Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The akaline conformational transitions of ferricytochrome c Rosell, Federico Ignacio 1999

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata

Download

Media
ubc_1999-389669.pdf [ 8.48MB ]
Metadata
JSON: 1.0089257.json
JSON-LD: 1.0089257+ld.json
RDF/XML (Pretty): 1.0089257.xml
RDF/JSON: 1.0089257+rdf.json
Turtle: 1.0089257+rdf-turtle.txt
N-Triples: 1.0089257+rdf-ntriples.txt
Original Record: 1.0089257 +original-record.json
Full Text
1.0089257.txt
Citation
1.0089257.ris

Full Text

THE ALKALINE CONFORMATIONAL TRANSITIONS OF FERRICYTOCHROME c by Federico Ignacio Rosell B.Sc, The University of British Columbia, 1990 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES THE DEPARTMENT OF BIOCHEMISTRY AND MOLECULAR BIOLOGY We accept this thesis as conforming to the required standard The University of British Columbia January 1999 ® Federico Ignacio Rosell, \W 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 The University of British Columbia Vancouver, Canada DE-6 (2/88) ABSTRACT The alkaline conformational transitions of several yeast iso-l -ferricytochrome c variants were characterized with a variety of spectroscopic and physical techniques. The first group of variants consists of cytochromes in which selected lysyl residues were replaced by alanine. On the basis of *H-NMR and EPR spectroscopy, it was shown that Lys73 is one of the residues that replaces Met80 as an axial ligand to the heme iron and results in the alkaline conformer IVa. Lys79 had been shown previously to coordinate the metal centre in a second alkaline conformer IVb (Ferrer, J.C., Guillemette, J.G.,Bogumil,R.,Inglis, S.C, Smith,M.&Mauk, AG. (1993)J. Am. Chem. Soc., 115, 7507-7508). The absence of both lysyl residues from the cytochrome sequence (i.e., Lys73Ala/ Lys79Ala double variant) results in the stabilization of the native conformation of the protein which persists at higher pH (pATa=10.5(3)). However, addition of amino group-containing compounds to solutions of this variant results in a protein adduct with spectroscopic features similar to those of the alkaline isomers. Furthermore, the conformational transition of the wild-type protein (pK=S. 70(2)) represents a weighted average of the transition of the native Met80-bound conformer to either the Lys73- (p£a=8.44(l)) or Lys79-bound (pATa=8.82(2)) conformers. pH-jump kinetics indicate that replacement of Met80 by Lys73 occurs with a greater rate constant (£f=160(5) s'1) than the replacement by Lys79 (kf=\ .51(5) s'1) despite the more significant structural changes that must take place on the distal side of the protein to accommodate coordination of the heme iron by Lys73. This rearrangement of the molecule, however, does not change significantly the exposure of the heme group to solvent relative to the Lys79-bound isomer. Both alkaline conformers, IVa and IVb, have a similar midpoint reduction potential (-190 mV vs. SHE, pH -8.7) suggesting that the dielectric constant of the heme cavity in the two conformers is comparable. The temperature and ionic strength dependence of the conformational equilibria of the Lys73 Ala and Lys79Ala variants was studied. The ii formation of the Lys73-bound alkaline conformer exhibits a stronger temperature dependence than the isomerization involving Lys79. 'H-NMR shows that salts can significantly affect the distribution of alkaline conformers so that Lys73 ligation to the iron is favoured at low salt concentrations, whereas at higher concentrations, Lys79 coordination is preferred. The second group of variants includes cytochromes with amino acid replacements at highly conserved sites in the sequence (Thr78Ala, Tyr67Phe, Phe82Gly, Phe82Lys, Phe82Ser, Phe82Asp, Phe82Met, Phe82His, Phe82Trp, and Phe82Ile). In general, the Lys73-bound conformer is the favoured alkaline species in the Phe82Xxx variants, thus accounting in part for the generally decreased pKap of the transition of these variants. Unconventional cases are the Phe82His variant in which a non-native, bis-histidine ligation is observed that interferes with the normal alkaline transition, and the Phe82Trp variant in which a high-spin species is trapped at mildly alkaline pH. There is no indication at this time that Lys82 coordinates the heme iron. The variant with Phe at position 67 is unique among all the other cytochromes investigated in that the significantly stabilized native conformer appears to undergo the transition to primarily the Lys79- bound conformer. Finally, the alkaline transition of a cytochrome expressed with a newly developed bacterial expression system was characterized. During bacterial expression, Lys72 is not trimethylated as it is in yeast. The lack of this post-translational modification results in a third alkaline conformer in which this lysyl residue replaces Met80 as a heme ligand. Although this isomerization occurs more readily than that involving Lys73 or Lys79, the absence of the methionyl ligand in Tml72Lys/Met80Ala variant does not facilitate the formation of any of the three alkaline conformers. In this variant, the lysyl residues must compete with water and hydroxide for the axial heme iron coordination site, and the Lys72-bound conformer appears to be preferred alkaline species. iii TABLE OF CONTENTS ABSTRACT ii TABLE OF CONTENTS iv LIST OF TABLES x LIST OF FIGURES x LIST OF ABBREVIATIONS xiii ACKNOWLEDGMENTS xv 1. INTRODUCTION 1.1 Overview 1 1.2 Mitochondrial Cytochromes c 1.2.1 The Native Structure of Cytochrome c 2 1.2.2 The Physiology of Cytochrome c 11 1.3 The Alkaline Conformational Transition of Ferricytochrome c 1.3.1 pH-Linked Conformational Heterogeneity 14 1.3.2 Identity of the Axial Ligand in Ferricytochrome c, State IV 17 1.3.3 The Mechanistic Features of the Isomerization 21 1.3.4 The Trigger Group 24 1.3.5 Functional Considerations 6 1.4 Spectroscopic Characteristics of Cytochromes c 29 1.4.1 Electronic Absorption Spectroscopy 31 1.4.2 Electron Paramagnetic Resonance Spectroscopy 33 1.4.3 Proton Nuclear Magnetic Resonance Spectroscopy 36 1.5 Thesis Objectives 40 iv 2. EXPERIMENTAL PROCEDURES 2.1 Production of Yeast wo-1-Cytochrome c Variants 42 2.1.1 Cytochrome c Expression in Yeast2.1.2 Cytochrome c Expression in E. coli 46 2.1.3 Purification of Cytochromes c 7 2.2 Spectroscopic Techniques 2.2.1 ^-NMR Spectroscopy 49 2.2.2 EPR Spectroscopy 52 2.2.3 Electronic Absorption Spectroscopy 52.2.4 Circular Dichroism and Magnetic Circular Dichroism Spectroscopies 53 2.3 Physicochemical Properties of Cytochrome c 2.3.1 Electrochemistry 54 2.3.2 Protein Stability 6 2.4 Alkaline Isomerization 2.4.1 pH Titrations 57 2.4.2 Exogenous Ligand Replacement 59 2.4.3 pH lump Kinetics 52.5 Numerical Analysis of Data 61 3. RESULTS 3.1 Production of Yeast iso-1 -Cytochrome c Variants 3.1.1 Construction of pBTRl 62 3.1.2 Expression and Purification of Cytochrome c 62 v 3.2 Native Cytochromes c 64 3.2.1 S pectrophotometry 6 3.2.2 Natural and Magnetic Circular Dichroism Spectroscopy 68 3.2.3 EPR Spectroscopy 71 3.2.4 ^-NMR Spectroscopy 3 3.2.5 Spectrophotometric pH-Titrations 77 3.3 Identification of the Axial Ligands in Alkaline Ferricytochrome c 3.3.1 Electronic Absorption Spectrophotometry 78 3.3.2 Circular Dichroism Spectroscopy 81 3.3.3 MCD Spectroscopy 83 3.3.4 'H-NMR Spectroscopy 5 3.3.5 EPR Spectroscopy 87 3.3.6 Modeling the Spectroscopic Features of the Alkaline Conformers 89 3.4 Physical Properties of the Lys^ Ala Variants of Cytochrome c 3.4.1 Direct Electrochemistry of Cytochrome c 92 3.4.2 Influence of Temperature on the Alkaline Isomers of Ferricytochrome c 98 3.4.3 Influence of Salt on Alkaline Ferriytochrome c 102 3.5 Dynamics of the Alkaline Transition 3.5.1 pH-Iump Kinetics 106 3.6 Yeast wo-1-Cytochrome c Expressed in E. coli 112 3.7 Effect of Other Point Mutations on the Alkaline Transition 3.7.1 Phe82Xxx 117 3.7.2 Tyr67Phe 138 vi 3.7.3 Thr78Ala 141 3.7.4 Tml72Lys/Met80Ala 143 4. DISCUSSION 4.1 The Alkaline Conformational Transition of Cytochromes c 4.1.1 Yeast iso-\-Ferricytochrome c, state IV 146 4.1.2 The Alkaline Isomers of Other Cytochromes c 149 4.1.3 Obstacles to the Identification of the Alkaline Ligands 152 4.2 Mechanism of the Alkaline Conformational Transition 4.2.1 Intermediate Conformational States 153 4.2.2 The Identity of the Trigger Group 155 4.3 Factors that Influence the Isomerization 4.3.1 The Driving Force of the Isomerization 158 4.3.2 Structural Elements 160 4.3.3 Effects of Salts on the Conformational Transition of Cytochrome c 163 4.3.4 Cytochrome Conformational Equilibria at Charged Surfaces 165 4.4 pH-Linked Conformational Transitions in Other Proteins 167 4.5 Possible Roles of the Alkaline Conformers 169 4.6 Concluding Remarks 172 Bibliography 173 Appendix 1 188 Appendix 2 192 vii Appendix 3 194 Appendix 4 1^5 Appendix 5 7 Appendix 6 *^9 Appendix 7 200 Appendix 8 2 viii LIST OF TABLES la. Chemically-modified lysine groups 18 lb. Literature review of chemical modification experiments 19-20 2. Mutagenic and sequencing oligonucleotide primers 45 3. UV/Vis spectroscopic characteristics of the Lys-Ala variants of cytochrome c 66 4. EPR parameters of the Lys-Ala variants of cytochrome c at neutral pH 73 5. Peak assignments of hyperfine-shifted resonances in the 'H-NMR spectra of the Lys-Ala variants of cytochrome c 76 6. Alkaline isomerization equilibrium constants of the Lys-Ala variants of cytochrome c . . 78 7. Midpoint reduction potentials of cytochromes c at alkaline and neutral pH 97 8. Temperature dependence of the alkaline transition equilibria of ferricytochromes c .... 98 9. The thermal stabilities of cytochromes c at pH 6.15 and 9.58 101 10. Gouy-Chapman analysis results of the salt dependence of the alkaline pKa of cytochromes c 104 11. pH-Jump kinetics parameters of yeast cytochromes c 108 12. The alkaline transition of various variants of yeast iso-l -cytochrome c 123 ix LIST OF FIGURES 1. Three-dimensional structure of yeast iso-\ -cytochrome c 3 2. Sequence comparison of cytochromes c 5 3. Illustration of the location of conserved lysyl, histidyl, and arginyl residues 7 4. Three-dimensional structure of the heme prosthetic group 8 5. The electron transport chain 12 6. The physiological electron transfer partners of cytochrome c 13 7. Crystal field splitting of a d5 cation and electronic transitions of heme proteins .... 30 8. "Truth" diagram of Blumberg and Peisach 35 9. Illustration of the lysyl residues of cytochrome c targeted for mutagenesis 41 10a. The sequence of CYC I 43 10b. Functional map of the plasmid pING4 44 11. 'H-NMR pulse sequences 50 12. 2-Compartment glass cell for cyclic voltammetry 55 13. Functional maps of pBPCYCl(wt)/3 and pBTRl 63 14. Hydrogen bond interactions of Lys79 65 15. Electronic absorption spectra of the native yeast iso- 1-ferricytochomes c 67 16. Far UV and visible CD spectra of the native yeast /so-l-ferricytochomes c 69 17. Near-IR MCD spectra of yeast /'so-1-ferricytochomes c at neutral pH 70 18. EPR spectrum of wild-type yeast /'so-1-ferricytochrome c at pH 6.6 . 72 19. 'H-NMR spectra of Lys-Ala variants of cytochrome c at pD ~7 74-75 20. pH titration curves of yeast /'so-1-ferricytochromes c 79 21. Electronic absorption spectra of yeast iso- 1-ferricytochomes c at pH 10.3 80 22. Far UV and visible CD spectra of alkaline yeast /'so-l-ferricytochomes c 82 23. Near IR MCD spectra of yeast /so-l-ferricytochomes c at alkaline pH 84 24. Downfield 'H-NMR spectra of Lys-Ala variants of cytochrome c at pD 9.3 86 25. EPR spectra of yeast /'so-l-ferricytochromes c at alkaline pH 88 26. Methylamine hydrochloride titration of the Lys73/Lys79 variant of cytochrome c 90 27. 'H-NMR spectra of the methylamine adduct of the Lys73Ala/Lys79Ala variant 91 28. EPR spectra of the methylamine adduct of the Lys73Ala/Lys79Ala variant 93 29. Cyclic voltammogram of the Lys79Ala variant of cytochrome c at pH 8.4 95 30. pH-dependence of the midpoint reduction potential of cytochromes c 96 31. van't Hoff plots of the alkaline transitions of several cytochromes c variants 99 32. Temperature dependence of the 'H-NMR spectra of wild-type ferricytochrome c 100 33. Influence of salt concentration on the alkaline transition equilibria of ferricytochromes c . . 103 34. 'H-NMR spectra of wild-type as a function of potassium chloride concentration 105 35. Kinetic profile of a pH-jump of the Lys79Ala variant from pH 6.0 to 9.45 107 36. pH-Dependence of the pH-jump kinetics of ferricytochrome c 109 37. Singular value decomposition analysis results from a pH-jump kinetics experiment above pHIO .111 38. Spectroscopic comparison of wild-type of yeast and bacterial cytochromes c 113 39. EPR spectra of alkaline yeast cytochromes c expressed in bacteria 114 40. XH-NMR spectra of alkaline yeast cytochrome c expressed in E. coli 116 41. Detail of the three-dimensional structure of cytochrome c near Phe82 118 42. Native state ^-NMR spectra of yeast cytochrome c variants at position 82 119 43. ^-NMR spectra of the Phe82Gly and Phe82Ser variants of ferricytochrome c 121 xi 44. EPR spectrum of the Phe82Asp variant of cytochrome c measured at pH 8.9 122 45. 'H-NMR spectra of the Phe82Asp and Phe82Met variants, pH > 8 124 46. Temperature dependence of'H-NMR spectra of the Phe82Ile variant of cytochrome c 126 47. EPR spectra of the Phe82Lys variant of cytochrome c 127 48. 'H-NMR spectra of the Phe82Lys variant of cytochrome c 129 49. Spectrophotometry titration of the Phe82His variant of cytochrome c 130 50. 'H-NMR spectra of the Phe82His variant of cytochrome c 132 51. Spectrophotometry titration of the Phe82Trp variant of cytochrome c 133 52. EPR spectra of the Phe82Trp variant of cytochrome c 135 53. 'H-NMR spectra of the Phe82Trp variant of cytochrome c 136 54. Detail of the structure on the distal side of wild-type cytochrome and the Tyr67Phe variant 139 55. 'H-NMR spectra of the Tyr67Phe variant of cytochrome c 140 56. 'H-NMR spectra of the Thr78Ala variant of cytochrome c 142 57. 'H-NMR spectra of the Tml72Lys/Met80Ala variant of cytochrome c 144 58. Comparison of the Lys73- and Lys79-bound alkaline cytochrome c models to the X-ray crystal structure of the wild-type protein 148 59. Diagram depicting the residues for which 'H-NMR resonances have been assigned 150 ABBREVIATIONS ATP adenosyl triphosphate CAPS 3-(cyclohexylamine)-l-propane sulfonic acid Cb5 cytochrome b5 CcO cytochrome c oxidase CcP cytochrome c peroxidase CcR cytochrome c reductase CD circular dichroism CHES 2-(/V-cyclohexylamino)ethane sulfonic acid CT charge transfer dipic dipicolinic acid DNA deoxyribonucleic acid DNase deoxyribonuclease DTT dithiothreitol DSS 2,2-dimethyl-2-silapentane-5-sulfonate duf dUTpase activity deficient EPR electron paramagnetic resonance e.u. entropy unit FID free induction decay flavo-CZ>2 flavocytochrome b2 IR infrared kpoint kilopoint MCD magnetic circular dichroism xni MES 2-(vV-morpholino)ethane sulfonic acid MRW mean residue molecular weight NMR nuclear magnetic resonance NOE nuclear Overhausser effect pge pyrolytic graphite electrode PMSF phenyl methyl sulfonic acid RNA ribonucleic acid RNase ribonuclease SCE standard calomel electrode seq sequencing SHE standard hydrogen electrode Tm melting temperature TAPS A^-tris^ydroxymethyOmethyl-S-aminopropane sulfonic acid Tml trimethyllysine TOC total overall correlation Tris Tris (hydroxymethyl) amino methane UAS upstream activating sequence ung' uracyl N-glycosylase activity deficient UV ultraviolet Vis visible YP yeast/peptone ACKNOWLEDGMENTS No endeavour like this, however small, can ever be the product of a single individual. Therefore, I express my sincerest gratitude to all the people whose help and support have proven invaluable. Thank you Marcia, Juan, Ralf, Emma L-R. and Emma B., Jordi, Paul and Paola, Jeanette, Greg and Lawrence, Guy, Kim, Mark B. and Mark T., Logan, Dean, Tom, Martin, Brent, Terry, Kimphrey, Susanne, and Christie. I have learned a lot from all of you, and each page of this manuscript is partly your work too. Although this is also true of Grant Mauk, my supervisor, you deserve special thanks for your support and for keeping me focused while still giving me the freedom to pursue new ideas regardless of how hair-brained some of these were. Many thanks to the members of my advisory committee, Lawrence Mcintosh and Chris Orvig, for looking after me and for reviewing my thesis. The friendships with other people in the Department are appreciated greatly also, you guys made working in the lab more fun. Finally, I would like to thank my family, particularly Mom, Erika, Carlos, Oma, Roger, Alex, and Christina. Los quiero mucho. xv 1 Introduction 1.1 Overview Regulatory responsiveness is an important feature of proteins which allows them to react appropriately to changes in solution conditions (Somero, 1986). pH, for example, can have significant effects on the function of a protein, and in many instances, the influence of pH on rates of enzymatic activity can be explained readily in terms of acid^ase catalysis. However, examples of proteins which exhibit distinct pH-dependent conformational states (Guss etal., 1986; Holzmanef a/., 1990; Dewan etal, 1993; Evans & Chan, 1994; Chaine & Pakdaman, 1995; Tatulian & Tamm, 1996) indicate that pH can also exert a significant influence over enzymatic activity by modulating protein structure. Although such a relationship is not surprising, our understanding regarding pH-linked conformational equilibria of proteins is far from complete. Mitochondrial cytochrome c has been used extensively to model important chemical and biological processes, most notably electron transfer reactions. Because this heme protein also exhibits not one, but rather five spectroscopically distinct forms depending on pH (Theorell & Akesson, 1941), cytochrome c serves also as a valuable model of the intricate relationship between pH and protein structure. Thus, the alkaline conformational transition has been the focus of many studies involving novel chemical and spectroscopic techniques (see for e.g., Moore & Pettigrew, 1990; Wilson & Greenwood, 1995 and references therein). Many of these approaches are used in concert with site-directed mutagenesis in the present study to define important aspects about the alkaline isomerization of ferricytochrome c that remain to be elucidated even after close to sixty years after the discovery of this phenomenon (Theorell & Akesson, 1941). 1 1.2 Mitochondrial Cytochromes c 1.2.1 The Native Structure of Cytochrome c Mitochondrial cytochromes c are soluble, globular proteins that consist of a single heme prosthetic group and 103-113 amino acids (108 amino acids in yeast; Figure 1). During protein assembly, apo-cytochrome c attaches to the outer mitochondrial membrane, and the N- terminus of the polypeptide is spontaneously inserted across to the intermembranal space (Nicholson etal., 1988; Stuart etal., 1990). Here, cytochrome c heme lyase (sometimes also called cytochrome c synthetase; Taniuchi et al., 1983) catalyzes the covalent attachment of a ferrous heme to the incoming apo-cytochrome c chain (Nicholson & Neupert, 1989). The newly incorporated prosthetic group then serves as the core around which the polypeptide chain folds (Dumont et al., 1994) to enclose the heme within a hydrophobic environment. This folding of the polypeptide is believed to drive the translocation of the apo-protein across the mitochondrial outer membrane (Stuart etal, 1990). Some cytochromes c undergo further post-translational modification such as specific trimethylation of lysyl residues (e.g., Lys72 in yeast cytochrome c) and /V-terminal acetylation (e.g., equine cytochrome c). Recently, Thony-Meyer presented an in-depth review on the biogenesis of respiratory cytochromes in bacteria (Thony-Meyer, 1997). Although some steps for the assembly of c-type cytochromes in bacteria are similar to those outlined above, it appears that threading of the apo-cytochrome across a membrane is not an essential requirement. Mauk and coworkers have successfully expressed holo-cytochrome c in the cytoplasm of E. coli by co-expressing cytochrome c heme lyase (CTC3) and mitochondrial cytochrome c from yeast (CYCl) (Pollock et al., 1998). 2 Figure 1. Ribbon representation of yeast wo-1-cytochrome c in the classical orientation of Takano and Dickerson (1980) (coordinate data set: 2ycc; Berghuis & Brayer, 1992). The heme group, illustrated in red with the iron in yellow, is at the center of the molecule and is viewed edge-on with Met80 (green) and His 18 (blue) on the distal (left) and proximal (right) sides respectively. 3 The amino acid sequences of more than one hundred cytochromes c have been published (Moore & Pettigrew, 1990 and references therein) and are available in a number of databases (e.g., Brookhaven Protein Data Bank; GenBank; PROMISE, Degtyarenko etal., 1997, 1998). Sequence comparison analysis of 96 mitochondrial cytochromes c reveals striking similarities between these proteins (Cutler al. 1987; Moore & Pettigrew, 1990; Louie, 1990; Brayer& Murphy, 1995). Many of these analogies, and presumably the most significant, persist if the sequences of some bacterial cytochromes and those of more recently published mitochondrial proteins are included. Comparison of yeast iso-1 -cytochrome c (reference sequence) to 137 other sequences reveals at least five regions bearing a high degree of conservation (residues 14-18, 29-38, 45-49, 70-80, and 84-87; Figure 2). From a structural standpoint, the most significant region is the consensus sequence Cys-Xxx-Zzz-Cys-His which is generally regarded as a requirement for lyase-catalyzed heme attachment to the apoprotein (Hampsey et al., 1988). A covalently-bound heme prosthetic group is, after all, the principal criterion for classifying a heme protein as a c-type cytochrome. Moreover, Hisl 8 is an axial ligand to the heme iron and is essential for the function of cytochrome c (Fumo et al., 1995). Notably, the majority of mitochondrial cytochrome c sequences can include as many as 19 lysyl, 4 arginyl, and 4 histidyl residues distributed throughout the polypeptide chain. The abundance of these residues reflects the unusual basicity of these proteins (pi 9.7-10.5) and makes them well-suited to interact electrostatically with a variety of acidic electron transfer proteins (vide infra). With the elucidation of the X-ray crystal structures of horse (Dickerson etal, 1971; Bushnell etal, 1990; Sanishvili etal, 1995), tuna (Takano etal, 1973, 1977; Swanson etal, 1977), bonito (Ashida et al, 1973; Tanaka, et al, 1975), rice (Ochi et al, 1983), and yeast (Louie et al, 1988; Louie & Brayer, 1990; Berghuis & Brayer, 1992) cytochromes c, it was found that in addition to high sequence conservation, these mitochondrial cytochromes c also exhibit a high degree of structural 4 100 80 60 40 20 •a 0 ll MTEFKAGSAKKGATLFKTRCLQCHTVEKGGPHKVGPNLHGIFGRHSGQAEGYSYT 10 20 30 40 PP P P 100 80H 60 40 20] 0 top Q Q right side Q Q floor li lull Jl DANIKKNVLWDENNMSEYLTNPKKYIPGTKMAFGGLKKEKDRNDLITYLKKACE 50 60 70 BO 90 100 Q lower lower left rear left side Q upper left side rear Figure 2. Comparison of the primary structure of yeast /so-1-cytochrome c to 137 known c-type cytochromes (GenBank). Variability at each position is defined as the percentage of sequences that bear a residue other than the one found in the sequence of yeast iso-1 -cytochrome c. Conservative mutations are depicted with a white bar, and non-conservative mutations are depicted in black. The green dots emphasize residues that line the heme cavity of the protein. The secondary structure elements of the protein are illustrated beneath the sequence where p and Q label the residues that form p-turns and Q-loops. The approximate location of these structural elements in the fully folded protein is quoted in relation to the perspective shown in Figure 1. 5 homology. Approximately 40% of the cytochrome polypeptide chain is organized into five a-helices: 2-14 (top), 49-55 (lower left), 60-70 and 70-75 (left face), and 87-102 (rear) (perspective according to Figure 1; Takano & Dickerson, 1980). The remainder of the protein is arranged primarily into type II p-turns (21-24, 32-35, 35-38, 39-42,43-46, and 75-78), located mainly on the floor and right face of the protein, and Q-loops (18-32, 34-43, 40-54, and 70-84) on both sides of the cytochrome. Mapping the most highly conserved lysyl (27, 72, 73, 79, 86, 87, and 89), arginyl (13, 38, and 91) and histidyl (26) residues on the protein surface reveals that they cluster asymmetrically with the majority arranged around the periphery of the exposed heme edge (Figure 3). This observation has been interpreted by many authors to suggest that this region is the preferred docking surface for the interaction of cytochrome c with its electron transfer partners (vide infra). The X-ray crystal structures also confirm that the cysteine residues in the consensus sequence are the sites of heme attachment in mitochondrial cytochromes c. Cysl4 and 17 form thioether bonds with the cc-ethylene carbon atoms of the heme pyrrole rings B and C, respectively (Figure 4). These linkages and other constraints within the heme cavity force the porphyrin ring to ruffle such that it is saddle-shaped with the top and bottom edges bending towards the proximal side, and the rear and front edges bending towards the distal side. Furthermore, it is clear that the e2 nitrogen atom of the Hisl8 imidazole ring coordinates the heme iron from the proximal side of the protein while the 8-sulfur atom of Met80 binds to the iron on the opposite side (Figure 1). An issue of long standing interest has been the identification of the structural properties that influence the electron transfer function of cytochrome c (Margoliash & Schejter, 1995). Therefore, significant efforts have been made to define structural differences between the ferrous (Fe2+) and the ferric (Fe3+) states of the protein. From their X-ray crystallographic analysis conducted on yeast iso-1 -cytochrome c, Berghuis and Brayer (1992) noted that only minor changes in the general protein fold 6 Figure 3. Spacefilling Model of wild-type, yeast wo-1-ferricytochrome c. The exposed edge of the heme is illustrated in red, and the highly conserved lysyl residues are shown in blue. Arginyl and histidyl residues are shown in green and cyan respectively. 7 Figure 4. Diagram of the heme prosthetic group of cytochrome c and the Fischer nomenclature used in the text. The axial ligands His 18 and Met80 are located below and above the plane of the page, respectively, in view (A), and to the right and left in view (B). 8 occur upon oxidation. Interestingly, these authors identified subtle structural changes in the area of the heme. The hydrogen bond formed between the hydroxyl group of Tyr67 and the sulfur atom of Met80 is severed, thus reducing the electron withdrawing ability of the axial ligand and weakening its bond strength to the iron atom. The increased mobility of Met80, as well as the mobilities of its preceding and succeeding polypeptide segments, is reflected in their greater thermal factor parameters. At the same time, the ruffling of the heme becomes more pronounced, and the plane of the imidazole ring of the axial His 18 ligand rotates slightly about the normal to the heme plane. Finally, the most significant structural change reported by Berghuis and Brayer (1992) involves an internal water molecule (H20 166) which moves -1.7 A closer to the heme iron to break the hydrogen bond between this water molecule and Asn52. This migration is accompanied by a rotation of the water molecule such that the oxygen atom is closest to the heme iron. The structural analysis described above was recently extended to yeast cytochrome c in solution. Using 'H-NMR, Baistrocchi et al. (1996) and Banci et al. (1997) confirmed many of the observations by Berghuis and Brayer (1992) and emphasized that two segments of the protein (residues 14-26 and 75-82, inclusive) experience greater dynamic freedom in the oxidized protein. Increased mobility was not an entirely unexpected result considering that oxidation state-linked structural differences were known from previous observations such as an increased susceptibility to proteases (Nozaki et al., 1958; Yamanaka et al., 1959), denaturing agents (Butt & Keilin, 1962; Dickerson & Timkovich, 1975), thermal denaturation (Pfeil, 1981; Moore & Pettigrew, 1991 and references therein; Nail, 1995), and earlier NMR experiments (Moore, 1983; Williams et al., 1985). The increased ruffling of the porphyrin ring upon oxidation could not be confirmed by the 'H-NMR studies because the heme was defined as a planar molecule in the computational determination of the solution structures. However, detection of structural changes about the heme, including the formation 9 of a hydrogen bond between heme 7-propionate and the amide nitrogen of Gly41, may indirectly confirm the enhancement of the porphyrin ruffling observed in the crystal structure. It is tempting to assume that because a high degree of sequence identity and structural homology exists among mitochondrial cytochromes c, the oxidation state-dependent structural differences observed in the yeast protein also occur in cytochromes from other species. Indeed, similar structural differences were observed previously in crystallographic studies conducted on tuna cytochrome c (Takano & Dickerson, 1980). However, comparison of reduced and oxidized horse heart cytochrome c by ^-NMR spectroscopy indicates that upon oxidation, this protein undergoes more significant structural changes than does the yeast protein (Qi etal., 1994, 1996). Whereas the secondary structure of the yeast protein is relatively unaffected by a change in oxidation state, Qi et al. report several significant differences in hydrogen-bond networks of the backbone of the equine cytochrome. Helices (310-helix: 60-63; a-helix: 49-54) that are detectable in the structure of the ferrous protein are not detectable in the oxidized protein. Furthermore, the highly conserved Phe82 is found to be further from the heme iron in the oxidized protein than in the reduced protein (C-C: -2.5 A; Y-C: -1 A). More recently, Dong and coworkers (Calvert et al., 1997) compared the effect of oxidation state on the structures of yeast, tuna, and some mammalian cytochromes c. In the amide I band region, the FT-IR spectra of these cytochromes c exhibit similar structural changes (in a-helix, P-turns, extended p, and disordered structure content) upon oxidation, but the magnitude of these differences is not the same for all species. Not surprisingly, Dong and coworkers found that the extent to which the structure of these proteins changes upon oxidation increases as the sequence identity between these proteins decreases. 10 1.2.2 The Physiology of Cytochrome c The primary function of cytochrome c is to transfer electrons between a variety of electron transfer partners. Cycling between the ferrous (Fe2+) and ferric (Fe3+) oxidation states, cytochrome c promotes a multitude of metabolically significant reactions. This heme protein resides primarily in the intermembranal space of mitochondria where it is most closely associated with ubiquinol-cytochrome c oxidoreductase (complex III, CcR) and cytochrome c oxidase (complex IV, CcO) as part of the electron transport chain (Figure 5). However, cytochromes c also interact with other proteins, including cytochrome b5, liver sulfite oxidase (SOx), and yeast flavocytochrome Z»2(flavo-Cb2) (Pettigrew & Moore, 1987) (Figure 6) Although all of these electron transfer reactions may ultimately result in the production of ATP by regenerating cytochrome c oxidase, the latter two also regenerate the electron donating enzymes for the breakdown of sulfur-containing amino acids (SOx) and the conversion of lactic acid to pyruvate (flavo-C62). A third detoxifying reaction that occurs in yeast includes cytochrome c as an electron donor to cytochrome c peroxidase (CcP) for the breakdown of potentially harmful peroxides. More recently, the involvement of cytochrome c in apoptosis (programmed cell death) has become evident to the extent that this heme protein is regarded as an apoptosis-triggering factor (Kluck et al, 1997; Yang et al, 1997; Skulachev, 1998). The exact role of cytochrome c in the complex chain of events leading to nucleo- and cytoplasmic condensation, cytoplasmic membrane blebbing, and cellular fragmentation is not known. However, it has been established that cytochrome c is a required component for the activation of some caspases (CPP32 and CPP32-like cysteine proteases) (Liu etal, 1996; Kluck et al, 1997; Yang et al, 1997). Antibody-mediated sequestration of cytochrome c from cell-free extract assays blocks the apoptotic response. Conversely, addition of exogenous cytochrome c to these cell-free extracts and micro-injection of the heme protein into live 11 Figure 5. Schematic Diagram of the energy transducing pathway of oxidative phosphorylation showing the path of electrons to the terminal oxidase where oxygen is reduced to form water. The curved arrows indicate the cycle between the redox species at each step, while the straight arrows denote the reactions that result in the phosphorylation of ADP to ATP. The scale on the right denotes the reduction potential involved in the electron transfer at each couple. 12 Figure 6. The cellular organization of cytochrome c (Cc; 2YCC) and its electron transfer partners cytochrome c reductase (CcR; 1QCR), cytochrome c oxidase (CcO; 10CC), cytochrome bs (Cb5. 1CYO), liver sulfite oxidase (SOx; 1SOX), yeast flavocytochrome b2 (flavo-C&2; 1LDC), and yeast cytochrome c peroxidase (CcP; 2PCC). Each structure is illustrated approximately to scale, and these were obtained from the Brookhaven Protein Data Bank (PDB). CcR and SOx are shown as monomers (as opposed to the whole complex) to reflect the data contained within the PDB files. Similarly, additional cofactors that do not appear in the PDB file are not included in this Figure. The arrows indicate the flow of electrons while IMM and OMM denote the inner and the outer mitochondrial membranes respectively. The heme, flavin, and molybdopterin prosthetic groups are illustrated in red, yellow, and cyan respectively, while copper and molybdenum atoms are shown in green and magenta respectively. 13 cells induces events associated with cell death (Liu et al, 1996; Li et al., 1997; Zhivotovsky et al., 1998). Cytochrome c-mediated caspase activation is specifically triggered only by the holoprotein. Related heme proteins are ineffective. Furthermore, caspase activation appears to involve electrostatically stabilized protein complexes because high salt concentrations interfere with the process (Hampton etal., 1998). Initial results also suggest that cytochrome c-mediated initiation of cell death occurs independently of the oxidation state of the heme protein (Kluck et al., 1997; Hampton et al., 1998). Newmeyer and coworkers note that although agents added to apoptosis assays to maintain cytochrome c oxidized (e.g., K3[Fe(CN)6]) also perturb downstream effectors to account for a reduced apoptosis triggering efficiency, it is likely that some step(s) in the activation of caspases by cytochrome c requires reducing conditions (Kluck et al, 1997). 1.3 The Alkaline Transition of Ferricytochrome c 1.3.1 pH-Linkage in Brief In 1936, eleven years after the rediscovery of cytochrome c by Keilin (1925), Theorell and Akesson reported the pH-linkage of the spectrophotometry characteristics of the protein (Akesson, 1936; Theorell & Akesson, 1939, 1941; Theorell, 1941). In the reduced state, cytochrome c exists in one of three spectroscopically distinct forms. The spectroscopic changes associated with the interconversion of the acidic (state I) and alkaline (III) forms to the intermediate, neutral state (II) are a consequence of specific amino acid ionizations which are coupled to structural changes. The equilibria are described as : I ** II III (1) Note that the pK values associated with these equilibria are only approximations because their exact 14 values depend on the species of the protein as well as other factors such as ionic strength and temperature. State II is the isomer of ferrocytochrome c that is functionally relevant, at least in terms of energy transduction. This species has been thoroughly characterized by a myriad of techniques because of its important physiological role. The structure and properties of state II, therefore, are well defined (see section 1.2.1) unlike those of states I or III which are largely unfolded. Ferricytochrome c exhibits at least five pH-dependent isomers and, therefore, four conformational changes (Moore and Pettigrew, 1990): pKv-0A PKV~2.5 pKv~9.3 pKv~12.S I II III IV ** V (2) As in the case of ferrocytochrome c, the exact values of these equilibrium constants are species-dependent and subject to the conditions of measurement. Of the five conformers, the functionally relevant isomer for oxidative phosphorylation is state III. This conformer is also very well characterized, especially because the paramagnetic nature of the oxidized heme centre makes this group an informative reporter group in the active site of the molecule (vide infra). EPR spectroscopy parameters, for example, are perturbed only by environmental changes in the immediate vicinity of the heme as are the hyperfine shifted regions in 'H-NMR spectra. States I and II of ferricytochrome c are the acid-denatured forms of the protein. Under some circumstances, His3 3 and His26 can replace Met80 as a heme ligand (Elove et al., 1994; Colon et al., 1997). With increasing proton concentrations, the neutralization of Glu and Asp residues results in a net increase of the positive character of the protein. In the absence of electrolytes to shield these positive charges, electrostatic repulsions induce the protein to adopt an extended conformation. Although the tertiary structure of the protein is lost under these conditions, localized regions of 15 secondary structure elements may remain in a moderately stable form (Jeng & Englander, 1991). It follows that the charge-shielding effect of salt results in a stabilized, molten-globule-like conformation of the protein that retains the secondary structure elements of the native form but only a range of dynamic tertiary interactions (Goto etal., 1990). The molten-globule forms of various proteins are generally regarded as models of protein folding intermediates. As such, cytochrome c at low pH (~2) and in the presence of salt (e.g., 2:20 mM NaCl) has been the focus of much interest insofar as it a useful model for our understanding of structural determinants in cytochrome c and other proteins. State IV is generally known as the alkaline conformer of cytochrome c, and the interconversion between the native form and this isomer has been the focus of much interest. A pivotal characteristic of this structural reorganization is the loss of the absorbance band centered at 695 nm from the electronic spectrum of the protein (Theorell & Akesson, 1941). Because this band arises from ligation of Met80 to the heme iron, the loss of this absorbance reflects the disruption of this bond (Harbury etal, 1965; Schechter & Saludjian, 1967; Eaton & Hochstrasser, 1967; Smith & Williams 1970). State IV also possesses a low-spin heme iron which suggests that the native Met ligand is replaced by another strong field ligand. At even higher pH, this new ligand is also removed from the coordination sphere of the iron to yield state V of ferricytochrome c. A species with spectroscopic characteristics similar to those of state V are observed when ferricytochrome c binds to cytochrome c oxidase (Dopner et al, submitted). Although this possibility had been proposed earlier (Weber et al, 1987; Alleyne & Wilson, 1987), there is a conflicting report in which no changes in the CD spectrum of the cytochrome were encountered that would confirm the proposal (Michel et al, 1989). 16 1.3.2 Identity of the Axial Ligand in Ferricytochrome c, State IV A heme prosthetic group coordinated axially by at least one histidyl residue is a common structural element in many of the known heme protein structures. Despite this common characteristic, heme proteins can execute a wide variety of functions and display a range of spectroscopic properties. Although the immediate environment around the heme can exert considerable influence, the physicochemical characteristics of the heme protein are determined in large part by the second axial ligand to the heme iron (Loew, 1982; Wallace & Clark-Lewis, 1992). Therefore, it is important in characterizing a heme protein and its various conformers to identify the axial ligands. This information about the heme core is an essential first step in understanding the structure/function relationships of the protein. Sulphur, nitrogen, and oxygen atoms are endogenous components of a polypeptide chain that can coordinate to a metal centre. With the exception of the amino terminus in cytochrome/ (Martinez et al, 1994), examples in which the backbone carbonyl or amide groups coordinate the heme iron are not known. Consequently, only a handful of the twenty naturally occurring amino acids are considered here as potential ligand replacements for Met80 in the alkaline conformer of cytochrome c. Furthermore, consideration of the protein fold limits the list of potential residues to those that are located in the upper left side of the molecule (i.e., residues 60-90; Figure 1). This observation is consistent with chemical modification studies conducted on cytochromes c in which various lysyl residues were derivatized with a variety of functional groups (Tables la & lb) (Moore & Pettigrew, 1990; Wilson & Greenwood, 1996; Rosell etal, 1998). In general, modification of all lysyl residues resulted in altered alkaline transitions and non-native cytochromes of mixed-spin state. Specifically, these anomalous properties were observed when lysines between residue 65 and the C-terminus of the polypeptide were modified. As modification and purification techniques became 17 Table la. Chemical modification of surface lysyl residues used in the study of cytochrome c. CF3 Trifluoroacetyl-Lys v^NH- C /CH3 Acetyl-Lys v\^-NH-CN p-Thipropionyl-Lys NH—C^^2 ^ Trifluoro(phenyl) NH_p^° carbamyl-Lys v~" ^-d^O^ Guanidyl-Lys ^^NH-cf NH2 Acetimidyl-Lys NH—cf (amidyl-Lys) ^ CH2 NH2+ Maleyl-Lys N/S_NH_C/ n NCH=CH-C^ ON ai rn \0 vo vo O ti O o 00 g 1 c 1 i .00 o •3 g c g cn .S CD co £-8 CU CD CD a is ,CD ta a. •tS D U •VOX 2 3 C oo 0\ "8 J- CO 3 ^ .a s J CD 3«3 o J ;3 <•> in o ° T - e A O lw ,o \n o o >-i 5 o 3 -2 - | tS C T3 I-CN r-!*> <W 00 O a, o CD s o o CD .3 CD ON J3 ^ C = A •a 3 o x> ic o CD I c/l a 3 C/3 .3, o o o £ '3 £ O o" o — CD ™ 8* o o 1 1 C/3 C/3 c/3 c/3 c/3 c/3 c/3 c/3 c/3 c c3 ^ O CJ o T3 o g CO U4 C/3 C/3 C/3 < < < WWW H w c/3" < W CZ) C/3 W W Cl in IT) o o p, d r-' od ON ON O cs r>" ON „ in oo r- oo m. w Os VO -H 00 00 ON ON ON p, ON' A in >n in o o o p' p' p, ON ON od ON od i—i in ON ( in in o p p' p, NO1 ON' ON od .3 CD o O ON NO oo in CO 22 oo oo *m in V» -2 S? .3 CD a. o "o J3 CO CN ON —• r- r-00 CO 00 ^> ^> www .3 1 8-1 CS ON r- r-00 00 C O "81 cj TD I O s C o 1 CD o I CD 60 o O C/3 S3 O 2 tS Ji >>c/3 c3 o CD JS Q. O 3 •c CD S o U a o CD o Ui o 3 53 •c C/3 •8 §• o V 1 I I" tJ o X «4-l o § a "2 a> M 3 o a o u s u J3 o O C/3 1. T3 O 1 .s 8 o _ CO 0\ l-H cN (N 3 C/D < • oo oo .s I ex o I "a, i a o 3 a I '3 -2 s* a. V") IO o o p', p' ON 00 oo' oo' 2 S VO VO VO VO u a 1 CN «G 11 C -3 .2 '•8 « ts o 5 Is 8 vi O-VO o 43 jo "8 C/i a o <i> o c o .SP > ? c o c '3 £ 8" ° crt S < 2 W w 1 ON t-- r--in V) >, >, J o "33 C/3 C o c o 'a, 2 ex o o CN CO ex R< ex ex O T3 rt 3 Q O i> A ON c2 < C available to generate and isolate singly modified cytochromes c, a variety of lysyl residues within the C-terminal region of the protein were targeted for scrutiny (Morton, 1973; Wilgus & Stellwagen, 1974; Smith & Millet, 1980; Bosshard, 1981; Wallace, 1984; Theodorakis etal, 1995). The results from trifluorophenyl-carbamylation and P-thiopropionylation of these amino acids suggested that Lys72 and perhaps Lys79 are the residues that replace Met80 in the alkaline transition of horse heart ferricytochrome c (Smith & Millet, 1980; Theodorakis et al, 1995). However, in differential chemical modification experiments, these residues were as susceptible to modification in the native and alkaline states as in the free amino acid form (Bosshard, 1981). It was concluded, therefore, that the alkaline conformer ligand can not be a lysyl residue because a heme-bound amino group is protected and should not be susceptible to attack. The apparently contradictory conclusions from the different chemical modification strategies can be reconciled considering the NMR results of Hong and Dixon (1984). At alkaline pH, horse heart ferricytochrome c, and presumably other eukaryotic and bacterial cytochromes, undergoes two independent alkaline conformational transitions rather than one. These two structural isomerizations result in different alkaline species that originate from one of two lysyl residues binding to the heme. In yeast iso-l -cytochrome c one of these lysyl residues is located at position 79 (Ferrer et al, 1993). Because these ligand exchanges are dynamic processes, one residue is susceptible to modification while the second is protected by coordination to the heme. When this interaction is severed, the second ligand also becomes susceptible to modification. 1.3.3 The Mechanistic Features of the Isomerization The alkaline isomerization of ferricytochrome c is typically quantified by a pK^ value determined under equilibrium conditions (see Appendix 1). However, the process consists of a 21 combination of events that is more complex than the simple ionization of a titratable group. The simplest model that more accurately takes into account an ionization-linked structural reorganization was described by Davis et al. (1974). Based on pH-jump kinetics experiments, these authors postulated Relation 3, where N—H and N denote, respectively, the protonated and deprotonated species that possess native structure and absorbance at 695 nm, and A is the alkaline species. IT ^s1 N-H ^ N === A (3) P^H=11 ^.OSs-1 KH is the equilibrium constant for the deprotonation, and kf and kb are the forward and reverse rate constants for the conformational reorganization. The entire process thus consists of two components (pKH and p[kf/kh]) that together comprise the pKap that is measured under equilibrium conditions. This analysis has also been applied to various species of cytochrome c and to variants of the yeast protein (Pearce etal., 1989; Nail etal., 1989; Saigo, 1981; Rafferty, 1992). In some of these proteins, the pKip for the transition was altered by changes in the dynamic properties (p[kf/kb]), the deprotonation component, or a combination of both. Although the model described above can account for the data collected in pH-jump experiments below pH 10, it does not explain a second kinetic phase observed at higher pH and at wavelengths other than in the region around 695 nm band. This additional level of complexity was addressed with Relation 4 (Kihara et al., 1976; Hasumi, 1980). It was proposed that below pH 10, the alkaline transition progresses essentially as first outlined by Davis et al. (1974), resulting in the alkaline species A—H. Above pH 11, however, an alternate path via the transient T is favoured. At a pH between these limits, the alkaline transition occurs via both pathways. A simpler alternative to the mechanism described by Davis et al. (3; 1974) incorporates a 22 transient intermediate into the pathway from the native cytochrome to its alkaline conformer (Mechanism 5; Saigo, 1981 a, b): N-H ^ N ^ I ^ A (5) The fast phase that results in pH-jump experiments to pH above 10 reflects the rapid formation of this intermediate which then decays to the stable alkaline conformer. Presumably, at lower pH this intermediate still forms; however, its concentration under these conditions is too low for detection. In theory, the ligand exchange could result in an SN1 fashion: (1) coordination of the incoming lysyl residue as (2) the Met80-Fe3+ bond is broken. The formation of a transient, 5-coordinate intermediate upon breaking the Met80 bond and before coordination of the incoming lysyl residue is likely included in this mechanism (Palmer, 1977). This 5-coordinate intermediate would be a common transient for the formation of both alkaline conformers discovered by Hong and Dixon (1989). It is noteworthy that a native or native-like species precedes a transient in experiments where guanidine-denatured yeast /so-2-ferricytochrome c is renatured at alkaline pH (Nail, 1986). This result suggests that the alkaline conformational transition of ferricytochrome c is an ordered event, and that native tertiary structure elements are required to direct the structural reorganization. 23 1.3.4 The Trigger Group In parallel to the search for the sixth axial ligand in the alkaline conformer of cytochrome c, the identity of the titratable or "trigger" group which, upon ionization, initiates the alkaline transition has been sought (Moore & Pettigrew, 1990; Wilson & Greenwood, 1996). Close examination of the available crystal and NMR structures of cytochrome c reveals a multitude of hydrogen bonds (main chain: main chain = 49; main chain—side chain = 38; side chain—side chain =10; Louie, 1990). The perturbation of many of these structure-stabilizing interactions could, in principle, trigger at least a local structural collapse with the ensuing reorganizations of the polypeptide. The titration of this group occurs with a pKH of ~ 11 in equine cytochrome c as determined by pH-jump experiments (Davis et al, 1974). Therefore, amino acids that have been considered to fulfill the trigger role include the lysyl residue that also acts as a heme ligand, Tyr67, an internal water molecule (H20166), His 18, and the heme 6-propionate group. A lysyl residue that replaces Met80 cannot be protonated if the e-amino group is to coordinate the heme iron. Therefore, the ionization of this group, followed by the ligand exchange, is consistent with the kinetic mechanisms noted above. Moreover, the high p^TA value associated with a primary amine is compatible with the pKH values determined by pH-jump experiments. However, chemical modification of all lysyl residues in cytochrome c does not stabilize the Met80-Fe3+ bond, suggesting that ionization of the true trigger group is still initiating an alkaline conformational transition (Fanger & Harbury, 1965; Hettinger & Harbury, 1965; Morton, 1973; Stellwagen etal., 1975). Nitration (Skow et al., 1969; Schejter et al, 1970; Pal et al, 1975) and iodination (Morton, 1973b; Pal et al, 1975) of Tyr67 in horse heart cytochrome c results in a significantly less stable protein with a pK for the alkaline conformational transition of ~6. Replacement of this residue with 24 phenylalanine, on the other hand, increases the pKap of the rat ferricytochrome by -1.1 pH units (Luntz et al, 1989). These observations suggest that ionization of the hydroxyl group of Tyr67 perturbs the local environment on the distal side of the heme to trigger the alkaline transition. This assumption is consistent with the known structures of ferrocytochrome c, which does not exhibit a stable alkaline conformation and in which Tyr67 hydrogen bonds the 5-sulfur atom of Met80. However, in a double variant where the Tyr67Phe mutation was combined with Pro30Ala (located on the proximal side of the heme), the pKap value was restored to -9.5 (Schejter et al, 1992). Therefore, it would appear that Tyr67 does participate in the alkaline conformational transition of cytochrome c but not in the capacity of the trigger group that initiates the process. Takano and Dickerson (1981) favour the ionization of an internal water molecule as the trigger of the structural reorganization. This water molecule is located close to the iron on the distal side of the heme where it forms hydrogen bonds a number of amino acids, including Tyr67. The migration of the water molecule away from the iron in the more stable ferrocytochrome, the absence of an equivalent water molecule in cytochrome c551 from Pseudomonas aeruginosa (pKap~ 11), and the perturbation of the alkaline pKap upon modification or replacement of Tyr67 have been interpreted to support this assignment. Moreover, Takano and Dickerson proposed that, upon ionization, the resulting hydroxide coordinates the iron to form an intermediate. This intermediate is the product when all lysyl residues in the protein are chemically modified and, therefore, incapable of coordinating the iron (Table lb). However, no unequivocal evidence is yet available to confirm whether or not this water molecule potentiates the alkaline transition. Ionization of His 18 has also been suspected to trigger the alkaline isomerization of cytochrome c (Gadsby et al, 1987). A pH titration of the methylimidazole complex of the protein, monitored by EPR, showed that the His^His" interconversion (pKap~ 11.6) is consistent with the high 25 pKH values reported in the mechanism of Davis et al. (1974). Based on these observations Thomson and coworkers proposed that (1) Hisl8 ionizes to histidinate, thus (2) causing the rupture of the Met80-Fe3+ bond. (3) Upon ligation of a deprotonated lysyl residue to the iron, (4) the pKap of Hisl8 increases, and (5) the histidinate reprotonates. The principal attraction of this scheme is the apparent consistency of the pK values and the net change of one protonation state. Finally, the heme 6-propionate group has also been identified as the functional group that induces the alkaline transition (Hartshorne & Moore, 1989; Tonge et al., 1989). This proposal is based on the abnormally high pKa (-10.4) value of the carboxyl group as monitored by the asymmetric carbonyl stretch at 1571 cm"1 in the FTIR spectrum of the protein. Like all of the other proposed trigger groups, the involvement of the heme 6-propionate group in the initiation of the alkaline transition remains to be established unambiguously. 1.3.5 Functional Considerations Discussion of the alkaline conformational transition of ferricytochromes c inevitably concerns the metabolic role(s) that states IV or V of the protein may fulfill. Considering that the alkaline transition is highly reversible and widespread among cytochromes c, it could be argued that the phenomenon does not occur without a useful purpose. On the other hand, it is possible that the alkaline isomers of the protein were metabolically relevant in the distant past, and their occurrence today is simply an evolutionary vestige. Generally, the pH required for the alkaline conformational reorganization of most cytochromes c in vitro is greater than the upper limit of what is normally accepted to be of physiological significance or even attainable. However, the occurrence of this phenomenon is very sensitive to conditions such as temperature (Taler etal., 1995) and ionic strength (Osheroffe/a/., 1988; Rosell etal., 1998). Therefore, it is conceivable that during various metabolic 26 states, regions within the intermembranal space of mitochondria provide a suitable microenvironment in which the transition may take place more readily. Heavy physical exertion, for example, can raise the core temperature of a horse to 45 °C (Hodgson et al., 1995). At this temperature, the pK^ decreases as much as one pK unit from the value that is normally observed in vitro: 9.3 at 25 °C (Davis et al., 1974; Talerefa/., 1995). Moreover, the interaction of cytochrome c with other proteins and membrane surfaces may also exert a destabilizing influence on the structure of the native protein thereby reducing the effective pKap of the alkaline transition. For example, Aviram and Schejter showed that removal of water from cytochrome c through lyophilization results in the disruption of the Met80-Fe3+ bond (1972). Although the environment within the intermembranal space of mitochondria is not severe as lyophilization, the reduced water content within this compartment, which may be comparable to that of the cytoplasm (-70% water; Alberts, 1994), may also exert some influence on the conformational equilibria of the protein. Barker and Mauk (1992) demonstrated that the midpoint reduction potential of state IV is almost 0.5 V lower than that of the native protein. This property of alkaline cytochrome c would ensure that the electron flow is unidirectional during the interaction of the protein with its oxidizing electron transfer partners. If, through this interaction, cytochrome c reorganizes to its alkaline conformation immediately after donating the electron, the markedly lowered midpoint reduction potential would make it unlikely for the electron to return to the donor. In the most extreme of cases, the drop in potential essentially blocks oxidative phosphorylation. The heme edge that is partially exposed to solvent on the surface of cytochrome c is generally considered to be the optimum site for electron transfer. Here, the heme comes closest to the protein surface where the sulfur atom of Cysl7, which is also partially exposed, can act as an electron conduit by coupling to the 7T-electron system of the heme group (Tollin et al., 1986). Forming a docking site 27 around this exposed heme are highly conserved, basic residues which can direct and stabilize the formation of protein: protein complexes (Salemme, et al., 1973; Meyer et al., 1994; Figure 3). Results from chemical modification experiments (Fergusson-Miller et al., 1978; Moore et al., 1998), Brownian dynamics simulations (Northrup et al, 1988), and X-ray diffraction studies of the cytochrome c:cytochrome c peroxidase complex (Pelletier & Kraut, 1992) suggest that Lysl3, 27, 72, 73, 79, and 87 are involved in establishing the protein:protein interface. Similarly, Lysl3, 72, 79, 86, 87, and probably 73, also appear to be involved in the complexation of cytochrome c to cytochrome b5 (Salemme, 1976; Rodgers et al., 1988; Burch et al., 1990; Northrup et al., 1993; Guillemette et al., 1994), and to cytochrome c oxidase (Dbpner et al., submitted). The alkaline conformational transitions of ferricytochrome c essentially remove Lys79 (Ferrer et al., 1993) and a second lysyl residue from the surface of the protein. Like chemical modification, or replacement of these amino acids by an aliphatic side-chain, the alkaline structural reorganization reduces the ability of the cytochrome to interact properly with its electron transfer partners. The level of oxidative phosphorylation would also be reduced as a result of the decreased affinity for cytochrome c in the alkaline state. The view expressed by Moore and Pettigrew is that non-native conformations of cytochrome c in vivo, including the alkaline conformers, constitute too small a percentage of the protein to be of physiological significance (1990). Moreover, these authors cite the slow isomerization of the protein to its native structure, and the acidic nature of the intermembranal space of mitochondria, as further indications that the alkaline conformers are not functionally significant. Wilson and Greenwood (1995), on the other hand, do not dismiss the possibility that the alkaline transition is a "design feature selected through evolution" rather than "a chance byproduct of molecular architecture." Indeed, the participation of cytochrome c in apoptosis would not have been considered 2-3 years ago, but 28 cytochrome c is now regarded as an integral part of programmed cell death (Kroemer et al., 1998; Mignotte & Vayssiere, 1998; Skulachev, 1998). In a similar way, our present inability to assign a function to the alkaline conformers of cytochrome c does not, by itself, constitute irrefutable evidence that these structures are physiologically inactive. 1.4 Spectroscopic Characteristics of Cytochromes c The iron porphyrin in heme proteins can provide valuable information about the active site of cytochrome c through the use of a wide variety of spectroscopic techniques. Spectroscopic characterization of heme-containing proteins typically includes electronic absorption spectroscopy, nuclear magnetic and electron paramagnetic resonance spectroscopies, conventional and magnetically-induced circular dichroism spectroscopy, and resonance Raman spectroscopy (Moore & Pettigrew, 1990; Scott & Mauk, 1996 and references therein). These techniques probe various aspects of the interaction between the protein matrix and the prosthetic group. Therefore, a multidisciplinary approach is often implemented in the thorough characterization of heme proteins. The most common oxidation states of the heme iron are the ferrous (Fe2+: [Ar] 3d64s°) and the ferric (Fe3+: [Ar] 3 d54s°) states, although examples of the ferryl form (Fe4+: [Ar] 3d44s°) are also encountered in the form of Fe4+=0 (for e.g., compounds I and II in peroxidases). As noted earlier, the electronic configuration of the metal centre depends in large part on the nature of the axial ligands. The heme iron of both ferri- and ferrocytochromes c is in a tetragonally distorted ligand field generated by the equatorial heme and the axial Met80 and Hisl 8 ligands. Under these conditions, the d-orbitals of the iron are split as shown in Figure 7a, and the iron has low-spin electronic configurations in both oxidation states (Fe3+: 5=1/2; Fe2+: S=0). High-spin hemoproteins are also common in nature (for e.g., 5=5/2: 5-coordinate cytochrome c peroxidase in the resting state; S=2: 29 11111 ^xy <^xz dyz dzi d%i-dyi dzi dyi-dyi 111 ^xy ^xz ^yz $ 1 ^xz ^yz d xy Free Ion Octahedral Tetragonal Field Field B egrc a a 2uTC luTC A A dx2~dyi d^. d i 1 xz •4 yz 4 xy 71 11 TC Porphyrin Metal Ligand Figure 7. (A) Schematic Diagram illustrating the effect of octahedral and tetragonal field distortions on the <i-orbital splitting of a low-spin ferric heme. (B) Schematic Diagram illustrating the commonly accepted electronic transitions between porphyrin and metal electronic orbitals that are detectable by electronic absorption spectroscopy. 30 5-coordinate deoxymyoglobin), and while an example of a quantum mechanically admixed spin state (QMAIS) has been described (5=3/2, 5/2 : 5-coordinate cytochrome c') (Maltempo et al., 1974; Maltempo, 1976 a and b), naturally occurring hemoproteins with intermediate-spin states (5=3/2) have not. In mitochondrial cytochromes c, however, high-spin states and admixtures are often evidence of gross structural perturbations (for e.g., acid or chaotropic denaturation) and require more thorough characterization to define the cause and consequences of the perturbation with some precision. Although resonance Raman spectroscopy has been used successfully in the characterization of heme proteins including cytochrome c (Spiro, 1985; Desbois, 1994), the bulk of this dissertation consists of data collected with CD, electronic absorption, EPR, MCD, and 'H-NMR spectroscopy. Therefore, only the latter techniques are considered briefly in the following sections. Much of this material is applicable to heme proteins in general, but the emphasis here is on the relevance of these techniques to the study of mitochondrial cytochromes c or on the issues that are relevant to this work. 1.4.1 Electronic Absorption Spectroscopy The UV-vis spectrum of cytochrome c is dominated by the Soret (y or B) band centred at -410 (Fe3+) or 416 nm (Fe2+). This intense absorbance (e~ 106100 M"1 cm-1 oxidized and e~ 129100 M"1 cm'1 reduced) arises from strongly allowed electronic transitions between bonding and antibonding molecular orbitals of the porphyrin (7i-7t*) (Figure 7b). At longer wavelengths, two less intense features (a and p, or Q0 and Q„ respectively) appear in the visible region (-530 nm) and are minor components of the same excitations. According to the four-orbital model of Gouterman (1959), TT-Tt* transitions interact unfavourably and mix, thus separating otherwise degenerate excited states. The Soret band and the a-P system are manifestations of the higher and the lower excited states, 31 respectively (Smith & Williams, 1970). Vibrational processes contribute further to split the transition of the lower excited state into the two bands that are labeled a and p. Additional absorption bands attributable to porphyrin TT-TC* orbital transitions can appear between 215 and 325 nm, but these features are weak and often masked by absorbances from peptide bonds (~ 190-230 nm) and aromatic groups (-280 nm) (Moore & Pettigrew, 1990). In general, displacement of electronic distribution to the periphery of the porphyrin ring causes the Soret and a-p bands to shift to lower energy. Reduction of the metal centre, for example, exhibits this spectroscopic result as does replacement of the iron by a metal that is less electronegative and the introduction of stronger electron withdrawing substituents to the porphyrin ring system. Concomitantly, these absorbances shift to higher energy as the electron withdrawing strength of the axial ligands increases. The spectroscopic influence of axial ligands is demonstrated by the alkaline conformer of ferricytochrome c which displays a blue-shifted Soret band as a result of the replacement of Met80 for a nitrogenous base (Theorell & Akesson, 1941). Figure 7b illustrates electronic transitions from the porphyrin TZ- to the lower available metal c/-orbitals. These charge transfer (CT) bands are associated with absorbances of ferricytochromes c in high- (-600 nm) and low-spin (1200-1500 nm) electronic configurations (Smith & Williams, 1970; Gouterman, 1978; Makinen & Churg, 1983; Spiro, 1983). CTs from the axial ligand u-metal d-orbitals are allowed for some ligand types but these bands are difficult to assign and predict. The present consensus is that the weak 695 nm band (e~ 1 mM^cm'1) arises from just such a CT (Smith & Williams, 1970; Makinen & Churg, 1983) based on the similar characteristics of this band and those observed in the spectra of the OH' and N3" adducts of metmyoglobin. This band is of particular interest because it is diagnostic of the Met80-Fe3+ bond (Harbury et al., 1965; Schechter & Saludjian, 1967; Eaton & Hochstrasser, 1967; Smith & Williams 1970) and its strength. Consequently, this 32 feature is often used in spectroscopic studies to assess the structural integrity of the heme cavity. Metal d-d transitions are generally forbidden. Therefore, if such transitions occur, they are nearly undetectable in visible spectra where they are largely obscured by the more intense porphyrin TT-metal ti-orbital CTs. 1.4.2 Electron Paramagnetic Resonance Spectroscopy One of the earlier applications of EPR spectroscopy to the study of heme proteins, was the elucidation of axial ligands of low-spin, ferrihemoproteins (Blumberg & Peisach, 1971). This technique is sensitive to structural perturbations that change the geometry of the iron core, and in particular, to ligand exchanges. These perturbations invariably affect the axial (A) and rhombic (V) symmetry elements that define the ligand field of the iron. These parameters reflect the energy separation between the metal t2% orbitals. Whereas axial distortions destabilize the and dyZ orbitals relative to d^, rhombic distortions lift the degeneracy of the former pair (Palmer, 1977). Furthermore, the t2g orbitals mix through spin-orbit coupling (X) so that three resonance conditions exist (Equation 1): hu = g/B0 (1) where hv is the energy of the excitation radiation (microwave region), g a proportionality factor, /? is the Bohr magneton, and B0 is the applied magnetic field, and. Each of these conditions is associated with a g-factor which is read from the EPR spectrum and which corresponds to each of the coordinate axes (gx, gy, g^) where the iron ion is at the origin. Mathematical relationships have been derived to determine the tetragonality and rhombicity normalized with respect to the spin-orbit coupling (Palmer, 1977) : 33 Sx , Sy (2) Sz Sy Sz Sx /k Sz+gy Sy+sx 2/ With these coordinates in hand, proteins of unknown coordination environment have been compared to structurally-characterized standards which are compiled in so-called "truth diagrams" (Figure 8) (Blumberg & Peisach, 1971). Proteins with equivalent axial ligation cluster into well-defined regions, so ligand assignment is often possible based largely on the similarity of these ligand field parameters. In the native state, cytochromes c correspond to group C in the truth diagrams and are characterized by the lowest tetragonal fields (Figure 8). At higher pH, the EPR spectra of the protein reveal the presence of at least two additional low-spin species (Morton, 1972; Brautigan etal, 1977; Gadsby et al, 1987; Gadsby & Thomson, 1990; Ferrer et al, 1993) with ligand field parameters similar to those of group H members. According to these EPR data, the alkaline forms of ferricytochrome c are similar to the N-butylamine adduct of leghemoglobin (Gadsby et al., 1987) and to cytochrome/(Rigby et al, 1988; Gadsby & Thomson, 1990). In the former example, the amino group of the exogenous ligand acts as the ligand, while in the latter protein, the A^-terminus of the polypeptide chain fulfills this role (Martinez et al, 1994). At first sight, the use of EPR spectroscopy in the identification of residues that are coordinated to the heme is a convenient approach because of the relative simplicity of the measurement. However, this method is not always accurate or precise. Group B in the truth diagrams, for example, is composed of heme proteins with bis-histidine ligation such as cytochrome b5. Member proteins of this family exhibit a large variation in ^-values (i.e., g^-3.0-3.6) depending on the orientation of the histidine ring planes in relation to each other and to the porphyrin ring (Gadsby and 34 Figure 8. Crystal-field correlation diagram denoting the five common families of low-spin ferric hemoproteins as defined by Blumberg and Peisach (1971). Rhombicity values of 0 and 0.66 corresponds to purely axial and purely rhombic systems, respectively. Cytochromes c (solid circle) reside in zone C (solid line). 35 Thomson, 1990). In many of these spectra, it is difficult to determine g^d and values experimentally when is greater than -3.4 (Salerno, 1984). In the worst of cases, the truth diagram approach of ligand identification can lead to the incorrect assignments in cytochromes (Gasby & Thomson, 1990; Teixeiraef a/., 1993; Arciero etal, 1994) exhibiting HALS (highly axial, low-spin)-type EPR spectra (Migita & Iwaizumi, 1981; Walker et al, 1986). As noted earlier, weak manifestations of CT bands from the porphyrin n (&lu and a2u symmetry) to an available metal t2g orbital appear in the near-IR region (700-3000 nm) of ferric, low-spin hemoproteins (Figure 7b). These transitions are diagnostic of heme coordination because when the axial ligands affect the axial and rhombic symmetry elements of the iron (as detected by EPR spectroscopy), they also affect the resonant frequency v for the CT (Gadsby &Thomson, 1990 and references therein). Although, under ordinary conditions vibrational bands from C—H, O—H, and N—H bonds overlap and obscure these CT transitions, application of a magnetic field considerably enhances the CT bands. This enhancement is a consequence of the stronger magnetic moment associated with electrons relative to nuclei. Optionally, freezing samples to cryogenic temperatures further enhances the CT bands because nuclear vibrations are discouraged. Thomson and coworkers (1990) have thus compiled an extensive database of near-IR MCD spectral parameters of well characterized heme proteins and model porphyrin compounds. Therefore, the presently accepted criteria for the spectroscopic identification of heme axial ligation in low-spin, ferrihemoproteins combines EPR and near-IRMCD spectroscopy (Gadsby & Thomson, 1990; Arciero etal, 1994; Hawkins etal, 1994). 1.4.3 Nuclear Magnetic Resonance Spectroscopy The value ofNMR spectroscopy for the investigation of macromolecules such as cytochromes c is well established (see for e.g., Wiithrich, 1996; Roberst, 1993; Pielak et al, 1995 and references 36 therein). As early as 1971, NMR was used to solve important questions about this protein. Redfield and Gupta, for example, demonstrated that Met80 is the sixth axial ligand in native ferricytochrome c (1971). Previous to this report it was commonly believed that the axial ligand was a histidyl residue or even a lysine. More recently, high resolution NMR spectrometers, and the increasing availability to obtain isotopically-enriched protein samples (13C and 15N) have facilitated the assignment of chemical shifts in the spectra of proteins of -120 amino acids, though the upper limit is -250 residues. Concurrently, innovative strategies for data acquisition and analysis frequently permit the elucidation of protein three-dimensional structures. NMR-active nuclei in paramagnetic proteins are subject to the same forces that affect nuclei in diamagnetic molecules (for e.g.,, ring current shifts, nuclear Overhausser effect, etc.). However, in the proximity of transition metal ions, the large magnetic moment associated with unpaired electrons of the metal ion exerts a significant influence over the chemical shift of nuclei (Equation 4a and b). ^observed — ^diamagnetic ^paramagnetic (4&) ^ paramagnetic — ^ dipolar ^ contact (4b) This physical influence has dipolar (or pseudocontact) and contact (or Fermi) components. As the names imply, the dipolar shift results from through-space, dipolar coupling of the magnetic moments of the unpaired electron(s) and the nuclei. The contact shift, on the other hand, is a consequence of direct derealization of the electron spin density through chemical bonds. Therefore, this influence affects primarily the nuclei that are directly bound to the paramagnetic centre. In the case of heme proteins, electronic spin density can be delocalized throughout the heme macrocycle and into the axial ligands (La Mar & Walker, 1979). Therefore, protons that are located relatively far away from the 37 iron centre (e.g., heme methyl groups and propionate groups) are susceptible to the effects of the unpaired electron(s) of the metal mainly through contact interactions. In ^-NMR spectra of ferricytochromes c, resonance frequencies of heme substituent groups are shifted considerably outside of the diamagnetic envelope. Heme methyl group protons, for example, exhibit chemical shifts between 0 and 40 ppm. Because of the near-C2 symmetry of the electronic distribution on the heme plane, these heme methyl groups cluster in pairs with the most deshielded pair (-20-40 ppm) reflecting the residence of high spin density in the corresponding pyrrole rings. The pattern of these heme group chemical shifts have been correlated with the stereochemistry of the axial ligands with respect to the heme plane (Satterlee, 1986; Marion, 1994). When the sp3 lone-pair orbital of the Met80 5-sulfur atom points towards the nitrogen of pyrrole ring A (as it does in yeast iso-1 -ferricytochrome c), heme 8- and 3-methyl group protons are the most deshielded. In cytochrome c from Pseudomonas aeruginosa, it is heme 5- and 1- methyl group protons that are most deshielded reflecting the orientation of the sulfur lone-pair orbital towards the nitrogen of pyrrole ring B (Marion, 1994). The relative orientation of the axial histidyl imidazole with respect to the heme plane is also known to steer electronic spin density in other heme proteins (e.g., myoglobin and cytochrome c peroxidase; Satterlee, 1986). However, the orientation of this side chain is constant in the known structures of cytochromes c so that it is not a significant factor in this case. The chemical shifts of axial ligand protons in low-spin cytochromes c are also influenced considerably by the paramagnetism of the iron. For example, the e-CH3 protons of Met80 resonate between -9 and -24 ppm while the Hisl8 el-CH appears typically at -25 ppm. Although the paramagnetically-induced shift is not as marked for protons of residues not directly attached to the heme, the protons near the prosthetic group are affected through dipolar interactions. Examples of this behaviour are protons of Leu48 and Pro30 which exhibit signals as low as —5 ppm. However, 38 the dipolar influence of the unpaired electron(s) on the chemical shift of nuclei falls off in proportion to the inverse cube of the distance that separates the paramagnetic center and the nuclei so that only residues in the immediate vicinity of the porphyrin are affected. In addition to being susceptible to the paramagnetism of the metal center, proton chemical shifts in heme proteins exhibit a considerable temperature dependence. Most resonances in cytochromes c obey Curie Law behaviour (i.e., the chemical shift of resonances decrease with increasing temperature), except for heme 5-methyl group protons and those of the Met80 e-CH3 group (Angstrom et al., 1982). The longitudinal and transverse relaxation rates (T{1, T{x) of nuclei are also affected significantly by the influence of the magnetic moment of unpaired electron(s). This influence, through pseudocontact and contact contributions, enhances the rates of nuclear relaxation to broaden resonances. Complex formulae (see reviews by La Mar and Walker, 1977, Satterlee, 1986, and Bertini et al, 1989) reveal that the rates of relaxation are a function of the distance separating the nuclei and the paramagnetic center (r "6), the spin transition frequencies of the nuclei/electron(s) involved, and the correlation time that modulates the dipole-dipole interaction. In cytochromes c, the electronic relaxation is fast in low- and high-spin species (ls~ 10"u-10"12 and 10"13 respectively) so that these formulae can be approximated by Equation 5 (Marion, 1994): r1-,«r2-,*(i^)2.^4)^.Ts (5) 1 2 MTC r6 s Line broadening caused by increasing relaxation rates is often an indication of changes in the environment of the nucleus and, therefore, reflects structural heterogeneity or chemical exchange. This phenomenon is not exclusive to paramagnetic molecules, but the large chemical shifts induced 39 by the unpaired electrons in such systems helps to resolve the magnetically non-equivalent environments when the rate of the chemical exchange is comparable to the difference in chemical shift between the species undergoing the exchange (La Mar & Walker, 1979). 1.5 Thesis Objectives The primary goal of the work presented in this dissertation was to identify the second lysyl residue in yeast wo-1-ferricytochrome c that replaces Met80 as an axial ligand to the heme iron at elevated pH (Hong & Dixon, 1989; Ferrer et al, 1993). Based on chemical modification results obtained using the horse heart cytochrome (Wilgus & Stellwagen, 1974), the search for this second ligand was narrowed to six out of eight lysyl residues that are located toward the C-terminus of residue 65 and which can potentially take part in the isomerization (Figure 9). Yeast cytochrome c is amenable to site-directed mutagenesis (Pielak et al. ,1985). This genetic technique was used instead of chemical modification to eliminate one alkaline conformer and thus isolate the other in a homogenous form. The second part of this work deals with the factors that influence the alkaline conformational equilibria of the protein. In particular, the influence of ionic strength, temperature, and protein structural elements were investigated to define the determinants of each transition at a molecular level. Finally, the determination of the three-dimensional structure of one of these alkaline conformers was initiated. Excerpts of this work were published prior to the completion of this dissertation (Ferrer et al, 1993; Pollock et al, 1998; Rosell et al, 1998), and others are in preparation. 40 Figure 9. Models of yeast wo-1-ferricytochrome c illustrating six of the eight lysyl residues (shown in green) that are located towards the C-terminus from residue Met65. Lys99 and Lys 100 are located at the back. The porphyrin ring is illustrated in red. The spacefilling model (A) and the ribbon representation (B) are rotated -45° relative to the perspective shown in Figure 1. 41 2. Experimental Procedures 2.1 Production of Yeast iso-l-Cytochrome c Variants All the yeast iso-1 -cytochromes c used in this work were expressed in either yeast or bacteria. Regardless of the expression host, all the proteins were encoded by the gene CYC\ (Figure 10a; Smith et al., 1979) cloned in the vector pING4 (Figure 10b) (Inglis et al, 1991). In this gene the codon for the wild-type Cys residue at position 102 (vertebrate numbering scheme) was mutated to encode for Thr to avoid protein autoreduction and dimerization (Cutler et al., 1987). Therefore, the variant Cysl02Thr is referred to as the wild-type protein, and this replacement is included in all other proteins even though it is not specifically listed in the text. In general, molecular biology techniques were carried out according to Sambrook and coworkers (1989) or as outlined in Current Methods in Molecular Biology on CD-ROM (Ausubel, 1993). All enzymes were obtained from Sigma-Aldrich, Promega, Gibco BRL, New England Biolabs, or Pharmacia Biotech. The entire open reading frame of all new mutants was sequenced with a Sequenase version 2.0 kit (Amersham) and <x-[3SS]-dATP (Dupont) prior to protein expression to ensure that only the desired mutations were present. The mutagenic and sequencing oligonucleotide primers (Table 2) were complementary to the coding strand and were synthesized with an Applied Biosystems 380A DNA synthesizer by the Nucleic Acid and Protein Synthesis Laboratory of the University of British Columbia. 2.1.1 Cytochrome c Expression in Yeast The mutants Lys79Ala and Phe82Xxx (where Xxx is Gly, Lys, Ser, Asp, His, Met, Tip, and He) were constructed in the phagemid pING4 by Dr. Stephen Inglis in the laboratory of Professor Michael Smith (Pielak etal, 1985; Inglis etal, 1991). Site-directed mutagenesis was performed by 42 Smal UAS 1 CCCGGGAGCAAGATCAAGATGTTTTCACCGATCTTTCCGGTCTCTTTGGCCGGGGTTTACGGACGATGACCGA GGGCCCTCGTTCTAGTTCTACAAAAGTGGCTAGAAAGGCCAGAGAAACCGGCCCCAAATGCCTGCTACTGGCT -380 ****** UAS 2 AGACCAAAGCGCCAGCTCATTTGGCGAGCGTTGGTTGGTGGATCAAGCCCACGCGTAGGCAATCCTCGAGCA TCTGGTTTCGCGGTCGAGTAAACCGCTCGCAACCAACCACCTAGTTCGGGTGCGCATCCGTTAGGAGCTCGT ******* GATCCGCCAGGCGTGTATATAGCGTGGATGGCCAGGCAACTTTAGTGCTGACACATACAGGCATATATATAT CTAGGCGGTCCGCACATATATCGCACCTACCGGTCCGTTGAAATCACGACTGTGTATGTCCGTATATATATA ******** GTGTGCGACGACACATGATCATATGGCATGCATGTGCTCTGTATGTATATAAAACTCTTGTTTTCTTCTTTT CACAC GCTGCTGTGTACTAGTATACCGTACGTACACGAGACATACATATATTTT GAGAACAAAAGAAGAAAA ******* CTCTAAATATTCTTTCCTTATACATTAGGTCCTTTGTAGCATAAATTACTATACTTCTATAGACACGCAAAC GAGATTTATAAGAAAGGAATATGTAATCCAGGAAACATCGTATTTAATGATATGAAGATATCTGTGCGTTTG ******* -5 10 MTEFKAGSAKKGATLF ACAAATACACACACTAAATTAATAATGACTGAATTCAAGGCCGGTTCTGCTAAGAAAGGTGCTACACTTTTC TGTTTATGTGTGTGATTTAATTATTACTGACTTAAGTTCCGGCCAAGACGATTCTTTCCACGATGTGAAAAG * * 1 * 20 * 40 20 30 KTRCLQCHTVEKGGPHKVGPNLHG AAGACTAGATGTCTACAATGCCACACCGTGGAAAAGGGTGGCCCACATAAGGTTGGTCCAAACTTGCATGGT TTCTGATCTACAGATGTTACGGTGTGGCACCTTTTCCCACCGGGTGTATTCCAACCAGGTTTGAACGTACCA * 60 * 80 * 100 * 120 40 50 IFGRHSGQAEGYSYTDANIKKNVL ATCTTTGGCAGACACTCTGGTCAAGCTGAAGGGTATTCGTACACAGATGCCAATATCAAGAAAAACGTGTTG TAGAAACCGTCTGTGAGACCAGTTCGACTTCCCATAAGCATGTGTCTACGGTTATAGTTCTTTTTGCACAAC * 140 * 160 * 180 * 60 70 80 WDENNMSEYLTNPKKY I PGTKMAF TGGGACGAAAATAACATGTCAGAGTACTTGACTAACCCAAAGAAATATATTCCTGGTACCAAGATGGCCTTT ACCCTGCTTTTATTGTACAGTCTCATGAACTGATTGGGTTTCTTTATATAAGGACCATGGTTCTACCGGAAA 200 * 220 * 240 * 260 90 100 GGLKKEKDRNDLI TYLKKATEO GGTGGGTTGAAGAAGGAAAAAGACAGAAACGACTTAATTACCTACTTGAAAAAAGCCACTGAGTAAACAGGC CCACCCAACTTCTTCCTTTTTCTGTCTTTGCTGAATTAATGGATGAACTTTTTTCGGTGACTCATTTGTCCG * 280 * 300 * 320 + CCCTTTTCCTTTGTCGATATCATGTAATTAGTTATGTCACGCTTACATTCACGCCCTCCCCCCACATCCGCT GGGAAAAGGAAACAGCTATAGTACATTAATCAATACAGTGCGAATGTAAGTGCGGGAGGGGGGTGTAGGCGA ******* CTAACCGAAAAGGAAGGAGTTAGACAACCTGAAGTCTAGGTCCCTATTTATTTTTTTATAGTTATGTTAGTA GATTGGCTTTTCCTTCCTCAATCTGTTGGACTTCAGATCCAGGGATAAATAAAAAAATATCAATACAATCAT ******** TTAAGAACGTTATTTATATTTCAAATTTTTCTTTTTTTTCTGTACAGACGCGTGTACGCATGTAACATTATA AATTCTTGCAATAAATATAAAGTTTAAAAAGAAAAAAAAGACATGTCTGCGCACATGCGTACATTGTAATAT ******* HindiII CTGAAAACCTTGCTTGAGAAGGTTTTGGGACGCTCGAAGGCTTTAATTTGCAAGCTT GACTTTTGGAACGAACTCTTCCAAAACCCTGCGAGCTTCCGAAATTAAACGTTCGAA * * * * 600 Figure 10a. DNA sequence of the Smal-Hindm fragment of pING4 (Inglis etal, 1992). The CYCX gene (1-330) encodes the expression of yeast iso-l-cytochrome c (one-letter code symbol for amino acids and sequence number in italics; Smal and Hindlll restriction sites and the upstream activating sequences (UAS) are indicated). This figure is reproduced from Inglis et al., 1992. 43 2213 Figure 10b. Functional map of the 10.5 kbp phagemid pING4. Selected restriction sites are shown as are the genes of the P-lactamase (Amp*), yeast /'so-1-cytochrome c (CYCl), the yeast (2u), bacterial (ori), and phage (fl ori) origins of replication, and the P-isopropylmalate dehydrogenase gene (Leu2) (Satyanarayana, 1968; Mortimer, 1980). The size of the genes and the distance between restriction sites are drawn approximately to scale. 44 the method of Zoller and Smith (1982 a and b, 1984) as adapted by Kunkel and coworkers for high efficiency (1985, 1987). Single-stranded, uracil-containing DNA was obtained using E. coli strain RZ1032 {dut; ung) and helper phage R408 (Russel et al., 1986). The mutants Lys73Ala, Lys86Ala, Lys87Ala, and Lys73Ala/Lys79Ala were prepared in the same manner. However, for the latter set of mutants, the protocol of Inglis and coworkers (1991) was modified in that transfection of mutated phagemid into S. cerevisia; strain GM3C-2 was achieved by enzymatic digestion (Glusulase, Dupont) of the cell wall prior to transfection by heat shock (Ausubel, 1993). The phagemid was then rescued from this host and transfected intois. coli Sure® cells (Stratagene) by electroporation (25 uF, 2.5 kV, 200 Q) for sequencing with a Sequenase version 2.0 sequencing kit. Table 2. Synthetic mutagenic and sequencing (seq) oligonucleotide primers used in the preparation of yeast /so-1-cytochrome c variants. The underlined bases are those that introduce the mutation. Primer Mutation Nucleotide Sequence FRA73 Lys73Ala 5'-P-AGG«AAT»ATA»GGOCTT»TGG»GTT»AGT-3 FRA86 Lys86Ala 5'-P-T«TTC*CTT'CGG«CAA'CCC'ACC»AAA-3' FRA87 Lys87Ala 5,-P-GTC^TTT^TTC^CGG^CTT^CAA•CCC•ACC-3, BTR-5A Ala-5Thr S'-P-C'CTT'GAA'TTC'AGT'CAT'GGATATATCT-S1 SCI33 Tml72Ala 5'-P-C*AGG*AAT*ATA*TTT*CGC*TGG«GTT»AG-3' JB1 seq 5'-GGGGAGGGCGTGAATGTAAG-3' CYC1X seq 5'-GAATTCGAGCTCGGTACCCG-3' BTR-SA seq 5'-AGCTAGCTTACTAAAAAGAT-3' Expression of variant cytochromes c with the S. cerevisice GM3C-2::pING-4 expression system entailed fermentation of the yeast in 50 L batches as described previously (Cutler et al., 1987), except that the duration of the incubation period after lactate addition was prolonged to a total of 72 h. The YP medium consisted of peptone (2% w/v), yeast extract (1% w/v), glycerol (3 % v/v), and sodium lactate (1 % v/v) pH -5.5. In addition, the cultures were supplemented with -40 ml Antifoam 45 C, and one dose of 2 g streptomycin sulphate and 0.5 g tetracycline hydrochloride at the time of inoculation and again at the time of the lactate boost. The resulting yeast were harvested using a CEP A Z41 continuous flow centrifuge (Carl Padberg Zentrifugenbau) operating at room temperature and -20000 rpm. 2.1.2 Cytochrome c Expression in E. coli pBPCYCl/3 (vide infra, Figure 13) is a vector that harbours both CYC\ and CYC3 for the co-expression in E. coli of yeast iso-1 -cytochrome c and cytochrome c heme-lyase. pBPC YC1/3 and the mutant that encodes the variant Tml72Lys/Met80Ala were constructed by Dr. Brent Pollock, and this work is described in detail elsewhere (Pollock et al., 1998). Briefly, a DNA cassette was constructed which contains the C7C1 gene. This cassette is amenable to mutagenesis in Ml3 phage and can be incorporated into the pUC18-based plasmid pBPCYCl(wt)/3 directly upstream from CYC3 by gene exchange using the unique Sacl and BamHl restriction sites. A second vector was constructed for the expression of cytochrome c in bacteria to facilitate mutagenesis by a variety of methods and to improve protein yields. This phagemid, pBTRl, was constructed by isolating the 1.6 kbp Hind III—Ncol, CYC\- and C7C3-containing fragment from pBPCYCl(T-5A)/3. The fragment was purified by agarose gel electrophoresis and ligated to the 4 kbp Nco I—Hind III fragment of pGYM (Guillemette et al, 1991). The resulting phagemid was analyzed by restriction mapping with Cial and Seal. pBTRl was mutated to encode the Lys72Ala variant of cytochrome c using the unique site elimination approach of Deng and Nickoloff (1992) whereby the Ncol site was eliminated while at the same time the Thr-5Ala mutation was reverted. Because this vector is a phagemid, it is also possible to isolate the single-stranded form for mutagenesis by the method of Zoller and Smith (1982 a and b, 1984). 46 Expression of recombinant cytochromes c in E. coli using either pBPCYCl/3 or pBTR was carried out essentially in the same manner as for other heme proteins (Guillemette etal., 1991; Lloyd etal, 1994). Large batch cultures consisted of 8-12 2 L Erlenmeyer flasks containing 1-1.5 LLuria broth (Sambrook etal, 1989) that was supplemented with 100 mg/L ampicillin (di-sodium salt) and 0.1% v/v glycerol. A notable exception was the expression of the Tml72Lys/Met80Ala variant for which 25 mM sodium nitrate was substituted for sodium chloride. The bacteria were harvested 36-48 hours after culture initiation by pooling the contents of every flask in a 50 L fermentor which acted as a reservoir to chill the culture and for centrifugation using the continuous flow centrifuge (vide supra). 2.1.3 Purification of Cytochromes c The cytochromes c used in this work were isolated in one of three ways depending on the expression host and cytochrome c variant. Yeast pellets were treated overnight with ethyl acetate and sodium chloride to extract the cytochrome from the cells (Sherman etal, 1968;Raffertyef or/., 1990). Reducing agents such as DTT or mercaptoethanol were excluded from all buffers. For proteins that were unstable in the presence of the organic solvent, the yeast cake was suspended in sodium phosphate buffer [460 mM, pH 7.2, 4 °C] to a concentration of 15-20 % w/v. This suspension was processed twice at 4 °C with a Dyno-Mill Type KDL bead mill (Willy A. Bachofen AG Maschinenfabrik) equipped with a water-jacketed 600 ml steel vessel and 0.5 mm acid-washed, lead-free, glass beads. Cell debris was removed by filtration through a bed of wet, shredded filter paper (Whatman BioSystems Ltd.) followed by centrifugation [Sorvall GS3 rotor, 8000 rpm, 4 °C, 30 min]. The protein in the resulting solutions of ethyl acetate treatment or bead mill grinding was extracted with cellulose CM52 (Whatman BioSystems Ltd.) in batch-wise fashion with approximately 150 ml 47 wet resin per 30 L of diluted protein (resistivity < 7000 Q^cm"1). The resin was transferred to a 10 cm diameter sintered glass funnel where it was washed with two column volumes of Buffer A [46 mM sodium phosphate buffer, pH 7.2, 4 °C] before elution [Buffer A plus 1 M sodium chloride]. The resulting eluate was dialyzed against cold, distilled water to a resistivity < 7000 uQ^cm"1 and the pH was adjusted to < 7 when necessary. E. coli suspended in a minimum volume of sodium phosphate buffer [460 mM, pH 7.2] supplemented with PMSF [0.1 mM] were treated with lysozyme, frozen in liquid nitrogen and thawed to lyse the cells. DNase I and RNase A [-20 mg and -5 mg, respectively] were added together with -2 mL of a 2 M solution of magnesium chloride to break down the DNA and RNA. Cell debris was removed by centrifugation (vide supra). Ammonium sulphate was added slowly to the supernatant fluid on ice to a concentration of 320 g/L, and the solution was stirred thoroughly. Debris was removed by centrifugation, and the supernatant fluid was dialyzed as above. After dialysis the protein solutions were centrifuged, and the protein in the supernatant fluid was oxidized with a small excess of [Co(dipic)2]NH4 (Mauk etal, 1979) prior to further purification by cation exchange chromatography on CM-Sepharose CL-6B (Pharmacia Biotech) equilibrated in Buffer A. The proteins were eluted with a linear salt gradient (0-125 mM sodium chloride) in a total volume of 300-400 ml Buffer A. The final purification step consisted of Fast Protein Liquid Chromatography using a MonoS F£R 10/10 column (Pharmacia Biotech) equilibrated with sodium phosphate buffer [20 mM, pH 7.2, 25 °C]. The proteins were again developed with a linear sodium chloride gradient of 1.0 mM/mL typically above 300 mM salt. Fractions with purity index ratios (A409 5/A2g0, pH 7.2) above 4.5 were pooled, concentrated by centrifugal ultrafiltration (CentripreplO and Centricon 10 units, Amicon), flash-frozen in liquid nitrogen, and stored at -80°C until needed for experimentation. 48 Protein concentrations were determined spectrophotometrically at 410 nm assuming that the extinction coefficient of the variants at this wavelength is the same as that of horse heart ferricytochrome c (e410106.1 mM^cm'1, pH 7.0) (Margoliash & Frohwirt, 1959). The concentration of the Tml72Lys/ Met80Ala variant of cytochrome c was estimated from the reported e406 of 115 mMW1 (Bren, 1996). 2.2 Spectroscopic Techniques 2.2.1 'H-NMR Spectroscopy One dimensional 'H-NMR spectra were recorded with a Bruker MSL-200 spectrometer operating in quadrature detection mode at 200 MHz. Some of these spectra were collected in close collaboration with Dr. Juan Ferrer. Also, a Varian Unity spectrometer operating at 500 MHz was used with the assistance of Prof. Lawrence Mcintosh, Dr. Greg Connelly, or Dr. Logan Donaldson. Protein samples (2-3 mM, 400 uL) were prepared in deuterated water and either sodium phosphate buffer (50 mM) or a solution of sodium phosphate and sodium borate buffers (25 mM each). The pH was adjusted to the desired value (6-11) by titration with ~1 M deuterated sodium hydroxide or deuterated phosphoric acid, and it was measured with a micro-combination electrode (Aldrich #Z 11,343-3). The pH values are direct meter readings and are, therefore, denoted as pH* (Glasoe & Long, 1960; Kalinichenko, 1976). Spectra were obtained using a superWeft pulse sequence (Inubushi & Becker, 1983) (Figure 11a) with a recycle delay of 220 ms and consisted of at least 2000 transients. The spectral resolution was enhanced by multiplication of the FID with a 90° phase shifted sine curve prior to Fourier transform, and baselines were corrected with an interactive polynomial function. In cases where spectra were integrated to estimate relative concentrations of a particular species within a sample, no window functions were implemented. Curve-fitting was 49 ^•""non-sel (5.8 US) ^non-sel (2.9 us) T (80-100 ms) A acq (100-120 ms) ^non-sel (2.9us) acq (100-120 ms) B selective saturation (-650 ms) Figure 11. Schematic Diagrams of the superWeft (Inubushi & Becker, 1983) (A) and the inversion transfer pulse sequences used for the investigation of the yeast iso- 1-ferricytochromes c in this work (B). The pulse durations quoted in these diagrams apply to the Bruker MSL 200 spectrometer. 50 performed using the routine of the program Grams386 (Galactic Software) to determine peak areas. For this procedure the algorithm optimized the proportion of Gaussian and Lorentzian peak character to minimize the x2 value for the fit. To determine the thermodynamic parameters associated with the alkaline transition of each of the Lys-Ala variants, protein samples were prepared (~ 3 mM protein, 50 mM sodium phosphate buffer, pH*~ 9.3), and the 'H-NMR spectra were measured as a function of temperature. The pH of the samples was assumed to remain the same at all temperatures. The areas of the resonances corresponding to the heme 8-methyl group protons of the cytochrome in the alkaline and native states were calculated, and the ratio was used to estimate the equilibrium constants. These values were plotted as a function of temperature, and the entropic and enthalpic contributions associated with each alkaline isomerization were calculated from the resulting van't Hoff plots. Magnetization transfer experiments were conducted to identify resonances that correspond to protons undergoing chemical exchange. These experiments were performed at elevated temperatures (~50°C) with selective presaturation (0.65 s) of known peaks applied through the decoupler channel (Figure lib). A variety of homonuclear, two-dimensional 'H-NMR spectra of the variant Lys79Ala (50 mM sodium borate buffer, pH 10.4) were recorded by Dr. Paola Turano in the laboratory of Professor Ivano Bertini (Dipt, di Chimica, Universita degli Studi de Firenze, Italy) with a Bruker AMX600 spectrometer. The panel consisted of 2x1 kpoint NOE (100, 30 , and 15 ms mixing time) and TOC (90 and 30 ms spin lock) spectra measured at 295 and 303 K in normal and deuterated water (Band et al., 1995). Spectral resolution was enhanced by multiplying the raw data in the FX and F2 dimensions by a pure cosine squared function prior to Fourier transform and polynomial baseline correction. The resulting 2D maps were analyzed with the program XEAS Y (ETH, Zurich) (Eccles 51 etal, 1991). All chemical shifts are referenced with respect to DSS using the residual water resonance as an internal standard. 2.2.2 EPR Spectroscopy X-band frequency (-9.5 GHz) EPR spectra were obtained with a Bruker ESP 300E spectrometer equipped with an Oxford Instruments ESR900 continuous flow cryostat, an ITC4 temperature controller, and a Hewlett Packard 5352B microwave frequency counter. Data were typically collected at 10 K with microwave power of 0.63-2 mW and with modulation frequency and amplitude of 100 kHz and 1 mT respectively. Protein samples (1-2 mM, 200 uL) included 25 mM MES and/or 25 mM CAPS buffer, and glycerol (50% v/v) as a glassing agent. The pH was adjusted by titration with 1 M sodium hydroxide (vide supra) and all samples were frozen immediately in liquid nitrogen prior to loading in the sample cavity. All spectra were corrected by subtracting the cavity signature measured without a sample tube in place but using the same collection parameters. Some of these spectra were collected with the assistance of Dr. Ralf Bogumil. 2.2.3 Electronic Absorption Spectroscopy UV/Visible absorption spectra were recorded at 25.0 °C with either a Cary 219 or a Cary 3 spectrophotometer. In general, spectra spanned the region between 700 and 260 nm with a minimum spectral resolution of 0.5 nm. Protein samples were prepared in 1.0 cm pathlength quartz cuvettes such that the absorption values were between 0.1 and 1.5 in the regions of interest. 52 2.2.4 Circular Dichroism and Magnetic Circular Dichroism Spectroscopies Circular dichroism (CD) spectra were collected in the region between 190 and 720 nm with a Jasco Model J720 spectropolarimeter equipped with a computer-controlled Neslab RTE-111 water bath. Sample concentrations (-10 uM) and path length (0.1 or 1 cm) were selected to maximize the CD signal and to minimize the photomultiplier gain. Molar ellipticities [6]^ were determined according to Equation 6 (Bychkova, 1996) where 0^ (mdeg) is the measured ellipticity, MRW is the mean residue molecular weight calculated from the protein sequence, and / (mm) and c (mg/ml) are the path length and protein concentration respectively. 9, -MRW [Gil = (6) Magnetic circular dichroism (MCD) spectra were collected in the near infra-red region (800-2000 nm) with a Jasco Model J730 spectropolarimeter equipped with an InfraRed Associates Mercury-Cadmium-Telluride solid state detector Model IS7 that was cooled with liquid nitrogen. The magnet was an Oxford Instruments SM4 superconducting magnet capable of exerting a field of 5 T. Samples were thermostatted at 298 K with an Oxford Instruments ITC4 temperature controller. Protein samples (1.5-3 mM) were prepared in deuterated buffer as for 'H-NMR and included glycerol 50 % v/v. A spectral band width of 10 nm was used together with a 1 nm data point interval, and spectra were collected at a rate of 100 nm/min between 1900 and 800 nm. All spectra are the average of at least three measurements and are corrected for the background CD signal prior to smoothing using the algorithm available in the Jasco software. The spectra are normalized with respect to protein concentration, magnetic field and path length to yield units of M"1T"1cm"1. These spectra were collected with the help of Dr. Dean Hildebrand. 53 2.3 Physicochemical Properties of Cytochrome c 2.3.1 Electrochemistry Cyclic voltammograms were obtained with a two compartment glass cell (Figure 12) that was thermostated at 25.0 °C. In all cases the reference and counter electrodes were saturated calomel (Radiometer, K401) and platinum mesh respectively. The calomel electrode was standardized relative to quinhydrone (E°'= 0.699 mV vs. SHE; Hill, 1956; Landrum et al., 1977), and potentials are referred to the SHE. Depending on the experiment, the working electrode was either a modified gold surface or edge-oriented pyrolytic graphite. The gold electrode was polished on Mastertex polishing cloth (Buehler) with slurries of increasingly finer grades of alumina to 0.05 microns. The excess alumina was removed by sonication in distilled water and thorough rinsing. The newly exposed gold surface was then electrochemically cleansed (deoxygenated 100 mM sodium perchlorate, 20 mM sodium phosphate buffer, pH 7.0) by repeated cycling between -1200 and 900 mV vs. SCE at a rate of 50 mV/sec. The surface was subsequently modified by dipping the electrode for 5 min in a saturated solution of 4,4'-dithiodipyridine [a few crystals in a 1 mL volume] (Eddows & Hill, 1977; Taniguchi etal., 1984, 1985;Elliot et al., 1986). The edge-oriented graphite electrode was polished with 0.3 micron alumina, sonicated and rinsed prior to each measurement without electrochemical cleansing or further surface modification (Panzer, 1975; Armstrong etal., 1989). The oxygen content in sample solutions (-0.3 mM protein, 0.5 ml) was reduced as much as possible by flushing the samples thoroughly with 02-depleted N2 or Ar. Because these gases were not "prepurified" grade, 02 contamination was reduced by passing the N2 or Ar gas through a 4x50 cm column packed with Catalyst R3-11 pellets (Chemalog) that had been thermally activated with a stream of H2 gas. After passage through this catalyst the gases were bubbled through a thermostated solution of 0.5 M sodium phosphate buffer, pH 7.0 to neutralize and humidify the gas. 54 working electrode Purging gas Pt wire sample level Luggin capillary Pt mesh counter electrode calomel reference electrode water bath level salt bridge Figure 12. Schematic Diagram of the two compartment electrochemistry cell used for cyclic voltammetry experiments. 55 Potentials and sweep rates were controlled with an Ursar Electronics potentistat (Oxford, UK). Voltammograms were recorded directly with a Kipp and Zonen BD 90 X-Y flat bed recorder or with a Hitachi Model VC6050 digital storage oscilloscope. For later experiments, a microcomputer equipped with a National Instruments Model AT-MIO-16x analog-to-digital converter board was used to capture the output from the potentiostat. For the analysis of the pH-dependence of the midpoint potentials of wild-type cytochrome c and its Lys73Ala and Lys79Ala variants the proteins were exchanged into 0.1 M sodium chloride. Protein solutions [~ 0.3 mM, 3 mL] were prepared, and the pH was adjusted to 5.0. Samples from this stock (-0.5 mL) were withdrawn for each measurement and pooled back with the remaining solution to adjust the pH to the next value. The voltammograms were collected on the second cycle at a sweep rate of 2 V/s as described by Barker and Mauk (1992). 2.3.2 Protein Stability The thermal stability of the cytochromes was assessed under native and alkaline conditions by thermal titrations (Lloyd etal, 1994; Hildebrand etal., 1995). These measurements were carried out with the Jasco720 spectropolarimeter by monitoring the change in ellipticity at 222 nm as a function of temperature. The protein samples [-10 uM protein, ionic strength u=5 mM, sodium phosphate buffer, pH 6.1 or u=5 mM, sodium borate buffer, pH 9.3] in a cylindrical quartz cuvette (0.1 cm path length) were heated at a rate of 50 °C/hr from 25 to 70 °C. At least three measurements were averaged and smoothed with the Jasco software utility and the first derivative of the resulting trace was calculated to determine the inflection or midpoint (Tm) of the melting curves. The reversibility of this process was estimated by inverting the temperature gradient soon after the protein was unfolded and is expressed in terms of the percentage of ellipticity that was regained. 56 2.4 Alkaline Isomerization 2.4.1 pH Titrations In general, the pKap value quoted for the alkaline transition of each cytochrome c is the value calculated from spectrophotometric titration of the 695 nm band. The standard conditions for these determinations were u=0.1 M sodium chloride, 25.0 °C. Each variant was exchanged into 0.1 M sodium chloride through repeated ultrafiltration, and the resulting solution (-2.5 mL, 5-20 uM) was placed in a 3 mL cuvette equipped with a magnetic stirrer. The pH was measured with a Radiometer Model 84 pH meter fitted with a Radiometer combination electrode Model 2321C or pHC4406. The pH was adjusted to -5.5 by addition of 0.1 M hydrochloric acid, and the spectrum was measured in the appropriate wavelength range after each micro-addition of 0.1 M sodium hydroxide. The resulting single-wavelength profiles were fitted to Equation 7: B + A-10(p^-pH) ~ l-rl0(P^"PH) where Zx is the absorbance at a particular wavelength and B and A are the final and initial Z values, respectively. Note that this equation applies to the simple deprotonation of a titratable group. Under equilibrium conditions, however, Equation 7 also holds for the determination of the apparent pXfor the more complicated pH-linked conformational reorganization of ferricytochrome c (see Appendix 1). In some instances, the resulting spectra from a pH titration of ferricytochromes c were also subjected to singular value decomposition analysis with the program Specfit (Spectrum Software Associates). These data were fitted to the model for a single deprotonation process and is applicable with the same assumptions as above. The effect of ionic strength on the pATapof the alkaline transition of the wild-type, Lys73Ala, 57 and Lys79Ala proteins was investigated. Potassium chloride was selected to minimize any sodium-induced electrode artifact at higher pH. The resulting pKap values were analyzed with the Gouy-Chapman formalism (Koutalos et al, 1990; see Appendix 2) (Equation 8): Sinh[0.5 • Inl0 • (pK^ - p^apparent)] = A • a • VC (8) In this relationship, A = SeeoA^kT'72 = 76.3 M'/2 at 25.0 °C assuming that e, the dielectric constant for yeast cytochrome c (25.0 °C) is 25 ± 10 (Simonson & Perahia, 1995), o is the average surface charge density in the region of the titrating group which has an intrinsic pKa = pK^. The influence of ionic strength on the distribution of native and alkaline conformers of wild-type cytochrome c was investigated by 'H-NMR spectroscopy. The protein was exchanged into unbuffered, deuterated water, and the solution was divided into two equal portions (~300 uL each); the first was diluted with 0.5 ml D20 (A) and the second with 0.5 ml 2.0 M potassium chloride in the same solvent (B). The pH of both solutions was adjusted to 10.0, and D20 was added to A and B to a total volume of 1.0 mL. The 'H-NMR spectrum of A was measured as described above and again after each successive additions of B to vary the concentration of potassium chloride from 0-0.5 M. The spectrum of B was measured in a similar fashion after mixing with increasing volumes of A. Attention was focused on the hyperfine-shifted region of the spectra between 10 and 25 ppm which is diagnostic of the state III «* state IV equilibrium (Hong & Dixon, 1989; Ferrer etal, 1992; Pollock et al, 1998; Rosell et al, 1998). To quantify the relative populations of the native and the two alkaline conformers, the resonances corresponding to the heme 8-methyl group protons were integrated as described in the 'H-NMR section above. 58 2.4.2 Exogenous Ligand Replacement Spectrophotometric titrations of the Lys73 Ala/Lys79 Ala variant of cytochrome c were carried out with methylamine hydrochloride (Sigma) that had been dried over phosphorus pentoxide. Two solutions were prepared which contained the same concentration of protein [-10 uM each] and buffer [238 mM sodium borate buffer, pH 9]. Whereas concentrated sodium chloride was added to solution A to a final ionic strength of400 mM, concentrated methylamine hydrochloride was added to solution B [u=400 mM]. Solutions A and B were mixed in varying proportions, and the A^ was monitored. The solution was allowed to reach an equilibrium (-5 min) prior to measuring the electronic absorption spectra between 720 and 260 nm. The resulting data were analyzed according to Connors (1987) to determine the apparent dissociation constant (Kd) for methylamine binding (Equation 9). Abs = Abs00-[NH3CH3] + Abs0-^d [NH3CH3] + /:d () The EPR and ^-NMR spectra of the methylamine adduct of the Lys73 Ala/Lys79Ala variant were measured at pH 9 to confirm that the protein binds methylamine. The ^-NMR spectrum of the methylamine adduct of the Met80Ala variant of cytochrome c and horse heart myoglobin were also recorded for comparison. 2.4.3 pH-Jump Kinetics Unbuffered solutions of the wild-type, Lys73Ala, and Lys79Ala proteins [~ 9 uM] were prepared in 0.1 M sodium chloride and adjusted to pH 6.0. 40 mM buffer solutions [40 mM] of MES (pH 6.07), sodium phosphate (pH 7.07, 7.50, and 7.57), TAPS (pH 8.09, 8.19, 8.43, 8.76, and 8.94), CHES (pH 9.05, 9.54, and 9.80), sodium borate (pH 9.01, 9.28, 9.50, 9.70, and 9.89), and CAPS 59 (pH 10.17) were prepared with enough sodium chloride to a final ionic strength of 0.1 M. The pH-jump rapid mixing experiments were performed at 25.0 ± 0.2 °C with a Dionex Model S-100 stopped-flow spectrometer interfaced to a microcomputer. Absorbance changes were monitored at 390 nm which is reported to reflect the same event as that observed at 695 nm (Kihara et al., 1976). Kinetic data were fitted to a first-order exponential function, and the resulting rate constants (kobs) were fitted to the minimal model of Davis etal. (1974) (Equation 10): ^obS = K + kf K}i+ (10) KH+[H+] K ' where k{ and £bare the forward and reverse rate constants respectively, KH is the dissociation constant of the titrating group, and [FT] is the proton concentration calculated from the pH of the buffer. In the kinetic analyses, the rate and dissociation constants were constrained to agree with the equilibrium pK3p of the alkaline transition determined independently by titration of the 695 nm band. This constraint obeys the relation: P^ap = " lOg10 (^7^-) = VKC + P^H (11) where Kc = kfl kh. The pH-jump kinetics of selected yeast iso-1 -cytochrome c variants were also measured above pH 10 at 25.0 °C with an Olis RSM1000 rapid scanning stopped-flow spectrophotometer (On-Line Instrument Systems Inc.) which was equipped with a 2 cm pathlength cuvette. Spectra were collected between 400 and 630 nm with a period of 1 msec and a resolution of 0.7 nm. The results were fitted to two sequential monoexponential processes with the global fitting routine (Matheson, 1989) provided by Olis. 60 2.5 Numerical Analysis of Data Quantitative experimental results were generally analyzed by fitting the data to the appropriate model(s) with the program Scientist version 2.0 (Micromath Scientific Software Inc.). This program uses a least-squares minimization procedure based on a modification of Powell's algorithm (1970). The statistical criteria used to accept the results from individual fitting analyses included a correlation coefficient > 0.99 and low standard deviations relative to the absolute value of each parameter, typically less than 10%. These standard deviations are indicated in parenthesis. 61 3 Results 3.1 Production of Yeast /so-l-Cytochrome c Variants 3.1.1 Construction of pBTRl Ligation of the Ncol-HindlW fragments originating from pGYM and pBPCYC 1 (T-5 A)/3 AN yielded the 5.6 kbp phagemid pBTRl. Digestion of this phasmid with Cla\ results in two fragments of approximately 1.2 and 4.4 kbp. Digestion with Seal, on the other hand, yields two fragments of 2 and 3.5 kbp. Both results are consistent with the expected map of pBTRl as shown in Figure 13. In this phagemid, the f 1 origin of replication is oriented such that the coding strand of the CYC\ and CYC3 genes is packaged in the helper phage R408. Single-stranded DNA prepared using this helper phage was sequenced with primers that are complementary to the coding strand of CYC\. The 5' fusion region of pBTRl was unaffected by the ligation of the fragments. More importantly, the CYCl gene remains in frame for the correct transcription of cytochrome c using the lacZ promoter and the ribosome binding site that are exploited for myoglobin expression with pGYM (Guillemette et al, 1991). 3.1.2 Expression and Purification of Cytochromes c The source of the media used in yeast fermentation influences considerably the final yield of cytochrome c. Typically, yields of protein and semi-dry cell paste (3-4 mg/1 and 20-30 g/1, respectively) are highest when the peptone and yeast extract are from either Merck or Difco. When Sigma yeast extract is used, the protein yield decreases by as much as 90 % even though the cell mass obtained is comparable to cultures which contained Merck or Difco yeast extract. This outcome varies with the lot of the yeast extract, and may result from contamination by glucose which is known to inhibit cytochrome c expression in Saccharomyces cerevisiae (Zitomer et al, 1979; Laz, et al, 62 £C/136JJ 238 Sad 240 Kpnl 2 Fspl 35741 Pspl406l 3568-J Seal 3316 Ps/>14061 3195 .Sspl 2992 AatI 2878 iVifel 2623 Narl 2571 JfosI 2570 Fs/>I 2551 Xmril 52oi Seal 5082 Fspl 4824 £coRI Smal Aval BamHl Xbdl 3326 NC0l3291 EcdRI 3283 5cal 3073 Kpnl 3039 ito»lHI 2934 Nhel2922, 2891 //pfifl 340 ATasI 437 Narl 438 .Sl^pl 745 5cal 1052 Kpnl 1086 BamHl im Nhel 1203, 1234 Cldl 1298 Ball 1523 Msd 1523 Afcol 1688 PmH 1773 ^4pal 1991 Pmll 2022 ATjal 2049 &/I 2392 Nj4val 2097 Pstl 2402 HindHI 2406 PssI 5579 DmJJ 5578 7//M 154 Cial 656 Drain 980 Banl 1055 JVael 1083 HindTSI ni9 Ps/I 1723 &7/I 1733 BsrEl 2017 ^4val 2018 ATjtfl 2076 Pmll 2103 2352xj4/>al2i34 MscI 2352 Ball 2602 '.EcoRI 2610 B Figure 13. Functional map of (A) the 4.8 kbp plasmid pBPCYCl(wt)/3 and (B) the 5.6 kbp phagemid pBTRl. Selected restriction sites are shown as are the genes that encode p-lactamase (Amp*), yeast /'so-1-cytochrome c (CYC\), cytochrome c heme lyase (CIO), lacZ and Trc promotors, and the origins of replication of bacteria (ori) and the phage (fl ori). Vector distances are drawn approximately to scale. The Afcol site in pBTRl is shown in bold to emphasize its use for mutagenesis by the protocol of unique site elimination (Deng & Nickoloff, 1990). 63 1984; Preznat etal., 1987; Amegadzie etal., 1990; Trumbly, 1992). Furthermore, visual estimation of the cell pellet colour suggested that the optimal culture schedule for cytochrome c production requires the lactate boost 48 h after inoculation and 72 h before harvesting as opposed to 72 and 48 hours respectively (Cutler et al, 1987). Fermentation of E. coli strain JM101 containing the phagemid pBTR results in the expression of -15 mg of purified cytochrome c per litre of medium scale culture as defined by Pollock et al. (1998). These authors report a yield of approximately 7 mg/L of culture when bacteria containing the plasmid pBPCYCl(wt)/3 are also cultivated in 2L flasks. Cytochrome c purification is achieved primarily through cation exchange chromatography. At the last chromatographic step the protein samples typically resolve into three bands. The bulk of the protein (> 90 %) elutes after a small fraction that is believed to be deamidated cytochrome c (Brautigan et al, 1978). Some protein batches also contain a third minor component which elutes after the main fraction and is a form of cytochrome c in which the N-terminal threonyl residue has been cleaved (Marcia Mauk, unpublished results). Neither the deamidated nor the cleaved protein were used in the present experiments. 3.2 Native Cytochromes c Of the five Lys-Ala variants of cytochrome c included in this study (Lys73Ala, Lys79Ala, Lys73 Ala/Lys79Ala, Lys86Ala and Lys87Ala), those modified at position 79 are the most susceptible to structural changes that could influence the alkaline transition characteristics of the protein. Replacement of Lys79 disrupts a limited hydrogen bond network involving the e-amino group of Lys79, the carbonyl group of Ser47, and the C-hydroxyl group of Tyr46 (Figure 14). The Lys79-Ser47 hydrogen bond has been purported to play an important role in stabilizing the closed 64 Figure 14. Structure of yeast wo-1-ferricytochrome c in the vicinity of Lys79. Potential hydrogen-bonding interactions of Lys79 with Tyr46, Ser47, and the heme 6-propionate group are shown, and the interatomic distances (A) are noted. 65 conformation of the cytochrome (Osheroff et al, 1980). The disruption of this so-called locking mechanism could weaken local structural elements to increase the mobility of the two halves of the protein. Nevertheless, spectroscopic and physicochemical scrutiny of the variants indicate that under native conditions (pH~6-7.0, 25 °C) the Lys-Ala replacements do not affect significantly the protein structure or its physical characteristics. 3.2.1 Spectrophotometry Under native conditions (5 mM each, MES and CAPS buffers, pH 6, 25 °C), the electronic absorption spectra of the five Lys-Ala variants of ferricytochrome c are similar to that of the wild-type protein (Figure 15). Table 3 summarizes the relative intensities of the absorption maxima with respect to the Soret band (y). Whereas the Soret band of all the proteins is centered at approximately 409 nm, the two cytochromes bearing the Lys79Ala replacement display a small blue shift (~ 1 nm) in the a/p band. The weak charge transfer band near 695 nm is present in the spectra of the five variants, indicating that the Met80-Fe3+ ligation is conserved. A higher y:695 ratio suggests that the amino acid replacement at positions 73 and 79 may have weakened the bond. However, the y:695 ratio and the position of the a/p band maxima are identical for all the spectra when measured in 20 mM sodium phosphate buffer, pH 7.2 (u~50 mM). Table 3. Ratios of electronic absorption maxima intensities of yeast /so-l-ferricytochromes c measured in 5 mM each, MES and CAPS buffers, pH 6.1, 25 °C. Ratio Wild-type Lys73Ala Lys79Ala Lys73Ala/ Lys79Ala Lys86Ala Lys87Ala y : 695 119 131 130 136 125 127 y : a/p 9.6 9.6 9.8 9.8 9.5 9.6 Y : 360 3.7 3.7 3.7 3.7 3.7 3.6 Y : 280 4.5 4.5 4.4 4.6 4.5 4.5 66 Figure 15. Electronic absorption spectra of yeast /so-l-ferricytochome c: wild-type ( ), and the variants Lys73 Ala ( ), Lys79Ala ( ), Lys73 Ala/Lys79Ala ( ), Lys86Ala ( ), and Lys87Ala ( ) (12-16 uM protein in 5 mM each MES and CAPS buffers, pH 6.1, 25 C, pathlength 1.0 cm). 67 3.2.2 Natural and Magnetic Circular Dichroism Spectroscopy In the far-UV region, circular dichroism spectroscopy results indicate that the secondary structures of the Lys-Ala variants are similar to that of the wild-type protein (Figure 16, A). The minima at -222 and 208 nm are diagnostic of alpha helical content (40-42 % for equine cytochrome c, Dong et al, 1992), and the molar ellipticities ([©]) of the variants at these wavelengths (—10° •cnvVdmol) are in close agreement with the values reported for other cytochromes c (Looze et al., 1976; Myer, 1985; Betz & Pielak, 1992). At longer wavelengths, only minor differences in the magnitude of the negative Cotton effects at 282 and 289 nm are encountered. These bands are associated with the single Trp residue at position 59 in yeast /so-1-cytochrome c (Davis et al., 1993). Because the intensities of these bands are associated with the mobility of the Trp side chain in particular (Bychkova et al, 1996), the small differences observed suggest that the protein environment in the vicinity of this tryptophanyl residue is not affected significantly by the mutations. The positive and negative Cotton effects at 404 and 417 nm, respectively, are characteristic of an intact Met80-Fe3+ bond in cytochrome c (Looze et al, 1978; Santucci & Ascoli, 1997). Disruption of the metal-sulphur interaction causes these features to degrade into a single positive Cotton effect around 408 nm. The Soret region in the CD spectra of the Lys-Ala variants changes little relative to the spectrum of the wild-type protein (Figure 16, B), suggesting that the higher y :695 ratios observed by electronic absorption spectrophotometry do not reflect a significant structural change. The room-temperature, near-infrared MCD spectra of the Lys73Ala, Lys79Ala, and Lys73Ala/Lys79Ala variants of cytochrome c (Figure 17) provide further evidence that the Met80-Fe3+ ligation is maintained by these proteins at pH* 6.0. The variants display maxima at 1752 (Lys73Ala), 1761 (Lys79Ala), and 1762 nm (Lys73Ala/Lys79Ala) in good agreement with 68 o cVo -10' ^- 1 1 1 1 1 1 8 200 220 240 260 cp o 1—1 X 0.2 0 f H \\ \\ \ B -0.2 v.., / I -0.4 i i i i 250 300 350 400 450 500 Wavelength (nm) Figure 16. (A) Far-UV and (B) near-UV/visible circular dichroism spectra of yeast /'so-l-ferri-cytochomes c: wild-type ( ), and the variants Lys73 Ala ( ), Lys79Ala ( ), Lys73Ala/ Lys79Ala ( ), Lys86Ala ( ), and Lys87Ala ( ) (12-16 uM protein, in 5 mM each, MES and CAPS buffers, pH 6.1, 25 °C, measured with (A) 0.1 cm and (B) 1.0 cm pathlengths). 69 50 40 30 o S5 < 20 10 0 i 1 800 1000 1200 1400 1600 1800 Wavelength (nm) Figure 17. Near-IR MCD spectra of yeast /'so-l-ferricytochromes c: variants Lys73Ala ( , A.^1752 nm), Lys79Ala( , Amax=1761 nm), and Lys73Ala/Lys79Ala ( , 1^= 1762 nm) (1.5-3 mM protein, 25 mM each, sodium phosphate and CAPS buffers, pH* 6.0, 50% v/v glycerol, 0.1 cm pathlength, 4.83 T, 25 °C). 70 previously published values for wild-type yeast cytochrome c measured under similar conditions (1756 nm; Simpkin et al, 1989; 1750, Hawkins et al, 1994). A maximum between 1740 and 1950 nm arises from a charge transfer from a porphyrin alu (TT) orbital to the heme iron dyZ orbital in oxidized, low-spin heme proteins with His-Met ligation (Gadsby & Thomson, 1990; Cheesman et al, 1991). Neither the relative intensities of the transitions, nor the small differences in the maximum wavelength are suggestive of significant structural perturbations. 3.2.3 EPR Spectroscopy At pH 6.0-6.6 (50 mM MES, 50 % v/v glycerol, 10K), the EPR spectra of the Lys-Ala variants of yeast /'so-1-ferricytochrome c are nearly the same as that of the wild-type protein (Figure 18). The gz- values of these variants range from 3.07 to 3.10 whereas the gy- values vary between 2.22 and 2.24 (Table 4). Note that when the gx values could not be determined accurately from the spectra due to g-strain, these values were calculated assuming that the sum of the squared ^-values equals 16 (Palmer, 1985). The observed and calculated ^-values range from 1.15-1.26. The rhombicity (V/A) and tetragonal field (A/A) calculated from these g-values are included in Table 4 and place the five variants and the wild-type protein within region C of Blumberg and Peisach's crystal field correlation (Figure 8; 1971a, b). In combination with the MCD results above, these EPR data confirm that the Met80-Fe3+ bond remains intact despite the Lys - Ala replacements. 71 Figure 18. X-band EPR spectrum of wild-type yeast /'sol-ferricytochrome c (~ 1.2 mM protein in 20 mMMES buffer, pH 6.6, 50% glycerol, 10 K). This spectrum is representative of the spectra of the Lys-Ala variants measured near neutral pH. Calculation of the rhombicity and the tetragonal field from these g-values reveals that these cytochromes correspond to group C (solid circle) as shown in the "truth" diagram of Blumberg and Peisach (Figure 8; 1971). 72 Table 4. X-band EPR parameters of Lys-Ala variants of yeast wo-1-cytochrome c (~ 2 mM samples in 50 % glycerol plus 20 mM each, potassium phosphate and CAPS buffers, pH 6.6 or 20 mM each, MES and CAPS buffers, pH 6.0-6.6 (Lys86Ala and Lys87Ala), 10 K, -9.45 GHz). wild-type Lys73Ala Lys79Ala Lys73Ala/ Lys79Ala Lys86Ala Lys87Ala g. 3.07 3.07 3.09 3.10 3.08 3.09 Sy 2.24 2.24 2.22 2.22 2.24 2.23 gx -1.24 -1.26 -1.25 -1.26 -1.26 -1.25 V/Aa 0.775 0.782 0.814 0.824 0.789 0.803 A/Aa 2.381 2.422 2.431 2.455 2.424 2.417 a. Note that the original axis system of Blumberg and Peisach was obtained by transforming the proper axis system according to Hall et al. (1974) (g^gy, g/*-g„ and gx-*-g;, see Palmer, 1979). 3.2.4 'H-NMR Spectroscopy At neutral pH (50 mM sodium phosphate buffer, pH* 7.0, 20 °C), ^-NMR spectroscopy reveals that of the five variants Lys73 Ala and Lys86Ala are the most similar to the wild-type protein in terms of the frequency of the hyperfine shifted resonances (Figure 19, a and b). The peaks corresponding to the heme methyl groups and the axial heme iron ligands are all within 0.2 ppm from the frequency at which the corresponding protons resonate in the spectrum of the wild-type protein (Table 5). However, the heme 3-methyl group resonance in the spectra of these two proteins is less intense than the neighboring resonance of the heme 8-methyl group. It is possible that these two proteins display the broadening phenomenon described by Burns and La Mar (1981). Cytochromes c from certain higher eukaryotes experience some localized structural heterogeneity in the vicinity of the heme 3-methyl group that results from Phe82 residing in two non-equivalent environments. At lower temperatures the exchange between these two conformations is sufficiently slow to be detected. However, the 'H-NMR spectra of the Lys73Ala and Lys86Ala variants were not measured at I 1 1 I 1 I I I I 40 30 20 10 Chemical Shift (ppm) Figure 19a. Hyperfine shifted region of the 200 MHz 'H-NMR spectra of native yeast iso-l-ferricytochromes c (2-3 mM protein, in 50 mM sodium phosphate buffer, pH* 7.0, 20 °C): (A) wild-type; (B) Lys73Ala; (C) Lys79Ala; (D) Lys73Ala/Lys79Ala; (E) Lys86Ala; (F) Lys87Ala. The resonances labeled (a), (b) and (c) correspond to heme 8-, 3-, and 5-methyl groups respectively, (d) HislS el-H, (e) and (j) propionate 7a-CH2, and (g) Hisl8 p-CH. 74 —I I I I I I I I 0 -10 -20 -30 -40 Chemical Shift (ppm) Figure 19b. Hyperfine shifted region of the 200 MHz 'H-NMR spectra of native yeast iso-l-ferricytochromes c (2-3 mM protein, in 50 mM sodium phosphate buffer, pH* 7.0, 20 °C): (A) wild-type; (B) Lys73Ala; (C) Lys79Ala; (D) Lys73Ala/Lys79Ala; (E) Lys86Ala; (F) Lys87Ala. The labeled resonances correspond to (g) Leu68 S-CH3, (h) Pro30, (/) Met80 e-CH3, (j) Hisl8 5- CH, (k) Met80 Y-CH. 75 temperatures lower than 20 °C to confirm whether or not this broadening occurs for these two proteins. Table 5. Hyperfine shifted resonances of the 200 MHz 'H-NMR spectra of wild-type ferricytochrome c and five Lys-Ala variants (2-3 mM samples in 50 mM sodium phosphate pH* 7.0, 20°C). 'H-NMR Shifts3 (ppm) WT Lys73Ala Lys79Ala Lys73Ala/ Lys79Ala Lys86Ala Lys87Ala 8-CH3 35.1 35.2 34.5 34.5 35.0 34.5 3-CH3 32.0 32.0 32.5 32.5 32.0 31.5 HAsl8, el-CH 24.8 25.0 23.5 23.6 24.8 24.2 Prop 7cc-CH 16.0 16.0 15.1 15.1 16.0 15.9 Hisl8, p-CH2 15.1 15.1 14.6 14.6 15.1 14.8 Prop 7a'-CH 12.7 12.8 12.9 12.9 12.8 12.8 12.1 12.1 12.1 11.8 11.0 10.9 10.9 10.9 5-CH3 10.3 10.3 11.9 11.8 10.3 10.6 -1.9 -1.9 -1.8 -1.6 -2.5 -2.6 -2.4 -2.4 -2.6 -2.4 Leu68 5-CHj -3.5 -3.5 -3.1 -3.1 -3.5 -3.3 Pro30, 8-CH2 -6.0 -6.0 -4.9 -4.9 -6.0 -5.7 Met80, e-CH3 -23.8 -23.8 -22.2 -22.2 -23.7 -23.2 Ksl8, 8-CH -26.9 -26.9 -24.9 -24.8 -26.9 -26.1 Met80, Y-CH -31.2 -31.3 -31.1 -31.1 -31.2 -30.4 a. Chemical shift assignments are according to Gao and Pielak (1990), and Moench and Satterlee (1989). The Lys79Ala and Lys73Ala/Lys79Ala variants exhibit more significant spectroscopic 76 differences from the wild-type protein. Some resonances in the downfield region are either absent from the spectrum of these variants, or they are shifted to a region where they are not sufficiently well resolved at 200 MHz. The aromatic region between 8 and 10 ppm in the spectra of these variants appears to be more crowded. Note also that while the resonance of heme 8-methyl group shifts upheld relative to its position in the spectrum of the wild-type protein, the signals of the heme 3- and 5-methyl group protons shift in the opposite direction. These observations could be a manifestation of the increased mobility of the protein as explained in the prologue to Section 3.2. The replacement of Lys87 with an alanine also causes some changes in the hyperfine shifted resonances of cytochrome c. However, because Lys87 is more distant from the heme environment as well as significantly exposed to the solvent, it is unclear what causes these perturbations. 3.2.5 Spectrophotometric pH Titrations One measure of the stability of the native conformation of ferricytochrome c is the pKap for the proton-linked isomerization of this species (state III) to its alkaline conformers (states IV). These values for wild-type yeast /so-1-ferricytochrome c and the Lys-Ala variants are listed in Table 6. The pKap values of the variants that contain a single Lys-Ala replacement are not too dissimilar from that of the wild-type protein. On the other hand, state III of the double variant is considerably more stable to alkaline pH and exhibits an estimated pKap~ 10.5. The accuracy with which this pKap value may be determined is compromised because the variant undergoes a spectroscopic transition that does not occur in any of the other proteins until pH~12. Whereas the ferric heme in all the other proteins remains low-spin, the native state of the double variant turns into a high- or a mixed-spin species at pH greater than 10. Consequently, the titration data for this variant could not be fitted with a model that assumes involvement of a single ionization. 77 Table 6. pK^ of the alkaline transition of wild-type yeast iso-1 -ferricytochrome c and the Lys- Ala variants (0.10 M sodium chloride, 25 °C). The numbers in parenthesis are the standard deviations obtained from non-linear least-squares fitting of the data to the Henderson-Hasselbach equation. Representative titration curves are illustrated in Figure 20. Wild-type Lys73Ala Lys79Ala Lys73Ala/ Lys86Ala Lys87Ala Lys79Ala pK 8.70(2) 8.82(2) 8.44(1) -10.5 8.60(4) 8.47(2) 3.3 Identification of the Axial Ligands in Alkaline Ferricytochrome c 3.3.1 Electronic Absorption Spectrophotometry At elevated pH (e.g., 5 mM each, MES and CAPS buffers, pH 10.3, 25 °C), the four ferricytochrome c variants that bear a single Lys-Ala replacement display nearly the same electronic absorption spectrum as the wild-type protein (Figure 21). The Soret maxima of these proteins shift to - 405 nm and increase in intensity (-19 %). The oc/p bands, centered at 528 nm, broaden and decrease in intensity (~ 12 %). The 695 nm band decreases in intensity and reflects the disruption of the Met80-Fe3+ bond (Moore & Pettigrew, 1990). The changes in the molar absorptivities of the Soret and the cc/p band, though not exactly the same, are in good agreement with published results (Rosell et al., 1998). The small discrepancy between these and the published values can be attributed to small difference in pH and other minor differences in buffer composition. Only the spectrum of the Lys73Ala/Lys79Ala variant differs from that of wild-type cytochrome (Figure 21). Spectroscopic features of the native conformation (e.g., the 695 nm band) persist even at high pH (~ 10) and only diminish to give rise to the spectrum of a high-spin species under more alkaline conditions. This high- or mixed-spin species displays a shoulder at -600 nm and a very weak maximum at -840 nm (inset). The band at -600 nm corresponds to charge transfers between the porphyrin alu and a^ TC orbitals to the ferric dz2, d^, and d^ orbitals. Although the Soret 78 pH Figure 20. (A) Representative pH titration of wild-type yeast iso-1 -ferricytochrome c. The arrows indicate the absorbance changes with increasing pH. (B) pH titration curves of yeast iso-\-ferricytochromes c monitored at 695 nm: wild-type ( •), Lys73Ala ( •), Lys79Ala ( A), Lys86Ala (--•), Lys87Ala ( o), and Lys73Ala/Lys79Ala ( •). The scale on the right corresponds to the titration data for the Lys73Ala/Lys79Ala variant only. For all titrations, protein samples (~ 150 uM) were prepared in 100 mM sodium chloride, 25 °C, and the data were fitted to the Henderson-Hasselbach equation for 1 deprotonation (2 deprotonations in the case of the Lys73Ala/Lys79Ala variant). 79 300 400 500 600 700 Wavelength (nm) Figure 21. Electronic absorption spectra of yeast iso-1 -ferricytochomes c: wild-type ( —), and the variants Lys73 Ala ( ), Lys79Ala ( ), Lys73 Ala/Lys79Ala ( ), Lys86Ala ( - •• -), and Lys87Ala ( ). The spectra were measured in 5 mM each MES and CAPS buffers, pH 10.3, 25 °C with protein concentrations of 12-16 uM in a 1 cm cuvette. Note that the spectra were normalized with respect to the Soret maximum of the wild-type protein to allow for differences in pKap. The absorbtivity values of the variants are, therefore, only approximations. The spectrum of the Lys73Ala/Lys79Ala variant in the near infrared region is compared to that of the wild-type cytochrome (inset). 80 maximum also shifts to shorter wavelengths (-3.5 nm), its intensity, and that of the a/p band are attenuated (-25 %) relative to those of the other variants. The oc/p region in the spectrum of the double variant broadens, and under some conditions, it also resolves into three maxima at 534.5, 523, and 550 nm The maximum at 550 nm is most likely a consequence of protein reduction that occurs under the conditions of the measurement (Aviram, 1972). 3.3.2 Circular Dichroism Spectroscopy Depending on the cytochrome, the far-UV circular dichroism spectra measured at pH 10.3 (5 mM each, MES and CAPS buffers, 25 °C) reveal that the variants undergo subtle changes in secondary structure. Figure 22 illustrates how the negative Cotton effects at 222 nm, and to a greater extent at 208 nm, increase in intensity. Similar changes have been reported to occur from methanol-induced denaturation of the cytochrome (Bychkovaef al, 1996;Kamatariefa/., 1996). Guanidinium hydrochloride and thermal denaturation, on the other hand, cause a reduction in optical activity at these wavelengths (Herrmann & Bowler, 1997). In the near-UV region, all the proteins retain optical activity at 289 and 282 nm Methanol-induced protein unfolding or denaturation results in the reduction of these bands (Bychkova et al, 1996; Kamatari et al, 1996). In contrast, the alkaline transition of these cytochromes results in an increased CD signal, particularly in the Lys79Ala variant, suggesting that the structure around Trp59 becomes more rigid in the alkaline form of this variant. In the Soret region of this protein, the negative Cotton effect at 417 nm is lost entirely, and the positive optical activity at 404 nm increases as expected to follow scission of the Met80-Fe3+ bond (Looze et al, 1978; Santucci & Ascoli, 1997). 81 200 220 240 260 ~ 0.5-B -0.5 250 300 350 400 450 Wavelength (nm) 500 Figure 22. (A) Far-UV and (B) near-UV/visible circular dichroism spectra of yeast zso-1-ferri-cytochomes c: wild-type ( ), and the variants Lys73 Ala ( ), Lys79Ala ( ), Lys73 Ala/ Lys79Ala ( ), Lys86Ala ( ), and Lys87Ala ( ) (12-16 uM protein in 5 mM each, MES and CAPS buffers, pH 10.3, 25 °C, (A) 0.1 cm and (B)1.0 cm pathlengths). The spectrum of the wild-type protein measured at pH 6.1 is shown for reference (heavy solid line). 82 3.3.3 MCD Spectroscopy The room temperature MCD spectra of the Lys73Ala, Lys79Ala, and Lys73Ala/Lys79Ala variants of yeast iso-1 -ferricytochrome c measured at pH 10.5 are shown in Figure 23. These data illustrate that under these conditions the heme iron in the two single variants is bound by a nitrogenous ligand, consistent with the work of Gadsby and Thomson (1990). The Lys73Ala and Lys79Ala variants display maxima at 1477 (e >65 M'1 cm'1 T"1) and 1473 nm (e~58 M"1 cm"1 T"1), respectively. The residual peak at ~ 1770 nm in the spectrum of the Lys73Ala variant reflects the higher pKap for the formation of the alkaline species of this protein. Note, therefore, that the intensity of the CT at 1477 nm in the spectrum of this variant is greater than that of the Lys79Ala cytochrome. The different intensities presumably reflect the different coordination geometry that results from ligation of Lys73, or Lys79 to the heme iron. As the angle formed by the coordinating amino group of the coordinating lysyl residue and the heme plane approaches 90°, the charge transfer associated with this band has a higher probability of occurring (Thomson & Gadsby, 1990). Hence, the lower intensity of the CT in the spectrum of the Lys79Ala variant relative to that of the Lys73Ala cytochrome suggests that the amino group of Lys73 coordinates the iron at a greater angle with respect to the normal to the heme plane. At high pH, the near-IR MCD spectrum of the Lys73 Ala/Lys79Ala variant differs from the corresponding spectra of the single variants. The band centered at 1766 nm is attributed to the residual Met80-bound heme which appears to comprise -50 % of the sample while the remainder is likely in a high-spin state. Note that the feature near 850 nm, and perhaps that above 1000 nm, support this conclusion (Lloyd etal, 1991). 83 60-7. 40 ID < 20 0 • / » • I » •I v. 'V / 800 1000 1200 1400 1600 1800 Wavelength (nm) Figure 23. Near-IR MCD spectra of yeast iso-1 -ferricytochrome c variants: Lys73 Ala ( ; Amax= 1477 run), Lys79Ala( ; = 1473 nm), andLys73Ala/Lys79Ala( ;Amax= 1766 nm)(l.5-3 mM protein, 25 mM each, sodium phosphate and CAPS buffers, pH* -10.5, 50% v/v glycerol, 0.1 cm pathlength, 4.83 T, 25 °C). 84 3.3.4 'H-NMR Spectroscopy The downfield region of the 'H-NMR spectrum between 10 and 25 ppm is diagnostic of the alkaline character of ferricytochrome c (Hong & Dixon, 1989; Ferrer et al, 1993; Rosell et al, 1998). Figure 24, b-d illustrate how at mildly alkaline pH the native isomer of yeast iso-\-ferricytochrome c (state III) is in equilibrium with two alkaline forms (states IV) reported by Hong and Dixon (1989). At 45 °C and pH* 9.3, the heme 8-methyl protons of the native state resonate at -32 ppm This group is in chemical exchange with two species (IVa, 23.5 ppm and IVb, 23 ppm) which increase in concentration with increasing pH or temperature and at the expense of the native protein. The remaining three heme methyl groups also display this conformational degeneracy which is consistent with the formation of two alkaline forms. Unlike the Lys86Ala and Lys87Ala variants, which have essentially the same 'H-NMR spectrum as the wild-type protein, the variants that contain Ala at position 73 or 79 display remarkably simplified spectra. Removal of either of these lysyl residues confines the protein to the formation of only one of the two possible alkaline conformers. Whereas one set of heme methyl group resonances is eradicated from the spectrum of the Lys73 Ala protein (Figure 24 E), it is the other set of peaks that is absent from the spectrum of the Lys79Ala variant (Figure 24 F). These observations unquestionably identify Lys73 and Lys79 as the two endogenous ligands in yeast iso-l-ferricytochrome c that replace Met80 as the sixth ligand to the heme iron in response to an increase in pH. The appearance of a signal corresponding to a high-spin species at the expense of a persistent resonance of the native form in the spectrum of the Lys73Ala/Lys79Ala variant (Figure 24 G) is further evidence of the involvement of these lysyl residues in the alkaline transitions of the yeast protein. In this cytochrome, neither Lys73, nor Lys79 is available to replace Met80, resulting in a stabilized native state and a destabilized state V. 85 I I I . I I I I I— 50 40 30 20 10 Chemical Shift (ppm) Figure 24. Hyperfine shifted region of the 200 MHz 'H-NMR spectra of yeast iso-1 -ferricytochrome c: wild-type pH* 7.0 (A) and pH* 9.3 (B), and the variants Lys86Ala (C); Lys87Ala (D); Lys73Ala (E); Lys79Ala (F); Lys73Ala/Lys79Ala (G). These spectra were measured at 45 °C with 2-3 mM protein in 50 mM sodium phosphate buffer, pH* 9.3 (B-F), and pH* 10.0 (G). The roman numerals refer to the native (III) or alkaline (IV) conformer of the protein as defined by Theorell and Akesson (1941), the subscripts a and b specify the Lys73- or Lys79-bound alkaline conformer respectively, and the number in parenthesis refers to the heme methyl group that gives rise to the resonance. 86 3.3.5 EPR Spectroscopy The EPR spectra of the Lys- Ala variants are consistentwith the 'H-NMR spectra. At alkaline pH wild-type ferricytochrome c consists of multiple low-spin species as shown by the EPR spectrum in Figure 25a. These low-spin species are characterized by Rvalues of 3.04 (state UJ), 3.33 and 3.53 in agreement with published resultsi (Brautigan e/ al, 1977; Gadsby et al, 1987). The two non-native forms exhibit EPR parameters that are similar to those of the Ar-butylamine adductofleghemoglobin (Gadsby etal, 1987) and to cytochrome/(Rigby etal, 1988) which havegz = 3.38 (pH 7.5) and 3.51, respectively. Raising the pH of the Lys86Ala and Lys87Ala variants results in essentially the sameEPR spectrum observed for the wild-type cytochrome. On the other hand, replacement ofLys by Ala at positions 73 or 79 eliminates one of the two non-native low-spin species from the EPR spectrum of these proteins (Figure 25 b and c). Whereas the gz = 3.33 component disappears from the spectrum by the removal of Lys73, removal of Lys79 eliminates the gz = 3.53 component. Addition of the spectra of the Lys73 Ala and Lys79 Ala variants measured at alkaline pH results in a spectrum (Figure 25d) which closely resembles the spectrum of the wild-type protein measured at high pH. The shoulder that is observed atgz ~ 3.45 in the spectra of the Lys73 Ala and Lys79Ala variants is attributed to less important subconformations or folding intermediates of the protein that are formed upon sample preparation. Further confirmation that Lys73 and Lys79 are involved in the formation of the two non-native low-spin species is found in the EPR spectrum of the double variant (Figure 25e). For this protein, the native conformation persists at higher pH as already described. Accordingly, the EPR spectrum of the Lys73 Ala/ Lys79 Ala variant is essentially that of the native cytochrome which coexists with ahigh-spin species that has a highly rhombic signature withg= 6.28. 87 3.33 3.53 I J L 100 200 300 400 500 Magnetic Field (mT) Figure 25. X-band EPR spectra of yeast 7'50-l-ferricytochromes c: wild-type (A); Lys73Ala (B); Lys79Ala (C); Lys73Ala/Lys79Ala (E) (~2 mM protein in 50 mM CAPS buffer, pH 10.5, 50% v/v glycerol, 10K). Spectrum D is the sum of spectra B and C. 88 3.3.6 Modeling the Spectroscopic Features of an Alkaline Conformer The Lys73Ala/Lys79Ala variant of cytochrome c does not form a conventional alkaline conformer with an endogenous lysyl residue bound to the heme iron. However, under mildly alkaline conditions the native Met80 ligation is perturbed by the addition of exogenous amine compounds. Figure 26 illustrates the results of a spectrophotometric titration of the Lys73Ala/Lys79Ala variant with methylamine hydrochloride at pH 8.9 and 25 °C. The spectroscopic changes observed upon addition of the ligand are similar to those seen in the spectra of the other cytochromes. The Soret band shifts to higher energy (-405 nm), and it increases in intensity. The a/p band becomes less intense, and the 695 nm band disappears. Analysis of the titration data yields an apparent dissociation constant of0.037(2) M"1 for binding of methylamine. Note that this Kd includes the binding equilibria involving the exogenous methylamine and Met80. Other amine-containing compounds such as ethanolamine and lysine were also tested. However, investigation of the affinity of the cytochrome for these ligands is complicated by sulphur-containing contaminants in the ethanolamine and the precipitation of the lysine at the elevated concentrations necessary for complex formation. The 'H-NMR spectra of the mixture of methylamine hydrochloride and the Lys73Ala/ Lys79Ala variant of cytochrome c (Figure 27) are consistent with the formation of an alkaline-like species upon addition of the ligand to the variant. At concentrations as low as 30 mM methylamine (i.e., ~ 10:1= ligand: protein), the intensities of the native heme 8- and 3-methyl groups decrease and give rise to new signals. New resonances at 26.9, 19.5, 10.5, and 9.2 ppm (20 °C) have intensities characteristic of heme methyl groups, and they increase with increasing methylamine concentrations. The resonance at 19.5 ppm is almost coincident with the heme 5-methyl group resonance of the Lys79-bound alkaline conformer. The signal at 26.9 ppm is believed to correspond to the heme 8-methyl resonance. However, the true assignment for these resonances have not been determined. The 89 1.40 J I L 300 400 500 600 700 Wavelength (nm) Figure 26. Spectrophotometric titration of yeast iso-\ -ferricytochrome c Lys73Ala/Lys79Ala by methylamine hydrochloride (~ 10 uM protein, 240 mM sodium borate buffer, pH 8.9, u = 0.5 M). The inset shows the change in absorbance at 405 nm as a function of ligand concentration. The dissociation constant calculated according to Connors (1987) for these data is 0.037(2) M"l. 90 Figure 27. Hyperfine shifted 200 MHz 'H-NMR spectra of yeast wo-1-ferricytochrome c: wild- type (WT: ~3 mM protein in 50 mM sodium phosphate buffer, pH* 9.3, 20 °C) and the Lys73Ala/ Lys79Ala variant (2-3 mM protein in 25 mM each sodium phosphate and sodium borate buffers, pH* 9.3, 20 °C plus (A) 300, (B) 200, (C) 70, and (D) 0 mM methylamine hydrochloride). Spectra A-D are normalized with respect to the intensity of the heme 8-methyl group resonance. 91 chemical exchange between the ligand-bound and the ligand-free species proceeds too slowly in the NMR time-scale to allow identification of these signals through magnetization transfer experiments. The yV-methylamine adduct of the Lys73Ala/Lys79Ala variant of cytochrome c results in a single species as demonstrated by the EPR spectrum of the complex (Figure 28). The complex has highly rhombic symmetry similar to the two alkaline conformers of ferricytochrome c. A sharp signal at gz=3.31 dominates the spectrum and resembles a feature in spectrum of the Lys73-bound alkaline conformer [g2=3.33]. Unlike the Lys73Ala and Lys79Ala variants, however, the //-methylamine adduct of the double variant does not display signs of the alkaline subconformations that form during the flash freezing process {i.e., the shoulder at approximately g=3.45). At the concentration of methylamine used (~ 3 M), ~ 50 % of the protein retains the ligand even when the pH is reduced to 6.0. 3.4 Physical Properties of the Lys-Ala Variants of Cytochrome c 3.4.1 Direct Electrochemistry of Cytochromes c Cyclic voltammetry at a 4,4'-dithiodipyridine-modified gold electrode indicates that the midpoint reduction potential of the five Lys-Ala variants of cytochrome c is unchanged from that of the wild-type protein [290 ±2 mV vs. SHE; u=0.1M sodium phosphate buffer pH 6.0] (Table 7). The small differences in the potential of the variants are within the experimental uncertainty of these measurements. The peak-to-peak separations of 55-61 mV at 25 °C and the linear response of the peak cathodic currents with respect to the square root of the scanning rate also indicate that the electrochemical reaction of each variant at this electrode surface is reversible and diffusion-controlled (Rafferty, 1992, Barker & Mauk, 1992). A pH titration of wild-type cytochrome c monitored by cyclic voltammetry reveals that a low 92 3.33 100 200 300 400 500 Magnetic Field (mT) Figure 28. Comparison of the X-band EPR spectra of the yV-methylamine adduct of the yeast iso-\-ferricytochrome c variant Lys73 Ala/Lys79Ala and the wild-type protein (~2 mM protein, 50 mMCAPS, 10K): (A) wild-type (pH 10.5, 50% glycerol), and the Lys73Ala/Lys79Ala variant: (B) pH 7.2, (C) pH 10.5, (D) pH 9.7 with ~3 M methylamine, (E) pH 6.0 with ~3 M methylamine 93 potential wave attributable to non-native protein emerges at —200 mV vs. SHE at the expense of the native protein [-290 mV vs. SHE] (Barker & Mauk, 1992) (Figure 29). The results from titrations of the Lys73Ala (a and c) and Lys79Ala (b and d) variants are shown in Figure 30 and are summarized in Table 7. Panels a and b in the Figure reflect the typical pH dependence of the peak cathodic currents measured at an edge-oriented pyrolytic graphite electrode. The pKap determined by least squares fitting of these data to the Hendersson-Hasselbach relation corresponds to the equilibrium pKap of the alkaline transition of the ferricytochrome under investigation (Barker & Mauk, 1992). The values of 8.66(5) and 8.40(7) for the Lys73Ala and Lys79Ala variants, respectively, are in good agreement with the corresponding values determined by spectrophotometric titration (Table 6). Note that the data collected above pH 10 (represented with a cross) deviate sharply from the fitted titration curve. This result may reflect a pH threshold above which the edge-Oriented pyrolytic graphite surface cannot interact favourably with the proteins (p/~9.5), thus eliciting a reduced current. These values were, therefore, excluded from the fitting routine. Panels c and d illustrate the pH-dependencies of the midpoint reduction potentials of the variants. The scatter in the potential for a given pH is on the order of 10-15 mV. Because the conditions of the measurement (i.e., 2 V/s, pge electrode, high pH, etc.) result in broad cathodic and anodic waves, the uncertainty in the determination of peak current potentials is estimated to be approximately the same magnitude as the scatter. This uncertainty accounts in part for the inconsistencies between the potentials calculated for the wild-type protein by Barker and Mauk (1992) and the values obtained for the alkaline species of the Lys73 Ala and Lys79Ala variants under equivalent conditions (Table 7). These differences are largest at low pH (~6) where the reported reduction potential of the wild-type protein was estimated to be —170 mV vs SHE (Barker & Mauk, 1992). At this pH, a wave attributable to the non-native cytochrome is present in the voltammograms 94 Figure 29. Representative cyclic voltammogram of the Lys79Ala variant of yeast iso-\ -cytochrome c (320 uM protein, in 0.1 M sodium chloride, pH 8.43, 25 °C) measured at an edge-oriented pyrolytic graphite electrode on the second scan (2 V/sec). The arrows denote the direction of the scan while A and B indicate the midpoint reduction potentials of the alkaline and native conformers of the cytochrome respectively. 95 Figure 30. Titration of the yeast iso-\-cytochrome c variants Lys73Ala (A and C) and Lys79Ala (B and D) (-0.32 mM protein, in 0.1 M NaCl, 25 °C) monitored by cyclic voltammetry at an edge-oriented pyrolytic graphite electrode (2 V/sec). Panels A and B illustrate the peak cathodic current generated on the second scan, whereas panels C and D depict the midpoint reduction potentials of the low-potential wave. Data sets from 2 (A and B) or 3 (C and D) independent studies are differentiated by different symbols). The crosses in panels A and B represent points that were excluded from the fitting of the data to the Henderson-Hasselbach model for the deprotonation of a single titrating group (solid curves). The dashed curves in panels C and D have no theoretical basis. 96 Table 7. Midpoint reduction potentials of the native and alkaline conformers of yeast iso-l-cytochrome c. Protein Midpoint Potential (mV vs SHE) Native (±2) Alkaline3 (± 10) wild-type 290 -17060 -205 845 -230 104 Lys73Ala 287 -160 ~61 -190 ~89 -240 ~10-6 Lys79Ala 289 -170 ~6-7 -190 ~84 -240 ~10-4 Lys73Ala/Lys79Ala 291 n/m Lys86Ala 289 n/m Lys87Ala 290 n/m a. Measured in u=0.1M sodium chloride, pH as indicated in the superscript, at an edge-oriented pyrolytic graphite electrode. of the Lys73Ala and Lys79Ala variants which is consistent with a midpoint potential of ~-140 mV. Although similar electrochemistry is observed for the two variants and the wild-type cytochrome, the results in the present work differ from those of Barker and Mauk (1992) in two ways. In the voltammograms of a solution of wild-type cytochrome c of comparable concentration only a high-potential cathodic wave (i.e., -260 mV) was measured below pH ~7. This wave had a magnitude of ~ 14 uA. On increasing the pH to ~ 10.2 a low-potential wave emerged (—230 mV) at the expense of the native signal and reached a maximum current of -10 uA. In contrast, the pH-dependence of the cyclic voltammograms of the Lys73 Ala and Lys79Ala variants are characterized by a low- and a high-potential wave even at pH~6 (Figure 29). The magnitude of the cathodic peak, which corresponds to the alkaline form of the variants is never below 4 uA even though at pH~6, there is less than 0.5 % of the protein in the alkaline state. Furthermore, the maximum current achievable at high pH for 320 uM protein (between 18 and 20 uA) is greater than the current measured for the cathodic peak of the wild-type protein in the native state (-16 uA). 97 3.4.2 Influence of Temperature on the Alkaline Isomers of Ferricytochrome c The stability of any organized structural element in a protein is invariably linked to temperature. In the case of cytochrome c, this interrelation has been observed to affect the strength oftheMet80-Fe3+bond(Wootene/a/., 1981; Santos & Turner, 1992; Taleretal., 1995). Consistent with these reports, increasing temperature shifts the alkaline transition equilibrium of yeast iso-\-ferricytochrome c toward the alkaline conformers (Figures 31 and 32). Figure 31 illustrates the temperature dependence of the state III state IV equilibrium for each of the Lys^ Ala variants and the wild-type protein at pH* 9.1-9.3. The equilibria were determined from the area ratio of equivalent heme methyl group resonances in the 'H-NMR spectra. Table 8 summarizes the associated thermodynamic parameters of these processes. Table 8. Thermodynamic characteristics of the state III ^ state IV equilibrium of yeast iso-l-ferricytochromes c monitored by 'H-NMR spectroscopy (50 mM sodium phosphate buffer, pH* 9.1-9.3). The equilibrium constants were calculated from the relative areas of the native and alkaline heme 8-methyl group resonances. Protein Wild-type93 Lys73Ala93 Lys79Ala93 Lys86Ala92 Lys87Ala91 AH° (kcal mol"1) 5.4(6) 5.9(9) 9.7(6) 6.6(4) 9.6(8) AS° (e.u.) 19(2) 19(3) 32(2) 23(1) 32(3) pKap 8.70(2) 8.82(2) 8.44(1) 8.60(4) 8.47(2) pKip values were determined at 25 °C (Table 6). The pH at which each experiment was conducted is indicated with the superscript. (1 e.u. = 1 cal mol"1 K"1) At this time it is unknown what role Lys87 may play in the alkaline transition of cytochrome c. The location of this residue at the surface of the molecule far from the heme center suggests that its influence is limited. However, the entropic and enthalpic changes associated with the structural reorganization of this protein are identical to those of the Lys79Ala variant. Not 98 Figure 31. van't Hoff plot of the alkaline transition equilibrium of yeast /so-l-ferricytochromes c: wild-type( •),Lys73Ala( •),Lys79Ala( A), Lys86Ala (-—•), and Lys87Ala( o). ^-NMR spectra (-2-3 mM protein, in deuterated 50 mM sodium phosphate buffer) were measured at pH* 9.3 (wild-type cytochrome, and Lys73Ala, and Lys79Ala variants), pH* 9.2 (Lys86Ala variant), and pH* 9.1 (Lys87Ala variant). Keq values were determined from the relative areas of the heme 8-methyl group resonances in the alkaline and native conformations. 99 IVb(8) IVa(8) IVb(5) IVb(3) ivb(i) IVa(3) 45 40 111(8) 111(3) 35 30 25 20 40 30 20 Chemical Shift (ppm) 10 Figure 32. Temperature dependence of the downfield hyperfine shifted region of the 200 MHz 'H-NMR spectra of wild-type yeast iso-l-ferricytochrome c (~3 mM protein in 50 mM sodium phosphate buffer, pH* 9.3). The number to the left of the spectrum indicates the temperature (°C) at which the data were collected. The peak assignment nomenclature is as in Figure 24. 100 surprisingly, these two proteins also have very similar pATaps, indicating that at mildly alkaline pH the Lys73-bound conformer is the preferred species formed by the Lys87Ala variant. Lys73Ala and Lys86Ala, on the other hand, more closely resemble the wild-type protein. The entropy and enthalpy of these three cytochromes are in good agreement with the values reported for the wild-type equine protein (AH° = 6.7(3) kcal mol-1 and AS° = 21.9 e.u.; Hong & Dixon, 1989). Table 9. Thermal stability of the overall protein fold in the native and alkaline conformers of wild-type and variant forms of yeast iso-\ -ferricytochrome c (u=5 mM sodium phosphate or sodium borate buffer). pH6.15 pH9.58 Protein Tm (°C) Reversibility (%) Tm (°C) Reversibility (%) wild-type 52.7 78 52.0 68 Lys73Ala 53.7 86 48.2 47 Lys79Ala 51.4 89 52.3 25 The estimated uncertainty in the Tm values is on the order of 0.3 °C. The thermal stabilities of wild-type ferricytochrome c and the Lys73 Ala and Lys79Ala variants were evaluated by CD spectroscopy at pH 6.15 (u = 0.01 M sodium phosphate buffer) and 9.58 (u = 0.01 M sodium borate buffer). Whereas the 'H-NMR analyses above focus primarily on the heme and its ligands, the thermal titrations monitored by optical activity at 222 nm focus on the protein secondary structure. The thermal stability of the Lys73Ala and Lys79Ala is compared to that of the wild-type protein in Table 9. The uncertainty in the Tm is estimated to be better than 0.3 °C, and that in the reversibility is ~ 10 %. These results indicate that the native Lys79Ala variant is less stable to thermal denaturation than the wild-type protein. The wild-type protein, in turn, is less stable than the Lys73 Ala variant. The relatively lower stability of the Lys79Ala variant may be a consequence of the 101 removal of the hydrogen bond network in which Lys79 participates (Figure 14). On the other hand, the increased stability of the Lys73 Ala protein is consistent with the results of Bowler and coworkers (Herrmann & Bowler, 1997).These authors have reported that replacing Lys73 with a neutral residue increases the stability of cytochrome c to guanidine hydrochloride denaturation. When the residue at position 73 is alanine, AG° associated with the folding of the protein decreases by ~ 1 kcal/mol relative to the wild-type protein. At alkaline pH, the Lys73Ala variant is clearly less stable than the wild-type protein or the Lys79Ala variant. The ~5°C decrease in the Tm of the Lys73Ala variant is a surprising result considering that the structure of the alkaline conformer should not be too is dissimilar from the structure of the native isomer. By contrast, the alkaline isomer that involves the more significant structural rearrangements to accommodate Lys73 binding to the heme becomes more stable than the native structure. Although this variant appears more stable, it also displays the lowest reversibility with only 25% of the circular dichroism at 222 nm being restored upon cooling the sample. At present, the mechanistic origins of the behaviour described above cannot be determined from the information available, but a more rigorous calorimetric study of the cytochrome in the different alkaline states may provide more information regarding the isomerization. 3.4.3 Influence of Salt on Alkaline Ferricytochrome c The stabilizing effect of some salts on the native conformer of ferricytochrome c has been known for some time (Osheroff et al, 1980; Andersson et al, 1980; Antalik etal, 1992). Figure 33 illustrates how increasing potassium chloride concentration stabilizes the native conformer of wild-type cytochrome c and of the Lys79Ala variant. In the case of the latter protein, the pKap increases by more than one pH unit over a range of 1 M potassium chloride. In contrast, the efficiency of this 102 0 0.2 0.4 0.6 0.8 1.0 KC1 Concentration (M) Figure 33. Influence of ionic strength on the alkaline transition of wild-type yeast iso-l-ferricytochrome c (25 °C). (A) Conformer distribution estimated on the basis of the relative peak areas of the native and alkaline heme 8-methyl group resonances (A Met80-, • Lys79-, and • Lys73-bound; pH 10.0), and (B) salt dependence of the pKap for the alkaline transitions of the wild-type protein ( 0) and the variants Lys73Ala ( •) and Lys79Ala ( •). The lines in (A) are drawn to illustrate the changes in isomer concentration with increasing ionic strength whereas the lines in (B) represent the non-linear least square fit of titration data to the Gouy-Chapman equation. 103 salt in the stabilization of the native Lys73 Ala variant is marginal with the pKap only increasing -0.2 pH units over the same range of potassium chloride concentration. However, it should be noted that this variant is already more stable (pKap - 8.7, 0.1 M potassium chloride) than the wild-type protein. Fitting the pKap values vs. potassium chloride concentration to the Gouy-Chapman relation yields the parameters shown in Table 10. Table 10. Gouy-Chapman parameters of wild-type yeast Lys73Ala and Lys79Ala variants. iso-\-ferricytochrome c, and the Average Surface Charge Density (A"2) wild-type 8.87 1.3 x 10"2 Lys73Ala 8.82 3.5 x 10"4 Lys79Ala 9.17 2.8 x 10"3 pKap values determined in varying concentrations of potassium chloride at 25.0±0.2 °C. The effects of increasing the potassium chloride concentration on wild-type ferricytochrome c at pH 10.0 were monitored by 'H-NMR spectroscopy as shown in Figure 34. The stabilization of the native conformer of the protein with higher salt concentrations is confirmed by the increasing intensities of the heme 8- and 3- methyl group resonances. This stabilization is maximal at approximately 0.6 M potassium chloride and then decreases. Concurrently, the Lys79-bound alkaline conformer, which represents only a small portion of the sample at low salt concentrations, becomes the major conformer at the expense of the Lys73-bound and the native forms at higher salt concentrations. The population distribution and the pKap dependence of the proteins on ionic strength are interpreted in terms of charge shielding which exerts a stronger influence on Lys73 than on Lys79. 104 KC1 111(8) 111(3) IVb(8) IVa(8) IVa(5) IVb(5) (M) :  I 35 30 25 20 15 10 Chemical Shift (ppm) Figure 34. Hyperfine shifted region of the 200 MHz 'H-NMR spectra of wild-type yeast iso-1-ferricytochrome c (~2 mM, pH* 10.0,20 °C) as a function of ionic strength. The number to the left of each spectrum indicates the concentration of potassium chloride (M). Resonance assignments are as in Figure 24. 105 Because Lys73 is highly exposed to the solvent, it is also highly susceptible to ionic strength. At low salt concentrations, a deprotonated form of Lys73 is stabilized to minimize charge repulsions on the surface of the cytochrome. Interaction of chloride ions with the protonated e-amino group of Lys73 reduces charge repulsion. The stabilization of a protonated Lys73, in turn, prevents this amino acid from replacing Met80 as an axial ligand to the heme. Consequently, with increasing potassium chloride concentrations, the native state of the cytochrome is stabilized. Lys79, on the other hand, is involved in a hydrogen bond network (Figure 14) which favours the protonated e-amino group of this residue even at low salt concentrations. Increasing potassium chloride concentrations disrupt this hydrogen bond network to release Lys79. The abundance of salt therefore resembles mutation of Lys79 to Ala in that the protein exhibits more mobility. The migration of the native heme 8- and 3-methyl group resonances with increasing salt content (Figure 34) to resemble the 'H-NMR spectrum of the Lys79Ala variant at pH 7.0 (Figure 19) is consistent with this observation. The effect of other salts was not investigated, but interaction of other salts with the cytochrome may influence the alkaline conformational equilibria in different ways. 3.5 Dynamics of the Alkaline Transition 3.5.1 pH-Jump Kinetics The pH-jump reaction profiles corresponding to the Lys73 Ala and the Lys79Ala variants (pH 6-10) are well described by a monophasic decay model (Figure 3 5). Similarly, the profiles of the wild-type protein can be fitted reasonably well to this model despite the two independent alkaline transitions that this protein can undergo. The pH-jump kinetics results for the wild-type and Lys73 Ala and Lys79Ala variants are shown in Figure 36 together with the statistical fits to the model described by Davis et al. (1974). This model accounts for a ligand exchange that is associated with the 106 ll 0.002 I | -0.002 0.460 0.450 -o co 0.440 0.430 5.0 10.0 15.0 20.0 Time (sec) Figure 35. Representative pH-jump kinetics trace of the yeast wo-1-ferricytochrome c variant Lys79Ala(~9 uM protein in 0.1M sodium chloride, pH6.0mixed with 40 mM sodium borate buffer, pH 9.45, -80 mM sodium chloride, 25 °C). The data were fitted to a monoexponential rate equation which yielded £obs=0.378(3) s"1. 107 deprotonation of a titratable group. The kinetic parameters derived by fitting the data for the wild-type, Lys73 Ala, and Lys79Ala proteins to the model are listed in Table 11 along with the parameters reported previously for wild-type yeast iso-l -ferricytochrome c (Pearce et al., 1989). Also listed in the table are the equilibrium pKap values determined by spectrophotometry titrations of the 695 nm band of each of the proteins. From the equilibrium pKap values, it appears that the wild-type protein behaves in a fashion that is intermediate between that of the two variants. All three parameters calculated (kt, kb, and pK^ for the Lys79Ala variant are greater than those for the wild-type protein which are, in turn, greater than those for the Lys73 Ala variant. This observation is also evident from Figure 36 in which the curve fitted to data obtained for the wild-type protein lies midway between the curves of the variants. This observation was explored further by reconstructing the kobs versus [H+] profile for the wild-type protein using a suitably weighted combination of the kinetic parameters of the single variants at positions 73 and 79 (Appendix 3). The inset in Figure 36 is the result of such a simulation for which the linear correlation coefficient is 0.91. Table 11. Equilibrium and kinetic parameters for the alkaline transition of yeast ferricytochromes c. Protein k( khx 102 P*H P*c + PKH P^ap" (s1) wild-type 48(2) 3.5(1) 13.9 (8) x 102 11.7(2) 8.6 (2) 8.70 (2) wild-type b 8.5 (3) 3.5(1) 244(11) 11.0(1) 8.6 (2) 8.5 Lys73Ala 1.51 (5) 1.6(1) 94 (7) 10.8(1) 8.8(1) 8.82 (2) Lys79Ala 160 (5) 4.0 (7) 4.0 (7) x 103 12.0 (3) 8.4 (4) 8.44 (1) a. Apparent p^TAP determined by non-linear least squares analysis of pH titration data measured at 695 nm. b. From Pearce et al., 1989. Note that in this reference, the values for wild type cytochrome c with Thr at position 102 and that with Cys at this position have been reversed (i.e., the line labeled "Cys-102 W.T." should bear the label "Thr-102" and vice versa). The standard deviation in the last digit of the value determined by non-linear regression is quoted in parenthesis. 108 JQ-10 1Q-9 1Q-8 JQ-7 [H+] (M) Figure 36. pH-dependence of the observed rate constants (&obs) for the alkaline transitions of yeast iso- 1-ferricytochromes c: wild-type ( ,•), Lys73Ala ( ,o), and Lys79Ala ( ,A). pH-jump kinetic experiments were carried out at 25 °C, with ~9 uM protein and a 2 cm pathlength cuvette. The curves shown are the non-linear least squares fits of the individual data sets to the model of Davis et al. (1974). The inset compares the pH-dependence of the kinetics of the wild-type protein (solid profile) to kinetics calculated from a weighted combination of the parameters determined for the Lys73Ala and Lys79Ala variants (dashed curve). 109 Despite the significant structural rearrangement that the Lys79Ala variant must undergo to introduce Lys73 into the coordination sphere of the heme iron, the stopped-flow kinetics data indicate that this isomerization is faster than the equivalent reaction involving Lys79. The pKH associated with this faster structural reorganization is also greater, thus offsetting the faster structural reorganization to give an overall equilibrium value (i.e., pA^ap) that is similar to that of the slower reaction observed for the Lys73Ala variant. Although the model of Davis et al. adequately describes the kinetic properties of the alkaline transition of ferricytochrome c, it does not take into consideration either the transition to two alkaline conformers nor the structural intermediates that must form during the ligand exchange reaction. Specifically, the Davis model does not consider the formation and decay of a 5-coordinate species that must occur during the reaction. Several authors have observed the formation of an intermediate when the pH-jump exceeds pH 10(Kiharaefa/., 1976; Saigo, 1981;Rafferty, 1992). Wild-type yeast ferricytochrome c and the Lys73 Ala and Lys79Ala variants also display this intermediate in stop-flow reactions above pH 10 (Figure 37). This transient is characterized by a band at -600 nm that is characteristic of a high-spin heme iron and that may correspond to the mandatory 5-coordinate intermediate. This species also occurs in pH-jump experiments of the Lys73 Ala/ Lys79Ala variant but not in the Tyr67Phe cytochrome (vide infra). In the case of the Lys73Ala/Lys79Ala variant, however, the species with the 600 nm absorption is more stable and persists between pH-10-11.5 under equilibrium conditions (Figure 21). 110 500 600 700 0 Wavelength (nm) 0.2 0.4 Time (s) Figure 37. Sample pH-jump reaction progress (pH 6 to pH 11.5) of wild-type yeast iso-l-ferricytochrome c. (A) The dotted lines correspond to selected spectra extracted during the first 45 msec of the reaction, and the heavy solid lines are spectra derived from singular value decomposition analysis for a model where 1 - 2 - 3. (B) Experimental, single wavelength reaction profiles measured at (a) 530 and (b) 600 nm. Ill 3.6 Yeast /so-l-Cytochrome c Expressed in E. coli Yeast iso-\-cytochrome c was recently expressed in the cytoplasm of E. coli (Pollock et al., 1998) together with the cytochrome c heme lyase that catalyzes the covalent attachment of the heme to the protein backbone (Dumont etal, 1987). At neutral pH the electronic absorption, 'H-NMR, and EPR spectra of the bacterially expressed yeast iso-\-ferricytochrome c (Figure 38) are indistinguishable from those of the yeast protein (Pollock et al, 1998): The midpoint reduction potentials of the two proteins measured at a pyrolytic graphite electrode are also indistinguishable [wild-type: 282(2) mV vs SHE, n = 60.3 mV; Tml72Lys: 283(2) mV, n = 61.1 mV], as are the thermal stabilities of the two cytochromes. The only difference between the protein expressed in yeast and E. coli that can be detected near neutrality is the lack of trimethylation of Lys72 in the bacterially expressed protein (Pollock et al, 1998). For this reason, the bacterially expressed protein is referred to from this point forward as the Tml72Lys variant of cytochrome c. The lack of trimethylation of Lys72 exerts a strong influence on the protein under mildly alkaline conditions. The pA^, of the alkaline transition of the Tml72Lys variant is 7.95(2) (u=0.1 M sodium chloride, 25 °C) which is significantly lower than that of the trimethylated protein [pATap=8.70(2)]. Furthermore, the EPR spectra observed at high pH for the proteins isolated from the two hosts (Figure 39) indicate that the cytochrome from£. coli has an additional low-spin species with a gz=3.20. Combined, these two results suggest that in the Tml72Lys variant, Lys72 can act as Overleaf Figure 38. Spectroscopic comparison of yeast iso-l -ferricytochrome c expressed in Saccharomyces cerevisiae ( Y), and Escherichia coli ( B). (A) Electronic absorption spectra (~9 uM protein in 20 mM sodium phosphate pH 7.2, 25 °C); (B) X-band EPR spectra (~2 mM protein in 20 mM sodium phosphate pH 7.2, 50 % v/v glycerol, 10 K); and (C) downfield hyperfine-shifted 200 MHz 'H-NMR spectra (~3 mM protein in 25 mM each sodium phosphate and sodium borate buffers pH* 7.2, 25 °C) 112 A 120 300 400 500 600 700 Wavelength (nm) B 300 C 40 400 500 600 Magnetic Field (mT) 700 30 20 Chemical Shift (ppm) 10 113 Figure 39. X-band EPR spectra of wild-type yeast wo-l-ferricytochromes c (~2 mM protein in 25 mM each, sodium borate and CAPS buffers, 50% v/v glycerol, 10 K): (A) cytochrome expressed in E. coli, pH 10.5; (B) cytochrome expressed in yeast, pH 10.5; (C) cytochrome expressed \nE. coli, pH 9.5; (D) cytochrome expressed in yeast, pH 9.5. 114 a heme iron ligand at a lower pH than the Lys73 or Lys79. Because the pKap of the alkaline transition of cytochrome c is a weighted average of all possible alkaline isomerizations, the overall reduced stability of the native protein suggests that the pKap for the formation of this new Lys72-bound state is less than 7.9. Cytochromes c with an alkaline transition pKap even lower than 7 are known to exist (for e.g., Phe82Gly, Pearce et al, 1989; Barker & Mauk, 1992; Pro71Thr, Wood et a/.,1987; Pro76Gly, Wood et al, 1988 a and b). The formation of a new alkaline conformer with Lys72 bound to the heme iron is confirmed by the ^-NMR spectrum of the Tml72Lys variant (Figure 40) which is significantly different from the spectrum of the trimethylated yeast cytochrome measured at alkaline pH. Magnetization transfer experiments reveal that the peaks at 22.4, 20.6, and 15.1 ppm correspond to protons that are in chemical exchange with the native heme 8-methyl group of the protein. The possibility exists that at least one of these peaks arises from a pure NOE interaction between the saturated protons and a nearby group; nevertheless, these peaks are tentatively assigned to heme 8-methyl groups of three alkaline isomers. Similarly, the resonance at 21.2 ppm is assigned to the heme 3-methyl group in one alkaline state based on the magnetization transfer observed upon inverting the corresponding resonance in the native conformer. The identification of a new alkaline conformer in the Tml72Lys variant is further supported by the restoration of the pKap of the protein to 8.65(2) when Lys72 is replaced with an alanyl residue. Comparison of the neutral and alkaline 'H-NMR spectra of the Lys72Ala variants expressed in bacteria and yeast reveals that these proteins are spectroscopically indistinguishable (Figure 40). Although these results support the formation of a Lys72-bound alkaline state conclusive evidence necessitates scrutiny of the Lys73 Ala/Lys79Ala variant of cytochrome c expressed in E. coli. 115 3 8 A B IV.(8). IVa(5) IVb(5) IVb(3) IVb(l) D 35 30 25 20 15 Chemical Shift (ppm) 10 Figure 40. Downfield hyperfine shifted 200 MHz 'H-NMR spectra of yeast iso-1- ferricytochromes c expressed in E. coli ((A) Tml72Lys ("wild-type"), (B) Lys72Ala)) and in yeast ((C) Tml72Ala variant, (D) wild-type). Protein samples (~3 mM) were prepared in 25 mM each, sodium phosphate and sodium borate buffers, pH* -9.3, and spectra were measured at 45 °C. The difference spectra obtained from inversion transfer experiments in which the heme 8-methyl (8) or the heme 3-methyl (3) groups of the native conformer were selectively inverted (indicated with an arrow) are also shown. Peak assignments are as in Figure 24. 116 3.7 Effect of Other Point Mutations on the Alkaline Transition 3.7.1 Phe82Xxx Phe82, a highly conserved residue in cytochrome c sequences, has been replaced with all naturally occurring amino acids (Pielak etal, 1985; Inglis etal, 1991). In the X-ray crystal structure of the wild-type yeast protein, the aromatic ring of Phe82 lies almost coplanar and adjacent to pyrrole rings B and C on the distal side of the heme (Figure 41), The alkaline transition of the variants with the Phe82-Gly, Lys, Ser, Asp, His, Met, Trp, and He replacements is characterized in this work. In all eight variants, the native state is destabilized significantly to alkaline pH (Table 12). The downfield hyperfine-shifted 'H-NMR spectra of the native forms of these variants are shown in Figure 42. Distinctive spectroscopic features of the native protein can be recognized. However, under more alkaline conditions, the spectra of some of these proteins bear little resemblance to those of the wild-type protein or the Lys-Ala variants (Figures 24). Peak assignments were not attempted. The only criteria used for assessing what alkaline isomers are formed and their approximate distributions are (i) similarities to assigned spectra, (ii) confirmation from EPR analysis, (iii) the relative peak intensities of groups of resonances, and (iv) the dependence of signals on temperature or pH. The 'H-NMR spectra of the Phe82Gly and Phe82Ser variants measured at mildly alkaline pH are compared to spectra of wild-type cytochrome c in Figure 43 (Ferrer, unpublished). As for the wild-type protein, it appears that each of these variants have two alkaline conformers. Whereas in the wild-type cytochrome the two alkaline isomers are formed at about equal proportions, the smaller side chain at position 82 favours one alkaline transition over the other. This bias is more obvious at higher temperature where one conformer is 3-4 times more abundant (Table 12). On the basis of this temperature dependence, it appears that the Lys73-bound conformer is the preferred species at alkaline pH for these two variants. This assignment is consistent with the relatively low pKip for the 117 Figure 41. Detail of the native three-dimensional structure of yeast iso-1 -ferricytochrome c showing the position of Phe82 (red) in relation to the heme prosthetic group (white). The iron is illustrated in orange. The native Met80 ligand and one of the alkaline ligands (Lys79) are also shown. 118 40 30 20 10 Chemical Shift (ppm) Previous Page Figure 42. Downfield hyperfine'H-NMR spectra of yeast /'so-l-ferricytochromes c measured at or near neutral pH (~2 mM protein 50 mM sodium phosphate buffer or 25 mM each sodium phosphate and sodium borate, 20 °C). (A) wild-type and the variants (B) Lys73Ala, (C) Lys79Ala, (D) Phe82Gly, (E) Phe82Lys, (F) Phe82Ser, (G) Phe82Asp, (H) Phe82His, (I) Phe82Met, (J) Phe82Trp, and (K) Phe82Ile. All spectra were measured at pH* 7.0 except for spectra E, G, and I which were recorded at pH* 6.5. The spectra are normalized with respect to the intensity of the heme 8-methyl group resonance except for the spectrum of the Phe82His variant for which the intensity of this peak is half that of the other spectra. The italicized labels denote the peaks that have been assigned in the wild-protein to (a), (b), and (c) heme 8-, 3-, and 5- methyl groups respectively, (d) Hisl8 el-H, (e) and (/) propionate 7a-CH2, and (g) His 18 p-CH. alkaline transitions of these proteins [-7.7; Pearce et al., 1989]. Replacement of Phe82 by Asp, Met, or He also destabilizes the native conformation of cytochrome c considerably (Table 12). However, these variants differ from those above in that the native state is in equilibrium with one principal alkaline conformer. The EPR spectrum of the Phe82Asp variant measured at pH 8.9 (Figure 44) displays one major, non-native, low-spin species [gz=3.29] and a minor component [gz=3A6]. While the major peak at gz=3.29 may correspond to Lys73 binding to the heme iron [gz=3.33], the shoulder is more similar to the preparation-induced artifact that was encountered in the spectra of the Lys73 Ala and Lys79Ala variants but still consistent with an amine-bound heme iron. The Lys72-bound alkaline conformer in the bacterially expressed cytochrome, and the A'-methylamine adduct of the Lys73Ala/Lys79Ala variant also display similar signals [gz~3.42 and 3.45, respectively]. 120 30 20 10 Chemical Shift (ppm) Figure 43. Comparison of the hyperfine shifted 200 MHz 'H-NMR spectra of yeast iso-l-ferricytochromes c: variant Phe82Gly at (A) pH* 7.0, 20 °C; (B) pH* 8.0, 20 °C; (C) pH* 8.0, 40 °C; variant Phe82Ser at (D) pH* 7.0, 20 °C; (E) pH* 8.5, 20 °C; (F) pH* 8.5, 40 °C; wild-type cytochrome at pH* 9.3, (G) 20 °C; (H) 40 °C. Protein samples (~3 mM) were prepared in 50 mM sodium phosphate buffer. 121 I I I I I 100 200 300 400 500 Magnetic Field (mT) Figure 44. Comparison of the X-band EPR spectra of yeast /'so-1-ferricytochrome c: (A) wild- type; and the variants (B) Lys79Ala and (C) Phe82Asp (~2 mM protein in 50 mM CAPS buffer, pH 8.9, 50% v/v glycerol, 10K(A and B), 4 K (C)). 122 Table 12. Alkaline transition of yeast iso-1 -ferricytochrome c variants. Variant Major Alkaline Ligand(s) wild-type 8.7 Lys73, Lys79 Lys73Ala 8.82 Lys79 Lys79Ala 8.44 Lys73 Phe82Gly 7.7" Lys73 Phe82Lys 6.5 Lys73 Phe82Ser 7.7a Lys73 Phe82Asp 7.4 Lys73 Phe82His 6.2, 9.9 His82, unknown Phe82Met 7.1 Lys73 Phe82Trp 6.9; 10.0 ROVOH"; Lys73, Lys79 Phe82Ile 7.2a Lys73 ? Tyr67Phe 10.7* Lys79 ? Thr78Ala 6.63 Lys73, Lys79 Met80Ala na Lys72 a Taken from Pearce et al. (1989). b Taken from Luntz et al. (1989). In the 'H-NMR spectra of the Phe82Asp variant (Figure 45a), four peaks appear upon increasing the pH* that can be attributed to heme methyl group protons [22.4, 18.1, 15.6, and 13.1 ppm at 20 °C]. The smaller peaks that arise at higher pH* presumably correspond to the Lys79-bound conformer of this variant [25.6, 19.3, and 14.0 ppm at 20 °C]. Similarly, the 'H-NMR spectra of the Phe82Met variant of cytochrome c (Figure 45b) are consistent with the formation of primarily one alkaline form. The four peaks at 23.2, 18.6, 13.0, and 11.6 ppm (20 °C) probably correspond to the heme methyl group protons. The peak at 13.0 ppm is coincident with that in the 'H-NMR spectrum 123 i i i i i i 30 20 10 Chemical Shift (ppm) Figure 45. Comparison of the downfield hyperfine shifted 200 MHz 'H-NMR spectra of yeast iso-\-ferricytochromes c: (A) Phe82Asp, pH* 8.9; (B) Phe82Met, pH* 8.2; and (C) wild-type, pH* 9.3. Proteins (~3 mM) were prepared in 25 mM each, sodium phosphate and sodium borate buffers (A and B) or 50 mM sodium phosphate buffer (C), and the spectra were measured at 20 °C. 124 of Phe82Asp; therefore, the major alkaline conformer of this protein is tentatively identified to have a Lys73-bound heme iron. This assignment is consistent with the observation that the less intense peak at 21.4 ppm (20 °C) is coincident with the peak corresponding to the heme 5-methyl group of the Lys79-bound alkaline conformer of the wild-type protein (Figure 24). The failure to align the remaining heme methyl group resonances in the spectra of the wild-type cytochrome or the Phe82 Asp variant with other peaks in the spectra of alkaline Phe82Met, however, argues against this assignment. The apparent deficit of peaks in the 'H-NMR spectra of the Phe82Ile variant measured at high temperature and pH suggests that this protein undergoes the alkaline transition to a single alkaline conformer (Figure 46). Furthermore, the extremely low pKap of this protein [~7.2; Pearce et al., 1989] also suggests that this alkaline conformer involves Lys73. Nevertheless, it is possible that the pair of peaks just above 20 ppm (20 °C; -20 ppm at 45 °C) correspond to two species rather than to two functional groups of a single species. Both peaks shift upfield with increasing temperature, and the intensity of one species appears to vary at the expense of the other. Note that these peaks bear some remarkable similarities to those observed in the 'H-NMR spectra of the Tyr67Phe variant (Figure 55) although thechemical shifts of the Tyr67Phe resonances exhibit a different temperature-dependence. The introduction of a lysyl residue at position 82 could mimic the behaviour of the bacterially expressed cytochrome c in that this new amino acid is available to replace Met80 as an axial heme ligand and could thus potentially generate a third alkaline species. The low pATap of this protein (Table 12) supports the likelihood of the formation of this new alkaline form. However, EPR spectroscopy of the Phe82Lys variant of cytochrome c (Figure 47) reveals that below pH 10.7, only two species result from the alkaline transition of this protein. These species have g=3A7 and 3.28 [gy~2.07 gx~1.3]. The more abundant species \g=3.2S] is that with Lys73 bound to the heme iron \g=3.33 125 30 20 10 Chemical Shift (ppm) Figure 46. Downfield hyperfine shifted 200 MHz 'H-NMR spectra of yeast iso-l -ferricytochrome c Phe82Ile (~3 mM protein in 50 mM sodium phosphate buffer) measured at (A) pH* 7.0,20 °C; pH* 7.5 (B) 20 °C; (C) 25 °C; (D) 30 °C; (E) 35 °C; (F) 40 °C; (G) 45 °C. 126 3.33 3.53 100 200 300 400 500 Magnetic Field (mT) Figure 47. X-band EPR spectra of alkaline yeast wo-1-ferricytochrome c: (A) wild-type, pH 10.5; and the Phe82Lys variant (B) pH 10.7; (C) pH 7.9; (D) pH 6.4 (~2 mM protein, 50 mM CAPS buffer, 50% v/v glycerol, 10K). 127 in the wild-type protein] while the second species resembles the minor component of the Phe82Asp variant [gz=3.46]. The 'H-NMR spectrum of the Phe82Lys variant is consistent with the formation of only two alkaline species (Figure 48). The signals assigned to the heme 1,3, and 8-methyl group protons of the Lys73-bound alkaline conformer in the wild-type protein spectra have close analogs in the spectra of the variant both in the high, and low temperature ranges. The larger peaks at 23.4, 14.0, and 12.3 ppm are thus in agreement with the EPR finding that Lys73 is the preferred alkaline conformer. Neither the EPR nor 'H-NMR spectroscopy data suggest that Lys82 acts as a heme iron ligand. If a Lys82-bound alkaline conformer does form, the lower pKip of the protein implies that this would be the favoured conformer. However, the spectroscopic data do not support this. Therefore, the lower pKap of the Phe82Lys variant represents a true destabilization of the native structure rather than the formation of an energetically more favourable alkaline conformer as in the case of the bacterially expressed cytochrome. In the oxidized state, the new histidyl residue in the Phe82His variant of cytochrome c is known to replace Met80 as a heme iron ligand even at neutral pH (Hawkins et al, 1994; Schejter et al, 1996; Feinberg et al, 1998). The near-IR MCD and EPR spectra of this protein were reported to be characteristic of a bis-histidine coordinated, low-spin heme iron (Hawkins et al, 1994). Although the alkaline transition properties of this variant were not reported, a pH titration of the protein (Figure 49) suggests that if a lysyl residue replaces the novel His ligand, the pKap for this transition is markedly higher (-10). Only minor spectroscopic changes are observed as the protein is titrated between pH 5.5 and 10. The data best satisfy a model consisting of four species and three deprotonations. This analysis yields values of pATap=6.2(3), 7.8(2) and 9.9(1) at 25 °C. While the first p^TAP would correspond to the replacement of Met80 by His82, the second may be linked to the / 128 1 I 1 I I I I I I l_ 40 ^ 30 20 10 Chemical Shift (ppm) Figure 48. Comparison of the hyperfine shifted 200 MHz 'H-NMR spectra of wild-type yeast iso-l -ferricytochrome c (A and D, pH* 9.3) and the Phe82Lys variant (B and C, pH* 8.4). Spectra (~3 mM protein in 25 mM each, sodium phosphate and sodium borate buffer) were measured at 20 (A and B) and 45 °C (B and C). The methyl group resonances corresponding to the Lys73-bound heme in the wild-type protein are indicated with arrows. 129 1.45 5 6 7 8 9 10 11 pH Figure 49. Single wavelength (A407) titration analysis of the yeast iso-l -ferricytochrome c variant Phe82His. The curve represents the mathematical fit of the data to three pH-dependent processes with pK values of 6.2(3), 7.8(2), and 9.9(1). 130 protonation state of the His82 ligand, and the third is probably associated with the formation of a Lys73- and/or Lys79-bound alkaline conformer(s). 'H-NMR spectroscopy reveals that at least a small proportion of the Phe82His variant exhibits normal His-Met ligation even at pH* 7.5. The peaks at -21.0 and -29.9 ppm (Figure 50) may correspond to protons of Met80 that is bound to the heme iron (see Table 5) while the peaks at 34.2 and 31.0 ppm correspond to 8- and 3-heme methyl group protons. Assignment of the very broad features around 25 and 14 ppm to heme methyl group protons in the imidazole adduct are consistent with the work of Tang and coworkers (Shao et al., 1995). The width of these features may imply that the histidine-bound species is experiencing rapid structural fluctuations. Furthermore, the peaks at 21 and 20 ppm are consistent with the 8- and 5-heme methyl group protons of a Lys-bound conformer as assigned in the equine, imidazole-bound cytochrome c (Liu et al., 1996). The sharper peaks at 22.9 and 15.2 may correspond to the second Lys-bound form that is not observed in the imidazole adduct of the horse protein. The Phe82Trp variant of cytochrome c undergoes two sequential heme-linked structural rearrangements between pH 5.5 and 10.5 (Table 12). Near neutral pH, the first process results in a high- or mixed-spin heme iron that is evident from the electronic spectra of the protein (Figure 51). An absorption maximum at -600 nm reaches maximum intensity at pH 8.5, and it increases with temperature. Together with the blue-shifted Soret band, the spectrum of this stable intermediate is similar to that observed for the Lys73Ala/Lys79Ala variant of cytochrome c at pH-10.5. Increasing pH above 8.5 leads to the conventional spectrum of alkaline cytochrome c where the band at 600 nm is not present. Inspection of the EPR spectra of the Phe82Trp variant (Figure 52) confirms that the protein is composed of multiple, pH-dependent species. At neutrality, two low-spin species [gz=3.08 and 131 A B x -10 -20 -30 x Y w 40 30 20 Chemical Shift (ppm) 10 Figure 50. Comparison of the (A) upfield and (B) downfield hyperfine shifted 'H-NMR spectra of (W) wild-type yeast iso-1 -ferricytochrome c (~3mM in 50 mM sodium phosphate buffer pH* 9.3, 20 °C) and the Phe82His variant (-0.8 mM in 25 mM each, sodium phosphate and sodium borate buffers, 20 °C) at (X) pH*~8.4; (Y) pH* -7.0; (Z) pH*~5.3. 132 500 550 600 650 700 Wavelength (nm) Figure 51. (A) Spectrophotometric titration of the yeast iso-1 -ferricytochrome c variant Phe82Trp (-150 uM protein in 0.1 M sodium chloride, 25 °C) and (B) deconvoluted spectra derived from singular value decomposition analysis for a model were n (the native conformer) - /' (a stable, high-spin intermediate) - a (the alkaline isomer(s)). In (A), the solid lines correspond to the spectra observed as the concentration of /' increases, and the dashed lines correspond to the spectra observed as this intermediate decays 133 gz = 3.39] coexist with a high-spin form of rhombic symmetry [g = 6.02 and 5.67]. The latter signals also appear in the EPR spectra of the Phe82Asp and Phe82Lys variants of cytochrome c, but these diminish with increasing pH such that at pH 9.1, they represent only a minor component. Increasing pH also results in the disappearance of the Met80-bound, native conformer [gz=3.08] and the emergence of a second highly rhombic, high-spin species (S=5/2) \g=1.2\ and gy=4.67]. The latter spectroscopic characteristics are present even at pH 10.4 and are not unlike those observed in the spectrum of the His93Tyr variant of horse-heart myoglobin [g=7.13, g=4.89; 20 mM Tris, pH 8.0, 4K] (Hildebrand et al., 1995). In this protein the proximal phenolate of Tyr93 binds to the oxidized heme iron to produce a 5-coordinate species that is spectroscopically similar to bovine liver catalase. This species also displays a 600 nm band in the electronic spectrum of the oxidized protein. For the Phe82Trp variant, the gz=7.21 peak diminishes with increasing pH as a pair of low-spin species [gz=3.38 and 3.27] emerge. The signal at g=3.2>9 probably corresponds to the Lys73-bound conformer of the protein, and that at g=3.21 may represent a subconformation of this alkaline conformer or alternatively, the Lys79-bound species. 'H-NMR spectroscopy of the Phe82Trp variant above neutral pH also reveals the presence of a high-spin form. In the downfield region a set of three broad peaks [46, 42, and 36 ppm, 20 °C] is observed (Figure 53). This 'H-NMR spectrum bears little resemblance to that of the His93Tyr variant of horse-heart myoglobin (Hildebrand, et al. 1995). However, it has been suggested that in wild-type myoglobin, a similar profile corresponds to heme methyl groups of a hydroxide-bound, low-spin Fe3+ species that is in rapid exchange with a water-bound, high-spin heme (Iizuka & Morishima, 1975). The resonance frequencies of the broad 'H-NMR peaks display a marked Bohr effect in myoglobin which results from the interaction of the water/hydroxide ligand and the distal His (E7) residue (Caughey, 1966). A six-coordinate, high-spin species is consistent with preliminary 134 Figure 52. Comparison of the X-band EPR spectra of yeast z'so-1-ferricytochromes c: wild-type ((A) 50 mM CAPS buffer, pH 10.5; (E) 50 mM MES buffer, pH 6.0), and the Phe82Trp variant (20 mM each, MES, TAPS, and CAPS buffers at (B) pH 10.8, (C) pH 8.3, and (D) pH 6.3). All samples (~2 mM protein) contained 50% v/v glycerol, and the spectra were measured at 10 K. 135 111(8)111(3) A ivb(i) IVa(3) B D 50 40 30 20 Chemical Shift (ppm) 10 Figure 53. Downfield hyperfine shifted 200 MHz 'H-NMR spectra of alkaline yeast iso-l-ferricytochrome c: (A) wild-type (50 mM sodium phosphate buffer, pH* 9.2), and the Phe82Trp variant (25 mM each, sodium phosphate and sodium borate buffer, (B) pH* 8.1; (C) pH* 9.1; (D) pH*; (E) pH* 10.9). Heme methyl group assignments are as in Figure 24. 136 resonance Raman analysis results of the Phe82Trp variant (Dopner & Hildebrandt, unpublished results). In contrast to myoglobin, however, the broad resonances in the spectra of the Phe82Trp variant of cytochrome c appear to be pH-independent. Presumably, the aqueous ligand may interact with the el nitrogen of the Trp in a similar fashion, but this interaction is impervious to pH over the range studied here. Note that the binding of an endogenous hydroxyl group to the heme iron is an alternative to the H20/OrT-bridged Trp82-heme iron interaction. Tyr67 and Thr78 could, in principle, participate in such a structure. At this time, however, the identity of the sixth ligand to the heme iron is unknown. A doublet at ~ 25.3 ppm that first appears at pH*~7 probably corresponds to alkaline heme 8-methyl group protons of two alkaline conformers while those at 20.5 and 21.2 ppm may be associated with the protons of the heme 5-methyl groups. More specifically, the peaks at 25.2 and 19.0 ppm (20 °C) are assigned to the Lys79-bound alkaline conformer of cytochrome c based on the similarity of the pattern of the resonant frequencies with corresponding peaks in the spectra of the wild-type protein. The presence of these alkaline features even at pH* 7.0 conflicts with the global analysis of the spectrophotometric pH titration data. Supposedly, at pH 8.5 most of the sample consists of the stable high-spin intermediate while only a small proportion of the protein is present in either the native or alkaline states. At this pH*, the total area subtended under the peaks that are tentatively assigned to heme 8-methyl group protons (native and alkaline) in the 'H-NMR spectra is approximately one third the area bound by the corresponding peak of the native protein in the spectrum measured at pH* 7.0. On this basis, it is estimated that the remaining 70 % of the protein is composed of the H20/OH"-bound species. Above pH* 10, the 'H-NMR spectra of the Phe82Trp variant illustrate the persistence of the high-spin species. Interestingly, the distribution of alkaline conformers changes drastically from 137 apparently equal parts of Lys73- and Lys79-bound cytochrome c to almost entirely the Lys73-bound alkaline conformer by pH* 10.9. This result is in good agreement with the observation from EPR analysis and suggests that the low-spin species with g=3.27 (Figure 52) is indeed a strained subconformation of the Lys73-bound alkaline isomer. 3.7.2 Tyr67Phe Tyr67 is another highly conserved residue in cytochromes c that is located in the distal side of the heme pocket (Figure 54). Together with Thr78, Tyr67 forms a hydrogen bond network via a bridging water molecule (H20166) (Berghuis & Brayer, 1992). In rat cytochrome c, mutation of Tyr67 to Phe enhances the stability of the native conformer of the molecule both at the acidic and alkaline extremes (A pKap=0.7 and 1.2 pH units respectively) (Luntz et al., 1989). The increased stability of the Tyr67Phe variant of the yeast protein is evident in the 'H-NMR spectra (Figure 55). Even at pH* 10, 25 °C, features of the native conformer predominate. What is not obvious is whether one or two alkaline isomers are formed at this high pH. Some spectroscopic features make this protein remarkably similar to the Phe82Ile variant, namely the pair of peaks at -20.5 ppm which may correspond to heme 8-methyl group protons of two conformers. If speculation that Lys73 is the preferred ligand in the Phe82Ile variant at high pH is correct, then the opposite temperature dependence of the intensity of the peaks in spectra of Tyr67Phe indicate that Lys79 is the predominant ligand in this case. pH-Jump experiments with the Tyr67Phe variant above pH 10 fail to detect a transient band at 600 nm, at variance with the behaviour exhibited by other cytochrome c variants described thus far. Assuming that the high-spin species in the Phe82Trp variant arises from a H20/OFT or a phenolate-bound Fe3+, the failure to observe the transient in the high pH-jump experiments of the 138 B Met80 Phe67 Thr78 Figure 54. Comparison of the three-dimensional crystal structures of (A) wild-type yeast iso-1-ferricytochrome c showing the interatomic distances (A) between Asn52, Tyr67, Thr78, and H20166, and (B) the Tyr67Phe variant. Note that in the crystal structure of the oxidized variant (1 CTY.PDB), no water molecule equivalent to H20166 could be located in the electron density map. However, the presence of two water molecules at this site in the reduced form of the variant, suggests that these molecules are also present in the oxidized protein, but they are moving too rapidly for detection. 139 IVb(5) IVb(l) 111(8) 111(3) IVa(8) IVb(8) A B D 40 30 20 Chemical Shift (ppm) 10 Figure 55. Hyperfine shifted region of the 200 MHz 'H-NMR spectra of alkaline yeast iso-l-ferricytochrome c (A and E) wild-type, pH* 9.3 ((A) 20 °C and (E) 40 °C), and (B-D) the variant Tyr67Phe, pH* 10 ((B) 20 °C, (C) 30 °C, and (D) 40 °C) (~3 mM protein in 50 mM sodium phosphate buffer). The peak assignments are as in Figure 24. 140 Tyr67Phe cytochrome c is consistent with the involvement of the wild-type tyrosyl residue or its hydrogen-bonded water molecule in the alkaline transition of the wild-type protein. 3.7.3 Thr78Ala In addition to its interaction with H20166 (Figure 54), Thr78 is hydrogen bonded to Lys79 and the heme 6-propionate group. A crystal structure of the Thr78Ala is presently unavailable to define the consequence of removing the gamma hydroxyl group of Thr78 on the positioning of H20166. However, this replacement is likely to weaken the native conformation of cytochrome c because the hydrogen bond to the propionate group does not exist, and the hydrogen bond network that includes H20166 is perturbed. Therefore, the properties of the heme prosthetic group should also be affected. The midpoint potential of the Thr78Gly variant at pH 6, for example, is 40 mV more negative than the wild-type protein (Rafferty, 1992). A decrease in potential is generally a consequence of increased solvent exposure of the heme which affects the dielectric constant of the environment around the prosthetic group. Also, the native structure of this variant is destabilized significantly under alkaline conditions (A pATap=-1.4) (Rafferty, 1992). Singular value decomposition analysis of the pH titration data for the Thr78Ala variant reveals that the transition between states II and III occurs with a pATap~4.2. Spectroscopic features of a high-spin species [Xmax=623 nm] are detectable in the electronic spectra of the protein at mildly acidic pH. Furthermore state III isomerizes into the alkaline conformers with a pATap=6.63(7), 0.5 pH units lower than what is observed for the Thr78Gly variant. 'H-NMR spectroscopy illustrates that the Thr78Ala variant of cytochrome c undergoes two alkaline transitions (Figure 56). The downfield region of these spectra are not too dissimilar from that of the corresponding spectra of the wild-type protein measured under alkaline conditions (Figure 24). 141 IVb(5) IVb(l) 111(8). 111(3) IVa(8) A B D 40 30 20 Chemical Shift (ppm) 10 Figure 56. Downfield hyperfine shifted 200 MHz 'H-NMR spectra of yeast iso-l- ferricytochromes c: wild-type, pH* 9.3 ((A) 20 °C; (E) 40 °C), and the Thr78Ala variant, pH* 7.2 ((B) 25 °C; (C) 30 °C; (D) 35 °C) (~3 mM protein in 50 mM sodium phosphate buffer). Peak assignments are as in Figure 24. 142 Furthermore, the spectra of the Thr78Ala variant also reveal that at pH* 7.2, 25 °C, the preferred alkaline conformer is that with Lys79 bound to the heme iron. It is only at higher temperatures that the Lys73-bound conformer becomes apparent. 3.7.4 Tml72Lys/Met80Ala Not surprisingly, the variant of yeast iso-1 -cytochrome c with alanine at position 80 resembles myoglobin in that it can readily bind a variety of ligands in the oxidized and reduced states (Bren, 1996; Banci etal., 1995; Twitchett, 1998). In the absence of exogenous ligands, this ferricytochrome variant possesses a water molecule as the sixth axial ligand. This molecule deprotonates with a pK^ of 6.1(1) (Bren, 1996;Twitchett, 1998) thus causing a high- to low-spin state change. The results from a pH titration monitored by 'H-NMR spectroscopy (25 mM each sodium phosphate and sodium borate buffer, 25 °C) are consistent with this finding (Figure 57). At pH*~5.5 a multitude of broad resonances appears between 22 and 62 ppm. Presumably, these peaks correspond to the high-spin Fe3+ heme in the sample. As the pH is increased, the fast relaxing signals are replaced by those belonging to protons of low-spin forms of the protein. Coincidence of the peak at 19.5 ppm (20 °C) with that identified as corresponding to IVb(5) (nomenclature as in Figure 24) suggests that the Lys73-bound alkaline conformer is present in the sample at pH* as low as 6.8. In addition, there are some spectroscopic similarities between the Tml72Lys/Met80Ala cytochrome and the wild-type protein expressed in bacteria. Based on these similarities, the peak at 22.5 ppm is assigned to the heme 8-methyl group of an alkaline conformer of the Tml72Lys/Met80Ala variant. Note that there are significant spectroscopic differences between the Met80Ala variant expressed in yeast (Banci et al., 1995) and that expressed in bacteria much as there are in the case of the wild-type proteins measured at high pH. Whether these differences result from the lack of trimethylation of Lys72 in the 143 70 50 30 10 Chemical Shift (ppm) Figure 57. Downfield hyperfine shifted 200 MHz 'H-NMR spectra of yeast iso-\- ferricytochromes c: Tml72Lys/ Met80Ala variant (A) pH* 5.5; (B) pH* 10.0; (D) pH* 7.0 plus 450 mM methylamine hydrochloride, and the Tml72Lys variant (C) pH* 9.3; (E) ./V-methylamine adduct of the variant Lys73Ala/Lys79Ala, pH* 8.9. The peaks labeled IV(X) correspond to heme 3-and 8-methyl group protons as in Figure 24. The peak indicated with an asterisk in (C) coincides with the resonance that has been assigned to a heme methyl group in the Lys72-bound alkaline conformer of the bacterial protein. 144 bacterially expressed cytochrome or to the difference in the generation of the proteins in the two expression hosts remains to be determined. As for the Lys73Ala/Lys79Ala variant, the Tml72Lys/Met80Ala cytochrome binds methylamine near neutral pH (Figure 57 d and e). Note that in the presence of this exogenous ligand, nearly all of the high-spin species is abolished and the sample becomes more homogeneous. However, the 'H-NMR spectrum of the resulting cytochrome adduct bears little resemblance to the spectrum of the iV-methylamine adduct of the Lys73Ala/Lys79Ala variant of cytochrome c (Figure 27), or of any of the alkaline conformers (Figure 24). 145 4 Discussion 4.1 The Alkaline Conformational Transition of Cytochromes c 4.1.1 Yeast iso-l-Ferricytochrome c, State IV A significant obstacle in the characterization of the alkaline conformational transition of ferricytochrome c has been the conclusive identification of the residues that coordinate the heme iron at elevated pH. Although the existance of multiple forms of cytochrome c at alkaline pH had been discussed before (Greenwood & Palmer, 1965; Brautigan et al., 1977; Gadsby et al., 1987), the demonstration by Hong and Dixon that the alkaline conformational transition of ferricytochrome c results in not one, but actually two alkaline isomers of the protein (1989), represents a major development in the characterization of this dynamic process. With knowledge of this key property of ferricytochrome c, Ferrer et al. were able to confirm earlier speculations that Lys79 is one residue that coordinates the heme iron at pH above neutrality (1993).Together with Lys72, Lys79 is cited most frequently in the literature as the amino acid that replaces Met80 as an iron ligand in the equine protein. The present work establishes that Lys73 is the residue in the yeast protein that is involved in the alternate conformational transition. Lys79 is adjacent to the native heme axial ligand Met80. Near neutral pH, Lys79 forms part of a hydrogen bond network that closes the heme cavity and limits the "breathing" of the protein (i.e., the expansion and contraction of the protein that involves the motions of the proximal and distal sides). By comparison, Lys73 is located at the surface, on the distal side of the protein. In the X-ray crystal structure of the yeast protein (Berghuis & Brayer, 1992) the side chain of this residue is almost co-aligned with a vector drawn by extending the bond from the heme iron past the 6-sulfur atom of Met80 and the e-amino group is located -19 A from the iron atom. Therefore, whereas replacement of Met80 by Lys79 at alkaline pH can be described as a "frame-shift" of the polypeptide chain that 146 occurs as Lys79 rotates into coordination position, a 180° "swinging" motion of the side chain of Lys73 is required for the second isomerization process. Comparison of a model of the Lys73-bound alkaline conformer to the X-ray crystal structure of the native cytochrome illustrates how the protein architecture might rearrange, particularly between residues 70 and 76, to give rise to this alkaline isomer (Figure 58). The swinging of Lys73 appears to extend the H3 a-helix (residues 62-69) by half a turn, while the 71-76 helix between two highly conserved proline residues (Pro71 and Pro76) is disrupted almost entirely. Although the loop containing Met80 is pushed away from the protein core so that the backbone protrudes at the exposed heme edge (~6A), Lys79 remains engaged in a hydrogen bond network that is similar to that which favours the closed conformation of the native isomer. Ligation of Lys79 to the heme iron also extends the H3 a-helix, and the 71-76 helix relaxes but not to the same extent as in the Lys73-bound conformer. Met80 remains in the heme pocket with the side chain now occupying a position parallel to the heme plane adjacent to pyrrole rings B and C. Note that the structural adjustments that occur to accommodate coordination of Lys79 appear less significant than those involving Lys73. Lys79 is in a loop region that is presumably more flexible and allows a wider range of motion without significantly disrupting other structural elements of the protein. Upon coordination of Lys79 to the heme iron, no other residue on the distal side of the protein appears to hydrogen bond any residues on the proximal side. Without such a locking mechanism, the Lys79-bound alkaline conformer may exhibit more "breathing" motions than the other alkaline isomer because the distal side of the protein is not anchored to the proximal side. The three-dimensional structures of both alkaline conformers of ferricytochrome c remain to be elucidated. Presumably, part of the problem in determining these structures by NMR or X-ray diffraction techniques stems from the conformational heterogeneity observed for the wild-type protein 147 Figure 58. Comparison of the native structure of yeast iso-1 -ferricytochrome c (B; 2YCC.PDB; Berghuis & Brayer, 1992) to models of the (A) Lys73- and (C) Lys79-bound alkaline conformers of the protein calculated by energy minimization of the structure in which the e-amine-Fe3+ bond length was constrained to -1.95 A (energy-minimized structures courtesy of Dr. Terence P. Lo). Only residues 59-82 are included (secondary structure as indicated by the ribbon) with H3 helix shown in red. The side chains of residues Lys73, Lys79, and Met80 are shown in blue, green, and orange, respectively. 148 at high pH. Also, the optimum crystallization conditions at the elevated pH required to transform fully the native protein to either alkaline conformer remain to be defined. However, preliminary NMR results indicate that the structure on the proximal side of the protein of the Lys73-bound conformer of cytochrome c is very similar to that of the native protein. As noted earlier, fifty spin systems have been identified based almost exclusively on the similarity to reported assignments for the native protein (Gao etal., 1991). These residues are located mainly on the proximal side which is predicted to undergo the least structural perturbation in both alkaline conformers. Also, a few residues have been located which are on the distal side (Figure 59). In the model of the Lys73-bound conformer, carbon atoms of Pro76 appear at almost the same coordinates as in the native protein despite the disruption of the 71-76 helix that precedes this residue. Interestingly, a spin system consistent with that of a proline residue was identified which has the same resonance frequencies as Pro76. Further work is required to elucidate the structure of the alkaline conformers of yeast ferricytochrome c completely, and before the properties of these isomers can be linked with confidence to specific structural features. This work is still in progress. 4.1.2 The Alkaline Isomers of Other Cytochromes c Although cytochromes c exhibit a high degree of sequence and structural homology, it is premature to assume that Lys73 and Lys79 also participate in the alkaline conformational transition of all cytochromes. Selected surface lysyl residues in yeast, fungal, and some plant cytochromes c are post-translationally modified by a methyl-transferase which specifically methylates the e-amino group of certain lysyl residues. In the yeast protein, the trimethylated Lys72 lacks the lone pair of electrons that is necessary to coordinate the heme iron. When expressed in E. coli, the yeast protein does not undergo trimethylation, and the native state is destabilized so that it forms a third alkaline conformer 149 Figure 59. Space filling model of (A) the proximal, and (B) the distal faces of wild-type yeast wo-1-ferricytochrome c (2YCC.PDB; Berghuis and Brayer, 1992). The structures in green represent residues for which at least partial 'H-NMR peak assignments have been made from spectra of the Lys79Ala variant collected at pH 10.8. These assignments are based on the similarity of the amino acid spin systems to the published assignments in the native wild-type protein (Gao, 1991; Paola Turano, personal communications). 150 at a lower pH than the protein that is expressed in yeast. Therefore, the decreased stability of the native conformer suggests that trimethylation in the yeast protein evolved to block Lys72 from coordinating the heme at high pH. If this function is indeed the purpose for trimethylation of this lysyl residue in yeast cytochromes c, then it is possible that in the rice protein, replacement of the Met axial ligand by Lys86 would occur were it not for trimethylation of this residue (Mori & Morita, 1980; Ochi etal, 1983). In horse cytochrome c, Lys72 is not trimethylated, yet this protein exhibits only two alkaline forms (Hong & Dixon, 1989). This observation leads one to question whether Lys73 or Lys79 are also the ligands involved in the alkaline conformational transitions of this and other cytochromes. The recent work by Banci and coworkers shows that spectroscopic features attributed to a third alkaline species of horse heart cytochrome c can be resolved with stronger magnetic fields (1998). However, the work of Ubbink et al. shows that the alkaline transitions of a cytochrome do not necessarily include Lys79 (1994). In cytochrome c550 from Thiobacillus versutus, Lys99, the residue at the position equivalent to Lys79, was mutated to glutamate with the expectation that this site replacement would eradicate the single alkaline transition of this protein. However, the Lys99Glu variant continues to undergo an alkaline conformational transition although at lower pH. The three-dimensional structure of cytochrome c550 from Thiobacillus versutus has not been elucidated, but assuming that this protein is similar to cytochrome c550 from Paracoccus denitrificans, then either Lys83 or Lysl02 are the residues in these bacterial proteins that replace the axial Met ligand at high pH. This prediction is based on the residence of these lysyl residues in the equivalent region to that where Lys72 and Lys73 are located in the yeast protein. Note that in cytochrome c550, Lys83 is flanked by prolyl residues at positions 82 and 84; therefore, if Lys83 is the heme axial ligand in this protein at high pH, then it is possible that the high degree of conservation of the prolyl residues in this area is a significant 151 structural component for the pH-dependent conformational dynamics of ferricytochromes c in general (vide infra). 4.1.3 Obstacles to the Identification of the Alkaline Ligands Prior to the implementation of genetic techniques to produce cytochrome c variants, the most widely used approach to determine the lysyl residue that coordinates the heme iron at alkaline pH was chemical modification (Table lb). Some of these protocols were highly sophisticated in terms of their ability to yield singly-modified proteins. However, even the most elegant of strategies suffered in that the binary nature of the alkaline transition was not considered. As it turns out, the failure to discriminate between the two different alkaline conformers is what hindered the identification. Although one of the appropriate residues may have been blocked by chemical modification, the alternate lysyl residue remained free to replace Met80 in the coordination sphere of the heme iron. In general, spectrophotometric titration of the 695 nm was the preferred method to analyze the modified cytochromes (Table lb). However, there are no detectable differences between the electronic spectra of the Lys73- and Lys79-bound alkaline conformers. EPR spectroscopy had also been used in the characterization of the isomerization products (Greenwood & Palmer, 1965; Gadsby etal., 1987) and, although in retrospect these results are consistent with the formation of two protein conformers, the uncertainty associated with the collection of data under cryogenic conditions made such a conclusion speculative (Gadsby & Thomson, 1990). With the confirmation by 'H-NMR spectroscopy that two distinct protein structures occur at elevated pH (Hong & Dixon, 1989), it is now clear that the heterogeneity observed in EPR spectra truly reflects the formation of two related protein structures rather than freezing artifacts. Consequently, 'H-NMR spectroscopy should be used in assessing the conformational properties of other cytochromes c. 152 4.2 Mechanism of the Alkaline Conformational Transition 4.2.1 Intermediate Conformational States In the simplified mechanism of the alkaline isomerization of ferricytochrome c, Davis et al. envisioned the ionization of a titratable group (pK^) which triggers a subsequent iron ligand exchange (pKc) (1974). In general, this model describes satisfactorily the kinetics observed for the wild-type horse and yeast proteins as well as for several yeast cytochrome c variants bearing point mutations (Kihara et al., 1976; Saigo, 1981; Pearce etal, 1989; Nail etal, 1989; Rafferty, 1992). However, a more complete description of the events leading to the alkaline conformers of the protein includes an intermediate in which the heme is probably 5-coordinate (Lambeth et al, 1973; Saigo,1981) (Scheme 6; /' is the intermediate and should not be confused with the acidic conformer of ferricytochrome c, state I). This species is a common precursor to both alkaline forms, and bifurcation of the isomerization mechanism at this point into two parallel branches, explains in part why the binary nature of the process has not been detected by pH-jump kinetics experiments. In the majority of these investigations, the decay of the 695 nm charge transfer band was monitored. As noted earlier, this absorbance is a manifestation of Met80-Fe3+ coordination (Schechter & Saludjian, 1967), so its decay reflects the generation of the common intermediate but not necessarily the alkaline conformers. If under the experimental conditions (i.e., u~ 100 mM, pH < 10), this step is rate limiting (i.e., kx < k„ kh) and insensitive to the nature of the final product, then a single monophasic process is detected Ill If k. (6) 153 At pH >10.5, the intermediate in the isomerization process may also be the precursor to or the fifth conformer of ferricytochrome c (state V). However, within the pH range 8 to -10.5, this common intermediate turns into the two stable, lysyl-bound conformers of ferricytochrome c that are distinguishable by NMR spectroscopy (Hong & Dixon, 198x; Ferrer et al, 1993,Roselle/a/., 1998). Interestingly, there is no evidence from saturation transfer experiments that one alkaline conformer converts directly into the other. Presumably, the mechanism for the reversion from the two alkaline states to the native cytochrome is reversible as shown in Scheme 6. It is possible, therefore, that under the experimental conditions (50 mM sodium phosphate, pH~9.3, 55°C), the failure to detect saturation transfer between the two alkaline species results from a faster reversion to the native state than to the alternate alkaline conformer {i.e., > ka, kb). Therefore, the bulk of the magnetization is transferred to the native species. In revising the mechanism of Davis et al, special emphasis has been placed on the experimental conditions. Below a certain pH threshold (~ 10), the 5-coordinate intermediate is not present at a sufficiently high concentration to be detected. However, evidence from pH-jump experiments above this threshold suggests that the situation changes under these conditions so that kx > ka, kb or < ka, kbkv The result is the transient formation of a high-spin species in appreciable amounts (vide infra). The relative variation of ku k-x, ka, or kb with pH is consistent with the existence of a "trigger" group the protonation state of which controls the conformational transition of ferricytochromes. For example, a change in the deprotonation of a residue in the alkaline species following the conformational rearrangement could induce a shift in ka or kb. However, from the original analysis of Davis et al. (1974), such a scenario requires that the observed rate constant of the isomerization decreases with increasing pH. This is contrary to what is observed in all the pH-jump experiments. On the other hand, deprotonation of a group in the native cytochrome prior to the 154 conformational rearrangement would induce a shift in (or £-,), and the observed rate constant would increase with pH. Conclusive identification of the group that triggers the alkaline conformational transition of ferricytochrome c would greatly enhance our understanding of the mechanism. This subject is considered further in the following section. The electronic absorption spectrum of the transient species derived from singular value decomposition analysis of pH-jump kinetics data collected at pH > 10 resembles the spectra of two cytochrome c variants in which neither the native nor the alkaline heme ligation are in effect. In the Lys73 Ala/Lys79Ala variant, Lys73 and Lys79 are unavailable to coordinate the heme. Therefore, at pH ~ 10.5 a stable high-spin species is formed when Met80 is expelled from the coordination sphere of the heme iron. In the Phe82Trp variant, the large tryptophanyl side chain represents a steric obstacle to the incoming lysyl residues and thus stabilizes a high-spin species that reaches maximum concentration at pH -8.5. The spectroscopic similarities suggest that the alkaline transition intermediate is somehow trapped in the isomerization of the Lys73Ala/Lys79Ala and Phe82Trp variants. However, resonance Raman spectroscopy indicates that in both cases, the high-spin species consists of a 6-coordinate heme in which a hydroxide probably occupies the distal coordination site (Dopner, unpublished results). Notably, the spectrum of the Met80Ala variant at a pH where hydroxide occupies this site exhibits a maximum at -625 nm rather than at -600 nm (Bren, 1996; Twitchett, 1998). 4.2.2 The Identity of the Trigger Group The identity of the functional group which, upon ionization, triggers the alkaline conformational reorganization of ferricytochrome c remains unknown. According to the pH-jump kinetics protocol of Davis et al. (1974), a key property that must be exhibited by the trigger group 155 is a relatively high pKa for ionization (-10.5 to 12, depending on the cytochrome). Few amino acid side chains titrate in this pH range, and computational analysis of molecular dynamics and energy-minimized structures of cytochrome c (Zhou & Vijayakumar, 1997) identifies only lysyl, arginyl, and tyrosyl residues as potential trigger groups. The trigger residue may be located in an internal region of the protein that is pivotal for the stabilization of the native structure of the protein and which stabilizes the protonated form of this group. Alternatively, it may be located near the protein surface, but the intimate association of this residue with neighbouring amino acids results in an elevated pKap. Among the functional groups that have been considered as the "trigger" (see section 1.3.4), the lysyl residues that coordinate the heme at high pH represent the simplest candidates, particularly because the e-amino group of these amino acids must be neutralized before it can bind the iron and the stoichiometry of the isomerization consists of a single proton. Analysis of the pH-jump kinetics data collected for the Lys73 Ala and Lys79Ala variants of the yeast protein yields pKH values of 10.8 and 12.0, respectively. Assuming that the titration of these residues triggers the isomerization, these pK values correspond to Lys79 (10.8) and Lys73 (12.0). Although the former value is in reasonable agreement with the pKa determined for this residue by Bosshard (~ 10.4; 1981), it is not clear why the pK for the latter is as high as 12.0. It is more likely, therefore, that the amino acid replacement at either position 73 or 79 results in the perturbation of the environment around the true trigger group. Consideration of the different environments around Lys73 and Lys79 in the wild-type protein structure (Berghuis & Brayer, 1992) reveals that mutation of Lys79 to Ala can potentially result in such a mutation-induced artifact because of its close association with Ser47, Tyr46, and the heme 6-propionate group. In contrast, Lys73 is highly exposed to the solvent, and mutation of this residue is unlikely to perturb the local structure of the protein significantly. The role of the heme propionate groups in the alkaline conformational transition of 156 ferricytochrome c has been investigated by Moore and coworkers (Tonge et al., 1989). This study argued that the heme 6-propionate group in the equine cytochrome has an unusually high pATa (> 9), leading the authors to conclude tentatively that the ionization of this carboxylic acid could trigger the alkaline isomerization of the protein. Ser47 (Thr47 in the equine cytochrome) and Tyr46 are also in the proximity of the e-amino group of Lys79, and the titration of these hydroxyl groups could also destabilize the cytochrome enough to cause the isomerization. Nevertheless, these residues have escaped scrutiny thus far. Inspection of the amino acid variability in this region of the protein reveals that Tyr46, Ser47, and Tyr48 are highly conserved and suggests that they are indeed significant for the structural stability if not the function of the native conformer of cytochromes c. Moreover, the calculated range of pATa values calculated for the tyrosyl residues (Zhou & Vijayakumar, 1997) is consistent with the pA"as determined by pH-jump kinetics experiments. Upon ionization of either tyrosyl residue or the Ser47 y-OH group, Lys79 could be released from the hydrogen bonds that anchor it, thus opening the heme cavity and facilitating the ligand exchange. In essence, mutation of Lys79 to Ala mimics this chain of events and results in a more open protein conformation that facilitates the isomerization to an alkaline form. Furthermore, Nail and coworkers have shown that interaction of horse heart cytochrome c with the antibody 2B5 significantly stabilizes the native conformation of this protein (ApJ^ap~l; Raman et al., 1992). In contrast, interaction of this cytochrome with antibody 5F8 barely affects the isomerization (Ap/Tap~0.3; Ibid). The epitope of MAb 2B5 has been mapped to the immediate region surrounding residue 44, whereas that of MAb 5F8 is located on the opposite face of this cytochrome. Therefore, it may be that on binding to antibody 2B5, but not 2F8, the trigger group becomes isolated from the medium, and the protonated form of the isomerization trigger is stabilized to increase the pKap of the protein. The occurrence of the trigger group on the surface of the cytochrome would resemble the 157 situation reported for rhodopsin by Koutalos et al. (1990). In metarhodopsin, a retinal moiety is linked to the protein backbone through a Schiff base. Deprotonation of this group triggers the isomerization of the retinal molecule from the 1 \-cis- to the all trans- conformation. This ionization is dependent on the surface charge density in the immediate surroundings of the Schiff base and, therefore, on the ionic strength of the medium. Consequently, Koutalos et al. analyzed the ionic strength dependence of the apparent pKap of this isomerization using the Gouy-Chapman formalism. This analysis parametrizes the average surface charge density in the vicinity of the titrating group, as well as the pKa of the titrating group in isolation from the influences of the charged surface. Extension of this analytical approach to the alkaline isomerization of yeast ferricytochrome c results in a pK of -8.9 for the wild-type and the Lys73Ala variant. In contrast, an equivalent analysis yields a value of -9.2 for the Lys79Ala variant of the protein. These results suggest that replacement of Lys79 by an alanyl residue perturbs the environment around the trigger group and, therefore, the pKa for its ionization. However, these pK values differ substantially from the corresponding pKH for these variants determined by pH-jump kinetics suggesting that additional factors must be considered in the analysis of the ionic strength dependence of the alkaline conformational transition of ferricytochrome c. Alternatively, and assuming that the underlying premises of pH-jump kinetics analysis and the Gouy-Chapman theory apply to the conformational dynamics of cytochrome c, these pK values reflect different properties of the isomerization process. 4.3 Factors that Influence the Isomerization 4.3.1 The Driving Force of the Isomerization An important issue raised by this work concerns the driving force for the alkaline conformational rearrangement of ferricytochrome c. It could be argued that coordination of a 158 nitrogenous ligand to the heme iron is a principal factor driving the isomerization reaction, in particular, the reaction which generates the Lys73-bound conformer. The increased stability of the native isomer in the Lys73 Ala/Lys79Ala variant supports this argument. Despite the greater structural mobility that is expected in the distal half of the protein in the absence of Lys79 and the hydrogen bonds that stabilize the closed conformation of the protein, the Met80-Fe3+ bond is broken only at a pH (-10), well above the pKap of the alkaline transition in the wild-type cytochrome. However, addition of methylamine hydrochloride to the Lys73 Ala/Lys79Ala variant, even at a pH closer to neutrality, results in a protein adduct with spectroscopic features that closely resemble those of the alkaline conformers of ferricytochrome c. Although the formation of this adduct could result from an ionic strength-induced destabilization of the native state of the variant, the observation of a bis-histidine coordinated species at neutral pH in the Phe82His variant (Hawkins et al., 1994) confirms that the affinity of the ferric heme for a nitrogenous base is comparable, if not greater, than that for the methionyl 8-sulfur atom. On the other hand, the apparent inability of Lys72, 73, and 79 to coordinate the iron in the Met80Ala variant near neutral pH suggests that loss of Met80 from the coordination environment of the heme iron is not sufficient to induce the conformational rearrangement. In this variant, the lysyl residues compete with water and hydroxide for this coordination site. Furthermore, azide (Sutin & Yandell, 1972; Saigo, 1986; Rafferty, 1992), cyanide (George & Tsou, 1952), and imidazole (Schejter & Aviram, 1969) are known to replace Met80 in the coordination sphere of the iron. A common feature of these relatively facile displacement reactions is the invariant rate for the rupture of the Met80-Fe3+ bond. Not surprisingly, there appears to be a correlation between the susceptibility of various cytochromes c to replacement of Met80 by a particular type of exogenous ligand and the pK of the alkaline conformational transition associated with each protein (Saigo, 1986). Together, 159 these observations suggest that the expulsion of Met80 from the iron coordination sphere is another important determining factor of the alkaline isomerization. The entropic and enthalpic changes associated with the alkaline conformational transition of the Lys73Ala variant of cytochrome c (Table 8: 19(2) e.u. and 5.4(6) kcal mol"1, respectively) are comparable to those of the wild-type protein. In contrast, the entropy and enthalpy of the isomerization of the Lys79Ala variant are nearly doubled. The greater driving force for the isomerization associated with the Lys79Ala variant reflects the greater structural reorganization that the polypeptide must undergo to accommodate Lys73 in the coordination sphere of the heme iron. 4.3.2 Structural Elements The manner in which the primary structure of a protein encodes the information for the proper folding of polypeptides into functional proteins/enzymes continues to demand much attention. Typically, this "problem" is addressed by studying the folding of a protein that is denatured by some means (i.e., through the use of chemical agents or elevated temperatures), or by studying the reverse process. The alkaline conformational transition of ferricytochromes c exemplifies another facet of this so-called protein folding problem. Specifically, this isomerization illustrates how for some proteins, the amino acid sequence also includes essential instructions that permit multiple conformational states throughout the lifetime of the polypeptide. Scrutiny of the effects of point mutations on the alkaline conformational transition of ferricytochrome c identifies components of the polypeptide chain that influence the ordered progression from the native protein to its alkaline conformers. For example, Phe82 is not only important for the electron transfer function of cytochromes c, but it is also a significant structural element involved in the pH-dependent dynamics of these proteins. Replacement of this conserved 160 residue with other amino acids generally destabilizes the native conformation of the yeast ferricytochrome, favouring in many cases the Lys73-bound alkaline conformer. The higher incidence of this species is consistent with the lower pK3ps of variants like Phe82Gly and Phe82Ser. However, pH-jump kinetics experiments indicate that the reduced pKip for the alkaline transitions of these variants also results from a significant drop in the pKH of the trigger group (Pearce et al, 1989). Evidently, the structural changes induced by the replacement of Phe82 include a perturbation of the immediate environment around this trigger group. Comparison of the X-ray crystal structures of the Phe82Ser or Gly variants to that of the wild-type protein reveals that the most significant structural differences in these variants are the movement of a conserved water molecule (H20166) 0.7-1.4 A closer to the heme iron (Figure 51), the collapse of the protein backbone between residues 82 and 84 into the heme cavity and closer to the porphyrin ring, and a general reorganization of the distal side of the protein that increases the solvent exposure of the heme (Louie, 1990). Replacement of Thr78 by Ala also destabilizes the native conformation of the cytochrome significantly. The pK^ of this variant is ~2 pH units lower than that of the wild type protein, but unlike the Phe82 variants discussed above, 'H-NMR spectroscopy analysis indicates that the predominant alkaline conformer of this protein below 30 °C has Lys79 bound to the heme iron. Only at higher temperatures (~30°C) is the Lys73-bound conformer comparable in abundance to the Lys79-bound form. This marked temperature dependence of the alkaline conformer distribution makes the Thr78Ala variant unique among those described in this work. The structural basis for this property is not clear because the structure of the Thr78Ala variant has not been determined yet. Nevertheless, it is expected that removal of Thr78 will also perturb the protein in the region of the conserved, internal water molecule (H20166) because the hydrogen bond with the y-OH group of the threonyl residue cannot be formed (Figure 54). 161 In contrast to the destabilization of the native conformer of ferricytochrome c induced by amino acid replacements at positions 78 and 82, replacement of Tyr67 with phenylalanine considerably stabilizes the native isomer. With a reported pKap of 10.7 (Lutz etal., 1989), this protein is the most stable variant of cytochrome c. Without the Tyr67 OH group, the hydrogen bond between this oxygen and the Met80 5-sulfur atom is severed, which evidently strengthens the Met80-Fe3+ bond (Berghuis et al., 1994). Interestingly, there is no evidence that this variant exhibits a high-spin intermediate in pH-jump experiments above pH 10, or even as high as ~ 12. As indicated earlier, this observation suggests that the phenol group of Tyr67 is an integral part of the conformational transition and acts in one of at least three ways. The oxygen atom of this residue could coordinate the heme iron transiently, or it could hydrogen bond to an exogenous water/hydroxide ligand before coordination by a lysyl residue to the heme iron is established. Alternatively, the intermediate species detected in pH-jump experiments at high pH could result from a strengthened Tyr67-Met80 hydrogen bond and a concomitant weakening of the Met80-Fe3+ bond. Other residues in the sequence of cytochrome c that have been have been shown to be relevant to the alkaline transition are the highly conserved prolyl residues which flank Lys73 and which delimit the 71-76 helix. Mutation of Pro71 to Thr, for example, depresses the pKip of yeast iso-2-ferricytochrome c from 8.45 to 6.63 (White et al., 1987). Semisynthetic substitution of Pro71 by norvaline induces a greater shift in the conformational equilibria of the horse protein (Wallace & Clark-Lewis, 1997). In this protein, the pK^ of the alkaline transition is ~4 pH units lower than in the wild-type protein. Although this dramatic change may be an artifactual result from the semisynthetic method (i.e., through imperfect packing of the non heme-containing synthetic polypeptide to the heme-containing chain, etc.), it is clear that Pro71 plays a crucial role in the stabilization of the native conformer of ferricytochromes c. In the same manner, Pro76 appears to stabilize the native form of 162 cytochromes c. The pKap of the Pro76Gly variant is also shifted significantly (pK^-6.71; Wood et al., 1988 a, and b), and pH-jump kinetics indicate that this destabilization is due almost exclusively to a 70-fold acceleration of the forward isomerization process (i.e., kf- 3500 s"1), without affecting the pK of the trigger group (pKH~ 12; Nail et al., 1989). By comparison, the faster forward rate constant reported in the present work is that corresponding to the Lys79Ala variant (160 s"1). Nail and coworkers attributed the marked increase in rate of the isomerization of the Pro76Gly variant to destabilization of the native conformer (AG° -2.5 kcal/mol) rather than stabilization of the transition state or the alkaline species. Evidently, the function of Pro76, and perhaps Pro71, is to hinder the formation of the Lys73-bound alkaline conformer of cytochrome c much in the same manner that trimethylation hinders the formation of a Lys72-bound alkaline conformer in the yeast protein. However, this suggestion requires evaluation by 'H-NMR spectroscopy. 4.3.3 Effects of Salts on the Conformational Transition of Cytochrome c The pH-dependent conformational equilibria of wild-type ferricytochrome c clearly vary with temperature as well as with solution conditions. The pKv of the alkaline isomerization increases with potassium chloride concentration, and is maximal at -600 mM. At the same time, one alkaline conformer (Lys79-bound) predominates at the expense of the other alkaline form (Lys73-bound). These observations were interpreted in section 3.4.3 to occur in response to an increasing shielding effect of charged residues by the electrolyte. This shielding would help to reduce potentially unfavourable electrostatic interactions which involve Lys73 and/or Lys79. This view is consistent with a substantial expansion of the oxidized protein at lower ionic strengths that is observed by small-angle X-ray scattering experiments (Trewhella etal., 1988). However, it should also be noted that the cytochrome c surface has at least two chloride binding sites (Osheroff et al., 1978; Andersson 163 et al., 1979, 1980). The first of these binding sites is associated with lysines 86, 87, 89 (Feng & Englander, 1990), and Lysl3 (Morton & Breskvar, 1977), and the second site is associated with Lys60 (Stellwagen & Schulman, 1973). Although no significant structural differences were detected in the 'H-NMR spectra of the protein measured in the absence or the presence of 200 mM sodium chloride at pH 7 (Feng & Englander, 1990), at higher salt concentration and pH (-500 mM, 7.3 respectively) local structural changes appear to cause the aromatic ring of Phe82 to approach the heme 3-methyl group of the porphyrin (Moench et al., 1991). As shown above, perturbation of the protein structure in this region is closely associated with alterations to the alkaline conformational equilibria of ferricytochrome c. A variety of other ions are also known to bind to cytochrome c, yet not all of these interactions have the same conformational influence as potassium chloride. Perchlorate, for example, is an efficient competitor for one of the chloride binding sites, particularly in the oxidized protein (Andersson etal., 1980). Instead of stabilizing the native isomer, however, perchlorate depresses the pA^ of the equine protein by as much as one pH unit (-200 mM sodium perchlorate). Andersson et al. reported the emergence of a second perchlorate binding site in the alkaline conformer of the protein. Presumably, this binding site is located in the distal side of the Lys73-bound conformer, and interaction with perchlorate serves to stabilize this structure. Closer to neutral pH, perchlorate destabilizes the Met80-Fe3+ bond as evidenced by a high-spin component that is detected by EPR spectroscopy (Andersson et al., 1980). Carbonate (Osheroff et al., 1980), phosphate (Taborsky & McCollum, 1979), and various anionic and cationic coordination complexes (Otiko & Sadler, 1980; Arean etal., 1988 and references therein; Moore et al., 1995) also bind to ferricytochrome c. Every one of these interactions with charged species can be expected to influence different physicochemical and structural properties of 164 the protein depending on the site of interaction. Increasing salt concentrations, for example, depresses the reduction potential of various cytochromes c (Gopal et al, 1988), and addition of Et3PAuCl to protein solutions at neutral pH interferes with the heme core as evidenced by a spin-state change (Otiko & Sadler, 1980). Presumably, this gold compound binds ferricytochrome c at one of the ion binding sites and weakens the Met80-Fe3+ bond in a similar fashion as perchlorate. The result is a protein with spectroscopic features similar to those observed in the electronic spectrum of the Phe82Trp variant at pH 8.5. Furthermore, polyphosphates (Concar et al, 1991) and other polyions have been of particular interest because these molecules are believed to mimic to some degree the surfaces of the electron transfer partners of cytochromes c (vide infra) (Hildebrandt, 1990; Antalik et al, 1992; Sedlak, 1997; Sedlak & Antalik, 1998). 4.3.4 Cytochrome Conformational Equilibria at Charged Surfaces Within the intermembranal space of mitochondria, cytochrome c is inevitably associated with the membrane surface. As much as 10% of the cytochrome is reported to remain attached to the inner mitochondrial membrane at a variety of ionic strengths (Cannon & Erman, 1980; Rytomaa & Kinnunen, 1994; Cortese etal, 1995). In fact, cytochrome c has been reported to have at least two independent binding sites for acidic phospholipids (Rytomaa & Kinnunen, 1994). This association appears to increase the P-sheet character of the protein (-35%) at the expense of the a-helical component (Cortese et al, 1998), it causes a localized conformational change in the vicinity of the heme 6-propionate group (Soussie et al, 1990), and it disrupts the Met80-Fe3+ bond (Letellier & Schechter, 1973; Pinheiro & Watts, 1994). Although this destabilization of the native structure of ferricytochrome c yields a high-spin species, and thus should facilitate the alkaline conformational transitions, there are no indications that these isomers are formed. In one report, the interaction of 165 cytochrome c with cardiolipin is actually reported to increase the pK3p of the alkaline isomerization (Soussi et al, 1990). Note that the effect of lipid membranes on the conformation of ferricytochrome c may vary depending on the composition of the membrane. Consequently, the experimental results of the type described above must be considered carefully. Investigations involving protein:protein interactions also indicate that the Met80-Fe3+ bond in ferricytochrome c is perturbed upon interaction of this protein with cytochrome c oxidase (Weber et al., 1987; Michel et al, 1989). The resonance Raman studies by Hildebrandt and coworkers confirm that cytochrome c undergoes structural changes upon interacting with charged surfaces that lead to a thermal equilibrium between two spectroscopic states (Hildebrandt & Stockburger, 1989). While state I is identical to the native protein, state II is composed of a mixture of species: a high-spin, 5-coordinate species, and high- and low-spin, 6-coordinate species (states I and II are the original terms used by Hildebrandt (1990), and should not be confused with the pH-linked conformers of ferricytochrome c as defined by Theorell & Akesson, 1941). More recently, state II was redefined as state B2 which forms in both the Lys73 Ala and Lys79Ala variants of the yeast protein at pH >10.5 and in the Lys73 Ala/Lys79Ala variant at pH<10 (Ddpner et al, 1998). State B2 does not correspond to either alkaline conformer but is a species observed at higher pH, state V. In this conformation, the heme iron is most likely coordinated by a hydroxide because spectroscopic features of B2 are also observed in the spectra of the Tyr67Phe and Thr78Ala variants. Clearly, charged surfaces have the potential to alter the heme core of ferricytochrome c significantly. Whether or not the alkaline species of this protein play a metabolic role remains a matter of conjecture and requires further scrutiny. 166 4.4 pH-Linked Conformational Transitions in Other Proteins Proteins are generally regarded to have an optimum pH range over which they have a folded structure and can function optimally. Outside of this range, the properties of the protein are largely ignored unless attempts to extend the pH limits through protein engineering are undertaken. However, other proteins that display pH-linked conformational equilibria similar to those of ferricytochrome c are becoming recognized (see for e.g., Gerstein & Krebs 1997; http://bioinfo. mbb.yale.edu/MolMovDB/). A prime example of a family of proteins that undergo organized and reversible pH-linked conformational transitions as part of their physiological role are the serum transferrins (Aisen, 1989). Based on a structural analysis of ovotransferrin, Sacchettini and coworkers have proposed that the two domains of the JV-terminal lobe of the protein exist in an open or closed conformation depending on the protonation state of Lys209 and 301 which are close together in three-dimensional space (Dewan etal., 1993). In the closed conformation, one of these lysyl residues has an unusually low pK3p so that near neutral pH the residue is deprotonated and hydrogen bonded to the neighbouring lysyl residue. Acidification of the solution results in the protonation of both lysines, the concomitant electrostatic repulsion between these residues, and the opening of the protein in a hinge-like fashion to release iron (Dewan et al., 1993). Such a proton-assisted structural modification of the binding site is consistent with the observed rapid phase of iron release from holotransferrin (Chung & Raymond, 1993, Kretchmar-Ngyuen et al., 1993; El Hage Chahine & Pakdaman, 1995). Other examples of proteins that exhibit pH-linked conformational heterogeneity are phosphophoryn (Evans & Chan, 1994) and hemagglutinin (Tatulian & Tamm, 1996 and references therein). The former protein is isolated from tooth dentine and acts as an acidic template for biomineralization. According to a recent study by Evans and Chan, the calcium-depleted apo-protein 167 undergoes three distinct conformational reorganizations between pH 2 and 9. As the pH increases, the protein expands significantly in response to the electrostatic repulsion between an increasing number of Asp and phospho-Ser residues that are deprotonated. In the fully expanded form, the protein is believed to retain secondary structure elements as well as some of higher order, but it is an inefficient substrate for biomineralization. Hemagglutinin is a trimeric glycoprotein found in the protein coat of the influenza virus. At acidic pH, both the secondary and tertiary structures of the protein change (Wilson et al, 1981; Bullough, et al, 1994) so that the JV-terminus of one of the subunits is inserted into the target membrane as a prelude to fusion of the viral and host membranes. Like the alkaline transition of ferricytochrome c, the acidic conformational transition of hemagglutinin is reversible in the absence of target membranes (Tatulian & Tamm, 1996). Ligand shuttling is an elegant and efficient mechanism by which a metalloenzyme can carry out its metabolic fiinction(s). In the case of cytochrome c oxidase, for example, labile coordination chemistry has been proposed to play a key role in the proton pumping activity of this enzyme. In one instance, Woodruff and coworkers envisioned a residue alternating as a ligand to the CuB centre and the heme a3 iron (Woodruff, 1993). Binding of an exogenous ligand, such as dioxygen, to the CuB centre initiates a series of ligand exchanges which result in the transfer of the oxygen molecule to the heme a3 for reduction, and the translocation of protons from one side of the membrane to the other. In a similar model proposed by Rousseau and coworkers, a histidyl and a tyrosyl residue alternately coordinate the heme a3 iron from the same side (1993). In each phase of the axial ligand replacement cycle, protons are conveyed across the membrane and, at the same time, dioxygen is reduced via an oxyferryl heme intermediate. A third model favoured by Wikstrom and coworkers (1994; Morgan et al, 1994) involves the labile nature of a histidyl-Cu coordination bond. The oxidation and ligation states of the copper influence the pKas of a histidyl residue so that the side chain cycles between three 168 protonation states and ligation to the copper atom and thus pumps protons across the membrane. Although other mechanisms have been proposed (for e.g., Gelles et al., 1986) and refuted based on details of the X-ray crystal structures of cytochrome c oxidase from Paracoccus denitrificans (Iwata et al., 1995) and beef heart (Tsukihara et al., 1996), the current opinion is that a ligand shuttle mechanism of some sort is included in the activity of cytochrome c oxidase. 4.5 Possible Roles of the Alkaline Conformers The alkaline conformational transitions of ferricytochromes c are ligand shuttling processes that significantly alter the structural and functional properties of these proteins. A similar ligand shuttling event was shown recently to occur at the heme iron of the c domain of cytochrome cdx nitrite reductase from Thiosphaera pantotropha which converts nitrite to nitric oxide (Williams et al., 1997). In this case, the ligand exchange from His/His to His/Met coordination upon reduction of the heme iron appears to protect the metal centre from the irreversible ligation by nitric oxide. Whether or not the alkaline transitions of ferricytochromes c occur for a specific metabolic function remains to be determined. The predominant view in the literature is that alkaline ferricytochromes c are inactive and rarely found in vivo. Part of this argument is that a Met-Fe3+ bond is required to maintain the midpoint reduction potential of cytochrome c at 250-320 mV vs. SHE despite its relative weakness, and the transition to an alkaline species is an "accidental" phenomenon. Nevertheless, several properties of these conformational transitions and of cytochrome c suggest that they are not merely chance events. (1) The isomerizations are highly reversible despite the significant structural reorganization that must occur to swing Lys73 into the coordination sphere of the iron. (2) The alkaline conformational transition is a common process in cytochromes c from a wide variety of species including bacteria and higher eukaryotes. (3) Amino acids shown here to influence these 169 conformational equilibria and rates of reactions are highly conserved. (4) Mechanisms appear to be in place to prevent the premature or unwanted formation of the Lys-coordinated species (i.e., the prolyl residues that flank Lys73 severely restrict the spontaneous generation of the Lys73-bound alkaline conformer, and trimethylation of Lys72 in the yeast protein prevents this residue from replacing Met80 as a heme ligand) which argues for, and against the assumption that the alkaline isomerizations serve a metabolic function. Clearly, the pH-linked structural dynamics of ferricytochromes c are strongly dependent on environmental factors including temperature, ionic strength, specific binding to ions, proteins and lipids. Therefore, it is imperative to define the nature of the milieu within the intermembranal space of mitochondria where cytochromes c reside. In vivo, interaction with phospholipids and hydration state may promote the formation of the alkaline conformers. The effects of phospholipids have been noted above. Early evidence that dehydration of ferricytochrome c (i.e., lyophilization) destabilizes the Met80-Fe3+ bond to result in an alkaline-like species suggests a role for this consideration (Aviram & Schejter, 1972). Notably, upon rehydration at pH 7.0, the native conformation of the protein is reestablished over a period of a few minutes. Assuming that the water content in the intermembranal space of mitochondria is no greater than that of the cytoplasm (-70%; Alberts, 1994), the hydration state of the protein may be an under-appreciated physiological consideration. If, as discussed above, the formation of the alkaline conformers of cytochrome c in vivo is possible under certain conditions, then hypotheses regarding the function of these alkaline species can be formulated on the bases of their known properties. For example, Nail and coworkers have demonstrated that ferricytochrome c denatures at a higher guanidine hydrochloride concentration at pH 9.1 (C50~2.2 M guanidine hydrochloride) than at pH 7.2 (C50~1.5 M guanidine hydrochloride) (Osterhout et al, 1985). The 'H-NMR spectra of the unfolded proteins at either pH are almost 170 indistinguishable. These observations suggest that at least one alkaline conformer of yeast iso-2-ferricytochrome c is less susceptible to chemical denaturation than the native protein and, therefore, that the alkaline transition of this hemeprotein may offer a degree of stabilization. Whereas the dramatic decrease in the reduction potential upon isomerization (-450 mV; Barker & Mauk, 1993) reverses the thermodynamic driving force for the reduction of alkaline ferricytochrome c by cytochrome b5, cytochrome c reductase, and sulfite oxidase, it enhances substantially the oxidation of the ferrous form by the peroxidase and the oxidase. Therefore, while the transfer of electrons to alkaline ferricytochrome c is essentially halted, the transfer from ferrocytochrome c to the peroxidase or the oxidase becomes more favourable if these interactions induce the alkaline transition in the ferrous form of the cytochrome. Although this isomerization of the reduced protein is not unprecedented in vitro (Barker & Mauk, 1992), it has not been documented in vivo. Nevertheless, resonance Raman evidence presently under review suggests that the Met80-Fe2+ bond in cytochrome c is disrupted by the interaction of the ferrocytochrome with charged surfaces in a similar fashion as the ferric protein (vide supra; Dopner et al, submitted). On the other hand, if the conformational transition takes place immediately after the electron is transferred from the cytochrome to the oxidase/peroxidase, then the transition provides a conformational switch that gates the electron transfer. As noted in the introduction, various lysyl residues are known to form an integral part in the docking surface of cytochrome c with its electron transfer partners. Lys73 and Lys79 figure prominently in the complexes of this heme protein with cytochrome c peroxidase, cytochrome b5, and cytochrome c oxidase. Therefore, the role of the alkaline conformational transition may be to decrease the affinity of these proteins for the cytochrome, particularly if the "redox energy" is transformed into conformational energy as in the case of the cytochrome cdx nitrite reductase system 171 mentioned above (Williams et al, 1997), or as in a model system based on the Phe82His variant of cytochrome c (Feinberg et al, 1998). Such a mechanism is particularly useful for the decomposition of peroxides and the reduction of oxygen because these reactions are two electron processes that benefit from a low competition between ferri- and ferrocytochrome c for the binding site(s) in the partner protein. Therefore, isomerization to the alkaline conformers may facilitate the separation of the protein from the peroxidase/oxidase to promote docking of a second ferrocytochrome . 4.6 Concluding Remarks The alkaline conformational transition of ferricytochromes c is an intriguing feature of this protein that remains only partially understood after decades of study. These isomerizations are highly organized processes that share some mechanistic features with the structural dynamics of other proteins. Therefore, the complete elucidation of the mechanisms by which these structural changes take place in ferricytochromes c should also contribute to a general understanding of the conformational dynamics of other proteins. The identification of the trigger group remains the most fundamental enigma of the process. However, it may turn out that this trigger is a collection of events rather than the deprotonation of a single titrating group that induces the alkaline transition of the protein. The elucidation of the three-dimensional structures of the alkaline conformers of ferricytochrome c is likely to be the next major milestone in the characterization of these conformational transitions. Fortunately, with the recent development of the bacterial expression system of mitochondrial cytochrome c, it is now feasible to prepare isotopically-enriched cytochrome c to facilitate the determination of these structures by nuclear magnetic resonance spectroscopy. 172 Bibliography Aisen, F. (1989) Iron Carriers and Iron Proteins (Loehr, T.M. Ed.), VCH Publishers, New York, pp. 353-372. Alberts, B. (1994) Molecular Biology of the Cell, John Wiley & Sons, New York. Alleyne, T. & Wilson, M.T. (1987) Cytochromes Systems, Molecular Biology and Bioenergetics (Papa, S., Chance, B. & Ernster, L., Eds.), Plenum Press, New York, pp. 713-720. Amegadzie, B.Y., Zitomer, R.S. & Hollenberg, CP. (1990) Yeast, 6, 429-440. Andersson, T., Thulin, E. & Forsen, S. (1979)7. Am. Chem. Soc, 18, 2487-2493. Andersson, T., Angstrom, J., Falk, K.E. & Forsen, S. (1980) Eur. J. Biochem., 110, 363-369. Antalik, M., Bona, M. & Jaroslava, B. (1992) Biochem. Int., 28, 675-682. Arciero, D.M., Peng, Q., Peterson, J. & Hooper, AB. (1994) FEBSLett., 342, 217-220. Arean, CO., Moore, G.R., Williams, G. & Williams, R.J.P. (1988) £wr. J. Biochem., 173, 607-615. Angstrom, J., Moore, G.R. & Williams, R.J.P. (1982) Biochim. Biophys. Acta, 703, 87-94. Antalik, M., Bona, M. & Bagelova, J. (1992) Biochem. Int., 28, 675-682. Armstrong, F.A., Bond, A.M., Hill, H.A.O., Psalti, I.S.M. & Zoshi, C.G. (1989) J. Phys. Chem., 93, 6485-6493. Ashida, T., Tanaka, N., Yamane, T., Tsukihara, T. & Kakudo, M. (1973) J. Biochem., 73,463-465. Ausubel, F.M. (1993) Current Protocols in Molecular Biology on CD ROM, John Wiley & Sons, Inc., New York, NY. Aviram. I. (1972) Biochem. Biophys. Res. Commun., 47, 1120-1125. Aviram, I. & Schejter, A. (1972) Biopolymers, 11, 2141-2145. Aviram,!, Myer, Y.P. & Schejter, A. (1981)7. Biol. Chem., 256, 5540-5544. Baistrocchi, P., Band, L., Bertini, I., Turano, P., Bren, K.L. & Gray, H.B. (1996) Biochemistry, 35, 13788-13796. Baldari, C. & Cesarini, G. (1985) Gene, 35, 27-32. 173 Banci, L., Bertini, I., Bren, K.L., Gray, H.B., Sompornpisut, P. & Turano, P. (1995) Biochemistry 34, 11385-11398. Banci, L., Bertini, I., Spyroulias, G.A. & Turano, P. (1998) Eur. J. Inor. Chem., 5, 583-591. Barker, P.D. & Mauk, A.G. (1992) J. Am. Chem. Soc, 114, 3619-3624. Berghuis, A.M. & Brayer, G.D. (1992)7. Mol. Biol., 223, 959-976. Betz, S.F. &Pielak, G.J. (1992) Biochemistry, 31, 12337-12344. Blumberg, W.E. & Peisach, J. (1971) Probes of Structure and Function of Macromolecules and Membranes, 2, (Chance, B., Yonetani, T., and Mildvan, A.S., Eds.), Academic Press, New York, pp. 215-227. Blumberg, W.E. & Peisach, J. (1971a) Adv. Chem. Ser., 100, 271-291. Blumberg, W.E. & Peisach, J. (197lb) Magnetic Resonances in Biological Research (Franconi, C, Ed.), Gordon and Breach, New York, pp. 65-73. Bosshard, HR. (1981)/. Mol. Biol., 153, 1125-1149. Brautigan, D. L., Feinberg, B.A., Hoffman, B.M., Margoliash, E., Peisach, J. & Blumberg, W.E. (1977) J. Biol. Chem., 252, 574-582. Brautigan, D. L., Ferguson-Miller, S. & Margoliash, E. (1978) Methods in Enzymology, 53, 128-64. Bren, K.L. (1996) Ph.D. Dissertation, California Institute of Technology, Pasadena CA. Brittain, T. & Greenwood, C. (1975) Biochem. J., Ul, 175-177. Bullough, P.A., Hughson, F.M., Skehel, J.J. & Wiley, D.C. (1994) Nature, 371, 37-43. Burch, A.M., Rigby, S.E.J., Funk, W.D., MacGillivray, R.T.A., Mauk, M.R., Mauk, A.G. & Moore, G.R. (1990) Science, 247, 831-833. Burns, P.D. & La Mar G. (1981) J. Biol. Chem., 256, 4934-4939. Bushnell, G.W., Louie, G.V. & Brayer, G.D. (1990) J. Mol. Biol., 214, 585-595. Butt, W.D. & Keilin, D. (1962) Proc. Royal Soc. Lond, B152, 429-458. Bychkova, V.E., Dujsekina, A.E., Klenin, S.I., Tiktopulo, E.I., Uversky, V.N. &Ptitsyn, O.B. (1996) Biochemistry, 35, 6058-6063. 174 Caughey, W.S. (1966) Hemes andHemoproteins (Chance, B., Eastabrook, R.W. & Yonetani, T., Eds.), Academic Press, New York, pp. 276-277. Cannon, J.B. & Erman, J.E. (1980) Biochim. Biophys. Acta, 600, 19-26. Cheesman, M.R., Greenwood, C. & Thomson, A.J. (1991) Adv. Inorg. Chem., 36, 201-255. Chung, T.D.Y. & Raymond, K.N. (1993) J. Am. Chem. Soc, 115, 6765-6768. Colon, W., Wakem, L.P., Sherman, F. & Roder, H. (1997) Biochemistry, 36, 12535-12541 Concar, D.W., Whitford, D. & Williams, R.J.P. (1991) Eur. J. Biochem., 199, 569-574. Connors, K.A. (1987) Binding Constants, John Wiley & Sons, New York. Cortese, J.D., Voglino, A.L. & Hackenbrock, CR. (1995) Biochim. Biophys. Acta, 1228, 216-228. Cortese, J.D., Voglino, A.L. & Hackenbrock, CR. (1998) Biochemistry, 37, 6402-6409. Cutler, R.L., Pielak, G.J., Mauk, AG. & Smith, M. (1987) Protein Eng., 1, 569-574. Davies, A.M., Guillemette, J.G., Smith, M., Greenwood, C, Thurgood, A.G. P., Mauk, A.G. & Moore, G.R. (1993) Biochemistry, 32, 5431-5435. Davis, L.A., Schejter, A. & Hess, G.P. (1974) J. Biol. Chem., 249, 2624-2632. Degtyarenko, K.N., North, A.C.T. & Findlay, J.B.C. (1997) Protein Engineering, 10, 183-186. Degtyarenko, K.N., North, A.C.T., Perkins, D.N. & Findlay, J.B.C. (1998) Nucleic Acids Res., 26, 376-381. Deng, W.P. & Nickoloff, J.A. (1992) Anal. Biochem., 200, 81-88. Dente, L., Cesarini, G. & Cortese, R. (1983) Nucleic Acid Research, 11, 1645-1655. Dente, L. & Cortese, R. (1987) Methods in Emzymology, 155, 111-119. Desbois, A. (1994) Biochimie, 76, 693-707. Dewan, J.C, Mikami, B., Hirose, M. & Sacchettini. J.C (1993) Biochemistry, 32, 11963-11968. Dickerson, R.E., Takano, T., Eisenberg, D., Kallai, O.B., Samson, L., Cooper. A. & Margoliash, E. (1971)/. Biol. Chem., 246, 1511-1535. Dickerson, R.E. & Timkovich, R. (1975) The Enzymes (Boyer, P.D. Ed..), Academic Press, New York, pp. 397-547. 175 Dixon, D.W., Hong. X. & Woehler, SE. (1989) Biophys. 7, 56, 339-351. Dong, A., Huang, P. & Caughey, W.S. (1992) Biochemistry, 31, 182-189. Dopner, S., Hildebrandt, P., Rosell, F.I. & Mauk, AG. (1998) 7. Am. Chem. Soc, 120, 11246-11255. Dopner, S., Hildebrandt, P., Rosell, F.I., Mauk, A.G., von Walter, M., Soulimane, T. & Buse, G. (1998) Eur. 7. Biochem., submitted. Dopner, S. (1995)Ph.D. Dissertation, Max-Planck-Institut fur Strahlenchemie,MulheimanderRuhr, Germany. Dumont, M.E., Ernst, J.F., Hampsey, D.M. & Sherman, F. (1987) EMBOJ., 6, 235-241. Dumont, M.E., Corin, A.F. & Campbell, G.A. (1994) Biochemistry, 33, 7368-7378. Eaton, W.A. & Hochstrasser, R.J. (1967)7. Chem. Phys., 46, 2533-2539. Eccles, C, Guntert, P., Billeter, M. & Wuthrich, K. (1991) J. ofBiomol. NMR, 1,111-30. Eddows, M.J. & Hill, H.A.O. (1977) 7. Chem. Soc. Chem. Commun., 771-772. El Hage Chahine, J-M. & Pakdaman, R. (1995) Eur. 7 Biochem., 230, 1102-1110. Elliot, D., Hamnett, A., Lettington, O.C., Hill, H.A.O. &Watton, N.J. (1986) J. Electroanal. Chem., 202, 303-314. Elove, G.A., Bhuyan, A.K. & Roder, H. (1994) Biochemistry, 33, 6925-6935. Evans, J.S. & Chan, S.I. (1994) Biopolymers, 34, 507-527. Fanger, M.W. & Harbury, H.A. (1965) Biochemistry, 4, 2541-2545. Ferrer, J.C., Guillemette, J.G., Bogumil, R., Inglis, S.C, Smith, M. & Mauk, AG. (1993)7 Am. Chem. Soc, 115, 7507-7508. Feng, Y. & Englander, S.W. (1990) Biochemistry, 29, 3505-3509. Forsen, S. & Hoffman, R.A. (1963)7 Chem. Phys., 39, 2892-2901. Frew, DE. & Hill, H.A.O. (1988) Eur. 7. of Biochem., Ill, 261-269. Fumo, G, Spitzer, J.S. & Fetrow, J.S. (1995) Gem, 164, 33-39. 176 Gadsby, P. M. A., Peterson, J., Foote, N., Greenwood, C. & Thomson, A. J. (1987) Biochem. J., 246, 43-54. Gadsby, P. M. A. & Thomson, A. J. (1990) J. Am. Chem. Soc. Ill, 5003-5011. Gao, Y., Boyd, J., Pielak, G.J. & Williams, R.J.P. (1991) Biochemistry, 30,7033-7040. Gelles, J., Blair, DF. & Chan, S.I. (1986)Biochim. Biophys. Acta, 853, 205-236. George, P. & Tsou, CL. (\952) Biochem. J., 50, 440. Gerstein, M. & Krebs, W.G. (1997) "A Database of Molecular Motions," submitted. (Database located at http://bioinfo.mbb.yale.edu/MolMovDB/). Glasoe, P.K. & Long, F. A. (1960) J. Phys. Chem., 64, 188-191. Gopal, D., Wilson, G.S., Earl, RA. & Cusanovich, MA. (1988)7. Biol. Chem., 263,11652-11656. Gouterman, M. (1959) J. Chem. Phys., 30, 1139. Greenwood, C. & Palmer, G. (1965) J. Biol. Chem., 240, 3660-3663. Guillemette, J.G., Matsushima-Hibiya, Y., Atkinson, T. & Smith, M. (1991) Protein Eng., 4, 585-592. Guillemette, J.G., Barker, P.D., Eltis, L.D., Lo, T.P., Smith, M., Brayer, G.D. & Mauk, AG. (1994) Biochemie, 76, 592-604. Hall, P.L., Angel, BR. & Jones, J.P.E. (1974) J. Mag. Res., 15, 64-68. Hampsey, D.M., Das, G. & Sherman, F. (\9%%)FEBS Lett., 231, 275-283. Hampton, M.B., Zhivotovsky, B., Slater, A.F., Burgess, D.H. & Orrenius, S. (1998) Biochem. J., 329, 95-99. Harbury (1965) Proc. Natl. Acad. Sci. U.S.A. 54, 1658-1664. Hasumi, H. (1980) Biochim. Biophys. Acta, 626, 265-276. Hawkins, B.K., Hilgen-Willis, S., Pielak, G.J. & Dawson, J.H. (1994)/. Am. Chem. Soc, 116, 3111-3112. Herrmann, L.M. & Bowler, B.B. (1997) Protein Science, 6, 657-665. Hettinger, T.P. & Harbury, HA. (1965) Biochemistry, 4, 2585-2589. 177 Hildebrandt, P. (1990) Biochim. Biophys. Acta., 1040, 175-186. Hildebrandt, P. & Stockburger, M. (1989) Biochemistry, 28, 6710-6721. Hildebrandt, P. & Stockburger, M. (1989) Biochemistry, 28, 6722-6728. Hildebrand, D.P., Burk, D.L., Maurus, R., Ferrer, J.C., Brayer, G.D. & Mauk, AG. (1995) Biochemistry, 34, 1997-2005. Hill, R. (1956) Modern Methods in Plant Analysis, ( Peich, I.K. & Tracey, M.V. Eds.), Springer-Verlag, New York. Hodgson, D.R., Davis, R.E. & McConaghy (1995) Brit. Vet. J., 150, 219-235. Hong, X.L. & Dixon, D.W. (1989) FEBSLett, 246, 105-108 Inglis, S.C, Guillemette, J.G, Johnson, J.A. & Smith, M. (1991) Protein Eng. 4, 569-574. Inubishi, T. & Becker, E.D.J. (1983)/. Magn. Reson., 51, 128-133. Iizuka, T. & Morishima, I. (1975) Bioch. Biophys. Acta, 400, 143-153. Iwata, S., Ostermeier, C, Ludwig, B. & Michel, H. (1995) Nature, 376, 660-669. Jeng, M.F. & Englander, S.W. (1991) J. Mol. Biol., 221, 1045-1061. Kalinichenko, P. (1976), Stud. Biophys., 58, 235-240. Kamatari, Y.O, Konno, T., Kataoka, M. & Akasaka, K. (1996) J. Mol. Biol., 259, 512-523. Keilin, D. (1925) Proc. Roy. Soc. (London), B98, 312. Kihara, H., Saigo, S., Nakatani, H., Hiromi, K., Ikeda-Saito, M. & Iizuka, T. (1976) Biochim. Biophys. Acta, 430, 225-243. Kluck, R.M., Bossy-Wetzel, E., Green, D.R. & Newmeyer, DD. (1997) Science, 275, 1132-1136. KretchmarNgyuen, S.A., Craig, A. & Raymond, K.N. (1993) J. Am. Chem. Soc, 115, 6758-6764. Kroemer, G, Dallaporta, B. & Resche-Rigon, M. (1998) Ann. Rev. Phys., 60, 619-642. Kunkel, T.A. (1985) Proc. Natl. Acad. Sci. U.S.A., 82, 488-492. Kunkel, T.A., Roberts, J.D. & Zakour, RA. (1987) Meth. Enzymol. 154, 367-382. Koutalos,Y., Ebrey, T.G., Gilson, H.R. & Honig B. (1990) Biophys. J., 58, 493-501. 178 La Mar, G. & Walker, F.A. (1979) The Porphyrins, IV, 61-157. Lambeth, D.O., Campbell, K.L., Zand, R. & Palmer, G. (1973)7. Biol. Chem., 248, 8130-8136. Landrum, H.L., Salmon, R.T. & Hawkridge, F.M. (1977) J. Am. Chem. Soc., 99, 3154-3158. Laz, T.M., Pietras, D.F. & Sherman F. (1984) Proc. Natl. Acad. Sci. U.S.A., 81, 4475-4479. Letellier, L. & Schechter, E. (1973) Eur. J. Biochem., 40, 507-512. Li, F., Srinivasan, A., Wang, Y., Armstrong, R.C., Tomaselli, K.J. & Fritz, L.C. (1997) J. Biol. Chem., 272, 30299-30305. Liu, G., Chen, Y. & Tang, W. (1996) J. Chem. Soc.-Dalton Trans., 5,795-801. Liu, X., Kim, N.C., Yang, J., Jemmerson, R. & Wang, X. (1997) Cell, 86, 147-157. Lloyd, E., Ferrer, J.C, Funk, W.D., Mauk, M.R. & Mauk, AG. (1994) J. Am. Chem. Soc, 33, 11432-11437. Lloyd, E., Hildebrand, D.P., Tu, KM. &.Mauk, AG. (1995) J. Am. Chem. Soc, 111, 6434-6438. Lo, T.P. (1995) Ph.D. Dissertation, Department of Biochemistry and Molecular Biology, The University of British Columbia. Loew, G.H. (1982) Iron Porphyrins, Part One (Lever, A.B.P. & Gray, H.B., Eds.), Addison-Wesley Publishing Company, Don Mills, Ont, pp. 1-87. Looze, Y., Polastro, E., Gielens, C. & Leonis, J. (1976) Biochem. J., 157, 773-775. Looze, Y., Polastro, E., DeconinckM.C. & Leonis, J. (\918)Intl. J.Pept. &Prot. Res., 12,233-236. Louie, G.V., Hutcheon, W.L.B. & Brayer, G.D. (1988) J. Mol. Biol., 199, 295-314. Louie, G.V. & Brayer, G.D. (1990) J. Mol. Biol., 214, 527-555. Louie, G.V. (1990) Ph.D. Dissertation, Department of Biochemistry and Molecular Biology, The University of British Columbia. Luntz, T.L., Schejter, A., Garber, E.A.E. & Margoliash, E. (1989)Proc Natl. Acad. Sci. U.S.A., 86, 3524-3528. Maltempo, M.M., Moss, T.H. & Cusanovich, MA. (1914) Biochim. Biophys. Acta, 342, 290-305. Maltempo, M.M. (1976) Q. Rev. Biophys., 9, 181-215. 179 Maltempo, M.M. (1976) Biochim. Biophys. Acta, 434, 513-518. Margoliash, E. & Frohwirt, N (1959) Biochem. J., 71, 570-572. Margoliash, E. & Schejter, (1995) Cytochrome c. A Multidisciplinary Approach, (Scott, R.A. & Mauk, A.G., Eds.) University Science Books, Sausalito, CA, pp. Marion, D. (1994) Biochemie, 76, 631-640. Martinez, S.E., Huang, D., Szczepaniak, A., Cramer, W.A. & Smith, J.L. (1994) Structure, 2, 95-105. Matheson, I.B.C. (1989) Computers Chem., 13, 299-304. Mauk, A. G., Coyle, C. L., Brodignon, E. & Gray, H. B. (1979) J. Am. Chem. Soc, 101, 5054-5056. Meites, L. & Meites, T. (1948) Anal. Chem., 20, 984-985. Meyer, T.E., Tollin, G. & Cusanovich, MA. (1986) Biochimie., 76, 480-488. Michel, B., Proudfoot, A.E.I., Wallace, C.J.A. & Bosshard, HR. (1989)Biochemistry, 28,456-462. Migita, CT. & Iwaizumi, M. (1981) J. Am. Chem. Soc, 103, 4378-4381. Mignotte, B. & Vayssiere, J.L. (1998) J. Biochem., 252, 1-15. Mitchell, D.M., Adelroth, P., Hosier, J.P., Fetter, J.R, Brzezinski, P., Pressler, M.A., Aasa, R., Malmstrom, B.G., Alben, J.O., Babcock, G.T., Gennis, R.B. & Ferguson-Miller, S. (1996) Biochemistry, 35, 824-828. Moench, S.J. & Satterlee, J.D. (1989)7. Biol. Chem., 264, 9923-9931. Moench, S.J., Shi, T-M. & Satterlee, J.D. (1991) Eur. J. Biochem., 197, 631-641. Moore, G.R. & Pettigrew, G.W. (1990) Cytochromes c. Evolutionary, Structural and Physichochemical Aspects, Springer Verlag, New York. Moore, G.R. (19S3) FEBS Lett., 161, 171-175. Moore, G.R., Cox, M.C, Crowe, D., Osborne, M.J., Mauk, AG. & Wilson, J.T. (1995) Nuclear Magnetic Resonance of Paramagnetic Macromolecules (La Mar, G.N., Ed.), Kluwer Academic Publishers, Amsterdam, pp. 95-122. Moore, G.R., Cox, M.C, Crowe, D., Osborne, M.J., Rosell, F.I., Bujons, J., Barker, P.D., Mauk, M.R., Mauk, AG. (1998)Biochem. J., 332, 439-449. 180 Mori, E. & Morita, Y. (1980) 7 Biochem., 87, 249-266. Morton, R.A. (1973) Can. J. Biochem., 51, 465-471. Morton, R.A. & Breskvar, K. (1977) Can. 7. Biochem., 55, 146-151. Myer, Y.P. (1985) Current Topics in Bioenergetics, 14, 149-188. Nagendra, H.G., Sukumar, N.S. & Vijayan, M. (1998) Proteins: Struc. Func. Gen., 32, 229-240. Nail, B.T. (1986) Biochemistry, 25, 2974-2978. Nail, B.T., Zuniga, E.H., White, T.B., Wood, L.C. & Ramdas, L (1989) Biochemistry, 28, 9834-9839. Nail, B.T. (1995) Cytochrome c. A Multidisciplinary Approach, (Scott, R.A. & Mauk, A.G., Eds.), University Science Books, Sausalito, CA, pp. 167-201. Nicholson, D.W., Hergersberg, C. & Neupert, W. (1988) J. Biol. Chem., 263, 19034-19042. Nicholson, D.W. & Neupert, W. (1989) Proc. Natl. Acad. Sci. U.S.A., 86, 4340-4344. Northrup, S.H., Thomasson, K.A., Miller, CM., Barker, P.D., Eltis, L.D., Guillemette, J.G., Lo, T.P., Inglis, S.C. & Mauk, A.G. (1993) Biochemistry, 32, 6613-6623. Nozaki, M., Mutzushima, H., Horio, T. & Okunuki, K. (1958)7 Biochem., 45, 815-823. Ochi, H., Hata, Y, Tanaka, N., Kakudo, M., Sakurai, T., Aihara, S. & Morita, Y. (1983) 7. Mol. Biol., 166, 407-418. Osheroff, N., Koppenol, W.H. & Margoliash, E. (1978) Frontiers ofBiological Energetics (Dutton, P.L., Leigh, J. & Scarpa, A.), Academic Press, New York, pp. 439-449. Osheroff, N., Borden, D., Koppenol, W.H. & Margoliash, E. (1980)7. Biol. Chem., 255,1689-1697. Osterhout, J.J.Jr., Muthukrishnan, K. & Nail, B.T. (1985) Biochemistry, 24, 6680-6684. Otiko, G. & Sadler, P.J. (1980) FEBSLett., 116, 227-230. Pal, P.K., Verma, B. & Myer, Y. (1975) Biochemistry, 14, 4325-4334. Palmer, G. (1979) The Porphyrins, vol. IV (Dolphin, D. Ed.), Academic Press, New York, pp. 313-353. Palmer, G. (1985) Biochem. Soc. Trans., 13, 548-560. 181 Panzer, R.E. & Elving, P.J. (1975) Electrochim. Acta, 20, 635-647. Pearce, L.L., Gartner, A.L., Smith, M. & Mauk, AG. (1989)Biochemistry, 28, 3152-3156. Pettigrew, G.W. & Moore, G.R. (1987) Cytochromes c: Biological Aspects, Springer Verlag, New York. Pettigrew, G.W., Aviram, I. & Schejter, A. (1976) Biochem. Biophys. Res. Commun., 68, 807-813. Pfeil, W. (1981)Mol. Cell. Biochem., 226, 3-28. Pielak, G.J., Mauk, AG. & Smith, M. (1985) Nature, 313, 152-154. Pinheiro, T.J.T. & Watts, A. (1994) Biochemistry, 33, 2451-2458. Pollock, W.B.R., Rosell, F.L, Twitchett, M.B., Dumont, M.E. & Mauk, AG. (1998) Biochemistry, 37, 6124-6131. Preznat, T., Pfeifer, K. & Guarente, L. (1987) Mol. Cell Biol., 7, 3252-3259. Qi, P.X., Di Stefano, D.L. & Wand, A.J. (1994) Biochemistry, 33, 6408-6417. Qi, P.X., Beckman, R.A. & Wand, A.J. (1996) Biochemistry, 35, 12275-12286. Rafferty, S.P., Pearce, L.L., Barker, P.D., Guillemette, J.G., Kay, CM., Smith, M. & Mauk, AG. (1990) Biochemistry 29, 9365-9369. Rafferty, S.P. (1992) Ph.D. Dissertation, The University of British Columbia, Vancouver, B.C., Canada. Raman, C.S., Jemmerson, R, Nail, B.T. & Allen, M.J. (1992) Biochemistry, 31, 10370-10379. Redfield, A.G. & Gupta, R.K. (1971) Cold Spring Harb. Symp. Quant. Biol., 36, 405-411. Rigby, S.E., Moore, G.R., Gray, J.C, Gadsby, P.M.A., George, S.J. & Thomson, A. J. (1988) Biochem. J., 256, 571-577. Roberts, G.C.K. (1993) NMR of Macromolecules : A Practical Approach, JRL Press at Oxford University Press, Oxford, U.K. Rodgers, K.K., Pochapsky, T.C & Sligar, S.G. (1988) Science, 240, 1657-1659. Rosell, F.I., Ferrer, J.C. & Mauk, AG. (1998)7. Am. Chem. Soc, 120, 11234-11245. Russel, M., Kidd, S. &Kelley, MR. (1986) Gene, 45, 333-338. 182 Rytomaa, M. & Kinnunen, P.K. (1994) J. Biol. Chem., 269, 1770-1774. Saigo, S. (\9S\a) Bioch. Biophys. Acta, 669, 13-20. Saigo, S. (1981b) J. Biochem., 89, 1977-1980. Saigo, S. (1986) J. Biochem., 100, 157-165. Salemme, FR. (1976)7. Mol. Biol., 102, 563-568. Salerno, J.C. (1984) J. Biol. Chem., 259, 2331-2336. Sambrook, J., Fritsch, E.F. & Maniatis, T. (1989) Molecular Cloning: a Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York. Sanishvili, R., Volz, K.W., Westbrook, E.M. & Margoliash, E. (1995) Structure, 3, 707-716. Santos, H. & Turner, D.L. (1992) Eur. J. Biochem., 206, 721-728. Santucci, R. & Ascoli, F. (1997) J. Inorg. Biochem., 68, 211-214. Satterlee, ID. (1986) Ann. Rep. NMRSpec, 17, 79-178. Schechter, E. & Saludjian, P. (1967) Biopolymers, 5, 788-790. Schejter, A.& Aviram, I. (1969) Biochemistry, 8, 149-153. Schejter, A., Aviram, I. & Sokolovsky, M. (1970) Biochemistry, 9, 5118-5122. Schejter, A., Luntz, T.L., Koshy, T.I. & Margoliash, E. (1992) Biochemistry, 31, 8336-8343. Schejter, A., Taler, G., Navon, G., Liu, X.J. & Margoliash, E. (1996)/. Am. Chem. Soc, 118, 477-478. Scott, R.A. & Mauk, A.G. (1995) Cytochrome c. A Multidisciplinary Approach, University Science Books, Sausalito, CA. Sedlak, E. (1997) Biochem. Molec. Biol. Int., 41, 25-32. Sedlak, E. & Antalik, M. (1998) Biopolymers, 46, 145-154. Shao, W., Sun, H., Yao, Y. & Tang, W. (1995) Inorg. Chem., 34, 680-687. Sherman, F., Stewart, J.W., Parker, J.H., Inhaber, E., Shipman, N.H., Putterman, G.J., Gardisky, R.L. & Margoliash, E. (1968), J. Biol. Chem., 243, 5446-5456. 183 Simonson, T. & Perahia, D. (1995) Proc. Natl. Acad. Sci. U.S.A., 92, 1082-1086. Simpkin, D., Palmer, G., Devlin, F.J., McKenna, M.C., Jensen, GM. & Stephens, P.J. (1989) Biochemistry, 28, 8033-8039. Skov, K., Hofmann, T. & Williams, G.R. (1969) Can. J. Biochem., 47, 750-752. Skulachev, V.P. (1998) FEBSLett., 423, 275-280. Smith, D.W. & Williams, R.J.P. (1970) Struct. Bond., 7, 1-45. Smith, H.T. & Millet, F. (1980) Biochemistry, 19, 1117-1120. Somero, G.N. (1986) Am. J. Phys. Soc, 25, R197-R213. Soussi, J3., Bylund-Fellenius, A., Schersten, T. & Angstrom, J. (1990) Biochem. J., 265, 227-232. Spiro, TG. (1985) Adv. Prot. Chem., 37, 111-159. Stellwagen, E. & Schulman, R.G. (1973)7. Mol. Biol, 80, 559-573. Stellwagen, E., Babul, J. & Wilgus, H. (1975) Biochim. Biophys. Acta, 4, 115-121. Stuart, R.A., Nicholson, D.W. & Neupert, W. (1990) Cell, 60, 31-43. Sutin, N. & Yandell, J.K. (1972) 7. Biol. Chem., 247, 6932-6936. Swanson, R., Trus, B.L., Mandel, N., Mandel, G, Kallai, O.B. & Dickerson, R.E. (1977)7. Biol. Chem., 252, 759-775. Taborsky, G. & McCollum, K. (1979)7 Biol. Chem., 254, 7069-7075. Takano, T., Kallai, O.B., Swanson, R. & Dickerson, R.E. (1973)7. Biol. Chem., 248, 5234-5255. Takano, T., Trus, B.L., Mandel, N., Mandel, G., Kallai, O.B., Swanson, R. & Dickerson, R.E. (1977)7. Biol. Chem., 252, 776-785. Takano, T. & Dickerson, R.E. (1980) Proc Natl. Acad. Sci. U.S.A. 11, 6371-6375. Takano, T. & Dickerson, R.E. (1981a) 7. Mol. Biol, 153, 79-94. Takano, T. & Dickerson, R.E. (1981b) 7. Mol. Biol, 153, 95-115. Taler, G, Shejter, A., Navon, G, Vig, I. & Margoliash, E. (1995) Biochemistry, 34, 14209-14212. Tanaka, N., Yamane, T., Tsukihara, T., Ashida, T. & Kakudo, M. (1975)7. Biochem., 11, 147-162. 184 Taniuchi, H., Basile, G., Taniuchi, M. & Veloso, D. (1983) J. Biol. Chem., 258, 10963-10966. Taniguchi, I., Masahiro, I., Yamaguchi, I. & Yasukinicki, K. (1984)7. Electroanal. Chem., 1975, 341-348. Taniguchi, I., Iseki, M., Yamaguchi, H & Yasuko-uchi, K. (1985)7 Electroanal. Chem., 186, 299-307. Tatulian, S.A. & Tamm, L.K. (1996)7. Mol. Biol, 260, 312-316. Teixeira, M., Campos, A.P., Aguiar, A.P., Costa, H.S., Santos, H., Turner, D.L. & Xavier, A.V. (1993) FEBSLett., 317, 233-236. Theodorakis, J.L., Garber, E.A., McCracken, J., Peisach, J., Schejter, A. & Margoliash, E. (1995) Biochim. Biophys. Acta, 1252, 103-113. Theorell, H. (1936) Biochem. Z, 285, 207-218. Theorell, H. (1939) Science, 90, 67. Theorell, H. & Akesson, A. (1941)7. Am. Chem. Soc, 63, 1804-1811. Theorell, H. & Akesson, A. (1941)7. Am. Chem. Soc, 63, 1812-1814. Theorell, H. & Akesson, A. (1941)7. Am. Chem. Soc, 63, 1815-1820. Theorell, H. (1941)7. Am. Chem. Soc, 63, 1820-1827. Thony-Meyer, L. (1997)Microbiol. Mol. Biol. Rev., 63, 337-376. Tollin, G., Hanson, L.K., Caffrey, M., Meyer, TE. & Cusanovich, MA. (1986) Proc Natl. Acad. Sci. U.S.A., 83, 3693-3697. Tonge, P., Moore, G.R. & Wharton, C.W. (1989) Biochem. 7, 258, 599-605. Trewhella, J., Carlson, V. & Curtis, EH. (1988)Biochemistry, 27, 1121-1125. Trumbly, R.J. (1992) Mol. Microbiol, 6, 15-21. Ubbink, M., Warmerdam, G.C, Campos, A.P., Teixeira, M. & Canters, G.W. (1994) FEBS Lett., 351, 100-104. Uno, T., Nishimura, Y. & Tsuboi, M. (1984) Biochemistry, 23, 6802-6806. Walker, F.A., Huynh, B.H., Scheidt, W.R. & Osvath, SR. (1986)7. Am. Chem. Soc, 108, 5288-5297. 185 Wallace, C.J.A. (1984) Biochem. 7, 217, 604-604. Wallace, C.J.A. & Clark-Lewis, I. (1992)7 Biol. Chem., 267, 3852-3861. Wallace, C.J.A. & Clark-Lewis, I. (1997) Biochemistry, 36, 14733-14740. Weber, C, Michel, B. & Bosshard, HR. (1987)Proc. Natl. Acad. Sci. U.S.A., 84, 6687-6691. White, T.B., Berget, P.B. & Nail, B.T. (1987) Biochemistry, 26, 4358-4366. White-Dixon, D., Hong, X. & Woehler, S.E. (1989) Biophys. 7, 56, 339-351. Wilgus, H. & Stellwagen, E. (1974) Proc. Natl. Acad. Sci. U.S.A., 71, 2892-2894. Williams, G., Moore, G.R., Porteus, E., Robinson, M.N., Soffe, N. & Williams, R.J.P. (1985)7. Mol. Biol., 183, 409-428. Wilson, I.A., Skehel, J.J. & Wiley, D.C. (1981) Nature, 289, 366-373. Wilson, M.T. & Greenwood, C. (1995) Cytochrome c. A Multidisciplinary Approach, (Scott, R.A. & Mauk, A.G., Eds.) University Science Books, Sausalito, CA, PP- 611-635. Wood, L.C., Muthukrishnan, K., White, T.B., Ramdas, L. & Nail, B.T. (1988) Biochemistry, 27, 8554-8561. Wood, L.C., White, T.B., Ramdas, L. & Nail, B.T. (1988) Biochemistry, 27, 8562-8568. Woodruff, W.H. (1993)7. Bioenerg. Biomembr., 25, 177-188. Wooten, J.B., Cohen, J.S., Vig, I. & Shejter, A. (1981) Biochemistry, 20, 5394-5402. Wuthrich, K. (1986) NMR of Proteins and Nucleic Acids, John Wiley & Sons, New York. Wyman, F. (1948) Advances in Protein Chemistry, IV, (Anson, M.L. & Edsall, J.T., Eds.) Academic Press Inc., New York, pp. 407-531. Yamanaka, T., Mutzushima, H., Nozaki, M., Horio, T. & Okunuki, K. (1959)7. Biochem., 46,121-132. Yang, J., Liu, X., Bhalla, K., Kim, N.C., Ibrado, A.M., Cai, J., Peng, T., Jones, D.P. & Wang, X. (1997) Science, 275, 1129-1132. Zhivotovsky, B., Orrenius, S., Brustugun, O.T. & Doskeland, S.O. (1998) Nature, 391, 449-450. Zhou, H-X. & Vijayakumar, M. (1997)7. Mol. Biol., 267, 1002-1011. 186 Zitomer, R.S., Montgomery, D.L., Nichols, D.L. & Hall, BD. (1979)Prac. Natl. Acad. Sci. U.S.A., 76,3627-3631. Zoller, M.J. & Smith, M. (1982a) Nucleic Acids Res. 10, 6487-6500. Zoller, M.J. & Smith, M. (1982b) DNA 3, 479-488. Zoller, M.J. & Smith, M. (1983) Methods Enzymol. 100, 468-500. 187 Appendix 1. Definition of the model used to determine the Kap of the alkaline conformational transition of yeast /so-1-ferricytochrome c. P-H " \ P 1-1 H+ For the simple deprotonation of the molecule P-H as shown above, the Henderson-Hasselbach Relation is : K*= H]] 1-2 [P-H] = [P]-10^-pH) U5 log(2^S) = p^a - pH 1.4 where [T] is the total concentration of the molecule, therefore TT1 [P] = [T>[P-H] 1-7 [P]=(1 + 10(P^-pH)) [P] = [T]-[P]-10(p^-pH) I-8 The absorbance at a given wavelength, Z, reflects the contribution from all absorbing species Z = e(p_H)-[P-H] + e(p)-[P] 1-10 Substituting equations 1-6 and 1-9 into 1-10 yields 7_[T]-(e(p_H)-10(p^-pH) + B(p)) (l + 10(^-pH)) B + A-10(p^~pH) 1-12 (l + 10(p^-pH)) where A=[T]-8(P_H) l-l2a B = [T]-8(P) l-12b loo However, the alkaline conformational transition of ferricytochromes c consists of a deprotonation that is coupled to a con formational reorganization of the protein. These linked equilibria are illustrated by the thermodynamic cycle shown in Relation 1-13. Spectrophotometrically, only the con formational states can be distinguished, not the protonation states. Consequently : N-H N A-H N K0bs -[NH] + [N] [AH] + [A] 1-13 1-14 or because 1 + ^obs - Q "(" [H+] KA . 1 + [H+] N 1-15 1-16 At high pH : and at low pH : ^obs ~~* Q ^obs ~~* Q> At equilibrium, the fraction of the sample in the native state is given by 1-17a while that of the alkaline species is given by 1-17b /(NH + N) = /(AH + A) = l + K, obs K. obs \ + K, obs l-17a l-17b 1 - ' ^obs Z = %-(TT^) + ^-(TTl^)i-i8a Z 1 + ^obs 1 + ^obs £^+£A-Kohs 1-18b At any given wavelength, the measured \ + K absorbance is the sum of the contribution of all obs absorbing species as shown in Equations 1-18 and 1-19. Z = ?N.(i+Gp)+*A.cd KN 1-19 189 As Cp - 0 (i.e., the ligand exchange is strongly disfavoured prior to deprotonation), KA »1 so that *N-(1 + Gp) + *A-Cd Z-> 1-20 G+O+^P) N At low pH, [FT] KN, therefore ffN.(E!l)+ffA.cd G [H+] AT, N 1-21 At G« 1 at G = 1 and at CH» 1 N Z-> AN N Z->£, At high pH, [H+] « ATN, therefore 2 _ gN + gA ' Q Cd+1 1-22 AtCd« 1 : atCd=l and at Cd »> 1 Z->£ Z-> N fN +£,A (,A+^)+,N.([p) z= p 1 TH+1 Equation 1-20 can be rearranged to 1-23 (1 H ) + ± ± l-23a where ATap = K^-Cd. Note that Equation 1-23 G ' ATap is analogous to 1 -12 when Cd »> 1. In this case A and B, from Equation 1-12 are eN and eA, respectively. This assumption is justified s£ +^L) + E^ .\QPKV~PH considering that the measured values of Cd of 2 the yeast iso-l -ferricytochromes c studied here 1 inpK -pH range from 94-4000 (=KC in Table 11). Q l'23h 190 Applying the same assumption (Kap - K^Cd) to the thermodynamic cycle (Equation 1-13) leads to the same conclusion as above. K ^[N]-[H+] [A] = [A-H]-[H+] [A] i_25 ap [NH] [N] [NH] [A-H] ap [NH] This form of the apparent equilibrium constant is the same as that shown in Equation 1-2. Therefore, titration data can be fitted to the Henderson-Hasselbach model for the simple deprotonation of a group using Equation 1-12 provided that the protonation state of the native and alkaline species does not influence the corresponding extinction coefficients. 191 Appendix 2. Using the Gouy-Chapman equation to investigate the ionic strength dependence of the alkaline conformational equilibrium of cytochromes c. According to the Boltzmann equation, the local pH at a surface depends on the bulk pH of the solution and on the electrostatic potential of the surface, <p0. [H+]local=[H+]bulk-exp(^) (2-1) kT where e is the charge of the proton, (p0 is the electrostatic potential of the surface, k is the Boltzmann constant, and T is the temperature (K). Because pK^ = pHlocal at the midpoint of the titration, the pHbuik = P^apparent (Koutalos et al, 1990). Therefore, Equation 2-1 can be redefined as P^true " P^apparent = l0g[exp(^-)] (2-2) e§0 = 2.30.(p^trae-p^apparent) (2-3) According to the Gouy-Chapman equation, the surface potential (p0 is a function of the surface charge density and the ionic strength. ^ • —= sinhf ^-) (2-4) where o is the surface charge density, C is the salt concentration, z is the valence of the electrolyte (i.e., z = 1 for KC1), and A a the constant: ^ = (8-e-e0-NA*-T)'/2 (2-5) In Equation 2-5, e is the dielectric constant of the protein (e = 25±10; reported by Simonson and Perahia (1995) assuming that the radius of cytochrome c, including all protein atoms in 2YCC.PDB and 20 bound water molecules is 16.5 A), e0 is the permittivity of free space, and A^A is Avogadro's number. At 25 °C, A = 76.3 MVl for cytochrome c. Substitution of this constant and Equation 2-3 into the Gouy-Chapman Equation (2-4) yields the expression : 76 3 o —j=- = sinh[l.l 5 • (pK^ - p£apparent)] (2-6) ex-e~x sinh(x) = [— ] The p^apparen, values are determined according to Appendix 1 and at a variety of salt concentrations. 192 These values are then be fitted to Equation 2-6 to estimate the true pK3 of the titrating group and the average surface charge density in the immediate region around this titrating group. 193 Appendix 3. Simulation of the alkaline transition kinetics of wild-type yeast iso-\-ferricytochromec. At a given pH, the percentage of the total ferricytochrome c in the alkaline state (A) can be determined from the experimentally determined pKip using the Henderson-Hasselbach equation (3-1) A =10(PH-P^) (31) 1-A Assuming that in the wild-type protein the alkaline conformational rearrangement to each of the two Lys-bound states (A73 and A79) occurs independently from the other, these ratios can be used as weighing factors to reconstruct the kobs for the wild-type protein using the kob3 obtained for the Lys73Ala and Lys79Ala variants. The fraction of the wild-type protein that is in the LysX-bound alkaline state is : WTX%= Awt'Ax (3-2) A-73 + A79 Therefore CsT = *o7bs • WT73% + *Jb9s • WT79% (3-3) where kxobs is the observed rate constant observed in pH-jump experiments of cytochrome c X at a given pH. 194 Appendix 4. Preliminary residue-specific proton assignments for the yeast /'so-1-ferricytochrome c variant Lys79Ala/Cysl02Thr. These assignments were compiled from NOE and TOC spectra collected at pH* 10.8, 298 K in H20 and D20. Residue Proton (ppm) Phe -3 HD1 7.062 -3 CE1 7.227 -3 HZ 7.137 Ala -1 HN 7.936 -1 HA 3.634 -1 QB 1.125 Gly 1 HA1 3.553 1 HA2 4.400 Ser 2 HA 4.823 2 HB2 3.724 2 HB3 3.182 Ala 3 HA 4.036 3 QB 1.515 Lys 5 HN 7.706 5 HA 4.067 5 HB3 1.762 5 HG2 2.276 5 HG3 2.719 5 HD2 2:008 5 HD3 1.687 5 HE2 1.002 Gly 6 HA2 3.296 6 HA1 3.992 Ala 7 HA 2.398 7 QB 1.238 Thr 8 HA 3.948 8 HB 4.186 8 QG2 1.226 Leu 9 HA 3.824 9 HB3 1.877 9 QDl 0.649 9 QD2 0.895 Phe 10 HA 3.425 10 HB2 2.964 10 HB3 2.659 10 CD1 7.011 10 CE1 7.470 10 HZ 8.568 10 HE2 7.530 Leu 15 HA 6.026 Residue Proton (ppm) Leu 15 HB2 2.415 15 HB3 2.106 15 HG 2.337 15 QDl 1.446 15 QD2 2.051 Cys 17 HA 5.981 17 HB2 0.635 17 HB3 -0.880 Thr 19 HA 6.542 19 HB 5.692 19 QG2 2.495 His 18 HN 11.395 18 HB3 15.033 18 HB2 9.467 18 HA 10.743 18 HD2 6.222 18 HE1 6.882 Val 20 HA 4.199 20 HB 2.301 20 QG1 1.075 20 QG2 0.928 Gly 23 HA1 3.133 23 HA2 4.160 Gly 24 HA2 3.894 24 HA1 4.399 His 26 HD2 6.952 26 HE1 7.764 Val 28 HA 3.388 28 HB 1.133 28 QG1 -0.861 28 QG2 0.616 Asn 31 HA 5.724 31 HB2 3.475 31 HB3 2.492 31 HD1 8.550 His 33 HE1 8.060 33 HD2 7.250 Gly 34 HA2 3.700 34 HA2 3.840 Phe 36 CD1 7.618 Residue Proton 0>pm) 36 CE1 6.711 36 HZ 7.248 Gly 37 HA1 3.673 37 HA2 4.188 His 39 HD2 6.974 39 HE1 8.495 Ser 40 HA 4.467 40 HB2 4.201 40 HB3 3.768 Ala 43 HA 4.315 43 QB 1.361 Gly 45 HA1 3.623 45 HA2 4.247 Tyr 46 HD1 6.434 46 HE1 7.077 46 HD2 6.040 46 HE2 6.730 Tyr 48 HA 4.111 48 HB2 6.299 48 HB3 3.232 48 HD1 6.932 48 HE1 6.563 Asn 52 HN 8.445 52 HA 4.594 52 HB2 2.891 52 HB3 3.161 52 HD1 7.031 He 53 HA 3.405 53 HB 2.851 53 QG2 1.370 53 HG12 0.730 53 HG13 1.167 53 QDl 1.031 Asn 56 HA 4.317 56 HB2 2.314 56 HB3 3.146 56 HD1 6.788 Val 57 HA 3.821 57 HB 1.128 57 OG1 0.471 195 Residue Proton (ppm) Ala 101 HA 4.042 101 QB 0.622 Thr 102 HA 4.607 102 HB 1.420 102 QG2 0.976 Glu 103 HN 6.795 103 HA 4.093 103 HB2 1.952 103 HB3 2.057 103 HG2 3.140 103 HG3 2.299 Hem 104 QM1 13.834 104 QM3 12.271 104 QM5 20.566 104 QM8 24.828 104 HAM 5.301 104 HBM -3.320 104 HGM 4.897 Hem 104 HDM -4.771 104 QT2 -0.440 104 QT4 0.139 104 HT4 -0.040 104 HAP61 0.593 104 HAP62 -0.887 104 HBP63 1.123 104 HBP64 3.364 104 HAP71 -0.343 104 HAP72 -0.611 104 HBP73 1.359 104 HBP74 -0.255 Residue Proton (pom) Val 57 QG2 -0.158 Trp 59 HA 5.161 59 HB2 3.932 59 HB3 3.284 59 HE3 7.121 59 HZ3 6.533 Trp 59 HZ2 6.736 59 HH2 6.365 Asn 62 HA 4.558 62 HB2 2.858 62 HB3 2.918 Asn 63 HA 4.415 63 HB2 2.942 63 HB3 3.229 63 HD21 7.054 Tyr 67 HE1 6.318 67 HD1 7.226 Leu 68 HA 4.408 68 HB2 1.527 68 HB3 1.963 Leu 68 HG 1.088 68 QDl -1.550 68 QD2 -0.500 He 75 HA 4.974 75 HB 3.463 75 QG2 2.068 75 HG12 2.741 75 HG13 2.861 75 QDl 2.381 P 76 HA 5.426 76 HB2 2.194 76 HB3 2.494 76 HG2 2.304 76 HG3 2.939 76 HD2 3.892 Residue Proton (l)|)m) 76 HD3 3.045 Ala 81 HA 5.275 81 QB 1.448 Phe 82 HA 4.826 82 HB2 3.174 82 HB3 3.711 82 HD1 7.560 82 HE1 7.020 82 HZ 8.620 Leu 94 HN 8.152 94 HA 3.799 94 HB2 1.270 94 HB3 1.323 94 HG 0.907 94 QDl 0.370 94 QD2 0.039 He 95 HN 8.550 95 HA 2.940 95 HB 1.738 95 QG2 0.292 95 HG12 0.224 95 QDl 0.536 Thr 96 HA 3.622 Thr 96 HB 4.253 96 QG2 1.075 Tyr 97 HD1 6.712 97 HE1 7.676 97 HE2 7.185 97 HD2 5.738 Leu 98 HA 3.193 98 HB2 1.218 98 HB3 1.312 98 HG 0.897 98 QDl -0.431 98 OD2 -0.283 Note that these assignment are preliminary and require further confirmation by assigning the remaining protons. 196 Appendix 5. Electrochemistry of yeast iso-l -ferricytochromes c as a function of pH. Lys79AJa pH AE E AE E (mV) (mV vs SHE) (uA) (mV) (mV vs SHE) (uA) 6.73 59.8 -165.4 4.6 81.2 278.0 13.3 7.14 58.6 -168.1 5.6 80.2 275.4 11.6 7.46 74.6 -177.0 7.1 75.6 274.6 13.2 7.71 63.9 -181.8 9.5 77.8 272.4 11.4 8.13 62.1 -179.2 10.1 75.4 272.6 10.6 8.43 66.2 -180.9 11.6 80.4 271.2 9.4 8.57 62.8 -184.0 13.3 86.8 273.0 7.1 8.79 79.0 -180.1 15.4 86.1 272.6 6.0 9.04 105.3 -173.4 17.1 81.3 274.0 4.1 9.37 77.5 -185.3 18.8 82.9 280.9 3.4 9.57 101.6 -189.2 18.6 88.5 277.9 2.3 10.11 80.7 -214.6 16.3 64.1 287.9 0.6 10.66 70.4 -240.1 11.3 6.78 44.3 -179.8 4.8 7.27 46.5 -191.8 5.8 7.86 48.6 -192.8 6.5 8.33 55.9 -201.2 10.1 8.85 61.1 -193.4 11.9 9.04 58.7 -201.6 13.5 9.51 62.7 -197.3 14.4 10.01 62.0 -217.9 14.8 10.41 68.4 -226.7 13.0 4.39 44.8 -15.6 5.07 40.4 -60.4 5.72 43.9 -89.9 6.16 59.6 -128.2 6.62 54.5 -159.3 7.20 50.8 -183.3 7.70 61.9 -188.9 8.31 47.4 -200.5 8.71 54.1 -202.9 9.09 72.0 -199.4 9.61 64.5 -211.4 10.03 65.8 -223.0 10.53 61.0 -264.0 11.04 48.4 -288.1 197 Lys73Ala PH LE E AE E (mV) (mVvjSHE) (uA) (mV) (mV vs SHE) (uA) 6.79 59.6 -169.5 4.0 79.5 271.0 13.6 7.10 77.9 -172.2 4.2 86.7 266.9 16.0 7.47 54.3 -182.7 5.5 82.5 270.0 15.4 7.78 54.4 -185.9 7.3 82.2 262.8 16.4 8.01 72.7 -175.4 6.9 81.5 263.9 14.6 8.32 61.4 -180.6 8.7 81 263.9 12.9 8.50 71.1 -179.0 10.0 92.4 264.5 13.2 8.70 59.5 -184.1 11.7 91.2 264.6 12.9 9.04 87.2 -174.3 14.0 83.4 262.8 9.2 9.32 87.2 -180.5 15.0 90 263.9 5.8 9.63 93.1 -187.6 17.3 97.9 268.2 3.5 9.95 89.2 -198.1 18.1 100.8 265.0 2.7 10.60 68.8 -234.4 13.5 6.85 56.5 -156.0 3.5 7.32 53.2 -178.6 4.1 7.86 61.2 -178.1 5.0 8.10 58.1 -188.8 6.2 8.49 63.9 -185.9 8.0 8.81 58.0 -182.3 9.5 9.25 67.8 -193.6 12.5 9.60 78.0 -189.7 13.4 10.07 71.6 -200.7 12.0 10.50 54.8 -239.4 12.5 11.00 141.9 -288.8 4.38 54.5 -31.6 5.06 66.4 -66.0 5.72 72.5 -100.7 6.16 95.5 -119.0 6.64 46.7 -164.2 7.22 37.4 -187.0 7.67 44.8 -195.2 8.33 71.8 -195.6 8.71 77.2 -193.8 9.12 90.5 -200.5 9.62 98.5 -202.2 10.01 73.2 -217.1 10.59 70.0 -258.4 11.04 65.2 -281.9 198 Appendix 6. Temperature dependence of the equilibrium constants for the alkaline conformational transition of wild-type yeast /'so-1-ferricytochrome c and the variants Lys73Ala, Lys79Ala, Lys86Ala, and Lys87Ala. [alkaline] ^ area[IVa(8) + IVb(8)] [native] ~ area[III(8)] T 1/T -R-ln(ATeq) (J-mol" '•K-1) (°C) (K-1) Lys87Ala Lys86Ala Wild-type Lys79Ala Lys73Ala 20 0.00341 -1.9833 -2.0383 2.9940 0.6617 25 0.00335 -3.4858 -2.6655 1.1059 3.9336 30 0.00330 -0.8655 -5.2858 -3.5367 -0.4723 1.2424 35 0.00325 -1.6003 -6.3730 -4.6838 -2.9387 0.3864 40 0.00319 -3.3671 -7.3561 -6.5599 -5.0945 -0.2290 45 0.00314 -5.7476 -9.8393 -8.0356 -8.0199 50 0.00309 -8.5278 55 0.00305 -10.3496 199 Appendix 7a. Distribution of conformational states of wild-type ferricytochrome c measured at pH* 10.0, 25 °C, as a function of potassium chloride concentration. [KC1] 111(8) 111(3) (mM) 8 (ppm) Peak Area Peak Height 8 (ppm) Peak Area Peak Height 0 35.23 22.40 46.61 31.71 18.85 50.17 12 35.24 21.28 67.32 31.74 18.61 68.12 30 35.24 22.80 64.75 31.80 17.82 83.06 53 35.21 30.68 98.49 31.83 27.18 103.07 69 35.20 32.24 109.02 31.88 26.69 112.67 85 35.18 35.93 119.63 31.89 32.20 135.61 98 35.15 42.21 139.26 31.90 35.69 141.52 199 35.11 44.80 147.09 32.00 36.60 156.87 296 35.07 46.88 171.31 32.05 43.33 181.78 395 35.02 49.84 159.80 32.09 43.20 184.12 494 34.97 53.55 181.39 32.11 58.88 200.75 597 35.00 68.01 268.33 32.17 67.42 289.31 697 34.97 59.16 237.14 32.19 57.94 251.72 797 34.96 42.24 172.81 32.22 41.51 170.61 903 34.97 38.93 146.92 32.25 33.40 135.89 996 34.96 30.42 102.22 32.26 30.37 101.21 IV.(8) IV,(8) 0 24.97 56.73 125.09 24.75 203.08 447.64 12 24.94 68.31 150.60 24.71 196.79 434.00 30 24.93 94.04 207.32 24.69 151.88 373.46 53 24.82 123.70 217.68 24.59 112.07 303.86 69 24.81 146.02 252.00 24.58 104.75 274.85 85 24.80 149.74 271.76 24.55 100.41 276.25 98 24.77 152.26 289.08 24.52 95.34 281.53 199 24.72 183.50 377.55 24.45 74.46 257.91 296 24.69 181.53 429.58 24.41 73.36 240.40 395 24.66 176.94 460.53 24.38 71.49 222.16 494 24.64 170.48 516.38 24.34 80.37 231.04 597 24.63 168.71 571.95 24.33 65.29 179.87 697 24.62 177.85 602.85 24.33 73.70 193.92 797 24.62 176.06 608.69 24.34 66.09 185.30 903 24.62 191.20 612.09 24.35 40.09 165.91 996 24.63 192.08 625.03 24.37 62.70 187.39 IVJ5) IVK(5) 0 20.91 230.18 128.33 19.32 97.63 316.09 12 20.98 223.86 112.86 19.31 101.73 362.01 30 na na na 19.32 96.98 364.95 53 na na na 19.31 105.45 397.71 69 na na na 19.33 111.02 426.07 85 20.92 10.34 20.92 19.32 112.49 456.86 98 20.79 39.06 55.09 19.31 118.49 470.70 199 20.85 10.85 38.24 19.31 129.32 583.35 296 20.96 47.75 87.25 19.30 138.40 667.92 395 20.98 57.71 115.91 19.29 142.02 699.98 494 21.01 102.31 157.95 19.29 155.07 767.49 597 21.04 95.87 152.07 19.30 157.96 793.36 697 21.03 111.26 169.29 19.29 163.70 855.19 797 21.03 111.61 161.35 19.30 162.24 824.16 903 21.05 99.86 153.45 19.30 158.13 781.36 996 21.04 117.28 168.29 19.30 166.48 812.36 200 Appendix 7b. Influence of ionic strength on the pKap of wild-type yeast iso-1 -ferricytochrome c and the variants Lys73 Ala and Lys79Ala. [KC1] (M) P*ap Wild-type Lys73Ala Lys79Ala 0.01 8.17(4) 8.59(3) 7.89(4) 7.87(1) 0.04 8.49(9) 8.69(2) 0.1 8.57(1) 8.80(1) 8.54(1) 0.2 8.72(2) 8.77(2) 0.4 8.70(2) 8.87(2) 0.5 8.82(2) 8.89(3) 0.6 8.74(3) 8.72(2) 1.0 8.71(2) 8.69(1) 9.05(2) 8.86(3) 201 Appendix 8. pH-Jump kinetics data of wild-type yeast iso-\ -ferricytochrome c and the variants Lys73Ala and Lys79AJa. Wild-type Lys73Ala Lys79Ala pH ^-obs Cs"1) pH ^obs (s-1) PH ^obs (s1) 6.07 0.038(2) 6.07 0.0128(7) 6.07 0.04(2) 6.13 0.0083(1) 6.20 0.0076(8) 7.57 0.0403(7) 7.07 0.0152(6) 7.07 0.0103(5) 8.09 0.0709(3) 7.57 0.043(5) 7.57 0.0158(2) 8.35 0.081(2) 8.09 0.0472(2) 7.80 0.014(1) 8.43 0.0852(3) 8.43 0.0617(3) 8.09 0.0206(1) 8.54 0.099(2) 8.54 0.0651(2) 8.35 0.02196(7) 8.76 0.1185(4) 8.76 0.0909(4) 8.43 0.0223(2) 8.86 0.1552(4) 8.86 0.1014(6) 8.54 0.02546(7) 8.94 0.170(4) 8.94 0.1075(6) 8.76 0.0323(1) 9.01 0.1615(6) 9.01 0.1337(5) 8.86 0.0366(1) 9.07 0.2148(7) 9.07 0.1434(7) 8.94 0.03804(9) 9.20 0.28(1) 9.20 0.2041(9) 9.01 0.0372(2) 9.28 0.244(1) 9.28 0.2122(1) 9.07 0.0451(1) 9.38 0.41(1) 9.38 0.3133(9) 9.20 0.0603(1) 9.50 0.379(2) 9.50 0.335(1) 9.28 0.0567(1) 9.56 0.541(1) 9.56 0.474(1) 9.38 0.0817(1) 9.70 0.552(3) 9.70 0.52(2) 9.50 0.0781(2) 9.80 0.99(2) 9.80 0.802(3) 9.56 0.1084(1) 9.89 0.860(5) 9.89 0.814(4) 9.70 9.80 9.89 0.1135(3) 0.1630(3) 0.1781(3) 202 

Cite

Citation Scheme:

    

Usage Statistics

Country Views Downloads
United States 24 0
China 16 28
Russia 7 0
New Zealand 5 0
Canada 2 0
City Views Downloads
Washington 10 0
Unknown 9 8
Shenzhen 8 28
Beijing 5 0
Penza 4 0
Ashburn 3 0
Vancouver 2 0
Mountain View 2 0
Houston 2 0
Nanjing 2 0
Matawan 2 0
Wilmington 1 0
Hanover 1 0

{[{ mDataHeader[type] }]} {[{ month[type] }]} {[{ tData[type] }]}
Download Stats

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:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0089257/manifest

Comment

Related Items