmatter, P. (1964) in Rapid Mixing and Sampling in Biochemistry, (Chance, B . , Eisenhardt, R., Gibson, Q. & Lundber Holm, K . Eds), Academic Press, New York, 71-84 . Strittmatter, P. Spatz, L . , Corcoran, D . , Rogers, M . J . , Setlow, B. & Redline, R. (1974) Proc. Natl. Acad. Sci. U.S.A., 71, 4565 - 4569. Strottman, J . M . , Stellwagen, A . , Bryant, C . & Stellwagen, E . (1984)/. Biol. Chem., 259, 6931 - 6936. Summers, F . E . & Erman, J . E . (1988)/. Biol. Chem., 263, 14267 - 14275. Takano, T. & Dickerson, R. E . (1980) Proc. Natl. Acad. Sci. U.S.A., 77, 6371 - 6375. Takano, T. & Dickerson, R. E . (1981a)/. Mol. Biol, 153, 79 - 94. Takano, T. & Dickerson, R. E . (1981b)/. Mol. Biol, 153, 95 - 115. Tamburini, P. B . & Schenkman, J. B. (1988),4rcfc. Biochem. Biophys., 245, 512 - 522. Tanford, C. & Kirkwood, J. G . (1957)/. Am. Chem. Soc., 79, 5333 - 5339. Tanford, C . (1980) in The Hydrophobic Effect, John Wiley and Sons, New York. Thomas, M . - A . , Favaudon, V . & Pochon, F . (1983a) Eur. J. Biochem., 135, 569 - 576. Thomas, M . - A . , Gervais, M . , Favaudon, V . & Valat, P. (1983b) Eur. J. Biochem., 135, 577 - 581. Tollin, G . , Cheddar, G . , Watkins, J . A . , Meyer, T. E . & Cusanovich, M . A . (1984) Biochemistry, 23, 6345 - 6349. Tollin, G . , Meyer, T. E . & Cusanovich, M . A . (1986) Biochim. Biophys. Acta, 853, 29 - 41. Tollin, G . , Brown, K . , De Francesco, R. & Edmondson, D . E . (1987) Biochemistry, 26, 5042 - 5048. Utech, R. E . & Kurtz, D . M . (1985) Inorg. Chem., 24, 4458 - 4459. Utech, R. E . & Kurtz, D . M . (1988) Biochim. Biophys. Acta., 953, 164 - 178. Vaish, S. P. & Tollin, G . (1970) Bioenergetics, 1, 181 - 192. Vanderkooi, J . M . & Erecinska, M . (1974) Arch. Biochem. Biophys., 162, 385 - 391. Vanderkooi, J . M . , Adar, F . & Erecinska, M . (1976) Eur. J. Biochem., 64, 381 - 387. Vanderkooi, J . M . , Landesberg, R. , Hayden, G . W. & Owen, C . S. (1977) Eur. J. Biochem., 81, 339-347. Vanderkooi, J. M . , Glatz, P. , Casadei, J. & Woodrow, G . V . (1980) Eur. J. Biochem., 110, 189 - 196. 120 Van Leeuwen, J . W. (1983) Biochim. Biophys. Acta, 743, 408 - 421. Veitch, N . C , Concar, D . W . , Williams, R. J. P. & Whitford, D . (1988) FEBSLett., 21%, 49 - 55. Vitello, L . B . & Erman, J . E . (1987) Arch. Biochem. Biophys., 258, 621 - 629. von Bodman, S. B . , Schuler, M . A . , Jollie, D . R. & Sligar, S. G . (1986) Proc. Natl. Acad. Sci. U.S.A., 83, 9443 - 9447. Waldmeyer, B . & Bosshard, H . R. (1985)/. Biol. Chem., 260, 5184 - 5190. Wand, A . J . , D i Stefano, D . L . , Feng, Y . , Roder, H . & Englander, S. W . (1989) Biochemistry, 28, 186-194. Warshel, A . & Churg, A . K . (1983)/. Mol. Biol., 168, 693 - 697. Weber, C , Michel, B . & Bosshard, H . R. (1987) Proc. Natl. Acad. Sci., U.S.A., 84, 6687 - 6691. Wendoloski, J . J . , Matthew, J . B . , Weber, P. C . & Salemme, F . R. (1987) Science, 238, 794 - 797. Wherland, S. & Gray, H . B . (1976) Proc. Natl. Acad. Sci. U.S.A., 73, 2950 - 2954. Wherland, S. & Pecht, I. (1978) Biochemistry, 13, 2585 - 2591. Williams, G . , Moore, G . R. , Porteous, R., Robinson, M . N . , Soffe, N . & Williams, R. J. P. (1985a) / . Mol. Biol., 183,409-428. Williams, G . , Clayden, Moore, G . R. & Williams, R. J . P. (1985b)/. Mol. Biol., 183, 447 - 460. Wilson, M . T. , Greenwood, C , Brunori, M . & Antonini, E . (1975) Biochem. / . , 145, 449 - 457. Wood, P. M . (1974) Biochim. Biophys. Acta, 357, 370 - 379. Yonetani, T. & Anni , H . (1987)/. Biol. Chem., 262, 9547-9554. Yoshimura, T. , Sogabe, T. & A k i , K . (1981) Biochim. Biophys. Acta, 636, 129 - 135. 121 A P P E N D I X A Fluorescence Anisotropy The purpose of these calculations is to detennine the ratio of the volumes of the porphyrin cytochrome ccytochrome bs complex to porphyrin cytochrome c free in solution. The experimental value used in these calculations is the fluorescence anisotropy, r, of the species. Fluorescence anisotropy is defined by equation 12 where Ip is the intensity of the emission when the observing Ip-Io r = [12] Ip + 2I C polarizer is parallel to the direction of the polarized excitation and I 0 is the intensity of the emission when the observing polarizer is perpendicular to the direction of the polarized excitation. I 0 is corrected for the inability of the spectrofluorometer to detect I p and I c with equal efficiency. Anisotropy is related to the rotational correlation time, and the fluorescence lifetime, r , of the fluorophore by the Pen-in equation (13). R 0 r = [13] 1 + T / $ The rotational correlation time of the fluorophore is related to the volume, V , of the molecule in which the fluorophore is located by equation 14 where n is the viscosity of the solution (.00894 flV $ = — • [14] k B T poise), k B is Boltzmann's constant and T is the temperature. The relative volumes of free and complexed porphyrin cytochrome c wil l be the ratio of the rotational correlation lifetimes. The rotational correlation of free porphyrin cytochrome c, 3>f, may be calculated by first estimating the volume of the molecule. Approximating the molecule as a sphere of radius 16 A , $ f evaluates to 3.7 ns. The fluorescence lifetime of porphyrin cytochrome c, r f , is 6.5 ns (Vanderkooi, 19). Knowing the maximal steady state fluorescence quenching ( Q m a x = 0.25) of porphyrin cytochrome c by bound cytochrome b5, the fluorescence lifetime of bound porphyrin cytochrome c, T c , is calculated to be 4.9 ns by the equation 15. r f Q = — [15] Assuming that the orientation of the dipole in the free and complexed porphyrin cytochrome c is the same (i.e., r 0 is unaffected by cytochrome b5 binding) then the rotational correlation time of bound porphyrin cytochrome c may be calculated by taking the ratio of die Perrin equation of bound and free protein (equation 16). This yields a value of 8.5 ns. 122 r c 1 + 7 y $ f r f 1 + rc/<J>c The ratio of the rotational correlation times, and hence the volumes is 2.3. 123 A P P E N D I X B Second order rate constants for ferrocytochrome b5 reduction of ferricytochrome c Cytochrome bs pH T ( ° C ) I ( M ) k (x 10-6 M ' V 1 ) s.e.m.* 7.0 25.0 0.1 540 90 7.0 25.0 0.2 95 4 7.0 25.0 0.3 47 4 7.0 25.0 0.5 18.1 1.5 7.0 25.0 1.0 8.2 .8 6.0 25.0 0.5 18.3 .9 6.5 25.0 0.5 17.6 .9 7.5 25.0 0.5 18.7 .8 8.0 25.0 0.5 20.0 1.6 7.0 3.9 0.5 6.2 1.8 7,0 9.6 0.5 8.4 .7 7.0 15.1 0.5 11.3 • 1.1 7.0 32.3 0.5 24.0 .4 DME-Cytochrome b5 pH T ( ° C ) I ( M ) k (x 1Q-6 M ' V 1 ) 7.0 25.0 0.1 157 26 7.0 25.0 0.2 51 2 7.0 25.0 0.3 27 1 7.0 25.0 0.5 16.5 .6 7.0 25.0 1.0 10.7 .6 6.0 25.0 0.5 15.2 .1 6.5 25.0 0.5 15.7 .2 7.5 25.0 0.5 16.5 .5 8.0 25.0 0.5 17.8 .4 7.0 4.1 0.5 5.5 .3 7.0 10.1 0.5 7.1 1.5 7.0 14.8 0.5 9.9 .6 7.0 19.9 0.5 13.1 1.4 7.0 31.5 0.5 21.2 .8 'Standard error of the mean 124 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0098188/manifest

Comment

Related Items