UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Peroxidase activity of yeast iso-1-cytochrome c Villegas, Jose 2006

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

Item Metadata


831-ubc_2006-201145.pdf [ 10.79MB ]
JSON: 831-1.0093001.json
JSON-LD: 831-1.0093001-ld.json
RDF/XML (Pretty): 831-1.0093001-rdf.xml
RDF/JSON: 831-1.0093001-rdf.json
Turtle: 831-1.0093001-turtle.txt
N-Triples: 831-1.0093001-rdf-ntriples.txt
Original Record: 831-1.0093001-source.json
Full Text

Full Text

PEROXIDASE ACTIVITY OF YEAST ISO-1-CYTOCHROME C  by Jose Villegas M . S c , National Autonomous University of Mexico, 1999  A THESIS SUBMITTED IN PARTIAL F U L F I L L M E N T OF THE REQUIREMENTS FOR THE D E G R E E OF  DOCTOR OF PHILOSOPHY In  The Faculty of Graduate Studies Biochemistry and Molecular Biology UNIVERSITY OF BRITISH C O L U M B I A  April 2006  © Jose Villegas, 2006  ABSTRACT  Understanding the peroxidase activity o f cytochrome c is essential for the development of potential applications o f this useful property in diagnostics, chemical synthesis and bioremediation. Such knowledge may also be relevant for understanding the chemical basis o f the physiological consequences o f this reaction under conditions of oxidative stress induced by toxic species generated by various cellular processes. In the present work, the thermodynamic, spectroscopic and catalytic properties o f two classes of cytochrome c variants have been studied in an effort to gain greater insight into the structural basis for the peroxidase activity o f the cytochrome. The ultimate goal o f this was to identify a strategy for the design or identification o f new cytochrome c variants with significantly enhanced peroxidase activity. Development o f one class o f these variants was based on structural insight provided by X-ray crystallography, and the resulting variants were produced by site-directed mutagenesis. The other class o f variants were obtained by a directed evolution method in which randomly-generated variants with elevated peroxidase activity were identified by the use o f an activity screen. The results obtained suggest that the changes in the peroxidase activity observed for each variant are related not only to an overall destabilization of the protein structure but also to localized variations around the prosthetic group that function in combination to regulate the kinetic behavior o f this protein. Such structural factors are discussed here in the context o f the thermodynamic, spectroscopic and catalytic properties o f each variant.  11  T A B L E OF CONTENTS  ABSTRACT  .ii  T A B L E OF CONTENTS  ,  iii  LIST OF T A B L E S  viii  LIST OF FIGURES  xi  DEDICATIONS  xii  1. INTRODUCTION I. Cytochrome c  1 ;  1  A. History  1  B. Classification  2  C. Structural characteristics of c-type cytochromes  6  D. Function  10  II. Catalytic properties of cytochrome c  13  A . Overview and applications  13  B. Peroxidase activity of cytochrome c  15  C. Influence of cytochrome structure on peroxidase activity  19  D. Spectroscopic methods for characterization of cytochrome c  19  E. Practical advantages of a cytochrome c-based catalyst  21  •ILL Design of cytochrome c variants A. Mutation sites TV. Directed evolution A. Experimental strategies to identify functionally useful structural variants  22 23 • .26 27  B. Activity screening strategies  .  30  C. Examples of activity enhancement through directed evolution  32  V. Limitations of oxidative catalysis  35  A. Catalytic inactivation and degradation products  .  B. Principal  into  challenges to the conversion of cytochrome c  catalyst  35 an  efficient 38  2. METHODS  40  I. Computer simulation methods  40  II. Molecular biology methods  40  A. Expression of yeast iso-1-cytochrome c  40  B. Production and purification of recombinant cytochrome c  41  C. Site-directed mutagenesis of cytochrome c  42  D. Random mutagenesis  43  E. Screening assay  44  HI. Spectroscopic methods  45  A. Electronic absorption spectroscopy  45  B. Spectropolarimetry  45  C. Magnetic circular dichroism  45  D. NMR spectroscopy  46  E. Electron spin resonance spectroscopy  46  IV. Kinetic methods  47  A. Steady-state kinetics 1. Determination of k i and k  47 3  47  iv  2. Inactivation constant  47  3. Partition ratio  ...48  4. Heme decay  49  V . Physical methods  49  A . Thermal denaturation curves  49  B. p K a value  50  C. Direct electrochemistry  50  V I . Protein analytical methods  51  A . Electrophoresis  51  B. Exogenous ligand binding to the heme  51  C. Trypsin digestion  52  ;  D . Mass spectrometry  52  3. R E S U L T S  :  I. Kinetic parameters o f cytochrome c variants  54 54  A . Peroxidase substrate and progress curve  54  B. Steady state kinetics analysis  57  C. k  59  i n a c t  and the partition ratio  II. Stability o f variants towards temperature, p H and  62  H2O2  III. Reduction potential and thermodynamic parameters  65  IV. Spectroscopic characteristics o f cytochrome c variants  69  A . Electronic absorption spectroscopy  '.  69  B. Spectropolarimetry  71  C. Magnetic circular dichroism  74  v  D. N M R spectroscopy  76  1. 1H N M R spectra..  76  2. 2D N O E S Y spectra  78  3. HM8  80  4. HM3  81  5. HP7-lb  82  6. HP7-la  82  7. HM5  83  8..HM1  83  V . Random mutagenesis  84  A. Mutagenic reaction  85  B. Screening methods  88  C. G84E variant  92  1. Kinetic parameters  92  2. Stability parameters  94  3. Reduction potential and thermodynamic parameters  '.  95  4. Spectroscopic characteristics of G84E  96  VI. The mechanism of cytochrome c inactivation by hydrogen peroxide  97  A. Heme bleaching  98  B. Exogenous ligand binding to the heme  100  C. Peroxide-induced oligomerization of fenicytochrome c  101  D. Identification of inactivation reaction products.  105  4. DISCUSSION  109  vi  I. Site-directed variants II. Random mutagenesis A. The G84E variant  UO :  U8 120  TJJ. Inactivation reaction  125  IV. Future directions  129  5. R E F E R E N C E S  1 3 1  vii  LIST OF T A B L E S  Table I. Variety o f catalytic activities displayed by cytochrome c  14  Table II. Catalytic parameters derived from the peroxidase ping pong analysis for cytochrome c variants  59  ;  Table HI. Inactivation constant for cytochrome c variants at three concentrations o f H2O2  60  Table TV. Partition ratio for cytochrome c variants calculated from the coefficient o f product oxidized versus the amount o f catalyst at three concentrations o f H2O2  61  Table V . Parameters related to stability o f ferricytochrome c variants  64  Table V I . Electrochemical properties o f wild-type iso-1 -cytochrome c and variants  68  Table VII. Results from random mutagenesis o f the cytochrome c gene  87  Table VIII. Rate constants k i and k constants obtained from steady-state kinetics analysis o f the 3  peroxidase activity o f wild-type ferricytochrome c and the G 8 4 E variant  93  Table I X . Inactivation and partition ratio comparison o f the G 8 4 E variant versus the W T protein  94  Table X . Parameters related to stability o f the randomly generated ferricytoclirome c  95  Table X I . Reduction potential and thermodynamic parameters for the G 8 4 E variant  96  Table XII. Electronic absorption spectroscopy o f the G 8 4 E variant  97  vm  LIST OF FIGURES  Figure 1. Examples o f various heme groups distributed among the cytochrome c family  4  ;  Figure 2. Ribbon view o f iso-1-cytochrome c  7  Figure 3. Role o f cytochrome c i n mitochondrial electron transport  11  Figure 4. Peroxidase cycle showing the path for compound I (CI) and compound II (CII) starting from the ferric form o f the enzyme  16  Figure 5. Intermediates considered for the kinetic analysis o f the peroxidase cycle  17  Figure 6. Mutation region i n (A) W T , (B) N52I/Y67F, (C) A 5 9 , and (D) N 5 2 I W 5 9 A / Y F 6 F 7 variants o f yeast iso-1-cytochrome c in the oxidized state  24  Figure 7. T w o different strategies for the in vitro evolution o f enzymes  28  Figure 8. Substrates used to test the peroxidase activity o f cytochrome c  55  Figure 9. Typical reaction curves for A B T S and dicarboxidine substrates  56  Figure 10. Calculation o f ki and k3 kinetic constants  58  Figure 11. Thermal denaturation curves o f cytochrome c variants  63  Figure 12. Cyclic voltammograms o f cytochrome c variants using a three electrode system with a modified gold electrode as working electrode  66  Figure 13. Temperature-dependent variations o f the midpoint redox potentital for cytochrome c variants  67  Figure 14. Electronic spectra o f cytochrome c variants  70  Figure 15. Far U V circular dichroism spectra o f ferricytochrome c variants Figure 16. Visible circular dichroism spectra o f cytochrome c variants  :.. .72 73  ix  Figure 17. Near IR magnetic circular dichroism spectra of the proteins studied in this work  75  Figure 18. Nomenclature conventions for the heme group  .76  '...  Figure 19. 1H N M R spectra of ferricytochrome c variants  77  Figure 20a. Up field region of the N O E S Y spectra of wild-type and variant ferricytochrome c  79  Figure 20b. Expanded view of Figure 20a showing the area around H M 8 and HM3 for each variant  80  Figure 21. Agarose gels (2%) to estimate the amount of D N A produced per cycle of mutagenic P C R (A) and the yield at different concentrations of M n C b (B)  86  Figure 22. Minimum amount of cytochrome c for which peroxidase activity was detectable with peroxidase substrates ABTS and H2O2  ;  88  Figure 23. Principal steps in the initial peroxidase activity screening strategy  89  Figure 24. Modified screening protocol that permitted detection of peroxidase activity in E coli cells expressing cytochrome c variants  90  Figure 25. P C R plate showing the positive reaction of variant G84E after repeated dilutions of the substrate solution, indicated by a dark green color absent in all the other tubes  92  Figure 26. Decrease in absorbance at the Soret maximum ("bleaching") following addition of H2O2 to ferricytochrome c  :.99  Figure 27. Change in Soret spectrum following incubation wild-type ferricytochrome or the double variant with K C N  100  Figure 28. Oligorimezation of cytochrome c in the presence of H 0 observed by SDS P A G E 2  2  electrophoresis  102  Figure 29. SDS P A G E electrophoresis of the products resulting from reaction of wild-type ferricytochrome c with various ratios of [H 02]/[ferricytochrome c] 2  103  x  Figure 30. Absorbance change versus time for the heme absorbance of WT cytochrome c (2uM) in the presence of I X , 10X and 100X excess (2 u M , 0.2 and 20 u M , respectively) of H 0 2  2  104  Figure 31. General strategy to identify protein modifications resulting from the reaction of ferricytochrome c with H2O2 in the presence of a spin trap by H P L C / M S peptide mapping  106  Figure 32. Reaction of cytochrome c with H C>2 in the presence of P O B N analyzed by 2  LC/MS  107  Figure 33a. Stereo drawing showing the region around HM8 in the WT and the variant N52I/W59A/Y67F  117  Figure 33b. Stereo drawing of the region around HM3 and HP7 in the WT and N52I/W59A/Y67F proteins  117  Figure 34. Stereo-drawing from the region around Argl3 in the WT protein  123  Figure 35. Structural changes produced by the G84E as estimated by molecular simulation  125  xi  To my famity: To my invafua6Ce parents, Juan Jose and Luz Maria To my belovedgrandmother, Ccrrmen To my dear brother, filejandro To my Coving wife, j\driana To my littte treasure, C^teste  xii  1. INTRODUCTION I. Cytochrome c A. History The first report describing cytochrome c was published i n the 1880s by Charles A . MacMunn,  a practicing physician. M a c M u n n  identified what  he referred  to as  "myohaematin," which he studied initially as a "pigment" distributed i n several animal tissues that he had analyzed by microspectroscopy, one o f his hobbies (Scott 1996). M a c M u n n ' s work was the target o f intense criticism by his contemporaries and was disregarded for many years. The primary limitation o f M a c M u n n ' s work was i n considering myohaematin to be a single substance rather than a combination o f compounds. The "rediscovery" o f cytochrome c and the acknowledgment o f M a c M u n n ' s results i n the m i d 1920s by David K e i l i n , an Oxford parasitologist who was initially interested i n the fate o f hemoglobin during the life cycle o f the fly Gasterophilus, is a good example o f serendipity (Bendall 2004). In his initial investigations o f fly tissues with a low-dispersion spectroscope, K e i l i n identified four characteristic bands i n the spectrum at 604, 566, 550 and 520 nm. B y analogy with other heme proteins already known, he assigned the first three to the a-bands o f three different components a, b and c, while the last one was ascribed as the combination o f the a-bands o f all o f them (Keilin 1966). In addition, by observing the presence or absence o f such bands i n the spectra o f fly muscle exposed to different oxygenation conditions, K e i l i n demonstrated that this protein undergoes reduction and oxidation in vivo. K e i l i n not only introduced the term cytochrome (curiously as a "temporary" designation) which means cellular pigment but also proposed a nomenclature system based on the wavelength o f maximal absorption for  1  the reduced proteins (Keilin 1966). He was also the first one to obtain a pure preparation of cytochrome c (around 1930) isolated from yeast. A t that time, this protein was normally recognized by the unusual position o f its a-band, suggesting a modification o f the heme group. Such modification was later identified by Theorell (Ehrenberg and Theorell 1955) as the product o f reaction between the a-carbon o f the vinyl substituents o f the prosthetic group and the thiol sulfurs o f two cysteine residues. These results i n turn, led to the later finding o f the conserved protein motif C y s - X - X - C y s - H i s for this class o f proteins. A s a result o f these pioneering experiments and subsequent work by many others, the proteins involved in oxidative phosphorylation (Ferguson 2001) and the role o f cytochrome c as the penultimate electron acceptor in this process have been defined so that the prominent role o f cytochrome c in the "respiratory mechanism o f the cell" described by K e i l i n (Keilin 1966) is now generally recognized.  B. Classification Cytochromes can be found not only in cells o f higher organisms, where they were originally discovered, but also in yeast and bacteria. In the case o f bacterial cytochromes, a greater variety o f structures and wider range o f cellular roles is apparent relative to the mitochondrial cytochromes (Ferguson 2001). The term cytochrome, as introduced by K e i l i n , encompasses  a group o f  hemeproteins that undergoes reversible oxidation-reduction reactions that are readily detected by distinct changes in the visible electronic absorption spectrum. More recently, a more specific but essentially equivalent definition describes a cytochrome as a  2  "hemeprotein  whose  characteristic mode  o f action involves transfer  o f reducing  equivalents associated with a reversible change i n oxidation state o f the prosthetic group" (Palmer and Reedijk 1991). A t present this term is applied to most intracellular electron transfer proteins that possess a heme group. Heme proteins responsible for dioxygen transport or heme enzymes (e.g., peroxidases, catalase, sulfite reductase) are not regarded to be cytochromes (Palmer and Reedijk 1991). Cytochromes are generally classified into four general groups and are referred to as cytochromes a, b, c and d, depending on the type (Figure 1) and binding characteristics of their heme group: a-Type Cytochromes: The members o f this group contain a molecule o f Heme A (Figure 1) and some examples include cytochrome acii, a membrane bound protein complex also referred to as cytochrome c oxidase, and cytochrome ci\ present i n Nitrobacter agilis. b-Type  Cytochromes: Cytochromes  in  this  group  possess  an  iron  protoporphyrin LX prosthetic group. This prosthetic group is not covalently attached to the protein, so the binding to the apo-cytochrome is mediated by the protein ligands to the heme iron, by hydrophobic interactions o f the heme with the apo-protein, and by H bonding interactions o f the heme propionate groups with the protein. Some examples include cytochrome b, cytochrome bs, cytochrome bsa, cytochrome 6559, and cytochrome &562-  c-Type Cytochromes: Cytochromes c possess one or more iron protoheme groups covalently bound.to the apo-protein by thioether bonds. These thioether bonds form between the heme vinyl substituents and Cys residues i n the heme binding site.  3  Heme C  Heme D  Figure 1. Examples of various heme groups distributed among the cytochrome c family.  4  In most species, these bonds are formed by two highly conserved Cys residues though Euglena and Crithidia cytochromes c are notable exceptions (Pettigrew et al. 1975) insofar as they each possess just one thioether bond to the heme group. The mitochondrial cytochrome c group includes cytochrome c and cytochrome c\.  The  former is the protein described by K e i l i n and moves freely i n the intermembrane space, while the latter is a membrane bound electron transfer protein. The wide diversity of ctype cytochromes has now been recognized, particularly in bacteria, and has led to the identification o f four classes (or types) o f cytochromes c (Moore 1990) as described below. Class I cytochromes c are small, soluble proteins o f high midpoint potential that usually contain a low-spin iron. The heme group is attached to the protein near the Nterminus while the sixth ligand, normally Met80, is situated near the C-terminus. Three main a-helices form a "basket" surrounding the heme group which is exposed to the solvent at the surface o f the protein along one o f its edges. Mitochondrial cytochrome c and cytochrome c are included in this class. 2  Class II ferricytochromes c contain primarily a high-spin iron, but in a few members o f this family (e.g., cytochrome C556) is low-spin. The heme group is also attached to the apo-protein near the JV-terminus, and the main protein fold consists o f a four a-helical bundle (Ambler 1991). Class III cytochromes c are normally high molecular weight proteins with lowspin iron and multiple heme groups o f a wide range o f midpoint potentials. Examples of this class include cytochrome c-j and C3.  5  Class I V cytochromes c are tetraheme proteins that normally function as subunits and that have heme groups with either bis-His or His-Met coordination. The typical example of this class is the subunit o f the photosynthetic reaction center. Cytochromes d: This group o f cytochromes is distinguished by the presence o f a partially reduced tetrapyrrole prosthetic group that possesses additional substituents not normally observed in proteins possessing protoheme LX-based prosthetic groups. This class o f prosthetic group is characterized by the presence o f one or two reduced pyrrole rings and they are referred to as chlorins or bacteriochlorins, respectively. Representative members o f this group o f cytochromes include cytochrome d, cytochrome bd and cytochrome cd\ (nitrite reductase).  C. Structural characteristics of c-type cytochromes Despite this structural variability, elements o f the secondary structure (e.g., surface exposed residues, components o f the hydrophobic core) as well as the tertiary structure o f c-type cytochromes are conserved (Borsari et al. 2000). The backbone structure o f cytochromes c consists primarily o f four to five a-helices, varying with species, that are connected through several loops that together enclose a covalently attached protoheme group at the active site (Figure 2). A s mentioned previously, this prosthetic group is attached to the protein covalently by thioether bonds to a highly conserved C X X C H motif although some examples o f C X X X X C H and C X X C K have been found (Stevens 2004). The iron atom coordinated by this cyclic tetrapyrrole is coordinated to two axial ligands provided by amino acid side chains that are typically methionine and histidine but  6  other amino acid residues can be involved depending on the species and type of cytochrome. Methionine can be substituted by another histidine in bacterial cytochromes, while histidine can be replaced by lysine in some cases (Stevens 2004).  Figure 2 . Ribbon view of yeast iso-l-cytochrome c. The heme group in the center (lateral view) is coordinated to the axial ligands, His 18 (right) and M e t 80 (left).  7  The biological function o f c-type cytochromes is to act i n electron transfer reactions in which the oxidation state o f the heme iron atom undergoes a reversible change in oxidation state between F e  + 2  and F e . +3  In view o f the similarities i n prosthetic group and axial ligands between various classes o f c-type cytochromes, the range i n midpoint reduction potential exhibited by this family o f proteins is striking. For bacterial cytochromes c, this range covers - 8 0 0 m V , from - 4 0 0 m V to +400 m V (Moore 1990; Lett and Guillemette 2002). For mitochondrial cytochrome c (Class III), the high potentials of+200 to +350 m V result from stabilization of the ferrous state by the methionine ligand, the hydrophobic nature o f the heme binding site and the low dielectric constant o f this environment, and the limited solvent accessibility o f the heme (Moore 1983; Churg and Warshel 1986; M a u k and Moore 1997; Battistuzzi et al. 2002). Besides the type o f axial ligands, additional factors that influence the potential of heme proteins include substituents on the porphyrin ring, degree o f exposure o f the heme group to the solvent, and the hydrophobic environment and electrostatic effects, including the interaction with the propionate groups (Moore 1983; Lett and Guillemette 2002; Paoli et al. 2002). The contributions o f specific amino acid residues to the midpoint potential has been explored through the systematic study o f variants prepared by chemical modification (Rees 1980; Wallace 1984; Wallace and Corthesy 1987; L i u et al. 1995), site-directed mutagenesis  (Komarpanicucci et al. 1992; Lett et al. 1996; Lett and  Guillemette 2002), semi-synthesis (Wallace and Proudfoot 1987; Wallace 1993) or a combination o f mutagenesis and semi-synthesis (Wallace and Rose 1983; Wallace et al.  8  1986) by means o f spectroscopic, electrochemical and kinetic methods (Moore 1983; Moore et al. 1984; Cutler et al. 1989; Moore 1990). Other conserved structural features among class I c-type cytochromes include the presence  o f interacting TV-terminal and C-terminal helices, contiguous hydrophobic  patches i n the C-terminus between position 94-98, conserved hydrogen bonds between the heme propionate groups and several amino acids around the heme group (including residues  Arg38, Tyr48, Trp59 and Thr78), and a large number o f lysyl  residues  distributed asymmetrically on the surface o f the protein in the region where the edge o f the heme prosthetic group is partially exposed to solvent. The positively-charged electrostatic surface created by these lysyl residues is generally believed to facilitate electrostatic  interaction with the  other  electron transfer  proteins with which  the  cytochrome reacts physiologically (Scott 1996). A s mentioned above, the functional roles of many individual amino acid residues have been studied by characterization o f modified cytochromes in which specific residues have been chemically modified or replaced by mutagenesis  or semi-synthesis. The bases for selecting some o f these  substitutions are readily apparent (e.g., the residues providing axial ligands to the heme iron, residues that anchor or interact with the heme group), but other residues have been selected for study on the basis o f several factors. B y its proximity to the heme binding site, by sequence analysis or selected by screening (e.g., Tyr67, which forms an H-bond with the axial ligand Met80 (Frauenhoff and Scott 1992)); Trp59, Arg38 and Tyr48 which participates in an internal H-bond network that involves a heme propionate (Schweingruber et al. 1977; Caffrey and Cusanovich 1993); A r g 38, Tyr48 and Trp59, which participate i n an internal H-bond network involving a heme propionate (Thurgood  9  et al. 1991); Asp52, Ile75, Thr78 and Phe82, all o f which have been implicated in the stability or electron transfer reaction o f the protein (Rafferty et al. 1996). Other residues like Lys73, Lys79, Lys86 and Lys87, have been evaluated with regard to the alkaline conformational change (Rosell et al. 1998); Cys 102, which allows dimerization o f the yeast protein (Zuniga and N a i l 1983); Met65, which relates to use o f semi-synthetic strategies with yeast cytochrome c (Schejter and A v i r a m 1970; Bren and Gray 1993); C y s l 4 , C y s l 7 , Leu34, Leu37, Trp59 and Met80 i n the expression and in vivo function (Hampsey et al. 1986); and replacement o f the (Q)-loop (residues 18-32) on the overall protein stability (Fetrow et al. 1997). Despite this substantial body o f work, the specific roles o f many residues in determining various functional properties o f cytochrome c remain to be evaluated rigorously.  D. Function Cytochrome c is an essential component o f the respiratory chain that facilitates electron flow between redox proteins during cellular respiration, either i n the periplasm o f various bacteria or i n the intermembrane space o f the mitochondrion. In mitochondria (Figure 3), cytochrome c transfers  electrons  from Complex III (a coenzyme Q-  cytochrome c oxidoreductase) to complex I V (cytochrome c oxidase) both located in the inner membrane, h i the intermembrane space, cytochrome c also interacts with sulfite oxidase and cytochrome bs (in animals) and with cytochrome b2 and cytochrome c peroxidase (in yeast).  10  e" f r o m Complex I and II Lactate  so  Pyruvate  | ^ QH  2  2  Q  Complex 311111  NADH  NAD  +  •*©:  lliP  Complex IV  0  2  + 4H  +  2H 0 2  H 0 2  '.M'/^r':'. l' , 7  H 0 2  2  + 2H  +  ' >  Outer Membrane  Intermembrane space  Inner membrane  Matrix  Figure 3. Role of cytochrome c in mitochondrial electron transport. The direction o f electron flow is indicated by broken arrows. Cytochrome c (c) transfers  electrons  between  complex III and I V and also interacts  with  cytochrome b5, sulfite oxidase (SO), cytochrome b2 (b2) and cytochrome c peroxidase ( C C P ) . Modified from (Moore 1990).  Despite extensive study over 80 years, the physiological roles of cytochrome c continue to be a source o f enquiry. A s mentioned above, although cytochrome c is recognized as a crucial component in electron transfer reactions, the detailed mechanism by which electron transfer between proteins occurs is a topic o f continuing study (Moore 1990; Rafferty et al. 1996; Stevens et al. 2004). Similarly, cytochrome c serves as a model for the investigation o f the mechanisms o f protein folding and thermal stability  11  (Caffrey and Cusanovich 1994; Liggins et al. 1994; Fetrow et al. 1997; Winkler et al. 1997; Lett et al. 1999; Liggins et al. 1999; M i l n e et al. 1999; Yamamoto et al. 2002; Bhuyan et al. 2005; Sherman 2005). More recently, cytochrome c has been recognized in many cell lines as a major participant in. apoptosis. In these cell lines, signaling pathways utilize the mitochondrial machinery to effect apoptosis by releasing various proteins including cytochrome c (Iverson and Orrenius 2004; Orrenius 2004). Once released from the mitochondrion, cytochrome c takes part in a complex that is referred to as the apoptosome. In turn, this complex triggers an activation cascade involving several caspases that eventually leads to characteristic morphological changes such as cytoplasmic blebbing, and condensation and fragmentation o f chromatin. Cytochrome c has  also been implicated in other cellular processes  under  conditions such as oxidative stress where the concentration o f oxidizing species, including  H2O2,  are significantly elevated. Under such conditions, cytochrome c can  catalyze the oxidation o f model membrane  structures  liposomes, very likely by various mechanisms,  such as phosphatidylcholine  and, thus, promote "site-specific"  mitochondrial lipid peroxidation during oxidative stress (Radi et al. 1991b). Cytochrome c has also been reported to act as a scavenger o f reactive oxygen species in neurons under conditions o f oxidative stress and has been proposed to function as a defense against such species (Atlante et al. 2000). Similarly, the reduction o f ferricytochrome c by superoxide has been suggested to serve an antioxidant function (Korshunov et al. 1999).  12  II. Catalytic properties of cytochrome c A. Overview and applications In recent years, realization that cytochrome c possesses a variety o f catalytic activities in vitro has led to increasing interest i n understanding the mechanisms by which these activities arise and in the potential use of cytochrome c as an industrial catalyst. The catalytic activity o f cytochrome was reported as early as 1962 i n rat kidney and liver (Flatmark 1962), but more recent examples were proposed to be involved i n membrane peroxidation (Florence 1985; Radi et al. 1991b). Several catalytic activities that normally involve reaction with  H2O2  have been attributed to cytochrome c and some examples are  listed in Table I. h i addition to these examples, cytochrome c can promote the nitration o f proteins by nitrite through a mechanism that involves the formation o f tyrosyl radicals and nitrogen dioxide (Castro et al. 2004), and it reacts with hypochlorous acid (Prutz et al. 2001), ethanol radicals (Anni and Israel 1999) and peroxynitrite (Cassina et al. 2000). In sum, these demonstrate the catalytic potential and diversity o f cytochrome c and suggest several potential applications especially for the peroxidase activity o f cytochrome c. These  possible  applications  provide  considerable  motivation  for  an  improved  understanding o f the mechanism and structural basis for the peroxidase activity of cytochrome c and for the development o f new cytochrome derivatives with enhanced peroxidase activity.  13  Table I . Catalytic activities displayed by cytochrome c  Substrate A B T S , 4 amino antipyrine and luminol Tyrosine m-aminophenol o-phenylendiamine .  Reaction  Reference  Oxidation  (Radi etal. 1991a)  Oxidation  (Chen et al. 2002)  Hydroxylation and dimerization  (Zhu etal. 1998)  Polymerization  (Zlmetal.  1998)  (Ono et al. 2001)  o-phenylendiamine  Oxidation  N,N-dimethylaniline  N-demethylation  (Fujitaef a/. 1994)  Oxidation  (Nantes etal. 1998)  Diphenylacetaldehyde and 3methylacetoacetone Diphenylacetaldehyde and 3methylacetoacetone Catecholamines  Peroxidation Oxidation  (Roseie^fl/. 1998)  Anthracene, pyrene, anthraquinone, etc Pinacyanol chloride  Oxidation  (Torres etal. 1995)  Oxidation  Dibenzothiophene  Oxidation  7,12-dimethylbenzanthrace, azulene and anthracene Pho sphatidylcho line  Oxidation  (Vazquez-Duhalt 1993) (Vazquez-Duhalt R 1993) (Tinoco 1998)  Peroxidation  (Radi et al. 1991b)  N-methylcarbazole and thioanisole Oleofines N,N,N',N'-tetramethyl-1,4phenylendiamine  Oxidation Epoxidation Oxidation  (Akasaka et al. 1993) (Zhou et al. 2004)  14  B. P e r o x i d a s e a c t i v i t y o f c y t o c h r o m e c The catalytic mechanism o f peroxidases has been studied extensively, establishing many general principles o f the mechanism (Abelskov et al. 1997; Banci 1997; Dunford 1999; Moffet et al. 2000; Nielsen et al. 2001). Nevertheless, detailed understanding o f the contributions o f amino acid residues i n the active site or some distance from the heme prosthetic group to regulating the kinetic properties o f such enzymes, as well as alternative pathways in the mechanism, remains a subject o f continuing investigation (Rasmussen et al. 1995; Ghibaudi and Laurenti 2003). Similarly, understanding o f the precise mechanisms by which peroxidases become inactivated during turnover remains poor. (Valderrama et al. 2002). The  general  peroxidase  catalytic  cycle (Figure  4)  involves a  series  of  intermolecular electron transfer reactions that start with the conversion o f the ferric form of the prosthetic group to a higher oxidation state, ferryl heme (Fe (IV)=0) to produce a cationic porphyrin radical intermediate referred to as compound I (CI). This two electron transfer reaction requires  H2O2  as a substrate and results i n the oxidation o f the protein  and the reduction o f the peroxide to water. Compound I can react with an electron donor, usually an aromatic moiety, to recover one electron and produce a second intermediate, compound II (CH), and a free radical on the donor molecule. To complete the cycle, CII reacts with a second electron donor to regenerate the Fe(III)-heme state (Banci 1997; Dunford 1999).  15  Figure 4. Peroxidase cycle showing the path for compound I (CI) and compound II (CH) formation starting from the ferric form of the enzyme. A H represents any electron donor molecule, while ' A H is the corresponding radical.  16  Several steady-state kinetic analyses o f this mechanism have been reported, and more than one possible peroxidase mechanism is recognized. Perhaps the most widely accepted analysis is the one developed by Dunford (Dunford 1999) i n which the peroxidase reaction is regarded as a special (irreversible) case o f a ping pong mechanism. This concept o f irreversibility implies that the rate o f enzyme-complex formation is much faster than the rate o f its breakdown. For this reason, when the peroxidase cycle is represented in the form o f a ping-pong mechanism (Figure 5), the products are shown to be formed as soon as the substrates are in contact with the enzyme because the enzymesubstrate complexes are so short lived that they are extremely difficult to observe. A s a consequence, the rate o f the reaction seems to lack an upper limit. In other words, the rate of the reaction w i l l increase with increased substrate concentrations until saturation or inactivation of enzyme occurs (Dunford 1999).  H 0 2  V 2  Native  H 0 2  AH  Compound I  A  AH  V  C o m p o u n d II  A  Native  Figure 5. Intermediates considered for the kinetic analysis of the peroxidase cycle (Dunford 1999).  17  The individual steps o f this mechanism are shown below (modified from (Dunford 1999):  Fe(III)-cytochrome + H 0 2  Compound I + A H  Compound II + A H  ^  3  2  —-—•Compound I + H 0  (1)  — — • Compound II + A»  (2)  2  •Fe(III)-cytochrome + A» + H 0 2  (3)  Where A H denotes a reducing substrate and A» is the corresponding free radical species. These steps are consistent with the overall reaction for this process shown below:  H 0 2  2  + 2AH = 2 H 0 + 2 A . 2  (4)  If the reaction rate (v) is defined in terms o f disappearance o f reducing substrate and ki is assumed to be much smaller that fo, the steady-state approximation simplifies to the following expression:  2rEnl  0MH O ]) 2  2  (fc[AH])  Therefore, under conditions that the starting concentration o f reducing reagent is held constant, a plot o f E / v versus 1/[H 0 ] allows calculation o f the second order rate 2  2  constants k\ and £3 ( M " sec" ) for the rates of compound I formation and compound II 1  1  reduction, respectively.  18  C. Influence of cytochrome structure on peroxidase activity Recent studies have correlated the structural stability o f ferricytochrome c with the catalytic activity o f the protein (Diederix et al. 2001; Diederix et al. 2002; Diederix et al. 2003) and the effect o f specific mutations on the thermal stability o f the protein (Berghuis et al. 1994; Lett et al. 1996). A major objective o f the current work was to identify and correlate changes i n the structure o f designed variants o f yeast iso-lcytochrome c that- exhibit varying peroxidase activities. For example, the N 5 2 I / Y 6 7 A variant is known to exhibit increased stability to thermal denaturation (Berghuis et al. 1994) while the W 5 9 A substitution was expected to destabilize the protein by disrupting an internal hydrogen bonding network (Caffrey and Cusanovich 1993; Black et al. 2001). These mutations may also directly influence the catalytic mechanism o f the peroxidase reaction because the N 5 2 I / Y 6 7 F double substitution has been found to delay the bleaching o f the heme group in the presence o f  H2O2  without fully preventing the  inactivation o f the protein (Villegas et al. 2000). O n the other hand, replacement o f Trp59 eliminates a likely inactivation target i n the catalytic mechanism by removing a residue prone to free radical formation. Finally, the N 5 2 I / Y 6 7 F / W 5 9 A triple variant was constructed to determine whether the combination o f these two opposing but desirable properties in a single protein would result in a more active and robust peroxidase activity.  D. Spectroscopic methods for characterization of cytochrome c Spectroscopic methods provide powerful tools for the study o f subtle structural changes i n chemically and genetically modified proteins. In the present work, several spectroscopic techniques were used to characterize the kinetic behavior o f the variants  19  under study and to study the structural and spectroscopic consequences o f substitutions introduced i n proximity to the heme prosthetic group. Specifically, U V and visible electronic absorption spectroscopy, circular dichroism ( C D ) , magnetic circular dichroism ( M C D ) and nuclear magnetic resonance ( N M R ) spectroscopies are established methods for the characterization o f heme proteins in general and cytochrome c i n particular. Some examples o f the application o f these methods to cytochrome c include the following: the study o f the conformation o f cytochrome c from various species and the effect of denaturants and p H by C D spectroscopy (Myer 1968a; M y e r 1968b; Vinograd and Zand 1968; Greenwood and W i l s o n 1971); the study o f the effect o f mutations involving axial ligands to the heme iron and other substitutions by electronic absorption spectroscopy (Pielak et al. 1986; Santucci and A s c o l i 1997; Zheng et al. 2000); evaluation o f the effects o f denaturants on the conformation o f the protein by C D spectroscopy (Tsaprailis et al. 1998); assessment o f the coordination state o f cytochrome c variants by M C D spectroscopy (Gadsby et al. 1987; Gadsby and Thomson 1990; Hawkins et al. 1994; Fedurco et al. 2004). In addition, an extensive use o f the N M R technique has been applied for the assignment o f the ' H - N M R resonances o f several heme methyl and ligand protons i n horse and yeast ferricytochrome c (Mcdonald and Phillips 1973; Keller and Wuthrich 1978; Moore and Williams 1984; Williams et al. 1985; Feng et al. 1990); the study o f the effects o f temperature and ionic strength on the structural dynamics o f the protein (Burns and Lamar 1979; Burns and Lamar 1981; Moore 1990; Turner and Williams 1993; Yamamoto 1996); characterization o f the influence o f mixed solvent systems on the conformational equilibria o f ferricytochrome c (Sivashankar and Mabrouk 1998; Sivakolundu et al. 2001; Sivakolundu 2003); the correlation o f mutation-induced  20  structural changes to cytochrome c with the electrochemical properties o f the protein (Moore et al. 1984; Cutler et al. 1989; Moore 1990); determination o f the structure of yeast wo-1-ferricytochrome c in solution (Gao et al. 1990; Banci et al. 1997), including determination o f the structure o f one alkaline ferricytochrome conformer (Assfalg et al. 2003); and the assessment o f the structural consequences  o f several amino acid  replacements (Gao et al. 1991; Moench et al. 1991; Thurgood et al. 1991; Davies et al. 1993).  E. Practical advantages of a cytochrome c-based catalyst In comparison with true peroxidase enzymes, the main disadvantage o f using cytochrome c to catalyze oxidative reactions is the low activity. The rate o f formation of CI for a typical peroxidase is between 10 -10 M ' V (Abelskov et al. 1997; Moffet et al. 5  7  1  2000), while for some cytochromes a second rate constant in the order o f 10-200 M ' V  1  has been reported (Prasad et al. 2002). O n the other hand, cytochrome c possesses several advantages relative to true peroxidases, the most obvious o f which is the covalent attachment  o f the  prosthetic  group  to  the  protein.  Although the physiological  advantage(s) o f a covalently-bound prosthetic group in this protein is not yet apparent (Stevens 2004), this characteristic endows cytochrome c with significant stability to thermal denaturation and with considerable tolerance to exposure to organic solvents and extremes  o f p H (Vazquez-Duhalt 1999). Notably, cytochrome c can promote  the  oxidation o f dibenzothiophene over the p H range o f 2 to 10 with maximum activity at pH~7 (Vazquez-Duhalt 1993). Similarly, cytochrome c can catalyze the oxidation o f pinacyanol chloride, a model peroxidase substrate, i n the presence o f varying proportions  21  of water miscible solvents with tetrahydrofuran  18% o f activity retained i n a 90% solution of  (Vazquez-Duhalt et al.  1993). Finally, the peroxidase activity of  cytochrome c can be observed over the temperature range o f 30 to 80 °C (VazquezDuhalt 1999). h i comparison, exposure o f the widely available and readily accessible plant and fungal peroxidases to even mild forms o f these environmental perturbations normally results i n the loss o f the heme group, which is bound only through non-covalent interactions ( N . B . some mammalian peroxidases also possess covalently bound heme prosthetic groups, but these are not abundant proteins and often exhibit other properties that render them challenging to study (i.e., membrane binding)). To exploit the structural advantages  o f cytochromes  and  overcome  their  catalytic limitations, a detailed  understanding o f the catalytic oxidative mechanism is required. The current work begins this undertaking by characterizing the spectroscopic, catalytic and thermodynamic consequences o f selected amino acid substitutions that are expected to increase the probability o f developing an effective peroxidase based on the cytochrome c structural scaffold.  III. Design of cytochrome c variants. h i this work, we used as a starting point recent studies that have correlated the stability o f cytochrome c with its peroxidase activity by demonstrating an increased catalytic activity upon unfolding o f the protein (Diederix et al. 2001; Diederix et al. 2002; Diederix et al. 2003)  a s  w e  U  a s  studies concerning the effects o f specific mutations  on the thermal stability o f the protein (Berghuis et al. 1994; Lett et al. 1996). B y  22  designing variants with opposing effects on the stability of the protein and investigating the resulting consequences on the peroxidase activity as well as oh the overall structure of the protein, new and useful insight was sought into the mechanism by which cytochrome catalyzes these type of reactions.  A. Mutation sites The amino acid residues substituted in the variants studied in the present work are involved in a complex hydrogen-bonding network (Figure 6A) that includes residues Thr78, Met80, the heme propionate groups and at least two buried water molecules (Louie and Brayer 1990). This internal hydrogen-bonding network has been identified as a crucial region of the protein with regard to some of the structural differences between oxidation states of the wild-type cytochrome (Berghuis and Brayer 1992; Lett et al. 1996) and the double variant N52I/Y67F (Berghuis et al. 1994). In the oxidized wild-type protein, reorientation of an internal water molecule (WAT 166) eliminates the hydrogen bond between Ile52 and Watl66 that occurs in the wild-type protein, and the hydrogenbond that normally forms between heme propionate A (or HP7) and Ile52. This latter interaction has been associated with an increase in the dynamics of a substructure of the protein upon oxidation (Berghuis and Brayer 1992). Tyr67 on the other hand, is located on the same side of the heme, and normally forms a hydrogei>bond with the Met80 axial ligand to the heme iron in the oxidized cytochrome (Berghuis and Brayer 1992) though the structural contributions of this tyrosyl residue are only partially understood (Berghuis etal. 1994).  23  Figure 6. Mutation region in (A) W T , (B) N52I/Y67F, (C) W59A, and (D) N52I/W59A/Y67F variants of yeast iso-l-cytochrome c in the oxidized state. In A - D , mutated residues are represented as ball and stick models while nearby residues, including the heme group are shown as stick models. The heme group, in thinner black lines, is facing the plane with the propionate groups 6 and 7 in the bottom. The residue number in each structure, starting from the top residue (Met 80) right in front o f the heme group, in a clock wise direction are: Thr 78; residue 52 ( A s n i n A and C , He in B and D ) ; residue 59 (Trp in A and B , A l a in C and D ) ; residue 67 (Tyr in A and C , Phe in B and D). The water molecule W A T 1 6 6 (present only in A and C ) is represented as a big sphere. Hydrogen bonds are shown as thick broken lines. A and B structures were taken from the P D B bank (I.D. 2 Y C C (Berghuis and Brayer 1992) and 1RCI (Berghuis et al. 1994) respectively), while C and D were simulated by introducing the mutation W 5 9 A in A and B respectively.  24  One o f the main structural changes to occur in the N 5 2 I / Y 6 7 F variant is the loss of Wat 166, which is caused by the compression o f an internal cavity by as much as 8 cubic Angstroms, rendering it too small to accommodate a water molecule (Bergliuis et al. 1994). This structural change results in the disruption o f the interaction o f W a t l 6 6 with residues Ile52 and Phe67, leaving only the hydrogen bond between Thr78 and the heme 6-propionate and the interaction between Trp59 and the heme 7-propionate (Figure 6B). A s the consequence o f these rearrangements, Met64 is displaced toward Leu68 to result in formation o f a new hydrogen bond. The other residue substituted in the present work is the sole tryptophanyl residue in the protein, Trp59. This phylogenetically conserved residue is located at the bottom of the heme crevice on the distal site. The principal roles proposed for this unique residue are the stabilization o f the hydrophobic core o f the protein by means o f the aromatic indole side chain and stabilization o f the heme prosthetic group through formation o f a hydrogen bond with the heme 7-propionate group. (Hampsey et al. 1986; Caffrey and Cusanovich 1993; Black et al. 2001). Because crystallographically determined structures are not available for the W 5 9 A and N 5 2 I / W 5 9 A / Y 6 7 F variants, the likely structural changes resulting from these substitutions were initially simulated by introducing the W 5 9 A substitution into the corresponding proteins ( W T and double variant respectively) using molecular graphics software (Swiss P D B viewer v.3.7b2 and D S ViewerPro v.5.0). Consequently, the discussion o f this structure represents an approximation o f the true structure. In these models, the overall protein backbone is conserved, and many of the interactions observed i n the parent proteins are also preserved. In both, variants, one o f the  25  more evident alterations is the lack o f a hydrogen bond between the side chain o f Trp59 and the oxygen o f H P 7 . A s mentioned previously, this hydrogen-bonding interaction has been proposed to help interaction o f the heme group with the protein (Caffrey and Cusanovich 1993; Black et al. 2001). The replacement o f Trp59 by an alanyl residue (Figures 6C and D ) potentially introduces a substantial cavity in the region o f the substitution. Despite the extreme structural modification introduced, this variant was expected to be functional i n part because several studies have shown that this structurally important residue can be replaced by several natural and synthetic amino acids with retention o f function (Schweingruber et al. 1978; Caffrey and Cusanovich 1993; Black et al. 2001). Examples o f residues used for replacement o f Trp59 include amino acids incapable o f forming hydrogen bonds with the heme propionates. In addition, the introduction o f an alanine close to the heme iron and propionate groups should impose less electrostatic disorder than the introduction o f any charged residues.  IV. Directed evolution One highly effective means o f identifying useful amino acid substitutions for the modification o f catalytic activities or the introduction o f catalytic activities into proteins that has emerged i n recent years is an approach often referred to as directed evolution (Stemmer 1994; Brakmann 2001; K o l k m a n and Stemmer 2001). Most simply, directed evolution involves the combined use o f random mutagenesis  and  high-throughput  screening to identify amino acid substitutions that produce proteins with desired properties  (Bornscheuer  1998;  Harayama  1998;  Iverson  and  Breaker  1998;  Schellenberger 1998; Valetti and Gilardi 2004; Hibbert et al. 2005). . A m o n g the  26  advantages o f this strategy is the fact that knowledge o f the protein structure is not required and detailed insight concerning the catalytic mechanism o f the protein is not required. In some cases, this approach has been reported to result in enzyme variants with activities several orders o f magnitude greater than that exhibited by the wild-type protein (vide infra). A s a result, directed evolution has at times proven to be a very powerful means o f "customizing" either an existing catalytic or binding activity or, less commonly, to introduce novel functions into proteins. A crucial point o f any artificial evolution strategy is the availability or design o f the selection or screening method to be used. A s well, efficient expression o f the protein i n a suitable host and an effective means o f random mutagenesis are also important.  A. E x p e r i m e n t a l  strategies  t o identify  functionally  useful  structural  variants  Several methods have been proposed for efficient generation o f large numbers o f variants o f a target protein, but most can be divided into two major groups that differ in the underlying principle involved (Bornscheuer 1998). One general approach involves creation o f sequential generations o f mutated genes, and the other approach involves the recombination in vitro o f homologous genes (Figure 7). These strategies are usually complementary to each other. The primary approach to mutant generation i n the first group is to introduce sequence variability by random mutagenesis. This method has often proven to be a useful means o f improving pre-existing activities o f proteins, particularly for situations in which homologous genes are not available or i n which genes are too small to afford effective in  27  vitro recombination. One o f the simplest and perhaps the most commonly used method of random mutagenesis is error-prone P C R (Cadwell 1991). This method is based on the controlled introduction of random errors by a D N A polymerase during a non-optimal PCR  i n the  presence  of elevated concentration o f divalent metals  ions, usually  Directed evolution by production of sequential generations  i  i  Error-prone  Parent gene ^  1  • i  n  "  n  °  1  n  i•  Selection  •  i  o  o ~ m  Improved variant :  Mutated genes (library)  •  m  f•  •••••'•••••••••••••••••••••••••••••••••••••••*••••  Further cycles  D i r e c t e d e v o l u t i o n b y in v i t r o r e c o m b i n a t i o n  1. D N A d i g e s t i o n  •  IS  2. R e c o m b i n a t i o n Pools of  a? o n  i  ni' 0  ' i r~m  rr~~i 3 ii—P~ a  :  o  i 3.  fragments  parent genes  » (U~Dcr  i un  In v i t r o r e c o m b i n e d  homologous  Selection  Improved variant  •  Further cycles  Figure  7.  T w o strategies for the  in vitro e v o l u t i o n o f e n z y m e s .  Modified from .  magnesium and/or manganese (Bornscheuer 1998).  28  Once the sequence variation has been achieved, the modified sequences  are  expressed, and the resulting proteins are screened for the desired activity or property to identify, effective mutations. A n extremely critical point i n this strategy is the level of mutagenesis or mutation frequency, which ultimately means the average number of mutations introduced per gene. Typically, advantageous  mutations are rare, while  deleterious mutations are more frequent, so it is important to ensure that the number of beneficial mutations is maximized. If the mutation frequency is too high, deleterious substitutions may overwhelm the effects o f beneficial substitutions so that the beneficial substitutions are not recognized. A t the same time, i f the frequency o f mutation is too low, the wild-type sequence w i l l dominate, and too few variants w i l l be generated for screening. Usually, a rate o f 1.5-5 mutations per gene is preferred (Cadwell and Joyce 1991; Chen and Arnold 1993) ; however, this number also depends on the capacity o f the screening method, so the mutation rate must be adjusted to suit the capabilities o f the screening or selection method (Kuchner and A r n o l d 1997). The principal means of controlling this rate is to evaluate a variety o f conditions, particularly the concentration of magnesium and manganese, because the activity and fidelity o f most o f the D N A polymerases can be affected by the concentration of these ions. The rate of errors introduced by the enzyme can also be affected by the proportions o f nucleotides employed during the amplification reaction. Furthermore, these proportions can be also varied to provide a "bias" toward specific types o f desired substitutions. The second major approach to efficient generation o f mutant genes is frequently referred to as D N A shuffling or molecular breeding (Stemmer 1994). For use of this  29  approach, either naturally occurring or engineered homologous genes are required. As depicted in Figure 7, random segments are created from a pool of parent genes usually by digestion with the enzyme Dnase I. These fragments are recombined in vitro as the result of sequence homology during a polymerase chain reaction (PCR) and are then subjected to a selection step that produces improved variants that are able to enter subsequent cycles of recombination. Several variations of this general strategy have been reported. For example, methods have been devised for the use of single stranded D N A (Kikuchi et al. 2000) or for the addition of a peptide tail to enhance the sequence space that can be explored (Matsuura et al. 1999). Despite the wide range of successful examples obtained with this second general strategy (Minshull and Stemmer 1999), some limitations remain such as the need for pre-existing groups or families of related D N A sequences and occasional difficulty in controlling the size and specificity of fragments produced by the Dnase I. The latter problem is a particular concern in the use of small genes for which advantageous changes may be adjacent to or very close to each other in the sequence and are less likely to be combined in different proteins.  B. Activity screening strategies As mentioned previously, the screening or selection method is a critically important component of any directed evolution strategy. The set up and complexity of such methods clearly depend on the enzymatic activity or functional property (e.g., improved thermal stability, greater tolerance of extreme pH, improved binding affinity, increased or decreased substrate specificity, enhanced tolerance of organic solvents) for which improvement is sought. The complexity of the assay is greater when the library of  30  enzymes cannot be assayed directly, for example i f they are intracellular or i f a coupled activity assay is required. The methods for screening and/or selecting clones from a pool o f random mutants can be divided into three general groups (Schellenberger 1998), each o f which has advantages and drawbacks. The first group is the standard screening method that has been mentioned previously. This method is the most flexible and may have the principal limitations o f cost and the number o f clones that can be analyzed, with a limit o f around 10 clones. The second group o f screening methods involves the expression and display 4  (e.g., on the surface o f a phage (phage display)) o f a large number o f variant proteins. This approach is frequently used to screen large populations o f proteins for specific binding recognition capability and, thus, is often used for screening antibodies. The main limitation o f this method is that it is oriented primarily to the optimization o f binding rather that catalytic capability. Finally, the third group o f strategies involves selection o f useful, transformed cultures rather than screening o f enzymes or extracts, h i this way, i  large populations, usually in the order o f 10 or bigger can be evaluated with a high level 5  of sensitivity because only advantageous conditions, even i f they are marginal, w i l l allow growth o f colonies expressing proteins with activity characteristics that permit growth under a variety o f conditions. The main limitation o f this strategy is that it succeeds only when the desired catalytic activity is linked to the survival o f the cell that is expressing the protein o f interest. Although the current screening and selection methods are frequently able to produce the desired result, they are nevertheless limited to evaluation o f small numbers o f variants relative to the immense number o f variants that could, i n principle, be created  31  (Iverson and Breaker 1998). In time, perhaps additional strategies w i l l be identified that eliminate this remaining barrier to creation o f "custom made" enzymatic activities. Numerical methods have been described in some reviews (Kuchner and A r n o l d 1997; Moore et al. 1997) to determine the number o f clones that must be screened to assure that an optimized combination o f mutations has been evaluated. The complexity of this problem is indicated by the estimate that even for a protein o f - 2 5 0 amino acids, approximately 5,000 single amino acid substitutions and 22 million double mutations are possible. In spite o f these discouraging numbers, many successful examples o f the use of current strategies have been reported even with libraries as small as 10  4  clones  (Schellenberger 1998). h i addition, when previous information suggests that particular areas o f a protein may be important for the desired property, random mutagenesis can be delimited by the use o f specific D N A primers towards such areas to create "smart libraries" that can also be recombined (Chica et al. 2005).  C. Examples df activity enhancement through directed evolution A t present, an ever increasing number o f successful applications o f directed evolution are being published. Examples concerning such work with oxidative enzymes are relatively recent and involve primarily cytochromes P450, biphenyl dioxygenases and peroxidases (Valetti and Gilardi 2004). One o f the main constraints in the use o f many of these oxidative enzymes, particularly at the industrial scale, is the requirement o f expensive cofactors as electron donors (Hibbert et al. 2005). In the case o f the hydroxylation o f naphthalene by cytochrome P450, this problem was approached by  32  evolving this enzyme not only to exhibit a 20-fold greater activity relative to the wildtype enzyme but to use H 2 O 2 as an electron donor rather than NADH (Joo et al. 1999). Another example of directed evolution involving this same family of enzyme involves cytochrome P450 from Bacillus  This protein possesses a  megaterium.  monooxygenase and a reductase domain in a single polypeptide and is able to hydroxylate fatty acids (between C12 and Cis), amines and alcohols but not alkanes. By a combination of random mutagenesis focused around the heme domain sequence and an automated screening protocol, a variant capable of hydroxylating the hydrocarbon octane was identified (Farinas et al. 2001). The thermal stabilities of cytochromes P450 have also been improved by this approach. For example, studies where the thermostability of a cytochrome P450 variant, also known as BM-3 peroxygenase 21B3, was increased by several rounds of random mutagenesis to produce a variant containing 15 additional substitutions that retained much of the original peroxygenase activity (Salazar et al. 2003). Examples of the use of directed evolution to modify the activity of biphenyl dioxygenases include the use of DNA shuffling on the genes from Pseudomonas pseudoalcaligenes  KF707 and Burkholderia  xenovorans  LB400 to modify substrate  specificity to include benzene and toluene (Kumamaru et al. 1998). Similarly, DNA shuffling of specific segments of genes from Burkholderia Rhodococcus  globerulus  P6 and Comamonas  testosteroni  xenovorans  LB400,  B-356 produced biphenyl  dioxygenase variants that unlike the original enzymes, displayed greater ability to transform the 2,2'-, 3,3'-, and 4,4'- dichlorobiphenyls and were capable of oxidizing the very persistent compound 2,6-dichlorobiphenyl (Barriault et al. 2002).  33  In some applications, substrate specificity broader than that o f the wild-type enzyme is desired. This goal was achieved by use o f directed evolution to produce a variant amine oxidase from Aspergillus methylbenzylamine.  niger with altered affinity for the substrate L-ot-  Following random mutagenesis  and the use o f a colorimetric  activity screen, a single variant with increased activity and selectivity towards the L enantiomer was identified (Alexeeva et al. 2002). Peroxidases have also been engineered by directed evolution, especially for improved stability. In the case o f the heme peroxidase from Coprinus cinereus, variants were generated by random and site-directed mutagenesis and screened for improved thermal stability and resistance to inactivation by H2O2. Recombination o f the favorable mutations obtained by both methods produced a variant containing seven mutations that displayed dramatic improvements i n thermal and oxidative stability (Cherry et al. 1999). In related studies, horseradish peroxidase expressed i n yeast was subjected to random mutagenesis and recombination. This approached was used to obtain a double variant with a three-fold greater half-life at 60 °C relative to that o f the wild-type protein. This variant also exhibited increased stability toward the addition o f S D S , N a C l arid urea as w e l l as to conditions o f extreme p H (Morawski et al. 2001). A final example o f another heme protein subjected to directed evolution is horse heart myoglobin. Through combined use o f PCR-based random mutagenesis  and a  colorimetric plate assay using the substrate A B T S , a triple variant with 25-fold increased activity was identified. Characterization o f the mechanistic basis for this increase in activity through use o f steady-state kinetics and stopped-flow spectroscopy established that this increased activity resulted primarily from an increase in formation o f the  34  myoglobin equivalent o f compound I (i.e., k\) as represented in the Dunford peroxidase ping-pong mechanism discussed previously (Wan et al. 1998).  V. Limitations of oxidative catalysis A. Catalytic inactivation and degradation products A s described above, a major limitation i n the industrial application o f peroxidases is the inactivation o f the enzyme by  H2O2  or any organic peroxide and the limited  understanding o f the structural and mechanistic basis for this inactivation. Exposure o f the enzyme to peroxide is, o f course, unavoidable because peroxide is a substrate. Rigorous understanding o f the catalytic cycle and its relationship to inactivation during turnover is complicated by the variation o f the mechanism from peroxidase to peroxidase and its variation with reaction conditions. Several inactivation pathways have been proposed, and they vary with the enzyme, the reaction products obtained and the kinetic behavior observed (Nicell and Wright 1997; Prasad et al. 2002; Valderrama et al. 2002; Diederix et al. 2003). Nevertheless, from the regular mechanism described i n Figure 4, it has been proposed for some peroxidases, the conversion o f the intermediate CII into compound III (CIII) which is also known as oxyperoxidase (Hiner et al. 2002). The reactivity o f the CfJI intermediate has led to the proposal that it can react further to produce oxidative damage to the protein moiety and thereby inactivate the enzyme by various pathways (Valderrama et al. 2002). Such pathways include the opening o f the tetrapyrrole heme ring to produce a biliverdin-like compound by oxidizing one o f the meso carbon bridges. This reaction results i n bleaching o f the Soret absorbance band and loss o f the heme iron  35  as reported for several peroxidases (Valderrama et al. 2002). In an alternative pathway, CHI could return to the resting Fe(III)-enzyme following the catalytic oxidation of an amino acid residue. A third and final option is that CIH may decay spontaneously to produce a peroxyl radical that can be converted to an even more reactive hydroxyl radical that is capable of modifying any susceptible protein residue. Most of these intermediates have not been observed in the reaction of ferricytochrome c with hydrogen peroxide except in the case of carboxymethylated horse hearth cytochrome c. In this cytochrome derivative, the sulfur atom of the Met80 axial ligand to the heme iron is carboxymethylated and no longer binds the heme iron. As a result, the reactivity with peroxide is sufficiently altered that a Cl-like intermediate has been identified by U V - V i s spectroscopy after addition of an excess of H 2 O 2 (Prasad et al. 2002). While such an intermediate has never been observed to form following similar treatment of native ferricytochrome c, extensive studies on the reaction of this protein with various hydroperoxides resulted in the identification of protein-based free radicals centered on tyrosine residues around the heme group (Barr et al. 1996; Qian et al. 2002). In addition, it has been shown by mass spectrometry and the use of the spin trap D B N B S , that a free radical originated on horse heart cytochrome c by its reaction with H2O2,  can be transferred to tyrosine and tryptophan residues of synthetic peptides  (Deterding et al. 1998). In accord with this report, cytochrome c and several other heme proteins, including myoglobin and hemoglobin, can be radioactively labeled by [ H]benzo[a]pyrene and [ H]17(3-estradiol after treatment with 3  3  H2O2  (Rice et al. 1983). In  the same work, gel electrophoresis experiments established that some heme proteins treated with peroxide can form covalently crosslinked dimers and polymers, thereby  36  suggesting additional reaction products for the peroxidative reaction of cytochrome c. In other work, the extent of bleaching of the absorbance o f the heme group o f different heme proteins after exposure to  H2O2  was shown to be proportional to the amount o f iron  released from the protein during the reaction (Harel et al. 1988). In agreement with this information, and due to the possible differences between the inactivation mechanism of peroxidases and cytochrome c during oxidative catalysis, a general inactivation mechanism can be proposed for the latter. In this model, the decrease in activity as a function o f time is assumed not to result from substrate depletion, and the oxidized product produced by the peroxidase activity is assumed to be stable (N.B.: a condition not met by many substrates). With these assumptions, the reaction progress curve at a specific concentration o f  H2O2  can be fitted to the single exponential function  shown below: P = Pa(1-et  /(inactt  ) + Pi  (6)  Where P is the concentration of product at time t, P a is the concentration of x  product generated during the assay, P\ is the initial concentration o f product and k  inact  the inactivation rate at a given  H2O2  is  concentration in sec" . 1  Usually with this parameter and the value o f the catalytic constant, it is possible to calculate a partition ratio:  Partition ratio =/c t//Cinact c a  (7)  This ratio, which has no dimensions, represents the number o f effective catalytic turnovers per inactivation event (Tudela et al. 1987). However, in the particular case of peroxidases and peroxidase-like catalysts like cytochrome c, this analysis cannot be  37  applied so that an alternative approach is necessary.  The partition ratio can also be  calculated from the plateau in the progress curve, which represents the maximum amount of substrate transformed under specific reaction conditions. If the moles of substrate transformed are divided by the moles of catalyst used in the reaction, this ratio will yield the moles of substrate transformed before the enzyme is completely inactivated. A similar analysis has been applied for the case of cytochrome c 550 (Diederix et al. 2001).  B . Principal challenges to the c o n v e r s i o n of c y t o c h r o m e c into an efficient catalyst The principal challenges to the use of ferricytoclrrome c as an oxidative enzyme are as follows: (a) the low catalytic activity of the cytochrome relative to a true peroxidase is problematic, particularly on the scale of commercial applications; (b) the instability of the protein to the presence of H2O2, a characteristic exhibited by any peroxidase catalyst, shortens the life expectancy of the catalyst; (c) recombinant expression of cytochrome c as an intracellular protein complicates development of activity screening strategies required for application of directed evolution; and (d) the limited understanding of the catalytic and inactivation mechanisms compromises structure-based design of a more active and robust peroxidase based on the cytochrome c scaffold. Based on this background information, the objectives of the current work were to (a) construct cytochrome c variants by site-directed mutagenesis to study the peroxidase activity of this protein, (b) characterize kinetically and spectroscopically the resulting variants to understand further the catalytic mechanism, (c) improve the peroxidase  38  a c t i v i t y o f c y t o c h r o m e c b y directed e v o l u t i o n , and (d) study the m e c h a n i s m b y w h i c h w i l d - t y p e and variant f o r m s o f c y t o c h r o m e c are inactivated d u r i n g p e r o x i d a t i v e turnover.  39  2. METHODS I. Computer simulation methods Structures previously reported (Louie and Brayer 1990; Berghuis et al. 1994) for wild-type yeast wo-1-cytochrome c and a double variant ( P D B accession numbers 2 Y C C and 1 C R I respectively) were used to simulate the structures o f the two other variants W 5 9 A and N 5 2 I / Y 6 7 F / W 5 9 A , with the programs Swiss P D B viewer v.3.7b2 (Guex and Peitsch 1997) and D S ViewerPro v.5.0 (Accelerys) running on a personal computer.  II. Molecular biology methods A. Expression of yeast /'so-1-cytochrome c Yeast /so-1-cytochrome c was expressed i n Escherichia coli with the expression system ( p B P C Y C (wt)/3) originally designed by Pollock (Pollock et al. 1998) and modified by Rosell (RoselLef al. 1998). In both systems, the cytochrome c gene ( C Y C 1 ) is co-expressed with the gene encoding cytochrome c heme lyase ( C Y C 3 ) , and both proteins are produced i n the cytoplasm. In the case o f p B P C Y C (wt)/3, the two genes were cloned i n the HindUl/Smal sites o f the vector p U C 1 8 , while i n Rosell's system (pBTRI), the NcoVHindlll  fragment from p B P C Y C (wt)/3 was cloned into the p G Y M  vector (Guillemette et al. 1991). In the case o f the latter system, attempts to improve protein production b y induction with I P T G did not produce any change i n the yield o f the protein. H i g h expression levels o f the recombinant proteins were obtained from several E coli strains, the most efficient being the C D 4 1 D E 3 strain (Miroux and Walker 1996). These cells produced on the order o f 1-8 mg o f protein per L o f culture depending on the variant being purified.  40  B. P r o d u c t i o n a n d p u r i f i c a t i o n o f r e c o m b i n a n t c y t o c h r o m e c. Cytochrome c variant proteins were produced according to previous protocols (Pollock et al. 1998) with minor modifications. In brief, 1.5 L o f Y T  a m p  media (15 g o f  bactotfyptone, 11.25 g o f yeast extract, and 5.6 g o f N a C l dissolved i n 1.5 L o f distilled water) were inoculated with 150 m L o f overnight culture, and incubated at 37 °C for 2436 hours i n 2 L Erlenmeyer flasks under intense agitation (200-250 rpm). In some cases, the expression o f the proteins was so efficient that the E. coli cell's were visibly red in colour. Cells were recovered by centrifugation at 6000 rpm i n a SS-34 Sorvall rotor for 20 minutes i n a R C - 5 B centrifuge, and the pellet was resuspended i n 20 m M sodium phosphate buffer, p H 7.2 (8.90 g o f N a H P 0 and 18.95 g o f N a H P 0 , p H was adjusted 2  with 1 M  NaOH  4  2  4  and volume completed to 1 L ) and washed several times with the same  buffer. To disrupt the cells, two different methods producing similar yields were employed, h i the first method, repeated freeze-thaw (4 °C) cycles were used i n the presence o f an enzyme cocktail that included 2.5 mg o f ribonuclease A (Sigma, R-4875), 5 mg o f deoxyribonuclease I (Sigma, D-5025), and 3 g/ml o f lysozyme (Sigma, L-6876) as well as 0.1 m M P M S F (Sigma P-7626) dissolved i n ethanol as a protease inhibitor. Alternatively, cells were lysed with a French pressure cell at -2,000 psi with at least two passes. The cell debris were removed by centrifugation at 13,000 rpm for 30 minutes in 50 m L tubes, using the same rotor and centrifuge described above (Sorvall). The colored supernatant fluid was recovered, and ammonium sulfate (Sigma, A4915) was added to a final concentration o f 326 g/L to precipitate some o f the contaminant proteins. After  41  centrifugation (as above), the supernatant fluid was dialyzed overnight against 20 m M sodium phosphate buffer pH 7.2. Cytochrome  c  was  oxidized  in  a  desalting  column by  the  use  of  NH4(Co[dipicolinate]2) (Mauk et al. 1979). After concentrating, the sample was loaded onto a Pharmacia Mono-S H R 10/10 column in an F P L C system equipped with two p500 pumps and an LC-500 chromatography controller. Various gradient conditions were used for purification depending on the identity of the variant cytochrome. For the variants W59A and N52I/Y67F/W59A variants, the final purification step was performed using a CM-Sepharose column at 4 °C to prevent protein degradation. The proteins were loaded in the column in 20 m M sodium phosphate buffer p H 7.2, washed with the same buffer containing 75 m M sodium chloride, and eluted with 300 m M sodium chloride. After elution, the proteins were exchanged in buffer without salt, concentrated, flash frozen in liquid nitrogen and stored. The identity of each variant was confirmed by the presence of a single component of the expected atomic mass in an electrospray mass spectrometry analysis.  C . S i t e - d i r e c t e d m u t a g e n e s i s o f c y t o c h r o m e c.  The  main mutagenic method employed to obtain site-directed variants of  cytochrome c was the procedure commercialized by Stratagene. In this protocol, a thermostable D N A polymerase, Pfu turbo™, produced a full length D N A copy of the plasmid of interest in the presence of mutagenic primers by means of an optimized PCR. The clones were transformed by a standard heat shock method using competent cells treated with a C a C l (Sigma C-8106) solution 0.1 M (2.94 g of C a C l - 2 H 0 dissolved in 2  2  2  42  200 ml of water). The desired clones were screened by colony PCR and identified by the presence of additional restriction sites introduced by silent mutations in the mutagenic primer.  D. R a n d o m m u t a g e n e s i s  Several methods were evaluated for the introduction of random mutations into the cytochrome c gene; however, the best results were obtained using the error-prone P C R method, using different concentrations of M g  2 +  in the presence of M n C l (Cadwell 1991). 2  A conventional P C R reaction was first used to estimate the number of doublings after a given number of PCR cycles. This program consisted of 20 cycles of a 1 min denaturation phase at 95 °C, 1 min of annealing at 50 °C and 1 min of extension at 72 °C. The starting concentration of plasmid D N A template (WT cytochrome c) was 20 pmoles/uL with 50 uL of a solution containing 0.3 u M of primers, 7 m M M g C l  2  (prepared from dilution of a 100 m M stock, 0.197 g in 10 ml of water), 0.5 m M M n C l  2  (dilution from a 1 M stock, 19.79 g in 100 ml of water), 1 m M DNTPs, 50 m M KC1 (dilution from a 1 M stock, 7.46 g in 100 ml of water), and 20 m M Tris buffer. From this reaction, 10 uL samples were taken every 4 cycles. After this experiment, it was determined that 10 cycles were required to obtain 10 D N A doublings; therefore, the mutagenic P C R employed similar reaction conditions using only 10 cycles. The resulting product was amplified again by conventional P C R and cloned into the vector pDrive  1M  with a PCR cloning system (Qiagen), to facilitate the  recovery of the blunt PCR product obtained by random mutagenesis. After this, a second  43  cloning step in the expression vector pBTRI allowed the expression of the variants for screening.  E. Screening assay The screening protocol that eventually allowed identification of several variants employed small scale liquid cultures (0.5-1 mL of Y T media) grown in multi-well (1 mL) plates (Axygen). Each plate was covered with a sealing mat (Axygen), which had been punctured in the area above each well to allow some oxygen transfer while preventing excessive evaporation during the incubation period. After sterilization, Y T  a m p  media (500 uL) was placed into each well with a  multichannel pipette and inoculated with an overnight culture (5 uL) that had been diluted to contain a small number of cells (-5-10 cells per well). These mini cultures were grown for -24 hrs at 37 °C until the cells were visibly red, and a secondary culture was obtained by inoculating a small volume from each well into fresh media in a new plate. These plates were incubated again at 37 °C and incubated overnight to preserve the original culture. The cells in the primary culture were lysed by adding 100 ul of "Bugbuster" protein extraction solution (Novagen) and assayed for peroxidase activity using ABTS (Sigma, A-1888) and H 0 (Sigma, HI 009) as substrates. To do this, 200 ul of a solution 2  2  containing 2.5 m M A B T S and 1.25 m M H 0 (0.027g of A B T S and 12.5 ul of a solution 2  of H 0 2  2  2  125 m M , dissolved in 200 ul of 20 m M sodium phosphate buffer pH 7.2) were  added directly to the well. Due to the limitations of this assay and the failure to observe further positive clones upon dilution of the screening solution, only 20 plates were  44  screened. Based on these results, it can be estimated that in the best case -1.8 X 10 colonies were screened. The cells with positive reaction were plated in agar plates, and at least 10 colonies from each plate were assayed individually in the liquid culture assay.  III. Spectroscopic methods A. Electronic absorption spectroscopy The UV-Visible spectra of each variant protein (-1 u M in 20 m M phosphate buffer p H 7.2) were collected with a quartz cuvette (1 cm pathlength) and a Varian Cary 4000 spectrophotometer.  Sample temperature  was maintained at 25 °C with a  thermostated circulating water bath.  B. Spectropolarimetry The far U V (185-255 nm) and visible (350-550 nm) C D spectra were recorded with a Jasco J-810 spectropolarimeter equipped with a PFD-425S Peltier device to control the temperature. For the far U V spectra, the protein solution (10 u M in phosphate buffer (20 m M , p H 7.2) was placed into a water-jacketed, cylindrical quartz cell (0.1 cm cell path length). For spectra in the visible region, protein solutions (1 u M in phosphate buffer (20 m M , pH 7.2)) were placed into a rectangular, quartz cuvette (1 cm cell path length). Each measurement is an average of at least 3 scans.  C. Magnetic circular dichroism For M C D spectroscopy, protein solutions (0.1 mM) were exchanged into deuterated phosphate buffer (20 m M pD 7.2) by centrifugal ultrafiltration with a  45  Millipore concentrator. The spectra were collected in a 1 m m cell path length cuvette and a Jasco J-730 spectropolarimeter adapted with an electromagnet (Alpha Magnetics) operated at 1 T. A t least 3 scans were averaged for each variant.  D. NMR spectroscopy Protein solutions for I D and 2D N M R spectroscopy were prepared by exchanging the protein into deuterated phosphate buffer by centrifugal ultrafiltration in an A m i c o n Centricon-10 centrifugal device. A t least 3 cycles o f dilution and concentration with deuterated buffer were performed to ensure that most o f the exchangeable N H protons were replaced by deuterium atoms. The final sample concentration was ~3 m M . Spectra were collected at 25 °C except in the temperature-dependent experiments for which the sample temperature was adjusted to 16, 25, 35 and 45 °C ( W T only). Spectra were collected with a Varian Unity 500 M H z instrument.  E. Electron spin resonance spectroscopy The X-band E P R spectrum o f wild-type ferricytochrome (250 u M i n 20 m M sodium phosphate buffer p H 7.2) incubated with peroxide (0.25 M i n the same buffer) was collected at 4 K on a E S P 300 E Bruker instrument (Villegas et al. 2000). The sample was flash frozen i n liquid nitrogen within the first 15 seconds o f reaction and the spectra were collected under the following parameters: frequency 9.45 G H z ; microwave power 0.5 m W ; receiver gain 1 x 10 ; modulation frequency 100 K H z ; modulation 5  amplitude 3.89 G .  46  IV. Kinetic methods A. Steady-state kinetics 1. Determination of k\ and fo. The peroxidase activities o f the W T and protein variants were measured by following the absorbance change at 440 nm from the oxidation o f dicarboxidine by H2O2. The oxidation product o f this peroxidase substrate is more stable to the excess o f H 0 2  2  than the corresponding product o f A B T S oxidation (Paul et al. 1982). Assays were performed at 25 °C on a Cary U V - V i s 4000 or 6000 spectrophotometer and 1 m l quartz cuvettes (1 cm path length) containing 1ml o f 0.5 u M protein solution i n 20 m M sodium phosphate buffer p H 7.2 and 250 u M dicarboxidine substrate solution (diluted from a 250 m M stock solution prepared by dissolving 0.012 g o f dicarboxidine dihydrochloride, Sigma D-5907, in 25 m l o f 20 m M phosphate buffer p H 7.2). Reactions were initiated by addition o f H 0 2  2  to a final concentration o f 10-50 m M . Triplicates o f each reaction were  performed, and the initial reaction rate was calculated using software provided with the spectrophotometer. A n extinction coefficient s  4 4 0  o f 13.4 m M "  1  cm" (Paul et al. 1982) 1  was used to calculate the amount of product oxidized. The rate constants k\ and £3 were evaluated using the equations describing the peroxidase ping-pong mechanism (Equations 1-3) postulated by Dunford (Dunford 1999).  2. Inactivation constant During  catalytic  turnover,  proteins  exhibiting peroxidase  activity undergo  spontaneous inactivation by a process that is generally assumed to involve oxidation of  47  one or more amino acid residues in or near the active site (Barr et al. 1996; Valderrama et al. 2002). To calculate the inactivation rate during the peroxidase assays, the reaction progress curve was monitored for each variant (25 °C) until the absorbance change reached a maximum; such time depended on the properties of each variant. In these experiments, protein (100 nM) and dicarboxidine substrate (250 mM) solutions in 20 m M sodium phosphate buffer pH 7.2, were placed in a quartz cuvette (1 cm pathlength) (25 °C), and the reaction was initiated by addition of  H2O2  solution to three different  concentrations: 40 m M , 60 m M and 80 m M . This curve was fitted to a mono-exponential function  (Equation 6) according to the model of Duggleby (Duggleby 1986). Curves  from three independent assays for each  H2O2  concentration, were averaged and fitted to  the above equation using the program Origin (ver 7.020, OriginLab Corp).  3. Partition ratio The frequency with which the peroxidatic catalytic cycle is completed relative to the frequency which the protein or enzyme is inactivated can be expressed in terms of a partition ratio (Tudela et al. 1987). For the present experiments, this value was calculated by dividing the maximum product concentration, obtained from the reaction progress curve, by the molar concentration of the catalyst. This quotient reflects the amount of product generated by each variant prior to inactivation. The reaction conditions were identical to those used for determination of steady-state kinetics parameters described above.  48  4. Heme decay Addition of  H2O2  to solutions of cytochrome c is known to cause a decrease in the  intensity of the electronic absorption spectrum of the protein that is often referred to as "bleaching" (Florence 1985) The kinetics of this decay of absorbance induced by the addition of hydrogen peroxide was determined for the proteins studied in the current work by monitoring the spectrum (250 to 750 nm) of a cytochrome solution (5 uM) in a quartz cuvette (1 cm path length) for at least 30 minutes following addition of a' 100-fold molar excess of H 0 . In these experiments, a Cary 300 spectrophotometer equipped with 2  2  a circulating, thermostatted water bath was used. The trace of the absorbance decay at 408 nm was fitted to a single exponential function curve using the program Origin (ver. 7.0, OriginLabs, Inc.).  V. Physical methods A. Thermal denaturation curves The thermal stability of wild-type and variant forms of yeast /so-1-cytochrome c was evaluated by monitoring the change in ellipticity at 220 nm with a Jasco J-810 spectropolarimeter equipped with a PFD-425S peltier device. The temperature (T ) at m  which half of the protein was denatured was determined by monitoring the ellipticity (220 nm) for protein (20 uM) in phosphate buffer (20 m M , pH 7.2) placed in quartz cuvettes (1 mm path length) while the temperature of the sample was increased from 25 to 80 °C at a rate of 1 °C per minute under computer control. T values were calculated with m  49  software provided by the instrument manufacturer (Jasco spectrum analysis) by obtaining the first derivative of the curve.  B. pK value a  The pAT for the alkaline conformational transition of wild-type and variant forms a  of ferricytochrome c was studied by monitoring the pH dependence of absorption at 695 nm of cytochrome solutions (~1 mM) following exchange by ultrafiltration into unbuffered 0.1 M NaCl. In these experiments, the solution pH was titrated from pH 5 to 11 by adding small volumes of 0.1 M NaOH or HC1 solutions (Rosell et al. 1998), monitoring the pH with a IQ240 ISFET Bench top pH meter (IQ Scientific instruments) equipped with a stainless steel micro pH probe. The resulting data were analyzed by fitting to a function for a single titratable group to obtain the pK  a  value of the  conformational transition.  C. Direct electrochemistry Cyclic voltammetry experiments were performed in a non-isothermal, twocompartment glass cell with a three-electrode system as described previously (Rafferty et al. 1990; Rafferty 1992; Rosell 1998) using a calomel reference electrode held at 25 °C and a platinum counter electrode. A gold electrode modified with a saturated solution of 4,4'-dithiodipyridine (Aldrich, 143057) was used as the working electrode, and the protein solution (500 uL, 0.4 mM) was prepared in a 100 m M KC1 solution pH 6.0. Voltammetry was performed over a potential range of +/- 250 mV at a 20 mV s" sweep 1  rate. The temperature was controlled by immersing the sample compartment in a jacketed beaker connected to a thermostatted, circulating water bath. Thermodynamic parameters  50  for the reduction of cytochrome c were obtained from the temperature-dependent variation of the midpoint redox potential of cyclic voltammograms between 5 and 35 °C.  VI. Protein analytical methods A. Electrophoresis Products of the reaction between ferricytochrome c and hydrogen peroxide were analyzed by SDS-PAGE gel electrophoresis (Laemmli 1970). A reaction mixture containing cytochrome c (20 uL, 32 uM) and various concentrations (10-, 100- and 1000fold molar excess relative to cytochrome c) of  H2O2  were incubated for 30 minutes in the  dark. In each case, half of the resulting reaction mixture was mixed with the appropriate buffer (with or without 2-mercaptoethanol), and loaded into the gel (15% acrylamide) and separated at 120 V for -90 min. The gel was stained with Coomassie blue solution (50% methanol, 10% acetic acid, 40% water and 0.05% Coomassie brilliant blue R-250, Biorad) for about 4 hrs under moderate shaking and destained for at least 2 h in an aqueous solution of 5% methanol, 7% acetic acid. Finally the gel was photographed using a gel documentation system.  B. Exogenous ligand binding to the heme The binding of the cyanide anion to ferricytoclrrome proteins was studied by monitoring the change in electronic spectrum as a function of time following the addition of cyanide. Protein solutions (5uM) in 20 m M sodium phosphate buffer pH 7.2, were incubated with lOmM K C N and spectra were acquired every 2 min over a 40 min period on a Cary 4000 spectrophotometer at 25 °C.  51  C. Trypsin digestion Cytochrome c was hydrolyzed using trypsin as previously reported (Mauk and Mauk 1988) with few modifications. Specifically, 0.5 mg of protein were dissolved in 250 ul of 0.5 M ammonium bicarbonate buffer and the volume was completed with distilled water up to 500 ul. TPCK-trypsin solution (6.25 ul of a solution prepared with 1 mg of enzyme in 1 ml of HC1 0.1 mM) was added and the reaction was incubated for 6 h. at 37 °C. After this period, a second equal amount of TPCK-trypsin was added, and the reaction was allowed to continue for 18 h more. Hydrolysis was stopped by addition of HC1 (1 M , 50 ul), and the reaction mixture was taken to dryness with a Speed-Vac vacuum centrifuge. Water (250 ul) was added to the dried hydrolysate, and the entire digestion and drying procedure was repeated. The resulting dried sample was dissolved in T F A (0.05% (v/v)) to a concentration of ~1 umol/uL, and the sample was submitted to H P L C / M S analysis.  D. Mass spectrometry The analysis of peptide fragments derived from cytochrome c following reaction of the protein with H 0 in the presence of the spin trap N-/ er/-Butyl-a-(4-pyridyl)nitrone l  2  2  N'-oxide (POBN, Aldrich, 215030) was based on the procedure described by Filosa (Filosa 2001). In the current study, 10 ul of a 2.5 m M solution of cytochrome c was mixed with 2.5 ul of 100 m M H 0 2  2  and 25 ul of 0.5 M P O B N solution in a 50 ul final  volume. A l l solutions were prepared in 20 m M sodium phosphate buffer pH 7.2. The reaction was stopped by injecting the reaction mixture onto an H P L C reverse-phase C4 column (Econosphere 300 4.5 mm in diameter and 250 mm in length from Alltech) 52  attached to a Beckman System Gold H P L C system or by diluting with 20 m M phosphate buffer pH7.2 and concentrating the reaction mixture by centrifugal ultrafiltration prior to loading it onto the C4 column. The reaction products were separated with a gradient from 0 to 60% acetonitrile in water (both with 0.05% TFA) in 60 min., followed by a steeper gradient of 60 to 100% acetonitrile in 20 min. in a C18 reverse phase column, and the fractions isolated in this manner were collected, dried and submitted for mass spectrometry analysis.  53  3.  RESULTS  I. Kinetic parameters of cytochrome c variants A. Peroxidase substrate and progress curve The first step towards the kinetic characterization o f the cytochrome c variants, was to select a suitable peroxidase substrate. For this purpose, several chemicals (Figure 8)  including  methoxyphenol  2,2'-azino-bis(3-ethylbenzothiazoline-6-sulfonate (Guaicol),  3,3'-dimethoxybenzidine  tetramethyl-l,4-phenylenediamine  (TMPD)  (ABTS),  (o-Dianisidine),  and  biphenylylenedioxy) dibutyric acid (Dicarboxidine) were tested  2-  N.N.N'.N 1  y,y'-(4,4'-Diamino-3,3'in terms o f solubility,  absorbance change and product stability. A B T S is usually the substrate o f choice to characterize this type o f catalysis, but it presents the disadvantage o f an oxidation product unstable to the excess o f hydrogen peroxide. On the other hand, in spite o f being less soluble and having a minor absorbance change upon oxidation compared to A B T S , dicarboxidine yields a product  stable  throughout the assay (Figure 9). This characteristic is very important for the analysis o f the reaction progress curve, as w i l l be discussed later, and therefore dicarboxidine was selected for subsequent experiments.  54  o  o  II HO-C-CH2-CH2-0  II  0-CH2-CH2-C-OH NH2  Dicarboxidine H3C—\  HsC  ,Crfe  s  N H3CQ (HCI)  tf-dianisidine  H3C'  N N  2  CH3  TMPD  Guaiacol  Figure 8. Substrates used to test the peroxidase activity of cytochrome c. A l l assays were performed in 20 m M phosphate buffer pH 7.2 using freshly prepared concentrated stocks solutions.  55  0  7  i  ABTS  'i  0  1  1  •  2  r  4  '-•  1  1  6  1  8  -•  1  10  Time (min)  Figure 9. Typical reaction curves for ABTS and dicarboxidine substrates Similar reaction conditions were used for each assay (100 n M cytochrome c, 250 u M substrate, 40 m M H2O2 i n 20 m M sodium phosphate buffer p H 7.2). The numbers represent the different phases o f the reaction (see text below).  A representative oxidation curve for dicarboxidine consisted o f three phases, as opposed to four when A B T S was used (Figure 9). In both curves, the first three phases are equivalent, with an initial relatively short linear segment (phase I), followed by a stage (phase II) where the reaction rate is lower compared to the initial rate and reaches a maximum (phase III). In the case of A B T S , the final stage (phase I V ) consisted i n a decline o f the absorbance, very likely due to product instability; this agrees with previous studies o f cytochrome c-550 (Diederix et al. 2001). However i n our case, no p'reactivation phase was observed in this curve with either A B T S or with dicarboxidine, even when different ratios o f protein versus  H2O2  and incubation times where used (data not  shown).  56  W i t h either dicarboxidine or A B T S as substrate, the initial rate o f the reaction increased with increasing concentration o f  and the rate depended linearly on the  H2O2  concentration of each cytochrome c protein. The initial rates were linear for at least 30 seconds (or longer depending on the concentration o f saturation was observed as the concentration of  H2O2  H2O2  and the variant) and no rate  was increased up to 100 m M (using  100 n M cytochrome c and 250 u M dicarboxidine).  B. Steady state kinetics analysis Based on the analysis proposed by Dunford for the peroxidase ping pong kinetics (Dunford 1999), the initial rate was plotted against the  H2O2  concentration (Figure 10),  and the following equation was fitted to the data:  k k 2rE ] 1  V  _  3  0  3  2  2  2  2  Where. [Eo] is the initial concentration o f enzyme; H2O2  (7)  [ H Q ] [Dc]  k [ D c ] + k! [ H 0 ]  [H2O2]  the concentration of  and [Dc] the initial concentration o f dicarboxidine. The fitting of this curve provided the two rate constants that characterize  peroxidases, k\ and fa. In this analysis k\ represents the rate o f formation of the intermediate Compound I (CI, Figure 4, p. 16) while £ 3 is a measure o f the rate of breakdown o f the second intermediate, Compound II (CH, Figure 4).  57  6.0x10"H  5.0x10' 7  0)  4.0x10"M  ns  a:  3.0x10"  2.0x10" H 7  1.0x10"  7  0.02  0.04  0.08  0.06  H 0 2  2  0.10  (M)  Figure 10. Calculation of k\ and £3 kinetic parameters. Conditions: 100 n M cytochrome c, 250 uM dicarboxidine, H 0-2 from 202  100 m M in 20 m M sodium phosphate buffer p H 7.2.  Each point is an  average o f at least three independent assays and the solid line represents the non-linear fitting o f the data to the rate equation with a R  2  value of  0.90.  A similar analysis was performed for all the variants, and the numerical results obtained in this manner are summarized in Table II. A s can be seen, considerable differences are evident when the two catalytic parameters are compared.  58  Table II. Catalytic parameters derived from the peroxidase ping pong analysis for cytochrome c variants.  (M'V )  h (M'V )  W T cytochrome c  20  6.5 x 10  3  W59A  130  3.3 x 10  4  N52I7Y67F  0.1  1.8 x 10°  N52I/W59A/Y67F  51  2.4 x 10  1  Protein  1  4  Conditions: 100 nM protein; 250 uM dicarboxidine; 20 to 100 mM H 0 in sodium phosphate buffer (20 mM, pH 7.2). The errors for both values were equal or minor to 5%. 2  2  The magnitude o f k\ for all proteins is considerably smaller than that o f £ , 3  especially in the case o f the N 5 2 I / Y 6 7 F variant. When this variant is compared to the W T protein, the low kj and k values reflect a very low rate o f formation o f the catalytic 3  intermediate C I and breakdown o f CII respectively. B y contrast, both the single and triple substituted variants have k i and k that are greater than that o f the W T protein. 3  C.  /Cjnact  and the partition ratio  To quantify the susceptibility o f the proteins to inactivation during catalytic turnover i n the peroxidase reaction, a rate o f inactivation,  &jact> n  was obtained from a  general inactivation analysis (Duggleby 1986). To rule out the contribution o f substrate depletion to the decrease in oxidation rate and therefore assure that such decrease was mainly caused by inactivation o f the protein, the reaction rate was calculated at the time  59  point were the enzyme is inactivated (maximum absorbance i n the progress curve) and it was shown to vary less than 10 % (~ 3%, data not shown). The resulting values for this parameter derived from numerical fits to the progress curves are shown i n Table III for all variants at three different concentrations of  H2O2.  Table III. Inactivation rate (s ) for cytochrome c variants at three concentrations of H 2 O 2 . 1  [H 0 ] 40 m M 2  2  WT  1.1 x 10"  W59A  [H2O2]  [H2O2]  60 m M  80 m M  1.7 x IO"  2.2 x IO"  5.7 x IO"  9.7 x IO"  1.4 x 10"  N52I/Y67F  1.1 x IO"  2.1 x IO"  2.3 xlO"  N52I/W59A/Y67F  2.3 x IO"  4.0 x 10"  5.3 x IO"  2  2  4  2  2  2  4  2  2  1  4  2  Calculated from the non-linear fit of the progress curve to a mono exponential function. The error was estimated to be equal or smaller than 10% with a R value of at least 0.91. Conditions: 100 n M protein; 250 u M dicarboxidine; H 0 40, 60 or 80 m M in sodium phosphate buffer (20 m M , pH 7.2). 2  2  2  The lowest value o f  &j  n a c t  at any concentration o f  H2O2  was obtained for the  double variant (1.1 to 2.3 x 10" s" ), indicating that this variant is the most resistant o f 4  1  these proteins to inactivation during catalytic turnover. W h e n comparing the rates at constant H 0 2  2  concentration (e.g., 60 m M ) , the value for the W T protein (1.7 x 10  s") is  increased 6-fold when the destabilizing substitution W 5 9 A . ( 9 . 7 x 10~ s" ) is introduced. 2  1  If the stabilizing substitutions N 5 2 I / Y 6 7 F are added to the destabilizing W 5 9 A mutation  60  iii the triple variant, the protein is still inactivated about twice as fast (4 x 10~ s" ) as the 2  1  W T protein but about half as rapidly as the W 5 9 A single variant. A partition ratio that provides a measure o f the number o f catalytic turnovers per inactivation event can be derived by dividing the total amount (moles) o f product oxidized  (obtained from the progress curve) by the amount o f catalyst (moles) used in  the assay. For the proteins studied in this work, the partition ratio varies slightly with H2O2  concentration (Table TV) however the differences do not appear to be significant  and essentially this value is independent o f peroxide concentration.  Table IV. Partition ratio for cytochrome c variants calculated from the coefficient of product oxidized versus the amount of catalyst at three concentrations of H 2 O 2 . [H 0 ] 40 m M  60 m M  [H 0 ] 80 m M  WT  178  155  150  W59A  385  387  389  N52I/Y67F  80  82  67  N52I/W59A/Y67F  186  171  160  2  [H2O2]  2  2  2  Conditions: 100 nM protein; 250 uM dicarboxidine; H 0 40, 60 or 80 mM in sodium phosphate buffer (20 mM, pH 7.2). The estimated error calculated from at least 5 independent experiments for each value was estimated to be equal or smaller than 6%. 2  2  Comparison o f partition ratios at constant peroxide concentration (e.g., 60 m M ) demonstrates that the double variant turns over the least number o f times (82) prior to inactivation while the single variant turns over the greatest number o f times (387) prior to inactivation. Interestingly, combining the stabilizing N 5 2 I / Y 6 7 F substitutions with the  61  destabilizing W 5 9 A substitution results in a turnover number (171) similar to that of the wild-type protein (155). Under these conditions, the most destabilizing substitution ( W 5 9 A ) produces a partition ratio more than double that o f the wild-type.  II. Stability of variants towards temperature, pH and H 0 2  2  To correlate the kinetic characteristics o f the variants with their structural properties, various parameters related to the stabilities o f the ferricytochromes were evaluated from three perspectives. First, the thermal stability o f each ferricytochrome was determined by spectropolarimetry (Figure 11) and represented by the T , the temperature m  at which half o f the protein is denatured. Second, the pAT for the alkaline conformational a  change o f each ferricytochrome was determined by spectrophotometric pH-titration (Greenwood and W i l s o n 1971). Finally, the percentage incubation with H2O2  o f the heme bleached by  (1:100 ratio) represents a measure o f oxidative stability. The  results o f these experiments are shown in Table V . From these results, it can be seen that the N 5 2 I / Y 6 7 F substitutions increase the T  m  by 11 °C i n agreement with a previous report (Berghuis et al. 1994). O n the other hand, the W 5 9 A substitution decreases the T  m  by 22 °C. Combining this destabilizing  substitution in a variant with the stabilizing N 5 2 I / Y 6 7 F substitutions results i n only partial restoration o f thermal stability as indicated by the observation that the T  m  for the  triple variant is ~18 °C lower than that o f the wild-type protein.  62  WT W59 N52I/Y67F N52I/Y67F/W59A  40  /  60  80  Temperature (°G)  Figure 11. Thermal denaturation curves of cytochrome c variants. Conditions: 20 u M protein on 20 m M sodium phosphate buffer pH 7.2. To facilitate comparison, the curves were normalized using the Jasco instrument software (Spectra analysis V . 1.53.04).  63  Table V . Parameters related to stability of ferricytochrome c variants.  T  Heme bleached  (°Q  (%)  m  a  WT  59.2 (0.8)  7.8 (3.8)  62 (9)  W59A  36.6 (1.0)  n.d.  78 (3)  N52I/Y67F  70.5 (0.7)  11.0 (2.7)  2 (6)  N52I/W59A/Y67F  40.8 (1.2)  8.4 (3.6)  74 (3)  Conditions used for each type of measurement are described in the Methods section. Calculated errors (%) are indicated in parenthesis and were calculated from the nonlinear fit (7m and p/C values) and from the results of 5 independent experiments (% of heme bleached) This value represents the percentage of decrease in intensity of the Soret maximum after incubation with 100-fold excess of H 0 for 30 minutes. a  a  2  2  A s thermal stability is often correlated with susceptibility o f the ferricytochrome to undergo the alkaline conformational transition (Berghuis et al. 1994) (exchange o f the Met80 axial ligand for Lys73 or Lys79 (Pollock et al. 1998; Rosell et al. 1998), the piC for this transition has  been  determined  by spectrophotometric  a  pH-titrations with  disappearance o f the absorption maximum at 695 nm that is characteristic o f native ferricytochrome c is monitored as p H is increased. A s expected (Berghuis et al. 1994), the pKa obtained for the N52I/Y67F double variant is significantly greater than that of the W T protein. O n the other hand, the value exhibited by the W 5 9 A variant could not be determined because the absorbance at 695 nm could not be detected at the p H values that were tried. Combining these stabilizing and destabilizing substitutions in one protein (the triple variant) restored the alkaline pK to a value very similar to that o f the wild-type a  protein.  64  The susceptibility o f the heme in each o f these proteins to oxidation by  H2O2  was  measured as the percentage decrease i n the Soret absorbance that resulted following incubation o f the protein with peroxide. A s seen, the W 5 9 A substitution significantly increases the vulnerability o f heme to destruction by peroxide as reflected by a decrease in the Soret absorbance by as much as 75%. Interestingly, a nearly identical result was obtained for the N 5 2 I / W 5 9 A / Y 6 7 F triple variant even though the N 5 2 I / Y 6 7 F double variant was more resistant to peroxide than was the wild-type ferricytochrome.  III. Reduction potential and thermodynamic parameters To investigate further the structural stability o f each variant, the reduction potential (Figure 12) and its dependence  on temperature and p H (Figure 13) were  determined for each o f the proteins studied in this work. From these results, the standard thermodynamic parameters A G , A H and AS were obtained (Table V I ) . These values are very useful to compare the effect o f mutations on the structure o f proteins, especially i f they have similar fold (Rafferty et al. 1990; Bixler et al. 1992; Komarpanicucci et al. 1992; Battistuzzi et al. 2001a).  65  ,~  N52I/Y67F/W59A N52I/Y67F WT  N  \ \  :'t  V  i -0.30  /  i  1  -0.15  x  —  0  r  ~ 0.15  0.30  Potential (Volts) Figure 12. Cyclic voltammograms of cytochrome c variants using a three electrode system with a modified gold electrode as working electrode. Measurements were performed at 25 °C over a potential range of 300 to +350 mV at 20 mVs" sweep rate. 0.4 m M protein solution were 1  prepared in 100 m M KC1 pH 6.0.  28(h  WT N52I/Y67F/W59A N52I/Y67F  260H  24(H  Figure 13. Temperature-dependent variation of the midpoint redox potential for cytochrome c variants. Conditions: 0.4 m M protein in 100 m M K C 1 p H 6.0. Temperature was varied from 5 to 35 °C. The data for each protein was fitted to a linear model with R value of at least 0.90 and PO.0001.  The reduction potentials o f double variant (218 m V vs. N H E ) and the triple variant (227 m V ) were both significantly lower than the potential o f the wild-type cytochrome (266 m V ) . A similar value has been reported previously for the double variant  (Berghuis  et. al.  1994).  Despite  repeated  attempts,  acceptable  cyclic  voltammograms could not be obtained for the W 5 9 A variant, presumably because this protein was not sufficiently stable to the conditions o f the measurement or as the result o f ineffective interaction o f the protein with the electrode surface. The greater changes in thermodynamic parameters were observed for the double variant, which exhibited an increase in the standard enthalpy o f ~4 kcal/mol and an increase in standard entropy o f - 1 0 eu. The less negative value of A H  observed for N 5 2 I / Y 6 7 F is consistent with a  0  significant increase i n stability (Berghuis et al. 1994; Terui et al. 2003).  Table V I . Electrochemical properties of wild-type /so-l-cytochrome c and variants. "Versus NHE, 298.15 K, 0.1 M KC1. Uncertainties: E°, 0.8%; AG°,1.9%; AS 5%; and AH , 5%. E° TAS/F -AH/F AH AG AS (mV) ( m V ) (kcal/mol) (eu) (kcal/mol) (mV) 0  0  0  0  0  a  WT  266  -6.1  -20.1  -12.1  +524.3  -258.6  W59A  n.d.  n.d  n.d.  n.d.  n.d.  n.d.  N52I/Y67F  218  -5.0  -10.2  -8.0  +348.2  -130.6  N52I/W59A/Y67F  227  -5.2  -15.5  -9.8  +426.0  -199.0  68  Previous authors have interpreted the thermodynamic parameters by factoring the contributions o f the enthalpic and entropic terms toward the midpoint potential i n each variant. When these values are compared between similar proteins (e.g. variants from a native protein) and significant changes are observed, this allows to determine whether the changes in the potential are mainly due to enthalpic or entropic effects (Battistuzzi et al. 2001b; Battistuzzi et al. 2002). A s observed i n Table V I , the contribution o f both terms changed to a considerable extent, especially when the W T protein is compared to the double variant in terms o f both, the contribution o f enthalpic and the entropic terms. The same is true for the comparison with N 5 2 I / W 5 9 A / Y 6 7 F , but the main difference is observed in the value o f the enthalpic term. These data suggest that the reduction in the potential o f N 5 2 I / Y 6 7 F is due to a contribution o f both the enthalpic and entropic terms, while the decrease in the value o f N 5 2 I / W 5 9 A / Y 6 7 is mainly enthalpic and may be produced by stabilization o f the oxidized form o f the protein (Battistuzzi et al. 2002). Finally for the case o f the tryptophan variant, the redox potential could not be determined in spite o f various attempts.  IV. Spectroscopic characteristics of cytochrome c variants A. Electronic absorption spectroscopy In an attempt to identify structural characteristics related to the differences in activity and stability exhibited by these cytochromes, these proteins were studied by a variety o f spectroscopic methods. For example, the electronic absorption spectra o f the ferricytochromes are shown in Figure 14.  69  0.3  I  '  240  1  '  1  260  r  280  •  500  '  525  300  320  ,  550  380  I  575  '  660  400  •  420  '  690  -  440  '  720  •  ->  750  X (nm)  Figure 14. Electronic spectra of cytochrome c variants (sodium phosphate buffer (20 m M ) p H 7.2, 25 °C). A varying ordinate scale was used for each region o f the spectrum to assist comparison o f the spectra. A l l spectra were normalized to the same absorbance intensity at the Soret maximum (12B).  70  The characteristic Soret band for heme proteins occurs at - 4 0 8 nm and is observed at this position for all the variants (12B). Nevertheless, several differences between these spectra (12A and 12C) are apparent, particularly i n the region o f the a and P bands (500-575 nm (12C). These latter bands are particularly sensitive to the environment around the heme group (Pollock et al. 1998). A s mentioned earlier, the main difference between these spectra observed i n the near-infrared region is the absence o f the 695 nm charge-transfer band in the spectrum o f the W 5 9 A variant, which is generally interpreted as indicating the absence o f the M e t 80-iron interaction.  B. Spectropolarimetry Comparison o f the far U V and visible C D spectra o f the ferricytochrome variants reveals differences from which consequences o f the mutations on the secondary and tertiary structures the proteins as well as modification o f the heme environment can be inferred. For example, the far U V C D spectra o f the variants reveals the two negative Cotton effects  (-210 and - 2 2 2 nm) and one positive Cotton effect (-195 nm) that are  characteristic o f the a-helix-rich secondary structure o f cytochrome c (Myer 1968c) (Figure 15). Notably, the far U V C D spectrum o f the W 5 9 A variant does not exhibit the negative Cotton effects, consistent with significant change i n the secondary structure o f this variant.  71  8.0  Figure 15. Far U V circular dichroism spectra of ferricytochrome c variants. 10 u M protein solutions were prepared in 20 m M sodium phosphate buffer p H 7.2 and spectra were measured at 25 °C on a 185-255 nm wavelength. The protein concentration was the same for each sample  The negative Soret Cotton effect present in the C D spectrum o f ferricytochrome c provides a sensitive indicator o f the integrity of the heme environment (Fetrow et al. 1997) in that this feature is less negative when the heme crevice is more open (Pielak et al. 1986; Chottard et al. 1987). The spectra o f the wild-type and N 5 2 I / Y 6 7 F and N52I/W59A/Y67F-variants exhibited a positive Cotton effect at -405 nm and a negative Cotton effect at -415 nm (Figure 16). This negative feature had different relative  72  intensities being less negative for the W T band, followed by N 5 2 I / W 5 9 A / Y 6 7 F and N 5 2 I / Y 6 7 F with the most negative value. However, the feature undergoes the greatest change in the spectrum of the W 5 9 A variant in which this feature has been replaced by a single, positive Cotton effect with a maximum at - 4 0 7 nm. Similar spectroscopic changes have been reported in the C D spectrum of ferricytochrome c that has been partially unfolded by the action o f a denaturing agent (Myer 1968b; Fetrow et al. 1997; Santucci and A s c o l i 1997), suggesting a similar structural state for this variant without addition of a denaturant.  -  24  • -  • W59A WT N52I/Y67F N52I/Y67F/A59  \  /  O) T3  E  04  CD  2A  4A  350  400  450  500  A, (nm) Figure 16. Visible circular dichroism spectra of cytochrome c variants. 1 u M protein solutions in 20 m M sodium phosphate buffer p H 7.2 at 25 °C  were scanned on a wavelength between 350-550 nm. (sodium  phosphate buffer (20 m M ) p H 7.2, 25 °C).  73  C. Magnetic circular dichroism The near-infrared M C D spectra o f low-spin, six-coordinate ferriheme proteins provide a good indication o f the Fe(III) coordination environments Thomson  1990; Hawkins et al.  1994). The near-infrared  (Gadsby and  M C D spectra  for  the  ferricytochromes studied i n the current work are shown in Figure 17. The N 5 2 I / Y 6 7 F and N 5 2 I / W 5 9 A / Y 6 7 F variants, as well as the W T protein, displayed a maximum at -1700 nm  that is consistent  with Met-His  axial  coordination o f the  heme iron.  The  corresponding spectrum o f the W 5 9 A variant, however, exhibited a peak at -1500 nm and generally observed for heme proteins and model compounds with o/s-nitrogen coordination (Gadsby and Thomson 1990; Fetrow et al. 1997; Fedurco et al. 2004).  74  1200  1400  1600  1800  2000  A, (nm)  Figure 17. Near I R magnetic circular dichroism spectra of the proteins studied in this work. 0.1 M protein solutions were exchanged i n duterated 20 m M sodium phosphate buffer p D 7.2 and measured on a wavelength between 1300 and 2000 rim using an electromagnet operated at 1. T. obtained in deuterated phosphate buffer (20 m M , p H 7.2, 25 °C).  D. NMR spectroscopy 1. H NMR spectra 1  Another highly sensitive means of assessing the heme environment, especially for the region around the ligands and the propionate groups, is the ' H N M R spectrum. The chemical shifts observed in this type of experiment have been thoroughly characterized by others and can be used to assess some of the changes introduced by the mutations in this region of the protein (Moore and Williams 1980; Cutler et al. 1989; Gao et al. 1990; Gao et al. 1991b; Thurgood et al. 1991). In Figure 18, the nomenclature convention for the heme group is shown (Nc-Iub 1991) and it w i l l be used further for the description of the N M R results. The I D ' H N M R spectrum for each of the ferricytochromes studied here is shown in Figure 19. In the particular case of the W 5 9 A mutation (Figure 19, last trace), very broad signals were obtained that complicated the assignment of resonances.  Figure 18. Nomenclature conventions for the heme group. The numbers denote the substituents in each pyrrol ring indicated by the letters.  76  35  30  25  20  15  10  5 (ppm)  Figure 19. H N M R spectra of ferricytochrome c variants. (20 mM. 1  phosphate in D 0 , pD 7.2, 25 °C. (A) wild-type cytochrome c; (B) the N52I/Y67F 2  variant; (C) the N52I/W59A/Y67F variant; (D) the W59A variant. Only the down field region of the spectrum is shown. Assigned resonances are labeled above each spectrum  and correspond to protons of heme methyls H M 8 and H M 3 , heme  propionate-7 (HP7-la and b) and heme methyl 5 (HM5). Equivalent signals in the spectra of each variant are connected by dotted lines. The main differences between consecutive spectra (starting with spectrum A) are highlighted with asterisks while the signals that could not be assigned were not labeled.  77  Some resonances observed in these spectra were assigned initially by comparison with previous reports of the spectra of the wild-type yeast cytochrome (Bums and Lamar 1981; Williams et al. 1985; Gao et al. 1990; Turner 1993; Banci et al. 1991a) and included the hyperfme shifted heme methyl resonances HM8, H M 3 , and HM5 as well as the broad signal for the s i proton of His 18 and some of the protons for heme propionate 7. Other characteristic signals of cytochrome c were also observed in the high field region (data not shown) and consisted of protons  S-CH2  of Pro30 and the y and s protons of  Met80 (Gao et al. 1990; Banci et al. 1997a). Notably, the spectrum of the W59A variant exhibited the greatest difference from the other three spectra. The spectrum observed for this protein suggests considerable structural perturbation relative to the other proteins and suggests that the much of the protein is significantly unstructured under these conditions. Following acquisition of this spectrum, no further N M R studies of the W59A variant were undertaken.  2. 2D NOESY spectra To confirm the assignments and to study these signals further, additional spectra were recorded to consider the dependence of the spectrum on temperature and pH (data not shown) and 2D N O E S Y spectra were obtained (Figure 20a and 20b). The H ^ H N O E S Y (Nuclear overhause effect spectroscopy) experiment, is a two dimensional N M R technique that is useful for determining i f signals arising from specific protons are close to each other in space even i f they are not bonded.  78  1  WT  E a a I i  30-  3  3234-  Bf  3635  30  25  20  15  10  w - H (ppm) 1  2  Figure 20a. Upfield region of the N O E S Y spectra of wild-type  and  variant ferricytochrome c . Only a section o f the spectra is shown for clarity. The heme methyl resonances, H M 8 and HM3 are indicated by numbers, and the cross peaks with specific protons are indicated by letters and arrows. Each mark represent a cross peak o f HM8 with (A) H P l - a ; (B) H P l - b ; and (C) with proton C7 of residue W59. Equivalent resonances in each spectrum are connected by dotted lines  79  WT  I52F67  I52F67A59  31 32  to  T3 T3  3  33  8  34  -8  35H  36  36  ~$5  S  3Y  32  31 36  3s  S  3Y  32  31  36  35  34  33  f\  32  w1- H 1  Figure 20b. Expanded view of Figure 20a showing the area around H M 8 and H M 3 for each variant. The labels are the same as for Figure 16a, except for R 8 and R3 in N 5 2 I / W 5 9 A / Y 6 7 F and their corresponding cross peaks ( A and B , see text).  Comparison o f these ' H and N O E S Y spectra reveals several differences among them; however only significant variations around specific protons are considered below.  3. HM8 The H M 8 resonance observed in the spectra o f all three proteins exhibits normal Curie behavior (i.e., upfield displacement as temperature is increased), and undergoes small shifts as p H is changed from 7 to 6 or to 9 (data not shown) that agree with previous reports (Moench et al. 1991). The N O E S Y spectra (Figure 20a) establish that these protons interact (form a cross peak) with protons o f heme propionate-7 ( H P 7 - l b and H P 7 - l a ) and with the proton C2  from Trp59 (Thurgood et al. 1991), except in the triple variant where this latter  residue has been replaced. T w o additional, unidentified signals overlap the peak for  80  H M 3 , but not that o f H M 8 (Figure 19C). However, in the N O E S Y spectrum o f the N 5 2 I / W 5 9 A / Y 6 7 F variant (Figure 20b), it is evident that the extra resonance at 32.9 ppm (designated R8) is in close proximity to H M 8 and results i n a strong cross peak. Moreover, R 8 also fonns a cross peak with both H P 7 protons, confirming the proximity of this proton to H M 8 . T w o more cross peaks are observed i n the N O E S Y spectrum of the triple variant (data not shown). One o f these cross peaks is an unassigned signal at 15.9 ppm that also interacts with protons H P 7 - l b and H P 7 - l a . The other cross peak is a signal at 10.9 ppm that very likely corresponds to the interaction with H M 1 (see discussion o f H M 1 below).  4. HM3 The resonance for H M 3 also exhibits normal Curie behavior (Turner 1993), and it shifts downfield as the p H is increased (Moench et al. 1991). From the N O E S Y experiments, it was concluded that the H M 3 protons o f both variants interact with protons CAH and PH (~6 and ~4 ppm, respectively) from Phe 82 (data not shown) and with protons 4 p C H (~3 ppm) i n the heme group (Moench et al. 1991). A s mentioned previously, the 3  spectrum o f the N 5 2 I / W 5 9 A / Y 6 7 F variant exhibits two additional resonances that overlap the resonance o f H M 3 and are absent from the corresponding spectra o f the wildtype protein and the N 5 2 I / Y 6 7 F variant. The resonance at 31.9 ppm (R3) forms a cross peak with H M 3 at 32.5 ppm (Figure 20a). In addition, R 3 forms a cross peak with the proton 4 p C H 3 (data not shown) that again suggests proximity o f this group to H M 3 .  81  5HP7-1b In the spectra o f all three cytochromes, the resonance for the H P 7 - l b proton exhibits Curie behavior and no change with p H over the range studied. Nevertheless, differences in the behavior o f this resonance are apparent i n the N O E S Y  spectra.  Specifically, in the spectrum o f the wild-type cytochrome, this proton exhibits a strong cross peak with another heme propionate proton, H P 7 - l a (at 13 ppm) and weaker interactions with the 8-2, C-2 r\-2 protons o f Trp59. A l l o f these interactions are maintained i n the spectrum o f the N 5 2 I / Y 6 7 F variant but are not evident i n the spectrum of the N 5 2 I / W 5 9 A / Y 6 7 F variant as expected by the lack o f Trp59 i n the latter protein, h i the triple, variant, however, additional, unassigned cross peaks are observed at 15.9, 9.7, 8.9, 8.0 and 7 ppm (data not shown) that are not observed i n the spectra o f either the wild-type or double variant protein.  6. HP7-1a - A s observed for H P 7 - l b , the chemical shift for H P 7 - l a exhibits normal Curie behavior and no dependence on p H . Interaction o f this group with Trp59 is evident in the spectra o f the wild-type and double variant proteins, and interaction o f H P 7 - l a with H M 1 is apparent with all three proteins. A s observed i n the spectrum o f the triple variant, the extra resonance R 8 interacts not only with this proton but also with H M 1 and  with  HP7- lb.  82  7. HM5 Two  significant differences  between  the  NOESY  spectrum  o f wild-type  cytochrome c and the corresponding spectra o f the variants, are the two strong signals that are most likely attributable to heme methyls H M 5 and H M 1 . A s seen i n Figure 19 (B and C ) , both signals shift downfield significantly. Specifically, the resonance o f H M 5 shifts from 10.5 ppm i n the spectrum o f the wild-type protein to 12.5 ppm in the spectrum of the variant. The resonance o f H M 5 is not readily observed i n the spectrum o f the w i l d type protein, but it appears at - 1 0 ppm in the spectrum o f both variants. In addition, both signals exhibit anti-Curie behavior (Turner 1993). Differentiation between H M 5 and H M 1 i n Figures 19 and 20 was based on the observation that i n the N O E S Y spectra the signal assigned as H M 5 exhibited no cross peaks with H M 8 , H P 7 - l a or H P 7 - l b and, therefore, could not be that o f H M 1 .  8. HM1 The resonance for H M 1 undergoes a substantial up field shift from 8 ppm in spectrum o f the wild-type cytochrome to 10.6 and 10.9 ppm in the double and triple variant respectively (Figure 19B and C ) . Analysis o f the N O E S Y spectrum o f the N 5 2 I / W 5 9 A / Y 6 7 F variant revealed strong cross peaks o f H M 1 with H M 8 , H P 7 - l a and H P 7 - l b (data not shown) as observed in the spectrum o f the wild-type protein, but in addition, a weaker cross peak is observed with H M 3 at 12.5 ppm (also observed in the spectrum o f the N 5 2 I / Y 6 7 F variant at 12.4 ppm) and with R 8 at 31.9 ppm, suggesting a significant modification in the environment o f this proton in the variant.  83  Summarizing these results, two significant changes are evident from the mutations as observed by the N M R experiments: First the effect o f substituting Asn52 and Tyr67 in both variants, N 5 2 I / Y 6 7 F and  N 5 2 I / W 5 9 A / Y 6 7 F , produced changes i n the electronic  environment o f protons H M 1 and H M 5 , in both proteins, since these signals are significantly displaced upfield i n comparison with the W T protein. The other evident effect o f the mutations is observed only in the more active variant N 5 2 I / W 5 9 A / Y 6 7 F , where additional cross peaks around H M 8 , H M 3 , and H P 7 protons are observed. These differences are not likely to be cause exclusively by the substitution o f Trp59, since the variant bearing only this mutation had very broad spectrum from which was not possible to obtain significant information by the N M R techniques employed.  V. Random mutagenesis A n alternative approach to studying the peroxidase activity o f cytochrome c is to create a collection o f random variants that are then screened for improved activity. Variants identified i n this manner have the potential o f being more useful i n a specific catalytic reaction (better activity, stability, expression, etc.). A t the same time, such variants can provide insight into the general mechanism o f enzymatic catalysis exhibited by the wild-type protein. In the case o f cytochrome c, two crucial limitations can be addressed i n this way: the low catalytic activity o f the cytochrome relative to authentic peroxidases and the low stability o f cytochrome c to heme oxidation by the substrate H 0 . In this work we employed random mutagenesis in combination with a plate 2  2  screening assay i n an attempt to isolate additional active variants.  84  A. Mutagenic reaction The first step i n this process consisted o f determining the number o f cycles o f a mutagenic P C R reaction that were required to achieve the desired level o f mutagenesis. The level o f mutagenesis required is a function o f the number o f times the starting amount o f D N A is doubled (number o f doublings) and o f the length o f the sequence to be mutated (Cadwell 1991). In this work, we chose to introduce an average o f 2 mutations in the 400 bp B a m H I - Nhe I fragment from our expression vector. Under the reaction conditions that we used for the error-prone P C R , an average o f 2 mutations per 300 base pair fragment would be observed after 10 D N A doublings (Cadwell 1991). Samples were withdrawn from the mutagenic P C R reaction taken after a given number o f cycles and were loaded onto an agarose gel (Figure 21 A ) . The product begins to be visible after 4 cycles, and it was estimated that after 10 cycles, ten D N A doublings were obtained. The next step involved promoting the introduction o f errors by the polymerase during the P C R reaction by the addition o f Mn(II). The optimal concentration of Mn(II) for this purposes was determined by adding varying concentrations o f M n C f : to the P C R reaction mixture and evaluating the effect on the amount o f product amplified. The results obtained with 0, 0.2, 0.5, and 1 m M M n C ^ are shown in Figure 21B.  85  B  A ?*> u  §  t*  **s| NI  i* u  8  #*  §•*  «*H **H L  HM  400  0  0.2  0.5  1.0  800 400 200  Figure 21. Agarose gels (2%) to estimate the amount of D N A produced per cycle of mutagenic P C R (A) and the yield at different concentrations of M n C h (B). In 21 A , the subscripts represent the number of cycles (c) while 400 is the approximate size in base pairs o f the D N A fragment.  In 2 I B , the different  numbers  i n the wells  denote the  concentration o f M n C b ( m M ) while L the ladder i n base pairs.  From this result, it is evident that the amount o f D N A produced decreased with increasing concentrations o f MnCl2, to the point that no product was observed at a concentration o f 1 m M . Based on this result, we subsequently used 0.5 m M M n C l  2  because at this concentration a significant effect is observed on the D N A yield, but a useful amount o f product is generated. This product was cloned back into the expression vector, transformed into E. coli cells (with a low transformation efficiency, estimated to be around 10 clones per jug o f D N A ) and purified. Fifteen clones were sequenced to 3  estimate the rate and type o f mutation generated by this approach (Table VII). A l l the clones sequenced contained insert, but not all were expressed.  86  As indicated, seven single mutations (-50%), two double substitutions (14%), one triple mutant (7%) and four clones with no mutations (29%) were obtained. Only two transitions were observed (-13%), and half of the mutations were missense.  Table VII. Results from random mutagenesis of the cytochrome c gene.  Original sequence GAC  Mutated sequence AAC  Amino acid substitution D90N  CCA  TCA  P25S  AAG  AGG  TTT  Type of mutation Type of mutation (A) .(B) Transition Missense a  a  K22R  a  a  CTT  F36L  a  tt  GGT  TGT  G23C  a  Transversion  GCC  ACC  A81T  a  Transition  GGG  GAG  G84E  u  AAA  GAA  K5E  a  AAG  AAA  K-2  Silent  AAC  AAT  N31  a  AAG'  AAA  K54  a  AAC  AAT  N56  a  "  TTC_  TTT  F10  u  "  ACT  ACC  T69  a  ACC  ACA  T78  "  "  Transversion  Underlined letters indicate the base that was changed. Single letter amino acid codes are used to indicate the original residue (first letter) and the new amino acid (letter after the number) resulting from mutagenesis. .  87  B. Screening m e t h o d s To establish the activity screening method, several parameters were evaluated to determine  the  optimal  assay  conditions.  Specifically,  a  serial  dilution  of  a  ferricytochrome c solution was adsorbed onto a filter paper, and several combinations o f peroxidase substrate and  H2O2  were evaluated. The reagent concentrations identified in  this manner (Figure 22) to permit detection o f the smallest amount o f wild-type cytochrome (1 u M ) were 12.5 m M A B T S and 25 m M o f H 0 . 2  Figure 22. Minimum  2  amount of cytochrome c for  which  peroxidase activity was detectable with peroxidase substrates A B T S and  H2O2.  The numbers  written on the  filter paper  indicate  the  concentration o f ferricytochrome c that was placed on the paper at that position. C" is the control without protein and C indicates a solution with +  a small amount o f myoglobin as positive control.  88  On the basis o f this information and the screening protocol reported previously for detecting the peroxidase activity o f myoglobin (Wan et al. 1998), the screening method used in the current study was designed. A schematic diagram o f this method is provided in Figure 23.  Expression cells  Filter paper  Figure 23. Principal steps in the initial peroxidase activity screening strategy.  In this method, a suspension culture o f E. coli cells expressing the protein was spotted onto a disc o f filter paper and exposed to chloroform fumes. In principle, this treatment induces cell lysis and the release of the cytochrome variants to permit detection of peroxidase activity. This activity assay produced a dark green colour (for a positive clone) that results from the oxidation o f A B T S , so variants with greater activity produced detectable amounts o f this green colour as the concentration o f peroxide in the reaction mixture was decreased. After repeated attempts involving a variety o f conditions (e.g., various lysis reagents, incubation conditions, peroxidase substrates, and E coli strains  89  used to express the protein), no clones with detectable levels of peroxidase activity were identified. This result led to the conclusion that the expression level o f the cytochrome combined with the intrinsically low peroxidase activity o f the protein prevented detection of a positive clone. Thus, an alternative screening strategy was required. In an attempt to improve the culture conditions that would allow a better expression of the protein, very small volumes of liquid culture (150 ul) were grown in P C R microplates. A general scheme o f this method is presented in Figure 24.  PCR Microplates  Culture dilutions  . w  .... . . Small liquid cultures  ^  p  .., . . Released protein  w  1  ABTS +  H0  Individual colonies  ^ ^  Culture in agar plates  ^  C  o  [  o  u  r  r  e  a  c  t  2  i  o  2  n  ^  Figure 24. Modified screening protocol that permitted detection of peroxidase activity in E coli cells expressing cytochrome c variants.  After transforming E. coli with the products o f the mutagenic P C R mixture, consecutive dilutions were used as inocula to set up "minicultures" in microtitre plates containing L B media. H a l f of each plate was used as a "duplicate" and was not assayed for peroxidase activity. Each plate was sealed with a special plastic mat to prevent excessive evaporation during incubation at 37 °C. After incubation, the cells were isolated by centrifugation, excess media was discarded and a commercial lysis reagent  90  was added. The resulting lysate could then be assayed for peroxidase activity by addition of a mixture of A B T S and  H2O2.  When a positive reactions was observed, the equivalent  tube in the duplicate cultures were spread on agar plates to obtain individual colonies that would be subjected to repeated activity assay by the same procedure. Finally, D N A isolated from individual colonies that exhibited activity was submitted for sequence analysis. In several cases, the positive clones turned out to be producing wild-type protein, so the stringency of screening was increased by using diluted substrate solution which would produce a positive result only for those clones expressing more active variants. A n average of 15 microplates per concentration of substrate solution was screened and only two clones, both with a single mutation, were identified as exhibiting improved activity by this method. D N A sequence analysis identified the variant cytochromes leading to this improved activity as G84E and K 5 E . Both mutations are situated in the surface of the protein, above the heme group. G84E is located relatively close to the heme group (on the top part of the heme cavity), while K 5 E is localized on the opposite side as part of the first a-helix of the protein. G84E was chosen for further characterization because it was the only clone that gave a positive reaction after repeated dilutions of the assay solution (Figure 25) and because of its proximity to His82 and Argl3, two residues considered to be relevant for the structure and function of cytochrome c (Rafferty et al. 1990; Berghuis and Brayer 1992). In Figure 25, a representative plate clearly shows a dark green colour only in the tube with the E84 variant, while the other samples have no reaction.  91  WT  A  G84E  C  Figure 25. P C R plate showing the positive reaction of variant G84E after repeated dilutions of the substrate solution. The positive reaction o f G 8 4 E was indicated by a dark green colour (darker tone i n the Figure) absent i n all the other tubes. The W T protein was included for comparison, while A and C indicate colonies with negative reaction.  C. G84E variant The G 8 4 E variant was characterized further i n terms o f its kinetic and structural properties for comparison primarily with the wild-type cytochrome.  1. Kinetic parameters The same analytical procedure used to calculate the kinetic parameters for the previously discussed variants (page 57) was applied to the G 8 4 E variant. The kinetic parameters derived i n this manner  for the peroxidase ping-pong mechanisms are  compared with the corresponding values for the wild-type protein i n Table VIII.  92  Table VIII. Rate constants k\ and & constants obtained from steady-state kinetics analysis of the peroxidase activity of wild-type ferricytochrome c and the G84E variant. 3  Error  &3 ratio (Variant/WT)  (%)  k\ ratio (Variant/WT)  1.4 X 10  4  2  2.9  2.2  6.4 X 10  3  5  —  • —  Error  (M'V )  (%)  (M'V)  G84E  59  3  WT  20  5  1  Conditions are the same as in Table II, page 59.  A s is apparent from these results, the catalytic parameters resulting from the single G 8 4 E substitution are distinctly different from those o f the wild-type protein. The value o f k\ is ~3-fold greater than observed for wild-type protein and & is about doubled, 3  consistent with the greater catalytic activity detected for the cell lysate in the screening assay. Nevertheless, the rate enhancements observed for the G 8 4 E variant are not as great as those observed for the variants described earlier i n this work (Table II, page 59). In terms o f inactivation, this protein resembles the W 5 9 A and N 5 2 I / W 5 9 A / Y 6 7 F variants in that it is less stable than the wild-type cytochrome to inactivation during turnover. This characteristic is reflected in the values o f  ^ ct Iia  obtained at three concentrations o f  H2O2  (Table IX). In this case however, the differences are marginal, and therefore the partition ratio observed in the same Table, is still higher due to an increased activity (Table IX). In a similar way, i f this ratio is compared against the other variants, is interesting to notice that G 8 4 E is more efficient that the triple variant N 5 2 I / W 5 9 A / Y 6 7 F showing a higher ratio at the three concentrations o f W2O2 tested.  93  Table IX. Inactivation and partition ratio comparison of the G84E variant versus the WT protein. [H 0 ] 40 m M 2  G 8 4 E fc  (s" ) 1  inact  1.7 x l O "  G 8 4 E Partition ratio WT *  (s" ) 1  i n a c t  2  [ H 0 ] 60 mM  [H 0 ] 80 m M  2.4 x 10~  2.8 x 10"  193  183  2  2  2  2  206  2  2  2  1.1 x IO"  1.7 x IO"  2.2 x IO"  178  155  150  2  W T Partition ratio  2  2  Conditions described i n Table III (for k , page60) and Table I V (for partition ratio, page 61). Error were calculated to be no more than 3% (partition ratio) and no more than 8% i n a c t  (kinact) •  2. Stability parameters The parameters  determined  to  characterize  the  structural  stability o f the  cytochrome variants produced by site-directed mutagenesis (Table V , page 63) were also determined for the G 8 4 E variant identified by random mutagenesis (Table X ) . From the 0.5 decrease in the pi^a for the alkaline conformational transition, it is apparent that this variant is less stable toward formation o f one or both o f the alkaline conformers than is the wild-type protein. Similarly, the T  m  for thermal denaturation o f the protein as  determined by spectropolarimetry is significantly lower than that o f the wild-type protein. Finally, the G 8 4 E substitution also renders the cytochrome more susceptible to oxidation of the heme prosthetic group by H 0 . A l l three parameters indicate that this variant is 2  2  significantly less stable than the wild-type cytochrome. These observations extend the general correlation o f decreased protein stability with increased activity i n the peroxidase activity.  94  Table X. Parameters related to stability of the randomly generated ferricytochrome c.  G84E  54.3  Error (%) 0.2  WT  59.2  0.8  CO  7.3  Error (%) 1.1  Heme (%) 50  Error (%) 1  7.8  3.8  38  9  a  Conditions used for each type o f measurement are described i n the Methods section and are similar to those for related values shown i n Table V .  3. Reduction potential and thermodynamic parameters The reduction potential of the G84E variant and the related thermodynamic parameters are compared with the corresponding values for the wild-type cytochrome in Table XI. Notably, the reduction potential of the variant is significantly lower (48 mV) than that of the wild-type protein, and the potential obtained is very similar the potential obtained for the double variant N52I/Y67F (Table VI, page 66). Despite the significant decrease in reduction potential, there is no significant difference in the entropy of reduction between proteins; however, there are differences of around 1 kcal/mol in the values of AG and AH. In terms of the factorization of the entropic and enthalpic terms, is possible to infer that the differences in the redox potential are mainly due to the later term (Battistuzzi etal. 2001b).  95  potential  Table XI. Reduction parameters for the G84E variant. E° (mV)  thermodynamic  AG (kcal/mol)  AS (eu)  AH (kcal/mol)  AH/F (mV)  TAS/F (mV)  0  a  and  0  0  G84E  218  -5.03  -20.3  -11.1  +480.5  -262.5  WT  266  -6.1  -20.1  -12.1  +524.3  -258.6  Versus N H E , 298.15 K , 0.1 M KC1. Uncertainties: E°, 2%; A G , 1%; A S , 2%; and A H , 1%. 0  0  0  4. Spectroscopic characteristics of G84E Despite the altered kinetic and stability properties exhibited by this variant and i n contrast the variants discussed earlier, the spectroscopic properties o f the G 8 4 E variant were found to be essentially identical to those o f the wild-type cytochrome. For example, the electronic absorption spectrum o f the ferricytochrome exhibited absorption maxima and intensities unchanged from those o f the wild-type protein (Table XII). The only changes evident in the U V or visible C D spectra o f the variant relative to the spectrum o f the wild-type protein were some small shifts in the visible region. Specifically, a shift in the positive Cotton effect at 409 nm was observed (to 405 nm) and i n the negative Cotton effect at 417 (to 415 nm). A minor shift in the absorbance maximum observed at 1744 nm in the near infrared M C D spectrum o f the wild-type protein to 1741 nm in the spectrum of  the  variant  was  also  observed.  The  ID  'H-NMR  spectrum  o f the  variant  ferricytochrome was unchanged from that o f the wild-type protein. Taken together, these results suggested the absence o f dramatic changes i n the overall structure o f the variant.  96  Table X I I . Electronic Absorption Spectroscopy of the G84E Variant Circular Dichroism far-UV  Electronic Absorption Protein  Soret  3  visible  <x,p  C-T  ^Max  'Winl  ^Min2  MCD  ^Max ^Min  ^Max  G84E  409  529,560  695  196  209  219  409  417  1741  WT  408  529,560  695  195  210  221  405  415  1744  Conditions for each measurement were similar to ones for the WT protein (page 67-72) and therefore were included for comparison.  VI. T h e m e c h a n i s m o f c y t o c h r o m e c i n a c t i v a t i o n b y h y d r o g e n p e r o x i d e  As discussed above, the peroxidase activity of cytochrome c is inactivated during turnover, presumably owing to oxidative modification of the cytochrome by hydrogen peroxide. As was discussed before, several inactivation pathways have been proposed, and the kinetics of such pathways seem to vary with the enzyme and the reaction conditions involved (Nicell and Wright 1997; Prasad et al. 2002; Valderrama et al. 2002; Diederix et al. 2003). Therefore, to gain insight into the mechanism by which this inactivation reaction occurs, the reaction of ferricytochrome c with hydrogen peroxide was studied by using several spectroscopic and chemical methods. The objective of this work was to identify the products of the reaction and to provide sufficient insight into the inactivation mechanism that strategies could be developed to prevent this inactivation and improve the catalytic performance of the cytochrome for subsequent applications.  97  A. H e m e b l e a c h i n g It has been known for many years that the electronic absorption spectrum o f ferricytochrome c and other heme proteins {e.g., myoglobin, hemoglobin) decreases dramatically on incubation o f the protein with  H2O2  (Florence 1985; Valderrama et al.  2002). Because this change in spectrum can be observed as a loss in colour o f the protein solution, this process is often referred to as "bleaching." The susceptibility to bleaching varies from protein to protein presumably because the mechanism o f the reaction is characteristic o f each protein. The change i n electronic spectrum induced by incubating wild-type yeast iso-l-ferricytochrome  c and the N 5 2 I / Y 6 7 F double variant with peroxide  is shown in.Figure 26. The inset shown i n the upper panel o f this Figure is the X-band E P R spectrum (4 K ) o f wild-type ferricytochrome (250 u M ) incubated with peroxide (0.25 M ) and frozen within 15 seconds from the start o f the reaction. A s seen, this spectrum clearly demonstrates the presence o f a free radical (Villegas et al. 2000) that was not observed prior to addition o f peroxide (data not shown). The corresponding  spectrum  obtained  following  identical treatment o f the  N 5 2 I / Y 6 7 F , is shown in the inset for the lower panel. This result combined with the resistance o f the double variant to bleaching by peroxide suggests that the bleaching mechanism involves a radical intermediate.  98  0.25  -,  0.20  A  0.15  0.10  0.05  o  o c  JS  0.00  O £  [*10 , A  N52I/Y67F  0.25  <  0.20  A  0.15  0.10  A  0.05  0.00 350  400  Figure 26.  450  500  550  550  X (nm)  Decrease in absorbance  at the  Soret maximum  ("bleaching") following addition of H2O2 to ferricytochrome c. A 2.5 u M  p r o t e i n s o l u t i o n i n 2 0 m M s o d i u m phosphate buffer p H 7.2, w a s  incubated f o r 2 0 m i n i n the presence o f 2 5 0 u M H2O2, a n d scans were t a k e n every t w o m i n u t e s . T h e arrows represent the d i r e c t i o n o f the change. T h e inset o n the right s h o w s the presence o f a free r a d i c a l , detected b y l o w temperature E P R , o b s e r v e d o n l y i n the W T reaction but not w h e n the d o u b l e variant i s used ( V i l l e g a s et al. 2 0 0 0 ) .  99  B. Exogenous ligand binding to the heme To investigate further the involvement o f a free radical i n the bleaching reaction, experiments o f the type shown i n Figure 27 were performed. In this experiment, the binding o f the cyanide anion to the wild-type ferricytochrome and to the double variant was studied by monitoring the change in electronic spectra as a function o f time following the addition o f cyanide. Ferriheme proteins are well known to have high affinities for binding o f CW (Horecker 1946; Wallace and Clarklewis 1992; Y a o et al. 2002). The formation o f this complex can be observed by monitoring changes in the Soret absorbance as a function o f time (Blumenthal and Kassner 1980) as shown i n Figure 27. Notably, the spectrum o f the double variant is virtually unchanged following identical  (H  350  .  ,  400  .  ,  '  450 350  •  '  1  400  '  '  450  X (nm) Figure 27. Change in Soret spectrum following incubation wild-type ferricytochrome or the double variant with K C N . 5 u M protein solutions in 20 m M sodium phosphate buffer p H 7.2, were incubated with l O m M K C N . Spectra were acquired every 2 m i n over a 40 m i n period. The direction o f the change i n the spectra as the reaction progressed is indicated (arrows). The initial and final scans are labeled as I and F , respectively.  100  treatment, indicating that this protein binds little or no cyanide under these conditions. Exposure of either protein to excess  H2O2  after incubation with CN" completely prevents  bleaching of either protein (data not shown). These results suggest that heme bleaching and formation of the free radical intermediate (indicated by the EPR results) are inhibited as the stability of axial ligand binding increases.  C. Peroxide-induced oligomerization of ferricytochrome c In addition to these results, it has been reported that, as a consequence of the inactivation reaction, horse cytochrome c forms protein oligomers as observed by SDSP A G E electrophoresis (Harel et al. 1988). The possibility that yeast ferricytochrome c also forms larger species upon treatment with peroxide was evaluated in experiments represented by. Figure 28, which provides the results obtained from experiments with the wild-type cytochrome. As can be appreciated from this result, dimers, trimers, and larger oligomers were obtained. It was also apparent, that under the conditions used in this experiment (2-20 min reaction time) that the identity and concentration of products were independent of the incubation time, suggesting that the products are formed rapidly. Notably, identical results were obtained in the presence of the chelating agent D T P A or buffer prepared after removing trace metal ions, thus ruling out the involvement of metal ions (e.g., Cu(II)) and radical species resulting from the presence of metal ions in this reaction.  101  10  20  Otfftk  jflUfc  Trimer  Dimer  Monomer  Figure 28. Oligomerization of cytochrome c in the presence of H 0 2  2  observed by SDS P A G E electrophoresis. 250 u M protein  solutions i n 20 m M sodium phosphate buffer p H 7.2 were incubated with 100 molar excess peroxide (2.5 m M ) . The numbers on the left represent the sizes (in kDa) of the protein ladder (lane L ) , and the apparent oligomer size assigned to the reaction products are indicated on the right, to, t , ts, tio and to t o were samples taken at 0, 2,5 10 and 2  2  20 min. respectively.  Further studies demonstrated that the extent o f oligomer formation depends on H 0 2  2  concentration (Figure 29). A s is apparent from this Figure, only a small amount of dimer was produced with  a 10-fold excess o f peroxide, and the yield o f dimer and trimer increased with higher oxidant concentrations (100- and 1000-fold excess).  102  L  OX  10X  100X  1000X  56 38 30 21  Figure 29. SDS P A G E electrophoresis of the products resulting from  reaction of wild-type ferricytochrome c with various ratios of  [Ff202]/[ferricytochrome c] . Solutions of protein 250 u M in 20 m M sodium phosphate buffer p H 7.2, were incubated with the corresponding molar excess o f H C>2 (0-0.25 M ) . The excess o f 2  H2O2  was removed by centrifugal  ultrafiltration prior to loading the gel. The ratio o f reactants relevant to each lane (0 to 1000 molar excess peroxide) is indicated at the top o f the photograph. The molecular weights (kDa) for the reference protein ladder (lane L ) are indicated on the left. The difference in the yield o f oligomers at equivalent H2O2  molar ratios (compare lanes to to  t2o  in figure 26 versus figure 27 lane  100X) can be explained by the chemical incompatibility between H C>2 and the 2  membrane o f the filtration device, producing a lost o f material.  Interestingly, kinetic analysis o f the same experiment monitored by electronic spectroscopy (Soret absorbance as a function o f time (Figure 30)) showed that the shape of the curve for the reaction varies with peroxide concentration. These observations may  103  suggest the involvement o f different reaction intermediates and or mechanisms in the inactivation reaction depending on the concentration o f the oxidant.  I — ' — - i — • — i — ' — i — • — i — '  0  10  20  30  40  • — i — ' — i — i — i — i — i — • — i — i  0  10  20  30  ^ — i — , — , — , — i — , — , —  40  0  10  20  30  Time (min)  Figure  30. Absorbance change  versus  time  for  the  heme  absorbance of W T cytochrome c (2 uM) in the presence of I X , 10X and 100X excess (2 u.M, 20 and 200 u M , respectively) of  H2O2.  The insets on  the right are the individual traces at different scale for the y axis.  104  D. Identification of inactivation reaction products Based on these results, an attempt was made to prepare modified cytochrome products resulting from treatment o f the protein with hydrogen peroxide and to identify the chemical modifications involved. Chemical definition o f these modifications could, in principle, provide an understanding o f the inactivation mechanism that would be crucial to the design o f cytochrome variants that are less vulnerable to inactivation during catalytic turnover. One strategy for understanding such modifications that has proved useful  in  the  identification o f  transiently  formed,  protein-centred  free  radical  intermediates (Fenwick and English 1996; Zhang et al. 2002; Davies and Hawkins 2004) is described in Figure 31. In this strategy, peroxide is added to a solution containing the ferricytochrome and a free-radical trapping reagent. In principle, the trapping reagent w i l l react with protein-centred radical intermediates that form even transiently as the result o f the reaction. M a n y radical trapping reagents have been reported, and a large number o f them are available from commercial sources (Rosen 1999). In the current study, 7V4ertbutyl-a-(4-pyridyl) nitrone N'-oxide  (POBN)  was  selected  as the  trapping agent  (Stoyanovsky and Cederbaum 1998) based on a previous report that this reagent is effective in such experiments with cytochrome c (Filosa 2001b). Modified cytochrome produced by this method could, i n principle, be characterized by tryptic peptide mapping and mass spectrometry to identify the residue(s) modified by hydrogen peroxide i n this reaction.  105  Protein + H 0 2  Trypsin digestion  (Spin trap) ^  Peptides  Protein-adduct  HPLC  2  Mass values  MS  Isolated peaks  Figure 31. General strategy to identify protein modifications resulting from the reaction of ferricytochrome c with  H2O2  in the  presence of a spin trap by H P L C / M S peptide mapping (see text).  The first step i n this strategy was to establish the formation o f the spin-labeled protein adduct (Figure 32) and to optimize the yield o f the modified protein. Also shown in Figure 32 is the formation of modified cytochrome produced in this manner as illustrated by the mass spectrum obtained for the reaction o f wild-type yeast iso-1ferricytochrome c. The mass difference obtained (+138 m.u.) agrees with previous reports in which the spin-trap looses the t-butyl group (-57 m.u.) after reaction with radical centre (Stoyanovsky and Cederbaum 1998). Several additional attempts were made to increase the amount o f product obtained with the spin-trap by varying reaction conditions; however, the best yield (estimated from relative peak intensities) was - 1 3 % conversion to spin-trapped adduct.  106  Figure 32. Reaction of cytochrome c with H 1 O 2 in the presence of P O B N analyzed by L C / M S . In the upper part, the general reaction is defined, while in the bottom, the result of the L C / M S analysis of the reaction (0.25 m M ferricytochrome c; 2.5 m M H 0 2  2  and 50 m M P O B N in 20 m M sodium phosphate buffer p H 7.2) in a  Perkin Elmer SCIEX API 300 Triple Quad spectrometer, is observed.  107  Unfortunately, repeated attempts to isolate modified peptides for further analysis by H P L C tryptic peptide mapping failed to detect any modified peptides in digests of the modified protein. This difficulty could have resulted from the involvement of multiplemodified peptides of which none formed i n sufficient amount to permit detection- or from instability of the modified peptides to the conditions required for analysis.  108  4. DISCUSSION A m o n g the numerous catalytic activities o f cytochrome c (Table I), the peroxidase activity is one o f the most interesting owing to the enormous potential for application o f this activity i n diagnostic tests, chemical synthesis and bioremediation (Conesa et al. 2002; Burton 2003; Torres et al. 2003). The potential advantages o f using cytochrome c as a peroxidase i n such applications arise from the stability o f the protein to extremes o f p H and temperature and to organic solvents. This potential, however, is hindered by two main factors: low catalytic activity compared to natural peroxidases and an oxidative instability towards one o f the substrates, hydrogen peroxide, that is inherent not only to cytochrome c but to any peroxidase-like catalyst (Seelbach et al. 1997; Valderrama et al. 2002; Diederix et al. 2003). To address these limitations, it is essential to understand the relationship between the protein structure and the catalytic mechanism. In this work, we used as a starting point recent studies that have con-elated the stability o f cytochrome c with its peroxidase activity by demonstrating  an increased catalytic activity upon  unfolding o f the protein (Diederix et al. 2001; Diederix et al. 2002; Diederix et al. 2003) as well as studies concerning the effects o f specific mutations on the thermal stability o f the protein (Berghuis et al. 1994; Lett et al. 1996). B y designing variants with opposing effects on the stability o f the protein and investigating the resulting consequences on the peroxidase activity as well as on the overall structure o f the protein, new and useful insight into the mechanism by which cytochrome catalyzes these type o f reactions was sought.  109  I. Site-directed variants Three variants were designed for the purpose of comparing the contribution of structural stabilization on cytochrome peroxidase activity to that of structural destabilization and the ability of stabilizing mutations to offset the contributions of destabilizing mutations on this activity. Yeast z'so-1-cytochrome c was used in this work because a large number of relevant variants of this cytochrome have been characterized structurally by X-ray diffraction analysis. The structurally stabilized variant selected for this work was the N52I/Y67F double variant (Berghuis et al. 1994; Lett et al. 1996) while the W59A variant was selected as the structurally destabilized variant (Caffrey and Cusanovich 1993; Black et al. 2001). Notably, the current study represents the first successful isolation of the W59A variant of cytochrome c. Previous attempts to purify this variant in this laboratory following expression in transformed yeast failed, presumably owing to the instability of the protein to exposure to ethyl acetate as required in the standard protocol for purification of cytochrome c from yeast. The relative contributions of stabilizing and destabilizing mutations on peroxidase are of interest in light of studies by Canters and co-workers concerning the effect of denaturing agents on the peroxidase activity of the bacterial cytochrome  C550  (Diederix et al. 2001; Diederix et  al. 2002). The N52I/Y67F variant was selected for this phase of the work because the threedimensional structure of this protein has been determined by X-ray diffraction analysis (Berghuis et al. 1994) in an effort to understand the structural basis for the greater thermal stability and elevated pAT for the alkaline conformation transition that result from a  these substitutions (Das et al. 1989; Luntz et al. 1989), two parameters often used as  110  measures of cytochrome c stability. Moreover, initial studies of the catalytic properties of this variant indicated that these substitutions stabilize the heme group to bleaching by H2O2  (Villegas et al. 2000), suggesting an involvement of these residues in the oxidative  mechanism. Replacement of the phylogenetically invariant Trp59 with A l a was undertaken with the expectation that this substitution would destabilize the protein by removing a key internally-located residue that is involved in a crucial and complex hydrogen bonding network (Cutler et al. 1989; Davies et al. 1993; Liggins etal. 1994). The kinetic analysis reported in this study demonstrates clear differences between the peroxidase activities of each variant. In addition, the absence of a pre-activation phase in our oxidation curve suggests that the model developed for the same reaction catalyzed by cytochrome c o (Diederix et al. 2001) is not appropriate for the yeast cytochrome 55  used here. Species-specific differences in cytochrome functional properties have been noted previously (Wan et al. 1998; Prasad et al. 2002), so it is not altogether surprising that the peroxidase activity also varies with the species of the cytochrome. In the case of the stabilizing mutation N52IY67F, the values for k\ (rate constant for formation of the first catalytic intermediate) and k (rate constant for oxidation of the 3  second intermediate) are very low. When the destabilizing substitution W59A was introduced into either the wild-type protein or the double variant, both rate constants increased significantly. However, even in the case of the most active variant, W59A, these values are much lower than those exhibited by true peroxidases such as horseradish peroxidase (k\ = 9 x 10 M s ; (Hiner et al. 1996)) and even other non-peroxidase heme 6  _1  _1  proteins such as myoglobin (^=540 M ' V ; (Wan et al. 1998)). At the same time, it is 1  interesting to observe that the value of k for all variants studied in this work was 3  111  considerably greater than the corresponding value for k\, contrary to the relative values of these rate constants that are generally observed for true peroxidases (Hiner et al. 1996; Moffet et al. 2000; Nielsen et al. 2001). In the specific case of the W59A variant, the magnitude of k is surprisingly high not only with dicarboxidine as reducing substrate but 3  also with A B T S as substrate. In this latter case, the value for k  3  (9.5 x 1(T M " V ; data  not shown), is four-fold greater than the same parameter determined for horseradish peroxidase (2 x 10 M ' V ; (Moffet et al. 2000)). These results imply that the active site 3  1  of the tryptophan variant is highly accessible to the reducing agent, as has been observed for other synthetic peroxidases (Moffet et al. 2000). In this regard, it was unfortunate that a direct comparison between the catalytic activity of the proteins studied in this work and other cytochromes was not possible, not only because different substrates were used (dicarboxidine versus guaiacol or ABTS) but also because the kinetic analysis was different for each case (Diederix et al. 2001; Prasad et al. 2002). As demonstrated by the results reported above, the rate constants k\ and k are 3  frequently insufficient to describe fully the peroxidase activity of this protein under a variety of conditions, thus necessitating the consideration of additional parameters. Notably, the inclusion of a rate constant for inactivation of the cytochrome peroxidase activity is crucial owing to the instability of these proteins to reaction with hydrogen peroxide. Nevertheless, the complex and incompletely understood chemical nature of this reaction has resulted in the interpretation of such results through use of various models (Nielsen et al. 2001; Prasad et al. 2002; Diederix et al. 2003). In the present work, the good fits obtained to the data allowed us to interpret the results in terms of a general inactivation model (Duggleby 1986). In this work, it is  112  essential that attention be drawn to careful selection of the peroxidase substrate. Normally, ABTS is by far the first choice for the kinetic characterization of peroxidases and peroxidase-like catalysts because it is highly soluble, the product exhibits high molar absorptivity, and it is available commercially at modest cost. Unfortunately, the product of ABTS oxidation exhibited considerable instability under the conditions of greatest interest to the current study. While this characteristic is not a problem for analysis of initial rate constants, it is a critical problem in determination of progress curves because it adds an additional, complicating phase to the reaction curve. In our case, the use of dicarboxidine, an established (Paul et al. 1982) but less-widely used peroxidase substrate than ABTS eliminated this problem, and we submit that this substrate should be the reagent of choice in any future studies of this catalytic activity of cytochrome c. The calculation of kinetic parameters is very important to compare the different behavior among variants during the oxidative reaction. However, as discussed before, this task is intricate for several reasons: First, the complexity of the peroxidase reaction among the different enzymes, due mainly to the high reactivity of the chemical species involved, makes it difficult to identify the common intermediates and reaction products as well as the proposal of a general mechanism, with some reaction products  still under investigation, this  complicates the analysis not only in terms of effective catalysis but also in terms of catalytic inactivation, with several models proposed to date (Rasmussen et al. 1995; Nicell and Wright 1997; Diederix et al. 2001; Valderrama et al. 2002; Ghibaudi and Laurenti 2003).  113  Second, one of the most generalized kinetic analysis of this reaction, the one proposed by Dunford (Dunford 1999) proposed that the typical kinetic parameters used to characterize enzyme catalysis, like k , k , kj, etc and their correlations, have no obvious cat  M  meaning in peroxidase chemistry. This is caused by the intrinsic characteristics of the reaction and is experimentally observed by the apparent absence of an upper limit in the reaction rate when the concentration of any of the substrates is increased (Dunford 1999). Third, even though cytochromes from different species have been reported to catalyze peroxidative reactions, no intermediates similar to the ones observed for peroxidases (e.g. compound I, compound II, etc) have been found except for the case of chemically modified horse heart cytochrome, where a compound I-like species has been identified (Prasad et al. 2002). In spite of these limitations, the inactivation constant determined by this study is useful in assessing the behavior of our catalysts during turnover where factors such as enzyme protection by the reducing substrate or the partition ratio play an important role in determining the overall catalytic performance of the protein. From the results obtained, it is apparent that the activity of the variant is directly related to the instability of the variant as indicated by thermal stability or pK for the alkaline conformational transition. a  Thus, additional parameters are required to define the relative contributions of the catalytic and inactivation rates of peroxidase-like catalysts. One such parameter is the partition ratio, which reflects the number of effective catalytic turnovers per inactivation cycle (Tudela et al. 1987). From the values obtained in this study, including those from the variant G84E obtained by random mutagenesis (discussed later), it can be concluded that the single variant (W59A) is the most efficient catalyst of those proteins studied. This  114  result reflects a significant increase in the rate of catalysis resulting from the destabilizing mutation despite the corresponding increased rate of inactivation. The structural and. electronic differences identified for this variant by U V - V i s , C D , M C D and N M R spectroscopy suggest an overall more disordered structure for the W59A variant as expected from the relatively drastic nature of this substitution. This disorder correlates with an increase in the catalytic efficiency during the peroxidase reaction, as observed for denatured cytochrome c-550 (Diederix et al. 2002). On the other hand, the stabilizing substitutions in N52I/Y67F produced a protein with increased stability as indicated by the values of pK , T and thermodynamic parameters that can be associated a  m  with a very low catalytic efficiency. The N52I/Y67F/W59A variant represents an interesting situation in which the effect of the destabilizing substitutions and the stabilizing effects of the double mutation are combined in one protein. The spectroscopic results and the higher pK and A H values relative to those of the wild-type protein a  indicate that the N52I/Y67F substitutions can stabilize the Fe-Met linkage and the heme environment of the protein significantly even in the presence of the highly destabilizing W59A substitution. Nevertheless, the triple variant does exhibit diminished thermal stability (lower T ) and therefore the proposed influence of the Fe-Met linkage on this m  value (Yamamoto et al. 2002) is less significant in this protein. In addition, the considerable difference in AS for the reduction of the proteins suggests that structural disorder is one of the factors responsible for the observed differences in reduction potential (Komarpanicucci et al. 1992). The results of this balance, in kinetic terms, are higher values of fa, fa, partition ratio and  &; ct na  constants relative to the corresponding  values for the wild-type protein. In this context, it is interesting to observe the presence of  115  additional resonance peaks observed exclusively in the H N M R spectrum of the !  N52I/Y67F/W59A variant. Such variations suggested a modified electronic environment for the H M 3 , H M 8 and HP7 heme substituent protons that could be related to the modified kinetic characteristics of the protein. Interestingly, the computer  analysis of the  structure  simulation for the  N52I/W59A/Y67F variant (Figures 33a and 33b), suggest changes in the electronic environment of such protons that could potentially be related to the differences in the structure and/or catalytic behavior. Specifically in the region around H M 8 (Figure 33a) the orientation of Argl3 and the absence of a water molecule (W107) produces the rupture of a hydrogen bond between those molecules in the N52I/W59A/Y67F variant. In the other region, around H M 3 and HP7 (Figure 33b) the replacement of Trp59 by an alanine, in addition to the creation of a cavity as discussed earlier, causes the side chain of Ala59 to be in place to form a novel hydrogen bond interaction with Arg38. These changes are also likely to affect the interaction between these residues and the propionate groups, which as observed also in the same Figure, are slightly changed in their orientation within the heme group. Notably, electrostatic interactions between the heme propionates, Arg38 and the iron atom have been argued to influence the reduction potential of the heme iron and related conformational changes of cytochrome c (Moore 1983). However, more extensive structural studies employing X-ray crystallography or more detailed N M R experiments are required to characterize the structure and heme electronic environment of N52IY67FW59A in sufficient detail to permit definitive structural correlations with the catalytic properties reported here.  116  Figure 33a. Stereo drawing showing the region around H M 3 in the W T and N52I/W59A/Y67F variants. The structure for the variant (dark gray) was simulated by computer analysis, while the structure o f the W T protein (light gray) was taken from the PDB bank (file 2 Y C C ) . A broken line indicates the hydrogen bond discussed in the text.  Figure 33b. Stereo drawing of the region around H M 8 and H P 7 in the W T and N52I/W59A/Y67F proteins. Same structure analysis as before (fig 31a), showing some o f the potential differences upon the introduction o f the mutations.  117  II. Random mutagenesis Optimization of enzyme catalytic activity by site directed mutagenesis requires at the least a detailed understanding of the structural elements that are required for efficient catalysis as for maintenance of the three-dimensional structure of the protein. Even with this knowledge, optimal design requires simultaneous adjustment of a great number of structural and functional issues that is, realistically, not possible to achieve. This situation is further complicated by the fact that it is highly unlikely that all relevant structural and functional issues can even be identified. The identification of useful mutations by generation of a large number of randomly-generated variants followed by selection of desirable variants by activity screening provides a strategy by which the limitations in knowledge, time, and labor that impede the use of structure-based design can be overcome. The success of this approach, often referred to as directed evolution, as represented by the work of Arnold and co-workers (Kuchner and Arnold 1997; Joo et al. 1999; Salazar et al. 2003) has resulted in its adaptation for use with a variety of enzymes. The goal of the present implementation of directed evolution in the use of cytochrome c as a scaffold for development of a robust peroxidase was to overcome two crucial limitations exhibited by the peroxidase activity of the cytochrome. The first of these limitations is the instability of the protein to inactivation by hydrogen peroxide, and the second is the intrinsically low peroxidase activity of cytochrome c. Considering that even true peroxidases are inactivated during turnover by oxidative damage resulting from exposure to the peroxide substrate, this first goal was regarded as the more challenging to overcome. In part for this reason, we decided to focus on selection of variants with greater catalytic activity. Optimizing this characteristic of the protein seemed more  118  promising owing to the great range of chemical activities exhibited by the heme prosthetic group in various protein environments (Dawson 1988). In addition, success with screening assays to detect peroxidase activity has been reported previously in the literature (Wan et al. 1998; Cherry et al. 1999; Joo et al. 1999; Farinas et al. 2001) while the design of an assay for stability to heme degradation by peroxide was less readily apparent. The first step in this procedure involved selecting and adjusting the desired mutagenesis level. A n average of 2 mutations per gene was chosen based on the small size of the cytochrome c gene and from concern regarding the high probability of introducing non-fuctional mutations in such a short D N A sequence. A low mutagenesis rate is normally preferred in such work because the number of variations in the sequence is greater that the actual number of variants that can be screened (Valetti and Gilardi 2004). The results obtained from the sequencing of representative clones exhibited a high degree of silent mutations and transitions, suggesting creation of a biased library of variants. These limitations could be overcome by using various DNTP mixtures to prevent the bias towards transitions and by increasing the mutagenesis rate by increasing the concentration of M n C b or by related methods (Vartanian et al. 1996). After achieving the desired number of mutations per variant, the next challenge was to obtain a positive result from the screening reaction. Consequently, a series of peroxidase substrates were tested, and it was concluded that in terms of sensitivity, solubility, and colour difference, the best combination consisted in the use of H C>2 and 2  ABTS. Other substrates such as o-dianisidine provided sensitivity comparable to that of ABTS, but the latter substrate exhibited a lower background colour. Even though this  119  substrate has limitations for kinetic analyses (as discussed previously), it was the best choice for the screening reaction. A n additional challenge to the development of an effective screening assay was related to the expression of the protein and the necessity of releasing the proteins from the cells to allow reaction with the (extracellular) substrate to proceed. The initial screening protocol for cytochrome c was designed on the basis of the previous use of the method with myoglobin (Wan et al. 1998). However, after repeated failed attempts, it was concluded that this procedure is inappropriate for use in the current application as the combined result of a very low catalytic activity and a relatively low protein expression level. Thus, the screening protocol required modification. The screening protocol presented several limitations especially in terms of the Tabor required and the number of clones that could be screened; however a similar method for evolving the hydroxylation activity of cytochrome P450 has been applied successfully (Schwaneberg et al. 2001). In addition, the restrictions imposed by our method could potentially be solved through use of robotics, which have been effectively applied in other screening strategies (Valetti and Gilardi 2004). Despite these problems, this method permitted identification of two variants of cytochrome c with increased peroxidase activity. One of these variants, G84E, produced a positive reaction after repeated dilutions of the screening solution and was, therefore, selected for further study.  A. The G84E variant To gain insight into the origin of the greater peroxidase activity of the G84E variant, the same kinetic and spectroscopic methods used to study the previous site-  120  directed variants, were used to characterize this variant and compare it with the wild-type cytochrome. The first comparison was made in terms of the kinetics constants, in part to confirm that this variant is, in fact, more active than the wild-type protein. The values obtained for the rate constants k\, k-$, the partition ratio as well as k  mzct  established that,  indeed, this variant is more active than the wild-type protein. The increases in the rate of formation of compound I and breakdown of compound II are not as great as those observed for W59A; nevertheless, the value of k for the G84E variant is much greater 3  than the value observed for the wild-type protein, suggesting an increase in the accessibility of the active site of this variant to the reducing substrate (Moffet et al. 2000). It is also interesting to note that the G84E variant also exhibits greater values for k\ and the partition ratio than does the N52I/Y67F/W59A triple variant but lower values for k and &j . The larger partition ratio observed for the G84E variant probably results 3  nact  from a combination of the increase in the rate of formation of compound I and the decrease in the inactivation rate constant relative to the corresponding values for the triple variant. Comparison of the spectroscopic properties of the G84E variant and the wild-type protein that were evaluated in the current study reveals no obvious differences. The electronic absorption, C D , M C D and even the N M R spectra of the two proteins were virtually identical, implying that the overall conformation of the polypeptide chain, the coordination of the iron atom and the electronic environment of the prosthetic heme group are not perturbed significantly by this substitution. On the other hand, the G84E substitution did result in some distinct changes in the stability of the protein towards thermal denaturation, altered pH, and the presence of  121  H2O2,  such that the variant exhibited a decreased value for T and the pK for the alkaline m  conformation transition and a decreased  a  stability toward exposure to peroxide.  Surprisingly, despite the notable difference in the reduction potential of the variant, no significant change was observed in the thermodynamic parameters for the oxidationreduction equilibrium. The small difference in the standard free entropy (AS) caused by the substitution suggests that the altered potential does not result from altered protein structure or disorder (Komarpanicucci et al. 1992) in agreement with the small change observed in T . The decreased reduction potential (which is of the same magnitude m  observed for the N52I/Y67F variant) suggests a relative stabilization of the oxidized form of the protein that presumably results from introduction of a negatively charged residue adjacent to the heme prosthetic group (Lett and Guillemette 2002). Overall, these results point toward small, discrete changes around specific areas of the protein rather than major structural differences as the result of the G84E substitution. This absence of significant structural changes as suggested by these results could be explained at least partially by the location of Gly84 on the protein surface in a region where the heme becomes partially exposed to the solvent; this is located near Phe82 and Argl3 and on the same side of the heme as the Met80 ligand (Figure 34). In the wild-type protein, Gly84 forms a hydrogen bond with Argl3 and possibly to a solvent water molecule. Introduction of a glutamyl residue at position 84 will alter this H-bonding interaction and could result in formation of alternative interactions.  122  Figure 34. Stereo-drawing from the region around A r g l 3 in the W T protein. A r g l 3 (top part of the figure) hydrogen bonds (broken lines) to the oxygen atom o f W107 (black molecule) and to the oxygen o f the carbonyl group (amide backbone) o f Gly84 (gray residue on the left). A section o f the heme group (lighter gray on the right) is shown for orientation purposes.  Simulation o f the effects of G 8 4 E on the structure o f cytochrome c, and its comparison with the W T protein (Figure 35), raise important points especially i n terms of the differences around three main areas of the protein. First, the angle o f Pro71 in G 8 4 E is significantly altered in comparison with its position in the W T protein residue (Figure 35). This amino acid is directly adjacent to residue Lys73 which has been identified as an alternative ligand to the heme during the alkaline transition of ferricytochrome c (Rosell 1998). Therefore, modifications in the orientation o f Pro71 may affect the flexibility o f this polypeptide region (required for the rearrangement of the ligand) that could be reflected i n the differences of p ^ a value observed in our experiments.  123  Second, the orientation of the carbonyl group in the amide backbone of Tyr67 changed notably (Figure 35). Even though only the peptide backbone of this residue seems altered, and the role of Tyr67 in the peroxidase reaction is still not clear, such changes could potentially affect the behavior of this amino acid in the catalytic mechanism and explain some of the kinetic differences observed for G84E. In addition, and as a consequence of the introduction of a much bulkier residue in G84E, the orientation of the two water molecules around Argl3 is expected to change considerably (Figure 35). This anticipated change is very likely to affect the hydrogen network around these groups, perhaps modifying the access of the substrate during the catalytic reaction, as suggested by our kinetic data. Whatever structural change results, it is reasonable to propose that one outcome would be a modified environment for the region around Argl3 that results in improved access of a reducing substrate to the heme prosthetic group. Notably, the region around Argl3 has been proposed to be involved in some of the oxidation-state linked structural differences of the cytochrome (Berghuis and Brayer 1992) that could also be related to the observed alteration in catalytic behavior of this variant.  124  Figure 35. Structural changes produced by the G84E as estimated by molecular simulation.  The residues for the G 8 4 E variant are colored in black  while the W T amino acids are represented in gray. A ) Superposition o f residues Gly84 and Glu84 showing the similar backbone alignment o f both residues. B ) Difference in the orientation of Pro 71. C) Changes in the backbone of T y r 67. D) Different orientation of the water molecules H O H 3 0 3 (bottom left) and H O H 1 0 7 (right) close to A r g 13 (top right).  III. Inactivation reaction To characterize the peroxidase activity o f cytochrome c or any heme-containing peroxidase catalyst adequately, the inactivation reaction that occurs i n the presence o f hydrogen peroxide must be understood. A s discussed above, although the inactivation of  125  many heme enzymes that use  H2O2  as a substrate has been recognized for many years, the  specific products and detailed mechanisms for their generation remain uncharacterized. For this reason, one of the objectives of this work was to identify at least some of these products in an effort to gain greater insight into the mechanism of inactivation and the manner in which this process influences the overall catalytic mechanism of the protein. One of the consequences of the oxidative reaction on the catalyst is the significant bleaching or decrease in absorbance of the heme prosthetic group as reflected by the absorbance intensity of the Soret band (Florence 1985). Notably, heme proteins and heme enzymes appear to vary considerably in their susceptibility to heme bleaching (Valderrama et al. 2002). In the case of cytochrome c and other non-natural peroxidases, the influence of this reaction on the overall catalytic activity of the protein is usually more significant than for true peroxidases owing to the low level of catalytic activity exhibited by the non-natural peroxidases. In general, peroxide-induced heme bleaching is assumed to be the result of direct modification (presumably oxidation) of the heme prosthetic group, particularly for those proteins such as cytochrome c in which the heme group is bound covalently to the protein and, thus, is not readily removed from the protein. This expectation is consistent with the finding that iron is released from cytochrome c in the presence of a peroxide-generating system (Harel et al. 1988; Villegas et al. 2000); however the chemical nature of the modification to the heme group and/or to the protein induced by exposure to peroxide has not been clearly established. In the present work, reaction of hydrogen peroxide with ferricytochrome c resulted in the formation of a free radical species during the oxidation of the protein that is absent (or non-detectable) under identical conditions when the N52I/Y67F double  126  variant is substituted for the wild-type protein (Villegas et al. 2000). Several previous reports have, in fact, demonstrated that free radical species can be formed at tyrosyl residues following exposure of horse heart cytochrome c to  H2O2  (Barr et al. 1996; Qian  et al. 2002). Even though the radical species observed in our work could be similar to those observed for the horse heart protein, the use of slightly different reaction conditions, the use of yeast rather than horse heart cytochrome, and the inability to observe an immobilized nitroxide by ESR spectroscopy after performing this reaction in the presence of various nitroso spin-traps reagents (data not shown) could also imply that the chemical functionalities involved in this case are not the same. It is worth noting that chelating agents and care in eliminating or avoiding contamination by trace metal ions made no difference to our attempts to identify a peroxide-induced protein-centered radical. The fact that the bleaching reaction is inhibited by the presence of the CN" ion suggests a direct involvement of the heme iron in this process. This latter point, however, needs to be examined more carefully by performing additional experiments. Another reported consequence of oxidizing horse heart cytochrome c with  H2O2  is  the oligomerization of the protein (Harel et al. 1988). This result was also observed for the yeast cytochrome in the current work, and it was observed that the extent of oligomerization depends on the molar ratio of  H2O2  used. This finding was consistent  with the kinetics experiments where the shape of the reaction progress curve, and presumably the mechanism of the bleaching reaction, also varies as the ratio of H 2 O 2 to cytochrome is varied. This latter conclusion is consistent with the report that the identity of the radical centres generated in ferricytochrome varies depending on whether  H2O2  is  generated enzymatically (low concentration) or added directly to the reaction (high  127  concentration) (Harel et al. 1988). For this reason, a detailed kinetic analysis of this reaction under a variety of conditions that includes the characterization of the free radical centres is essential for a comprehensive understanding of the reaction of peroxide and ferricytochrome c. In the present studies, numerous attempts were made to isolate the free radicals produced by the reaction with diverse spin traps. The most promising results were obtained with the spin-trapping reagent POBN, which led to the observation of a protein adduct of the anticipated molecular weight. Unfortunately, repeated attempts to isolate a modified tryptic peptide containing this modification were unsuccessful. This difficulty is presumably a consequence of the low yield of POBN-modified cytochrome c under the wide range of conditions evaluated in this work or it results from unexpected instability of the modification to the conditions of the analytical strategy used in this work. A n optimized protocol that allows the isolation of the adduct from the unmodified protein would very likely solve this problem.  128  IV. Future directions Even though some insight into the peroxidase reaction of wo-1-yeast cytochrome c has been obtained from this work, several important aspects require additional investigation for the catalytic activity of this protein to be understood in a more comprehensive way. One of the aspects that may require further work is the kinetic analysis of the peroxidase reaction, especially to unify or to clearly establish the differences between various heme proteins. Our analysis seemed adequate to compare the variants studied in this work and to compare them with "natural" peroxidase enzymes. However, differences with other examples of peroxidase activity among cytochromes and heme proteins suggest that such a mechanism could be more complicated and requires further study. This complexity is also reflected in the inactivation reaction. Again, even though our analysis allowed us to compare the properties of the different variants, the lack of results in the isolation of radical reaction products did not permit us to go further into the study of such reaction. The spin-trapping analysis, combined with liquid chromatography and mass spectrometry has proven to be very useful in the case of other cytochromes. In our case, perhaps the use of alternative spin-traps or a more detailed EPR analysis, including measurements at very low temperature, would provide more insight into this very important aspect of the peroxidase chemistry. In terms of the details in the protein architecture that dictate the differences in the catalytic activity of each variant, more specialized structural methods such as 2D N M R and crystallography are likely to provide a further understanding of the basics of such  129  dissimilarities. Either of the variants N52I/W59A/Y67F or G84E would be suitable candidates for such studies. Finally, regarding the possible physiological relevance of this peroxidase reaction involving cytochrome c would require modified set of experiments. In general the reaction conditions used in this study were chosen to optimize the catalytic process, and they were not selected to resemble the conditions encountered inside the mitochondrial membrane. Therefore, the careful selection of reaction conditions, in conjunction with more physiological models, for example, cell cultures, is required to allow a more detailed understanding of the role of cytochrome c in peroxidative reactions in a cellular environment.  130  V. REFERENCES Abelskov, A . K.; A . T. Smith; C. B. Rasmussen; H . B . Dunford; Welinder, K . G. (1997). pH dependence and structural interpretation of the reactions of coprinus  cinereus  peroxidase with hydrogen peroxide, ferulic acid, and 2,2'-azinobis(3-ethylbenzthiazoline6-sulfonic acid). Biochemistry 36: 9453-9463. Akasaka, R.; T. Mashino; Hirobe, M . (1993). Cytochrome-P450-like substrate oxidation catalyzed by cytochrome c and immobilized cytochrome c. Arch Biochem Biophys 301: 355-360. Alexeeva, M . ; A . Enright; M . J. Dawson; M . Mahmoudian; Turner, N . J. (2002). Deracemization of alpha-methylbenzylamine using an enzyme obtained by in vitro evolution. Angew Chem Int Edit 41: 3177-3180. Ambler, R. P. (1991). Sequence variability in bacterial cytochromes c. B B A 1058: 42-47. Anni, H . Y . Israel (1999). Characterization of adducts of ethanol metabolites with cytochrome c. Alcohol clin exp res 23: 26-31. Assfalg, M . ; Bertini, I.; Dolfi, A . ; Turano, P.; Mauk, A . G.; Rosell, F. I.; Gray, H . B . (2003). Structural model for an alkaline form of ferricytochrome c. J A m Chem Soc 125: 2913-2922. Atlante, A . ; Calissano, P.; Bobba, A.;. Azzariti, A . ; Marra, E. Passarella, S. (2000). cytochrome c is released from mitochondria in a reactive oxygen species (ROS)dependent fashion and can operate as a ROS scavenger and as a respiratory substrate in cerebellar neurons undergoing excitotoxic death. J Biol Chem 275: 37159-37166. Banci, L. (1997). Structural properties of peroxidases. J Biotechnol 53: 253-263.  131  Banci, L.; Bertini, L; Bren, K . L.; Gray, H . B.; Sompornpisut, P.; Turano, P. (1997a). Solution  structure  of  oxidized  Saccharomyces  cerevisiae  iso-1-cytochrome  c.  Biochemistry 36: 8992-9001. Banci, L.; Bertini, I.; Gray, H . B.; Luchinat, C.; Reddig, T.; Rosato, A . ; Turano, P. (1997b). Solution structure of oxidized horse heart cytochrome c. Biochemistry 36: 98679877. Barr, D. P.; Gunther, M . R.; Deterding, L. J.; Tomer, K . B.; Mason, R. P. (1996). ESR spin-trapping of a protein-derived tyrosyl radical from the reaction of cytochrome c with hydrogen peroxide. J Biol Chem 271: 15498-15503. Barriault, D.; Plante, M . M . ; Sylvestre, M . (2002). Family shuffling of a targeted bphA region to engineer biphenyl dioxygenase. J Bacteriol 184: 3794-3800. Battistuzzi, G.; Borsari, M . ; Cowan, J. A . ; Ranieri, A . ; Sola, M . (2002). Control of cytochrome c redox potential: Axial ligation and protein environment effects. J A m Chem Soc 124: 5315-5324. Battistuzzi, G.; Borsari, M . ; Loschi, L.; Menziani, M . C ; De Rienzo, F.; Sola, M . (2001a). Control of metalloprotein reduction potential: The role of electrostatic and solvation effects probed on plastocyanin mutants. Biochemistry 40: 6422-6430. Battistuzzi, G.; Borsari, M . ; Ranieri, A.; Sola, M . (2001b). Axial ligand and polypeptide matrix effects on the reduction thermodynamics of cytochrome c. J Inorg Biochem 86: 515-515. Bendall, D. S. (2004). The unfinished story of cytochrome f. Photosynth Res 80: 265276.  132  Berghuis, A . M . ; Brayer, G. D. (1992). Oxidation state-dependent conformationalchanges in cytochrome c. J M o l Biol 223: 959-976. Berghuis, A . M . ; Guillemette, J. G.; Mclendon, G.; Sherman, F.; Smith, M . ; Brayer, G. D. (1994). The role of a conserved internal water molecule and its associated hydrogenbond network in cytochrome c. J M o l Biol 236: 786-799. Bhuyan, A . K.; Rao, D. K.; Prabhu, N . P. (2005). Protein folding in classical perspective: Folding of horse cytochrome c. Biochemistry 44: 3034-3040. Bixler, J.; Bakker, G.; Mclendon, G. (1992). Electrochemical probes of protein folding. J A m Chem Soc 114: 6938-6939. Black, K . M . ; Clark-Lewis, I.; Wallace, C. J. A . (2001). Conserved tryptophan in cytochrome c: importance of the unique side-chain features of the indole moiety. Biochem J 359: 715-720. Blumenthal, D. C ; Kassner, R. J. (1980). Cyanide binding to the cytochrome c ferric heme octapeptide - a model for anion binding to the active-site of high-spin ferric hemeproteins. J Biol Chem 255: 5859-5863. Bornscheuer, U . T. (1998). Directed evolution of enzymes. Angew Chem Int Edit37: 3105-3108. Borsari, M . ; Dikaya, E.; Dikiy, A . ; Gonchar, M . V . ; Maidan, M . M . ; Pierattelli, R. A . ; Sibirny, A . (2000). Isolation and physico-chemical characterization of a cytochrome c from the methylotrophic yeast Hansenula polymorpha. BBA-Protein Struc M 1543: 174188.  133  Bren, K . L.; Gray, H . B . (1993). Structurally engineered cytochromes with novel ligandbinding sites - oxy and carbonmonoxy derivatives of semisynthetic horse heart ala80 cytochrome c. J A m Chem Soc 115: 10382-10383. Bums, P. D.; Lamar, G. N . (1979). Detection of localized conformational flexibility in horse heart cytochrome-c by proton nuclear magnetic-resonance. J A m Chem Soc 101: 5844-5846. Burns, P. D.; Lamar, G. N . (1981). Characterization of conformational heterogeneity in the heme pocket of ferricytochrome c using high-field proton nuclear magnetic-resonance spectroscopy. J Biol Chem 256: 4934-4939. Burton, S. G. (2003). Oxidizing enzymes as biocatalysts. Trends Biotechnol 21: 543-549. Cadwell, R., Joyce, G. (1991). Randomization of genes by PCR mutagenesis. PCR-Meth App 2: 28-33. Caffrey, M . S.; Cusanovich, M . A . (1993). Role of the highly conserved tryptophan of cytochrome c in stability. Arch Biochem Biophys 304: 205-208. Caffrey, M . S.; Cusanovich, M . A. (1994). Site-specific mutagenesis  studies of  cytochromes c. BBA-Bioenergetics 1187: 277-288. Cassina, A . M . ; Hodara, R.; Souza, J. M . ; Thomson, L.; Castro, L.; Ischiropoulos, H . ; Freeman, B. A . ; Radi, R. (2000). Cytochrome c nitration by peroxynitrite. J Biol Chem 275: 21409-21415. Castro, L.; Eiserich, J. P.; Sweeney, S.; Radi, R.; Freeman, B. A . (2004). Cytochrome c: a catalyst and target of nitrite-hydrogen peroxide-dependent  protein nitration. Arch  Biochem Biophys 421: 99-107.  134  Chen, Y . R.; Deterding, L . J.; Sturgeon, B. E.; Tomer, K . B . ; Mason, R. P. (2002). Protein oxidation of cytochrome c by reactive halogen species enhances its peroxidase activity. J Biol Chem 277: 29781-29791. Cherry, J. R.; Lamsa, M . H . ; Schneider, P.; Vind, J.; Svendsen, A.; Jones, A.; Pedersen, A. H . (1999). Directed evolution of a fungal peroxidase. Nat Biotechnol 17: 379-384. Chica, R. A.; Doucet, N . ; Pelletier, J. N . (2005). Semi-rational approaches to engineering enzyme activity: combining the benefits of directed evolution and rational design. Curr Opin Biotech 16: 378-384. Chottard, G.; Michelon, M . ; Herve, M . ; Herve, G. (1987). Modulation of the properties of ferricytochrome c by electrostatic interactions with heteropolytungstates. Reel Trav Chim Pay B 106: 285-285. Churg, A . ; Warshel, K . A . (1986). Control of the redox potential of cytochrome c and microscopic dielectric effects in proteins. Biochemistry 25: 1675-1681. Conesa, A . ; Punt, P. J.; van den Hondel, C. A . M . J. J. (2002). Fungal peroxidases: molecular aspects and applications. J Biotechnol 93: 143-158. Cutler, R. L.; Davies, A. M . ; Creighton, S.; Warshel, A.; Moore, G. R.; Smith, M . ; Mauk, A. G. (1989). Role of arginine-38 in regulation of the cytochrome c oxidation-reduction equilibrium. Biochemistry 28: 3188-3197. Das, G.; Hickey, D. R.; Mclendon, D.; Mclendon, G.; Sherman, F. (1989). Dramatic thermostabilization of yeast iso-1 -cytochrome c by an asparagine-]isoleucine replacement at position-57. P Natl Acad Sci U S A 86: 496-499.  135  Davies, A . M . ; Guillemette, J. G.; Smith, M . ; Greenwood, C.; Thurgood, A . G. P.; Mauk, A. G.; Moore, G. R. (1993). Redesign of the interior hydrophilic region of mitochondrial cytochrome c by site-directed mutagenesis. Biochemistry 32: 5431-5435. Davies, M . J.; Hawkins, C. L . (2004). EPR spin trapping of protein radicals. Free Radical Bio Med 36: 1072-1086. Dawson, J. H . (1988). Probing structure-function  relations  in heme-containing  oxygenases and peroxidases. Science 240: 433-439. Deterding, L . J.; Barr, D. P.; Mason, R. P.; Tomer, K . B. (1998). Characterization of cytochrome c free radical reactions with peptides by mass spectrometry. J Biol Chem 273: 12863-12869. Diederix, R. E. M . ; Fittipaldi, M . ; Worrall, J. A . R.; Huber, M . ; Ubbink, M . ; Canters, G. W. (2003). Kinetic stability of the peroxidase activity of unfolded cytochrome c: Heme degradation and catalyst inactivation by hydrogen peroxide. Inorg Chem 42: 7249-7257. Diederix, R. E. M . ; Ubbink M . ; Canters, G. W. (2001). The peroxidase activity of cytochrome c-550 from Paracoccus versutus. Eur J Biochem 268: 4207-4216. Diederix, R. E. M . ; Ubbink, M . ; Canters, G. W. (2002). Effect of the protein matrix of cytochrome c in suppressing the inherent peroxidase activity of its heme prosthetic group. Chembiochem3: 110-112. Duggleby, R. G. (1986). Progress curves of reactions catalyzed by unstable enzymes - a theoretical approach. J Theor Biol 123: 67-80. Dunford, B. (1999). Heme Peroxidases. New York, John Wiley & Sons. Farinas, E. T.; Schwaneberg, U . ; Glieder, A.; Arnold, F. H . (2001). Directed evolution of a cytochrome P450 monooxygenase for alkane oxidation. Adv Synth Catal 343: 601-606.  136  Fedurco, M . ; Augustynski, J.; Indiani, C ; Smulevich, G.; Antalik, M . ; Bano, M . ; Sedlak, E.; Glascock, M . C.; Dawson, J. H . (2004). The heme iron coordination of unfolded ferric and ferrous cytochrome c in neutral and acidic urea solutions. Spectroscopic and electrochemical studies. BBA-Proteins Proteom 1703: 31-41. Feng, Y . Q.; Roder, H . ; Englander S. W. (1990). Redox-dependent structure change and hyperfme nuclear-magnetic-resonance  shifts in cytochrome-c. Biochemistry 29: 3494-  3504. Fenwick, C. W.; English, A . M . (1996). Trapping and L C - M S identification of protein radicals formed in the horse heart metmyoglobin-H 02 reaction. J A m Chem Soc 118: 2  12236-12237. Ferguson, S. J. (2001). Keilin's cytochromes: How bacteria use them, vary them and make them. Biochemical Soc T 29: 629-640. Fetrow, J. S.; Horner, S. R.; Oehrl, W.; Schaak, D. L.; Boose, T. L.; Burton, R. E. (1997). Analysis of the structure and stability of omega loop A replacements in yeast iso-1 cytochrome c. Protein Sci 6: 197-210. Filosa, A . (2001a). Study of locally unfolded forms of cytochrome c b fourier transform infrared and of  H2O2  mediated oxidation of ferricytochrome c and metmyoglobin by mass  spectrometry. Chemistry and biology. Montreal, Concordia. Filosa, A . (2001b). Study of locally unfolded forms of cytochrome c by fourier transform infrared and of mass  H2O2  spectrometry.  mediated oxidation of feni-cytochrome c and metmyoglobin by Chemistry and Biochemistry. Montreal, Quebec,  Concordia  University: 188.  137  Flatmark, T. (1962). Demonstration of peroxidase activity in the rat kidney and liver: peroxidase effect of cytochrome c. Nature 196: 894-895. Florence, T. M . (1985). The degradation of cytochrome c by hydrogen-peroxide. J Inorg Biochem 23: 131-141. Frauenhoff, M . M . ; Scott, R. A . (1992). The role of tyrosine-67 in the cytochrome c heme crevice structure studied by semisynthesis. Proteins 14: 202-212. Fujita, A.; Senzu, H.; Kunitake, T.; Hamachi, I. (1994). Enhanced peroxidase-activity of cytochrome c by phosphate bilayer-membrane. Chem Lett: 1219-1222. Gadsby, P. M . A . ; Peterson, J.; Foote, N . ; Greenwood, C ; Thomson, A . J. (1987). Identification of the ligand-exchange process in the alkaline transition of horse heart cytochrome c. Biochem J 246: 43-54. Gadsby, P. M . A.; Thomson, A . J. (1990). assignment of the axial ligands of ferric ion in low-spin hemoproteins by near-infrared magnetic circular-dichroism and electronparamagnetic resonance spectroscopy. J A m Chem Soc 112: 5003-5011. Gao, Y . ; Boyd, J.; Pielak G. J. R. J. P. Williams (1991a). Comparison of reduced and oxidized yeast iso-1-cytochrome c using proton paramagnetic shifts. Biochemistry 30: 1928-1934. Gao, Y . ; J. Boyd; G. J. Pielak; Williams, R. J. P. (1991b). Proton nuclear-magneticresonance as a probe of differences in structure between the C102T and F82S, C102T variants of iso-1-cytochrome c from the yeast saccharomyces-cerevisiae. Biochemistry 30: 7033-7040. Gao, Y . ; Boyd, J.; Williams, R. J. P.; Pielak, G. J. (1990). Assignment, of proton resonances, identification of secondary structural elements, and analysis of backbone  138  chemical-shifts for the C102T variant of yeast iso-1-cytochrome c and horse cytochrome c. Biochemistry 29: 6994-7003. Ghibaudi, E.; Laurenti, E. (2003). Unraveling the catalytic mechanism of lactoperoxidase and myeloperoxidase - A reflection on some controversial features. Eur J Biochem 270: 4403-4412. Greenwood, C.; Wilson, M . T. (1971). Ferricytochrome c. 1. Effect of pH, ionic strength and protein denaturants on spectra of ferricytochrome c. Eur J Biochem 22: 5-10. Guex, N . ; Peitsch, M . C. (1997). SWISS-MODEL and the Swiss-PdbViewer: A n environment for comparative protein modeling. Electrophoresis 18: 2714-2723. Guillemette, J. G.; Matsushimahibiya, Y . ; Atkinson, T.; Smith, M . (1991). Expression in escherichia-coli of a synthetic gene coding for horse heart myoglobin. Protein Eng 4: 585-592. Hampsey, D. M . ; Das, G. Sherman, F. (1986). Amino-acid replacements in yeast iso-1cytochrome-c - comparison with the phylogenetic series and the tertiary structure of related cytochromes-c. J Biol Chem 261: 3259-3271. Harayama, S. (1998). Artificial evolution by D N A shuffling. Trends Biotechnol 16: 7682. Harel, S.; M . Salan, A . ; Kanner, J. (1988). Iron release from metmyoglobin, methemoglobin and cytochrome c by a system generating hydrogen-peroxide. Free Radical Res Com 5: 11-19. Hawkins, B. K.; Hilgenwillis, S.; Pielak, G. J.; Dawson, J. H . (1994). Novel axial ligand interchange in cytochrome c - incorporation of a histidine at position-82 leads to displacement of the wild-type methionine-80 ligand. J A m Chem Soc 116: 3111-3112.  139  Hibbert, E. G.; Baganz, F.; Hailes, H . C.; Ward, J. M . ; Lye, G. J.; Woodley, J. M . ; Dalby, P. A . (2005). Directed evolution of biocatalytic processes. Biomol Eng 22: 11-19. Hiner, A . N . P.; Hernandez-Ruiz, J.; Arnao, M . B . ; Garcia-Canovas, F.; Acosta, M . (1996). A comparative study of the purity, enzyme activity, and inactivation by hydrogen peroxide of commercially available horseradish peroxidase isoenzymes A and C. Biotechnol Bioeng 50: 655-662. Hiner, A . N . P.; Ruiz, J. H.; Lopez, J. N . R.; Canovas, F. G.; Brisset, N . C ; Smith, A . T.; Arnao, M . B.; Acosta, M . (2002). Reactions of the class IJ peroxidases, lignin peroxidase and Arthromyces ramosus peroxidase, with hydrogen peroxide - Catalase-like activity, compound III formation, and enzyme inactivation. J Biol Chem 277: 26879-26885. Horecker, B . L. K., A . (1946). The cytochrome c-cyanide complex. J Biol Chem 165: 1120. Iverson, B. L.; Breaker, R. R. (1998). Be careful what you wish for ... Trends Biotechnol 16: 52-53. Iverson, S. L.; Orrenius, S. (2004). The cardiolipin-cytochrome c interaction and the mitochondrial regulation of apoptosis. Arch Biochem Biophys 423: 37-46. Joo, H . ; Lin, Z. L.; Arnold, F. H : (1999). Laboratory evolution of peroxide-mediated cytochrome P450 hydroxylation. Nature 399: 670-673. Keilin, D. (1966). The history of cell respiration and cytochrome c. Cambridge, Cambridge U . P. Keller, R. M . ; Wuthrich, K . (1978). Assignment of heme c resonances in 360 mhz H  1  N M R spectra of cytochrome c. B B A 533: 195-208.  140  Kikuchi, M . ; Ohnishi K . ; Harayama, S. (2000). A n effective family shuffling method using single-stranded D N A . Gene 243: 133-137. Kolkman, J.A.; Stemmer, W.P.C. (2001). Directed evolution of proteins by exon shuffling. Nat Biotechnol 19: 423-428. Komarpanicucci, S.; Bixler, J.; Bakker, G.; Sherman, F.; Mclendon, G. (1992). Tuning the redox potential of cytochrome c through synergistic site replacements. J A m Chem Soc 114: 5443-5445. Korshunov, S. S.; Krasnikov, B. F.; Pereverzev, M . O.; Skulachev, V . P. (1999). The antioxidant functions of cytochrome c. Febs Lett 462: 192-198. Kuchner, O.; Arnold, F. H . (1997). Directed evolution of enzyme catalysts. Trends Biotechnol 15: 523-530. Kumamaru, T.; Suenaga, H . ; Mitsuoka, M . ; Watanabe, T.; Furukawa, K . (1998). Enhanced degradation of polychlorinated biphenyls by directed evolution of biphenyl dioxygenase. Nat Biotechnol 16: 663-666. Laemmli, U . K . (1970). Cleavage of structural proteins during assembly of head of bacteriophage^. Nature 227: 680-685. Lett, C. M . ; Berghuis, A . M . ; Frey, H . E.; Lepock, J. R.; Guillemette, J. G. (1996). The role of a conserved water molecule in the redox-dependent thermal stability of iso-1 cytochrome c. J Biol Chem 271: 29088-29093. Lett, C. M . ; Guillemette, J. G. (2002). Increasing the redox potential of isoform 1 of yeast cytochrome c through the modification of select haem interactions. Biochem J 362: 281287.  141  Lett, C. M . ; Rosu-Myles, M . D.; Frey, H . E.; Guillemette, J. G. (1999). Rational design of a more stable yeast iso-1-cytochrome c. BBA-Protein Struct M 1432: 40-48. Liggins, J. R.; Lo, T. P.; Brayer, G. D.; Nail, B. T. (1999). Thermal stability of hydrophobic heme pocket variants of oxidized cytochrome c. Protein Sci 8: 2645-2654. Liggins, J. R.; Sherman, F.; Mathews, A . J.; Nail, B . T. (1994). Differential scanning calorimetric study of the thermal unfolding transitions of yeast iso-1 and iso-2 cytochromes c and 3 composite isozymes. Biochemistry 33: 9209-9219. Liu, G. H . ; Shao, W. P.; Zhu, S. M . ; Tang, W. X . (1995). Effects of axial ligand replacement on the redox potential of cytochrome c. J Inorg Biochem 60: 123-131. Louie, G. V.; Brayer, G. D. (1990). High-resolution refinement of yeast iso-1-cytochrome c and comparisons with other eukaryotic cytochromes-c. J Mol Biol 214: 527-555. Luntz, T. L.; Schejter, A . ; Garber, E. A . E.; Margoliash, E. (1989). Structural significance of an internal water molecule studied by site-directed mutagenesis of tyrosine-67 in rat cytochrome c. P Natl Acad Sci U S A 86: 3524-3528. Matsuura, T.; Miyai, K . ; Trakulnaleamsai, S.; Yomo, T.; Shima, Y . ; M i k i , S.; Yamamoto,. K.; Urabe, I. (1999). Evolutionary molecular engineering by random elongation mutagenesis. Nat Biotechnol 17: 58-61. Mauk, A . G.; Coyle, C. L.; Bordignon, E.; Gray, H . B. (1979). bis-(dipicolinate) complexes of cobalt(iii) and iron(ii) as new probes of metalloprotein electron-transfer reactivity - analysis of reactions involving cytochrome c and cytochrome c551. J A m Chem Soc 101: 5054-5056. Mauk, A. G.; Moore, G. R. (1997). Control of metalloprotein redox potentials: What does site-directed mutagenesis of hemoproteins tell us? J Biol Inorg Chem 2: 119-125.  142  Mauk, M . R.; Mauk, A . G. (1988). Characterization of wild-type and mutant forms of yeast iso-1-cytochrome c by reversed-phase high-performance liquid-chromatographic tryptic peptide-mapping. J Chromatogr 439: 408-413. Mcdonald, C. C ; Phillips, W. D. (1973). Proton magnetic resonance studies of horse cytochrome c. Biochemistry 12: 3170-3186. M D B (2003). Metalloprotein Database and Browser. Milne, J. S.; X u Y . J.; Mayne L . C. S.; Englander, W. (1999). Experimental study of the protein folding landscape: Unfolding reactions in cytochrome c. J M o l Biol 290: 811-822. Minshull, J. ; Stemmer, W. P. C. (1999). Protein evolution by molecular breeding. Curr Opin Chem Biol 3: 284-290. Miroux, B.; Walker, J. E. (1996). Over-production of proteins in Escherichia coli: Mutant hosts that allow synthesis of some membrane proteins and globular proteins at high levels. J M o l Biol 260: 289-298. Moench, S. J.; Shi, T. M . J.; Satterlee, D. (1991). Proton N M R studies of the effects of ionic-strength and pH on the hyperfine shifted resonances  and phenylalanine-82  environment of 3 species of mitochondrial ferricytochrome c. Eur J Biochem 197: 631641. Moffet, D. A . ; Certain, L. K.; Smith, A . J.; Kessel, A . J.; Beckwith, K . A . ; Hecht, M . H . (2000). Peroxidase activity in heme proteins derived from a designed combinatorial library. J A m Chem Soc 122: 7612-7613. Moore, G. R. (1983). Control of redox properties of cytochrome c by special electrostatic interactions. FEBS Lett 161: 171-175.  143  Moore, G. R.; Harris, D. E.; Leitch, F. A.; Pettigrew, G. W. (1984). Characterization of ionizations that influence the redox potential of mitochondrial cytochrome c and photosynthetic bacterial cytochromes c2. B B A 764: 331-342. Moore, G. R., Pettigrew, G. W. (1990). Cytochromes c: evolutionary, structural and phvsicochemical aspects. Heidelberg, Springer-Verlag. Moore, G. R.; Williams, G. (1984). Assignment of H N M R resonances of the heme and 1  axial histidine ligand of mitochondrial cytochrome c. B B A 788: 147-150. Moore, G. R.; Williams, R. J. P. (1980). Stability of ferricytochrome c - temperaturedependence of its N M R spectrum. Eur J Biochem 103: 523-532. Moore, J. C ; Jin, H . M . ; Kuchner, O.; Arnold, F. H . (1997). Strategies for the in vitro evolution of protein function: Enzyme evolution by random recombination of improved sequences. J M o l Biol 272: 336-347. Morawski, B.; Quan, S.; Arnold, F. H . (2001). Functional expression and stabilization of horseradish peroxidase by directed evolution in Saccharomyces cerevisiae. Biotechnol Bioeng 76: 99-107. Myer, Y . P. (1968a). Conformation of cytochromes 2. Comparative study of circular dichroism spectra optical rotatory dispersion and absorption spectra of horse heart cytochrome c. J Biol Chem 243: 2115-2122. Myer, Y . P. (1968b). Conformation of cytochromes 3. Effect of urea temperature extrinsic ligands and pH variation on conformation of horse heart ferricytochrome c. Biochemistry 7: 765-776. Myer, Y . P. (1968c). Far ultraviolet circular dichroism spectra of cytochrome c. B B A 154:84-90.  144  Nantes, I. L.; Faljoni-Alario, A.; Vercesi, A . E.; Santos K . E.; Bechara, E. J. H . (1998). Liposome effect on the cytochrome c-catalyzed peroxidation of carbonyl substrates to triplet species. Free Radical Bio Med 25: 546-553. NC-IUB (1991). Electron transport proteins nomenclature. 2005. Nicell, J. A . ; Wright, H . (1997). A model of peroxidase activity with inhibition by hydrogen peroxide. Enzyme Microb Tech 21: 302-310. Nielsen, K . L.; Indiani, C ; Henriksen, A . ; Feis, A . ; Becucci, M . ; Gajhede, M . ; Smulevich, G.; Welinder, K . G. (2001). Differential activity and structure of highly similar peroxidases. lignifying  Spectroscopic, crystallographic, and enzymatic  Arabidopsis thaliana peroxidase  A2 and horseradish  analyses  of  peroxidase A2.  Biochemistry 40: 11013-11021. Oiio, T.; Kawakami, K . ; Goto M . ; Furusaki, S. (2001). Catalytic oxidation of ophenylenediamine by cytochrome c encapsulated in reversed micelles. J M o l Catal B Enzym 11: 955-959. Orrenius, S. (2004). Mitochondrial regulation of apoptotic cell death. Toxicol Lett 149: 19-23. Palmer,  G.; Reedijk, J.  (1991).  Nomenclature  of electron-transfer  proteins  -  recommendations 1989. Eur J Biochem 200: 599-611. Paoli, M . ; Maries-Wright, J.; Smith, A . (2002). Structure-function relationships in hemeproteins. D N A Cell Biol 21: 271-280. Paul, K . G.; Ohlsson, P. I.; Jonsson, N . A . (1982). The assay of peroxidases by means of dicarboxidine on enzyme-linked immunosorbent-assay level. Anal Biochem 124: 102107.  145  Pettigrew, G. W.; Aviram, I.; Schejter, A . (1975). Physicochemical properties of 2 atypical cytochromes c, crithidia cytochrome c-557 and euglena cytochrome c558. Biochem J 149: 155-167. Pielak, G. J.; Oikawa, K.; Mauk, A. G.; Smith, M . ; Kay, C. M . (1986). Elimination of the negative soret cotton effect  of cytochrome c by replacement  of the invariant  phenylalanine using site-directed mutagenesis. J A m Chem Soc 108: 2724-2727. Pollock, W. B . R.; Rosell, F. I.; Twitchett, M . B . ; Dumont M . E.; Mauk, A . G. (1998). Bacterial expression of a mitochondrial cytochrome c. Trimethylation of Lys72 in yeast iso-1-cytochrome c and the alkaline conformational transition. Biochemistry 37: 61246131. Prasad, S.; Maiti, N . C ; Mazumdar, S.; Mitra, S. (2002). Reaction of hydrogen peroxide and peroxidase activity in carboxymethylated cytochrome c: spectroscopic and kinetic studies. BBA-Protein Struct M 1596: 63-75. Prutz, W. A . ; Kissner, R.; Nauser, T.; Koppenol, W. H . (2001). On the oxidation of cytochrome c by hypohalous acids. Arch Biochem Biophys 389: 110-122. Qian, S. Y . ; Chen, Y . R.; Deterding, L. J.; Fann, Y . C ; Chignell, C. F.; Tomer, K . B.; Mason, R. P. (2002). Identification of protein-derived tyrosyl radical in the reaction of cytochrome c and hydrogen peroxide: characterization by ESR spin-trapping, H P L C and M S . Biochem J 363: 281-288. Radi, R.; Thomson, L.; Rubbo, H . ; Prodanov, E. (1991a). Cytochrome c-catalyzed oxidation of organic molecules by hydrogen peroxide. Arch Biochem Biophys 288: 112117.  146  Radi, R.; Turrens, J. F.; Freeman, B . A . (1991b). Cytoclirome c-catalyzed membrane lipid-peroxidation by hydrogen peroxide. Arch Biochem Biophys 288: 118-125. Rafferty, S. P. (1992). Functional properties of active site variants of yeast cytochrome c. Biochemistry and Molecular Biology. Vancouver, B.C., University of British Columbia. Rafferty, S. P.; Guillemette, J. G.; Berghuis, A . M . ; Smith, M . ; Brayer, G. D.; Mauk, A . G. (1996). Mechanistic and structural contributions of critical surface and internal residues to cytochrome c electron transfer reactivity. Biochemistry 35: 10784-10792. Rafferty, S. P.; Pearce, L. L.; Barker, P. D.; Guillemette, J. G.; Kay, C. M . ; Smith, M . ; Mauk, A . G. (1990). Electrochemical, kinetic, and circular dichroic consequences of mutations at position-82 of yeast iso-1-cytochrome c. Biochemistry 29: 9365-9369. Rasmussen, C. B.; Dunford, H . B.; Welinder, K . G. (1995). Rate enhancement of compound-I formation of barley peroxidase by ferulic acid, caffeic acid, and coniferyl alcohol. Biochemistry 34: 4022-4029. Rees, D. C. (1980). Experimental evaluation of the effective dielectric-constant of proteins. J M o l Biol 141: 323-326. Rice, R. H . ; Lee, Y . M . ; Brown, W. D. (1983). Interactions of heme-proteins with hydrogen-peroxide - protein crosslinking and covalent binding of benzb[a]pyrene and 17beta-estradiol. Arch Biochem Biophys 221: 417-427. Rosei, M . A . ; Blarzino, C ; Coccia, R.; Foppoli, C ; Mosca, L . ; Cini, C. (1998). Production of melanin pigments by cytochrome c  H2O2  system. Int J Biochem Cell B 30:  457-463.  147  Rosell, F. I. (1998). The alkaline conformation transition of ferricytochrome c. Biochemistry and Molecular Biology. Vancouver, B.C., University of British Columbia: 218. Rosell, F. I.; Ferrer, J. C ; Mauk, A . G. (1998). Proton-linked protein conformational switching: Definition  of the  alkaline conformational transition  of yeast iso-1-  ferricytochrome c. J A m Chem Soc 120: 11234-11245. Rosen, G. M . ; Britigan, B. E.; Halpern, H.J.; Pou, S. (1999). Free radicals: Biology and detection by sping trapping. New York, Oxford University Press. Salazar, O.; Cirino, P. C ; Arnold, F. H . (2003). Thermostabilization of a cytochrome P450 peroxygenase. Chembiochem 4: 891-893. Santucci, R.; Ascoli, F. (1997). The Soret circular dichroism spectrum as a probe for the heme Fe(III)-Met axial bond in horse cytochrome c. J Inorg Biochem 68: 211-214. Schejter, A.; Aviram, I. (1970). Effects of alkylation of methionyl residues on properties of horse cytochrome c. J Biol Chem 245: 1552-1557. Schellenberger, V . (1998). Directed evolution can improve subtilisin-secreting Bacillus strains. Asm News 64: 634-638. Schwaneberg, U . ; Otey, C ; Cirino, P. C ; Farinas E.; Arnold, F. H . (2001). Cost-effective whole-cell assay for laboratory evolution of hydroxylases in Escherichia coli. J Biomol Screen 6: 111-117. Schweingruber, M . E.; Sherman, F.; Stewart, J. W. (1977). Altered absorption-spectra of iso-1-cytochromes c from mutants of yeast. J Biol Chem 252: 6577-6580.  148  Schweingruber, M . E.; Stewart, J. W.; Sherman, F. (1978). Amino-acid replacements of evolutionarily invariant tryptophan at position 64 in mutant forms of iso-1-cytochrome c from saccharomyces-cerevisiae. J M o l Biol 118: 481-496. Scott, R. A . ;  Mauk, A . G. (1996). Cytochrome c: A multidisciplinary approach.  Sausalito, California, University Science Books. Seelbach, K . ; vanDeurzen, M . P. J.; vanRantwijk, F.; Sheldon, R. A . ; U . Kragl (1997). Improvement of the total turnover number and space-time yield for chloroperoxidase catalyzed oxidation. Biotechnol Bioeng 55: 283-288. Sherman, F. (2005). The importance of mutation, then and now: studies with yeast cytochrome c. Mut Res-Rev Mut 589: 1-16. Sivakolundu, S.; Pochapsky, S.; Mabrouk, P. A . (2001). N M R study of the cytochrome c conformers in mixed solvents. Abstr Pap A m Chem S 221: U743-U743. Sivakolundu, S. G.; Mabrouk, P.A. (2003). Proton N M R study of chemically modified horse heart ferricytochrome c confirms the presence of histidine and lysine-ligated conformers in 30% acetonitrile solution. J Inorg Biochem 94: 381-385. Sivashankar, S. G.; P. Mabrouk, A . (1998). The study of redox activity of cytochrome c in mixed solvents. Abstr Pap A m Chem 216: U154-U154. Stemmer, W. P. C. (1994). Rapid evolution of a protein in-vitro by D N A shuffling. Nature 370(6488): 389-391. Stevens, J. L., Daltrop, O., Allen, J. W. A . ; Ferguson, S. J. (2004). c-type cytochrome formation: chemical and biological enigmas. Accounts Chem Res 37: 999-1007.  149  Stoyanovsky, D. A . ; Cederbaum, A . I. (1998). ESR and H P L E - E C analysis of ethanol oxidation to 1-hydroxyethyl radical: Rapid reduction and quantification of P O B N and P B N nitroxides. Free Radical Bio Med 25(4-5): 536-545. Terui, N . ; Tachiiri, N . ; Matsuo, H.; Hasegawa, J.; Uchiyama, S.; Kobayashi, Y . ; Igarashi, Y.; Sambongi, Y . ; Yamamoto, Y . (2003). Relationship between redox function and protein stability of cytochromes c. J A m Chem Soc 125: 13650-13651. Thurgood, A . G. P.; Davies, A . M . ; Greenwood, C ; Mauk, A . G.; Smith, M . ; Guillemette, J. G.; Moore, G. R. (1991). N M R study of the structural characteristics of variants of yeast iso-1 -cytochrome c in which unvaried aromatic residues have been substituted. Eur J Biochem 202: 339-347. Tinoco, R.; Vazquez-Duhalt, R. (1998). Chemical modification of cytochrome c improves their catalytic properties in oxidation of polycyclic aromatic hydrocarbons. Enzyme Microb Tech 22: 8-12. Torres, E.; Bustos-Jaimes, I.; Le Borgne, S. (2003). Potential use of oxidative enzymes for the detoxification of organic pollutants. Appl Catal B-Environ 46: 1-15. Torres, E.; Sandoval, J. V . ; Rosell, F. I.; Mauk, A . G.; Vazquez-Duhalt, R. (1995). Sitedirected mutagenesis improves the biocatalytic activity of iso-1-cytochrome c in polycyclic-hydrocarbon oxidation. Enzyme Microb Tech 17: 1014-1020. Tsaprailis, G.; Chan, D. W. S.; English, A . M . (1998). Conformational states in denaturants of cytochrome c and horseradish peroxidases examined by fluorescence and circular dichroism. Biochemistry 37: 2004-2016.  150  Tudela, J.; Canovas, F. G.; Varon, R.; Carmona, F. G.; Galvez, J.; Lozano, J. A . (1987). Transient-phase kinetics of enzyme inactivation induced by suicide substrates. B B A 912: 408-416. 1  13  Turner, D. L . (1993). Evaluation of C  and H Fermi contact shifts in horse cytochrome  c - The origin of the anti-curie effect. Eur J Biochem 211: 563-568. Turner, D. L.; Williams, R. J. P. (1993). H N M R and C N M R investigation of redox1  1 3  state-dependent and temperature-dependent conformation changes in horse cytochrome c. Eur J Biochem 211: 555-562. Valderrama, B.; Ayala, M . ; Vazquez-Duhalt, R. (2002). Suicide inactivation of peroxidases and the challenge of engineering more robust enzymes. Chem Biol 9: 555565. Valetti, F.; Gilardi, G. (2004). Directed evolution of enzymes for product chemistry. Nat Prod Rep 21: 490-511. Vartanian, J. P.; Henry, M . ; WainHobson, S. (1996). Hypermutagenic P C R involving all four transitions and a sizeable proportion of transversions. Nucleic Acids Research 24: 2627-2631. Vazquez-Duhalt, R; Semple, K . M . ; Westlake, D. W.; Fedorak, P. M . (1993). Effect of water-miscible organic solvents on the catalytic activity of cytochrome c. Enzyme Microb Technol. 15: 936-942. Vazquez-Duhalt, R; Westlake, D. W.; Fedorak, P. M . (1993). Cytochrome c as a biocatalyst for the oxidation of thiophenes and organosulfides. Enzyme Microb Technol. 15: 494-499.  151  Villegas, J. A . ; Mauk, A . G.; Vazquez-Duhalt, R. (2000). A cytochrome c variant resistant to heme degradation by hydrogen peroxide. Chem Biol 7: 237-244. Vinogradov, S.; Zand, R. (1968). Circular dichroism studies I. Cytochrome c. Arch Biochem Biophys 125: 902-910. Wallace, C. J. A . (1984). The effect of complete or specific partial acetimidylation on the biological properties of cytochrome c and cytochrome c-t. Biochem J 217: 595-599. Wallace, C. J. A . (1993). Understanding cytochrome c function - Engineering proteinstructure by semisynthesis. FASEB Journal 7: 505-515. Wallace, C. J. A.; Clarklewis, I. (1992). Functional-role of heme ligation in cytochrome c - Effects of replacement of methionine-80 with natural and non-natural residues by semisynthesis. J Biol Chem 267: 3852-3861. Wallace, C. J. A.; Corradin, G.; Marchiori, F.; Borin, G. (1986). Cytochrome c chimerae from natural and synthetic fragments - Significance of the biological properties. Biopolymers 25: 2121-2132. Wallace, C. J. A . ; Corthesy, B . E. (1987). Alkylamine derivatives of cytochrome c Comparison with other lysine-modified analogs illuminates structure-function relations in the protein. Eur J Biochem 170: 293-298. Wallace, C. J. A . ; Proudfoot, A . E. I. (1987). On the relationship between oxidation reduction potential and biological-activity in cytochrome c analogs - results from 4 novel 2-fragment complexes. Biochem J 245: 773-779. Wallace, C. J. A . ; Rose, K . (1983). The semisynthesis of analogs of cytochrome c Modifications of arginine residue-38 and residue-91. Biochem J 215: 651-658.  152  Wan, L . L.; Twitchett, M . B.; Eltis, L. D.; Mauk, A . G.; Smith, M . (1998). In vitro evolution of horse heart myoglobin to increase peroxidase activity. P Natl Acad Sci U S A 95: 12825-12831. Williams, G.; Moore, G. R.; Porteous, R.; Robinson, M . N . ; Soffe, N . ; Williams, R. J. P. (1985). Solution structure of mitochondrial cytochrome c 1. H  1  nuclear magnetic-  resonance of ferricytochrome c. J M o l Biol 183: 409-428. Winkler, J. R.; WittungStafshede, P.; Leckner, J.; Malmstrom, B . G.; Gray, H . B. (1997). Effects of folding on metalloprotein active sites. P Natl Acad Sci U S A 94: 4246-4249. Yamamoto, Y . (1996). H N M R study of temperature-induced structure alteration at the 1  active site of horse heart cytochrome c. Journal Biochem 119: 16-22. Yamamoto, Y . ; Terui, N . ; Tachiiri, N . ; Minakawa, K , ; Matsuo, H . ; Kameda, T.; Hasegawa, J.; Sambongi, Y . ; Uchiyama, S.; Kobayashi, Y . ; Igarashi, Y . (2002). Influence of amino acid side chain packing on Fe-methionine coordination in thermostable cytochrome c. J A m Chem Soc 124: 11574-11575. Yao, Y . ; Qian, C. M . ; Ye, K . Q.; Wang, J. F.; Bai, Z. P.; Tang, W. X . (2002). Solution structure of cyanoferricytochrome c: ligand-controlled conformational flexibility and electronic structure of the heme moiety. J Biol Inorg Chem 7: 539-547. Zhang, H . M . ; He, S. M . ; Mauk, A . G. (2002). Radical formation at Tyr39 and Tyrl53 following  reaction of yeast cytochrome c peroxidase with hydrogen peroxide.  Biochemistry 41: 13507-13513. Zheng, J. W.; Ye, S. Y . ; Lu, T. H . ; Cotton, T. M . ; Chumanov, G. (2000). Circular dichroism and resonance Raman comparative studies of wild type cytochrome c and F82H mutant. Biopolymers 57: 77-84.  153  Zhou, S. P.; Yao, Y . ; Davidson, A . ; Vieira, A . (2004). A cytochrome c-enhanced peroxidation reaction with potential use in screening dietary antioxidants. J Pharmaceut Biomed 34: 1057-1061. Zhu, Y . M . ; L i , J. FL; Liu, Z. M . ; Cheng, G. J.; Dong, S. J.; Wang, E. K . (1998). Identification of o-phenylenediamine polymerization product catalyzed by cytochrome c. J M o l Catal B-Enzym 4: 33-39. Zuniga, E. H.; Nail, B . T. (1983). Folding of yeast iso-1-cytochrome c. Biochemistry 22: 1430-1437.  154  


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items