Citation
Proton Transfer in Catalysis by Iron and Manganese Superoxide Dismutase

Material Information

Title:
Proton Transfer in Catalysis by Iron and Manganese Superoxide Dismutase
Creator:
GREENLEAF, WILLIAM BRUCE ( Author, Primary )
Copyright Date:
2008

Subjects

Subjects / Keywords:
Active sites ( jstor )
Amines ( jstor )
Catalysis ( jstor )
Enzymes ( jstor )
Hydrogen bonds ( jstor )
Iron ( jstor )
Manganese ( jstor )
pH ( jstor )
Protons ( jstor )
Superoxides ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright William Bruce Greenleaf. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
4/30/2014
Resource Identifier:
73174874 ( OCLC )

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

PROTON TRANSFER IN CATALYSIS BY IRON AND MANGANESE SUPEROXIDE DISMUTASE By WILLIAM BRUCE GREENLEAF A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2004

PAGE 2

ACKNOWLEDGMENTS Foremost my gratitude goes to my wife Tricia for her love and support over the years. I thank my parents for their encouragement and for providing me with the opportunity to learn. I thank my mentor, Dr. David Silverman, for allowing me into his lab, and teaching me to think critically and thoroughly about data and experimental design. My appreciation also goes to my committee members Drs. Dan Purich, Harry Nick, and Bill Kem, who not only provided sound scientific advice but also encouraged and aided me in making important decisions for my scientific career. Over my graduate career I have been fortunate to receive excellent technical assistance from Kristi Totten, Dr. Ke Ren, and Max Iurcovich, which has made the journey significantly smoother. Finally, my appreciation goes to all the other members of the Silverman lab, who have been a pleasure to work with and have made my time at the University of Florida enjoyable. ii

PAGE 3

TABLE OF CONTENTS page ACKNOWLEDGMENTS..................................................................................................ii LIST OF TABLES...............................................................................................................v LIST OF FIGURES...........................................................................................................vi ABSTRACT.....................................................................................................................viii CHAPTER 1 INTRODUCTION........................................................................................................1 Formation and Reaction of Superoxide........................................................................1 Physiological and Pathological Role of Superoxide Dismutase...................................2 Classes of Superoxide Dismutases........................................................................2 Biological Relevance of Superoxide Dismutase...................................................3 Structural Characteristics of Iron and Manganese Superoxide Dismutase...................4 Redox Properties of Iron and Manganese Superoxide Dismutase................................5 Catalytic Mechanism of Iron and Manganese Superoxide Dismutase.........................6 Research Goals.............................................................................................................7 Activation of Catalysis by Iron Superoxide Dismutase........................................7 Hydrogen Bonding in the Active Site of Human Manganese Superoxide Dismutase...........................................................................................................8 2 ACTIVATION OF IRON SUPEROXIDE DISMUTASE.........................................13 Introduction.................................................................................................................13 Materials and Methods...............................................................................................14 Purfication and Analysis of Enzymes..................................................................14 Stopped-Flow Spectrophotometry.......................................................................15 Results.........................................................................................................................16 Discussion...................................................................................................................19 General Acid Catalysis and FeSOD....................................................................19 Free Energy Plot for General Acid Catalysis......................................................25 Uncompetitive Activation...................................................................................27 3 HYDROGEN BONDING IN THE ACTIVE SITE OF HUMAN MANGANESE SUPEROXIDE DISMUTASE....................................................................................29 iii

PAGE 4

Introduction.................................................................................................................29 Materials and Methods...............................................................................................30 Site-Directed Mutagenesis...................................................................................30 Protein Expression and Purification....................................................................31 Pulse Radiolysis...................................................................................................32 Differential Scanning Calorimetry......................................................................33 Stopped-Flow Spectrophotometry.......................................................................33 X-ray Crystallography.........................................................................................34 Results.........................................................................................................................34 Visible Spectrophotometry..................................................................................34 Catalysis..............................................................................................................35 Calorimetry..........................................................................................................38 Crystallography...................................................................................................38 Discussion...................................................................................................................39 4 KINETIC PROPERTIES OF Porphyromonas gingivalis CAMBIALISTIC SUPEROXIDE DISMUTASE....................................................................................57 Introduction.................................................................................................................57 Materials and Methods...............................................................................................58 Enzymes..............................................................................................................58 Stopped-Flow Spectrophotometry.......................................................................59 Pulse Radiolysis...................................................................................................59 Results.........................................................................................................................59 Discussion...................................................................................................................61 5 CONCLUSIONS AND FUTURE DIRECTIONS.....................................................71 Conclusions.................................................................................................................71 Future Directions........................................................................................................72 Activation by Exogenous Proton Donors............................................................72 Site-Directed Mutagenesis...................................................................................73 Redox Potential...................................................................................................75 X-Ray Crystallography........................................................................................77 LIST OF REFERENCES...................................................................................................79 BIOGRAPHICAL SKETCH.............................................................................................85 iv

PAGE 5

LIST OF TABLES Table page 2.1: Steady-state constants (Eqs. 2.1, 2.2) and solvent hydrogen isotope effects for activation of E. coli FeSOD catalysis by various primary amines.............................20 3.1: The zero-order rate constant k0/[E] describing a product-inhibited region of catalyzed decay of superoxide and the second-order rate constants k1 and k2 for human wild-type MnSOD and site-specific mutants.................................................40 3.2: Rate constants k3 and k4 and their solvent hydrogen isotope effects for steps in the formation and dissociation of the product inhibited complex (Eqs 3.3, 3.4) for human wild-type MnSOD and site-specific mutants.................................................41 3.3: Unfolding transition temperatures of wild-type human MnSOD and mutants at positions 34 and 123..................................................................................................42 4.1: Steady-state constants for the decay of superoxide catalyzed by FeSOD and MnSOD and for the activation of catalysis by 3-aminopropionate..........................................62 v

PAGE 6

LIST OF FIGURES Figure page 1.1: The active site region of E. coli FeSOD....................................................................10 1.2: Active-site region of human MnSOD........................................................................11 1.3: Superoxide decay catalyzed by 0.5 M wild-type human MnSOD as measured by pulse radiolysis, from Hsu et al. (1996).....................................................................12 2.1: The dependence on ethanolammonium ion concentration of the turnover number kcatobs for catalysis of superoxide decay by 0.5 M E. coli FeSOD...........................21 2.2: Activation of glycine of catalysis of superoxide decay by 0.5 M E. coli FeSOD...22 2.3: Variation with the pKa of proton donors of log[(kcat/Km)donor] for activation of E. coli FeSOD........................................................................................................................23 3.1: Change in molar absorptivity at 480 nm as a function of pH for wild-type (), W123F (),Y34F (), and Y34F-W123F () human MnSOD..............................43 3.2: Decrease in A250 (250 = 2000 M-1 cm-1) due to decay of superoxide catalyzed by 2 M W123F human MnSOD measured by stopped-flow spectrophotometry............44 3.3: Single turnover experiments describing rate constants k3 and k4 for Y34F-W123F MnSOD......................................................................................................................45 3.4: Spectrum of the product inhibited form of Y34F-W123F human MnSOD measured by pulse radiolysis......................................................................................................46 3.5: Dependence of k1 with pH for Y34F-W123F () and wild-type human MnSOD ().............................................................................................................................47 3.6: Dependence of k3 with pH for Y34F-W123F () and wild-type human MnSOD ().............................................................................................................................48 3.7: Dependence of k4 on pH for Y34F-W123F () and W123F () MnSOD..............49 3.8: Crystal structure of the active site of human Y34F-W123F MnSOD (black) super-imposed upon wild-type human MnSOD (blue) (Borgstahl et al., 1992)..................50 vi

PAGE 7

3.9: Double mutant cycles for the rate constants k1 and k3 in Tables 3.1 and 3.2.............51 4.1: The active site region of P. gingivalis FeSOD..........................................................63 4.2: The dependence on 2-aminopropionate concentration of the turnover number kcatobs for catalysis of superoxide decay by 0.3 M P. gingivalis FeSOD...........................64 4.3: Rate of superoxide decay catalysed by 0.5 M P. gingivalis MnSOD in the presence () and absence () of 40 mM glycine, measured by pulse radiolysis....................65 4.4: Increase and subsequent decrease of absorbance at 420 nm of 25 M P. gingivalis MnSOD after pulsing with 2.7 M O2•....................................................................66 4.5: Increase and subsequent decrease at 420 nm of 25 M P. gingivalis MnSOD that was incubated with 250 M H2O2 then pulsed with 2.7 M O2•.............................67 vii

PAGE 8

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy PROTON TRANSFER IN CATALYSIS BY IRON AND MANGANESE SUPEROXIDE DISMUTASE By William Bruce Greenleaf May 2004 Chair: David N. Silverman Major Department: Pharmacology and Therapeutics Iron (FeSOD) and manganese superoxide dimutases (MnSOD) are efficient redox enzymes that prevent oxidative damage by catalyzing the conversion of superoxide (O2•) to oxygen (O2) and hydrogen peroxide (H2O2). Differences in the catalytic pathways of Feand MnSOD were investigated using exogenous proton donors to activate catalysis of Escherichia coli FeSOD and Porphyromonas gingivalis FeSOD, as measured by pulse radiolysis and stopped-flow spectrophotometry. Activation of E. coli FeSOD was found to be saturable for most amines investigated, and a free-energy plot of the apparent second-order rate constant of activation was linear against the pKa of the proton donor, indicating activation by enhanced proton transfer. Solvent hydrogen isotope effects for activation were near unity, suggesting the pKa of the proton acceptor on the enzyme is outside the pKa range of the proton donors used, from 7.8 to 10.6. Activation was nonessential and uncompetitive, interpreted as proton transfer in a ternary complex of viii

PAGE 9

enzyme, amine, and bound peroxide dianion. Similar properties of activation were observed for P. gingivalis FeSOD, suggesting activation is a general property of FeSOD. A potential proton transfer pathway in the obligate MnSOD from human was investigated by examining effects of site-specific substitutions at Tyr34 and Trp123, which form hydrogen bonds with the critical active site residue Gln143. Mutation of these residues decreased the rate of exit from product inhibition, in which proton transfer is rate-limiting, by up to 5-fold. Other steps in the catalysis were decreased by as much as 50-fold by these substitutions, resulting in more product-inhibited enzymes. Simultaneous replacement of Tyr34 and Trp123 in a double mutant indicated these residues interacted in these effects. The crystal-structure of Y34F-W123F MnSOD at 2.2 resolution suggested these effects were not due to large structural changes but to subtle electronic effects around Gln143. Similar thermal stabilities for wild-type and mutants suggest the hydrogen bonds between Tyr34, Trp123, and Gln143 are functional but not structural in nature. Activation of proton transfer was not observed in wild-type or mutant MnSODs; however, amines potentially activated the less product-inhibited P. gingivalis MnSOD implying similarities in the catalytic pathways of Feand MnSODs. ix

PAGE 10

CHAPTER 1 INTRODUCTION Since the emergence of an oxygenated atmosphere, organisms have had to account for the toxicity of reactive oxygen species (ROS). In biological systems dependent on oxidative phosphorylation as an energy source, oxygen may react to form hydrogen peroxide, superoxide, and highly reactive hydroxyl radicals. These ROS can react with lipid membranes in lipid peroxidation chain reactions, with proteins, and with DNA; these reactions are detrimental to cell viability. Enzymes that catalyze conversion of ROS to less toxic products have therefore evolved to permit expansion of life into oxygenated environments. These enzymes include peroxidases and catalases, which prevent hydrogen peroxide toxicity, and superoxide dismutases (SODs), which convert superoxide to oxygen and hydrogen peroxide. Formation and Reaction of Superoxide The superoxide molecule is formed in several enzymatic and non-enzymatic reactions in the cell, including the xanthine oxidase reaction (McCord and Fridovich, 1968), autooxidation of hemoglobin (Misra and Fridovich, 1972), and the disproportionation of hydrogen peroxide (Wilshire and Sawyer, 1979). Perhaps the most important source of superoxide in the cell is from the mitochondrial respiratory chain in which inefficiency in electron transport facilitates reduction of oxygen to superoxide (Bannister et al., 1987; Fridovich, 1997). While superoxide is a potent oxidizing agent, its primary toxicity is likely due to the iron catalyzed Haber-Weiss reaction of superoxide 1

PAGE 11

2 with hydrogen peroxide to form the hydroxyl radical (Haber and Weiss, 1934), which in turn damages lipids, membranes, proteins, and DNA. Physiological and Pathological Role of Superoxide Dismutase Superoxide dismutases, including MnSOD, FeSOD, CuZnSOD and NiSOD, provide the primary defense mechanism against superoxide-mediated toxicity. These metalloenzymes rapidly catalyze conversion of superoxide to molecular oxygen and hydrogen peroxide, thereby limiting formation of the hydroxyl radical. Each class of SOD has distinct distribution and functionality. Classes of Superoxide Dismutases Four classes of SOD have been discovered to date. Though each apparently operates by the same mechanism, functional differences exist. McCord and Fridovich first discovered the CuZnSOD in bovine erythrocytes in 1969. Both extracellular and cytoplasmic CuZnSODs have been found in eukaryotes, and a periplasmic CuZnSOD is present in most gram-negative bacteria (Fridovich, 1997). The second class of SOD, found in both prokaryotes and eukaryotes, contains manganese and is not structurally related to CuZnSOD (Stallings et al., 1985). In eukaryotes MnSOD is targeted to the mitochondria by a leader sequence that is cleaved from the active enzyme upon passage through the mitochondrial membrane (Rosenblum et al., 1996). Therefore, MnSOD plays a significant role in preventing superoxide from damaging mtDNA and proteins essential to mitochondrial integrity. Typically found exclusively in prokaryotes is the iron-containing SOD, which is structurally very similar to MnSOD. This enzyme is constitutively expressed in E. coli, as opposed to MnSOD, which is inducible. Most recently discovered in Streptomyces was NiSOD, structurally unrelated to any of the other classes of SODs (Choudhury et al., 1999).

PAGE 12

3 Biological Relevance of Superoxide Dismutase Evidence of the importance of SOD comes from several systems in which SOD production has been disrupted. In Escherichia coli lacking both Fe and MnSOD, greatly increased mutational rates were observed during aerobic growth (Farr et al., 1986). In Drosophila, removal of CuZnSOD increased sensitivity to paraquat and attenuated lifespan (Phillips et al., 1989). In MnSOD knockout mice, death from severe dilated cardiomyopathy occurred within 10 days of birth, due to superoxide-mediated reduction of critical mitochondrial enzymes including aconitase and succinate dehydrogenase (Li et al., 1995). Further, heterozygous MnSOD knockout mice showed high sensitivity to pulmonary oxygen injury (Tsan et al., 1998). Conversely, overexpression of CuZnSOD and catalase in Drosophila melanogaster increased life span by as much as 30%, and simultaneously decreased oxidative damage of proteins (Orr and Sohal, 1994), providing strong evidence for the free radical theory of aging. Superoxide dismutases are also involved in several disease states. Ionizing radiation induces free radical production, including superoxide; thus it is not surprising that decreased resistance to radiation is seen in heterozygous MnSOD knockout mice (Epperly et al., 2000). Decreased expression of MnSOD has been observed in pancreatic cancers, and subsequent overexpression was found to slow growth of cancerous cells (Cullen et al., 2003). Mutations in human cytosolic CuZnSOD have been associated with the familial form of amyotrophic lateral sclerosis (FALS), and result in a toxic gain of function, rather than reduction of SOD activity. The diverse mutations in CuZnSOD associated with FALS are concentrated in areas that cause localized structural changes that promote aggregation of the enzyme (DiDonato et al., 2003).

PAGE 13

4 Structural Characteristics of Iron and Manganese Superoxide Dismutase Though iron and manganese superoxide dismutases form two distinct classes, the enzymes are structurally quite similar. Analysis of primary structure indicates 44% sequence homology between E. coli FeSOD and human MnSOD. The initial x-ray crystal structure of E. coli FeSOD revealed a dimeric molecule formed of identical two-domain subunits (Stallings et al., 1983). Subsequent refinements in structure showed a trigonal bipyramidal geometry about the active-site metal, with 3 histidines, an aspartic acid, and a solvent molecule as ligands (Figure 1.1) (Lah et al., 1995). While the crystal structure of human MnSOD indicated a homotetrameric enzyme, the active site structure was highly conserved compared with FeSOD (Figure 1.2), including conservation of the ligands and the prominent residues His30, Tyr34, Trp123, and Gln143 (Gln69 in E. coli FeSOD) (numbering according to human MnSOD sequence), which form an extensive hydrogen-bonded network (Borgstahl et al., 1992). Mutagenesis of residues His30, Tyr34, and Gln143 in this network has indicated these hydrogen bonds are critical for maintaining efficient catalytic activity (Silverman and Nick, 2002). Despite the structural similarities between Feand MnSOD, substitution of manganese in FeSOD or substitution of iron in MnSOD typically results in an inactive enzyme (Vance and Miller, 2001). However, SODs from Porphyromonas gingivalis and Propionibacterium shermanii as well as several others remain active when either metal is present in the active site; these are known as cambialistic SODs (Schmidt et al., 1996). While the origin of this metal-specificity is still a matter of some debate, several structural differences between Feand MnSODs exist which may help to explain this behavior. Studies in the MnSOD-like double mutant Gln70Gly-Ala142Gln P. gingivalis SOD have indicated a shift in metal-specific activity to favor manganese in the active site

PAGE 14

5 (Hiraoka et al., 2000). Substitution of iron-like Thr for Gly155, a residue 10 removed from the active site in P. gingivalis SOD, significantly increases the FeSOD/MnSOD activity ratio (Yamakura et al., 2003). However, a combination of these factors is likely responsible for metal specificity, since in E. coli MnSOD creation of the iron-like Gly77Gln-Gln146Ala double mutant (analogous to Q70G-A142Q P. gingivalis SOD) confers only a small degree of activity to the Fe-substituted MnSOD (Schwartz et al., 2000). Redox Properties of Iron and Manganese Superoxide Dismutase All superoxide dismutases are redox catalysts that convert superoxide to molecular oxygen and hydrogen peroxide in a two-stage reaction in which the metal ion (M = Mn or Fe) cycles between the +3 and +2 states: M-(III)SOD + O2•– 1k H M-(II)SOD(H+) + O2 (1.1) M-(II)SOD(H+) + O2•– 2k H M-(III)SOD + H2O2 (1.2) For each half reaction to be efficient, the protein must adjust the redox potential of the free metal to a point midway between the reduction potentials for the oxidation (E0 = –160 mV versus NHE) and reduction (E0 = +870 mV) of superoxide (Vance and Miller, 1998). This accounts for measured redox potentials between 200 and 500 mV for Feand MnSODs (Vance and Miller, 1998; Lvque et al., 2001; Hearn et al., 2003). For free manganese in solution (Mn2+/3+) E0 = 1510 mV while for Fe+2/+3 E0 = 770 mV (Sawyer et al., 1995). Because of this, MnSOD protein must reduce the potential of free manganese by approximately 900-1000 mV, whereas FeSOD protein needs to lower the potential of free iron by only 300-500 mV. In Fe-substituted MnSOD, the potential of Fe is lowered beyond a value capable of sustaining efficient oxidation of superoxide (Vance and Miller,

PAGE 15

6 1998). In Mn-substituted FeSOD, the midpoint potential of Mn is reduced to a level above the O2/H2O2 couple, thus explaining the inactivity of these metal-substituted SODs (Vance and Miller, 2001). Cambialisitic SODs likely lower the midpoint potential of either Mn or Fe by approximately 700 mV; reduction by this amount would leave both metals capable of carrying out each half reaction, though with varying efficiency. Catalytic Mechanism of Iron and Manganese Superoxide Dismutase Both Feand MnSOD carry out rapid catalysis, with kcat = 25-40 ms-1 and kcat/Km near the diffusion-controlled limit at 300-800 M-1 s-1. While both FeSOD and MnSOD carry out catalysis at similar rates and mechanisms (Eqs 1.1,1.2), differences exist in patterns of inactivation, activation, and inhibition. Addition of H2O2 inactivates FeSOD by modification of tryptophan, but does not inactivate MnSOD (Beyer and Fridovich, 1987; Beyer et al., 1989). Primary amines have been seen to activate catalysis by E. coli FeSOD (Bull and Fee, 1985); however, similar activation of MnSOD has yet to be observed. Activation of FeSOD has been proposed to be proton transfer dependent, based upon a dependence of the rate of activation on pKa of the primary amine. Azide inhibition patterns are markedly different for MnSOD and FeSOD and have been used to differentiate between them (Misra and Fridovich, 1978). Most significantly, catalysis by MnSOD, but not FeSOD, is complicated by the rapid appearance of a reversibly inhibited form the formation and disappearance of which is shown below (McAdam et al., 1977; Bull et al., 1991): Mn(II)SOD(H+) + O2•– Mn(III)(O2-)SOD(H+) (1.3) 3k Mn(III)(O2-)SOD(H+) 4k H Mn(III)SOD + H2O2 (1.4)

PAGE 16

7 No structure of the product-inhibited form of MnSOD has been presented. However, based on visible absorption spectra of inorganic complexes, Bull and coworkers (1991) proposed inhibition is due to addition of superoxide to Mn2+-SOD which then forms a side-on peroxo complex. Alternatively, Cabelli and coworkers (2000) have proposed an end-on peroxo complex to be responsible for product inhibition. Significantly, the visible absorption spectrum of the product-inhibited form is very similar to the spectrum of the end-on azide-inhibited form of MnSOD (Whittaker and Whittaker, 1996; Hearn et al., 1999). The decay of superoxide during product inhibition in MnSOD is zero-order (Figure 1.3). Research Goals The purpose of this work is twofold. First is the application of Brnsted and kinetic analysis to activation of E. coli FeSOD to develop a comprehensive mechanism and to estimate the pKa of the enzymatic proton donor in catalysis. Included in this are studies in Mnand Fe-substituted cambialistic SOD from P. gingivalis to confirm if the properties of activation observed in E. coli FeSOD are consistently seen in another FeSOD system, and further to test if these same studies can be expanded to an MnSOD that has the same protein backbone. Second is the examination of the role of hydrogen bonding in the active site of human MnSOD, specifically the hydrogen bonding between Trp123 and Gln143, and Tyr34 and Gln143. These studies include creation of a double mutant to test catalytic and structural interactions between Tyr34 and Trp123, and their role in proton transfer to facilitate exit from the product inhibited form. Activation of Catalysis by Iron Superoxide Dismutase This research involved a thorough kinetic analysis of activation by primary amines of catalysis by E. coli and P. gingivalis FeSOD. Stopped-flow and pulse radiolysis

PAGE 17

8 techniques were employed to develop a free energy diagram that showed a linear relation between pKa of the activating amine and rate of activation, indicating activation by proton transfer. This work expanded similar activation studies in carbonic anhydrase (Silverman et al., 1993), but showed significant differences from the carbonic anhydrase case, and from proton transfer between nitrogen and oxygen acids and bases in solution (Kresge, 1975). The mechanism of activation of FeSOD was shown to be nonessential and uncompetitive, and a model of activation was developed based upon the 5-6-5 scheme formulated by Lah and coworkers (1995). Significant evidence was provided for metal-bound peroxide dianion as the identity of the proton acceptor on the enzyme. Hydrogen Bonding in the Active Site of Human Manganese Superoxide Dismutase Previous studies in MnSOD have investigated the role of residues Tyr34, His30, and Gln143, which form an extensive hydrogen-bonded network in the active site of the enzyme (Silverman and Nick, 2002). This work expands these studies to Trp123, which forms a hydrogen bond with Gln143 opposite Tyr34, extending the hydrogen-bonded network in the other direction. Site-directed mutants were created at residue 123. Differential scanning calorimetry was used to determine the role of Trp123 in maintaining thermostability of the enzyme. Structural studies via X-ray crystallography aided in estimating the structural role of Trp123 and provided evidence that any changes observed in catalysis or other properties were not due to gross structural alterations. Pulse radiolysis and stopped-flow techniques were used to investigate the catalytic role of the hydrogen bond between Trp123 and Gln143, and specifically the role of Trp123 in maintaining product inhibition. Individual rate constants (k1-k4 of Eqs. 1.1-1.4) were measured in single turnover experiments, and pH profiles of these rate constants determined. Further, single turnover experiments were carried out in D2O to determine if

PAGE 18

9 rate limiting proton transfer was involved in any of these steps, and to confirm the catalytic mechanism was not changing in the mutant MnSODs investigated. Taken together with previous studies, this work suggests a likely nonspecific interaction site for superoxide when reacting with either oxidized or reduced enzyme, and further supports the idea that product inhibition in MnSOD is designed to limit the production of H2O2 so as to prevent toxicity (Davis et al., in press).

PAGE 19

10 Trp122 Tyr34 Gln69 Figure 1.1: The active site region of E. coli FeSOD. The iron is shown as an orange sphere. The four protein ligands, consisting of three histidines and an aspartic acid, are shown in blue. The fifth ligand, a solvent molecule, is represented by a red sphere. Hydrogen bonds are shown in yellow.

PAGE 20

11 Trp123 Tyr34 Gln143 Figure 1.2: Active-site region of human MnSOD. The manganese is shown as a purple sphere. The four protein ligands, consisting of three histidines and an aspartic acid, are shown in blue. The fifth ligand, a solvent molecule, is represented by a red sphere. Hydrogen bonds are shown in yellow.

PAGE 21

12 ) Superoxide (M Inhibited phase 15105 0 02 46Time (ms)Burst phase 8 Figure 1.3: Superoxide decay catalyzed by 0.5 M wild-type human MnSOD as measured by pulse radiolysis, from Hsu et al. (1996). Decay of superoxide during the burst phase and during the product-inhibited phase is indicated by the arrows.

PAGE 22

CHAPTER 2 ACTIVATION OF IRON SUPEROXIDE DISMUTASE Introduction Early studies of E. coli FeSOD by Bull and Fee (1985) developed a thorough mechanism for catalysis by this enzyme. As had been seen for Cu/ZnSOD and MnSOD, catalysis was described by a cyclic redox process represented by Eqs. 1.1 and 1.2 where M = Fe. However, no evidence of the product inhibition seen in MnSOD (McAdam et al., 1977) was observed for FeSOD, although FeSOD could be irreversibly inactivated by the Fenton reactions of H2O2. Stopped-flow measurements on E. coli FeSOD indicated an efficient enzyme with kcat = 26 ms-1 and kcat/Km = 320 M-1 s-1, near the diffusion controlled limit. While kcat was pH independent over a wide pH range, kcat/Km was found to be decreased by an ionization near 9.5, assigned to the deprotonation of the active-site residue Tyr34. Bull and Fee (1985) also observed that the maximal rate of catalysis for FeSOD could be enhanced by addition of millimolar amounts of primary amines, suggesting a rate-limiting proton transfer in the catalytic mechanism. Later studies by Fee et al. (1986) confirmed a solvent hydrogen isotope (SHIE) on kcat of 3, adding further evidence for a rate-contributing proton transfer in the catalytic mechanism of FeSOD. However, it is not clear if the proton transfer step is in the first half-reaction (Eq. 1.1) or the second half-reaction (Eq. 1.2). Further, the number of protons transferred to enhance catalytic rates is not clear. Depending on whether product dissociates as HO2 or H2O2, one or two protons must be provided from solution. Three groups capable of serving as proton 13

PAGE 23

14 shuttles to product are Tyr34, His30, and the solvent ligand of the iron, all of which form a hydrogen bonded network in the active site. In human MnSOD, mutagenesis of residues Tyr34 and His30 decreased maximal catalytic rates by at least one order of magnitude (Silverman and Nick, 2002). This strongly suggests the hydrogen-bonded network in MnSOD provides support for rapid proton transfer to product. The use of general acid catalysts such as primary amines to activate E. coli FeSOD allows for closer inspection of potential proton transfer-dependent steps in the catalytic mechanism, and allows questions such as the identity of proton transfer groups on the enzyme to be addressed. A similar approach has been used to answer these questions for other enzymatic systems, such as carbonic anhydrase (Kresge and Silverman, 1999). For this work, stopped-flow and pulse radiolysis techniques were employed to show activation was nonessential and uncompetitive with respect to superoxide. Taken with SHIE measurements on proton transfer by primary amines, this mechanism of activation suggests a model describing how the proton transfer from donor to enzyme takes place and suggests the identity of the proton acceptor. Materials and Methods Purfication and Analysis of Enzymes Purification of E. coli FeSOD was carried out by the method of Slykhouse and Fee (1976) with some modifications. Each step was tracked by gel electrophoresis. Briefly, an isopropyl-b-D-thiogalactopyranoside (IPTG) inducible plasmid containing E. coli FeSOD (gift of Dr. Anne-Francis Miller) was transformed into SodA/SodB E. coli (Carlioz and Touati, 1986). After reaching OD600 = 0.4, cells were supplemented with 500 M Fe2(SO4)3, induced with IPTG and allowed to express protein for 4 hours. Bacteria were then centrifuged to form a pellet, and lysed using lysozyme. The lysis

PAGE 24

15 solution was then heated to 60 C for 10 minutes, then centrifuged to remove any precipitate, omitting addition of streptomycin sulfate. A 50 to 75% ammonium sulfate precipitation step was carried out, with the 75% precipitate dissolved in 0.5 M potassium acetate at pH 5.5. Dialysis proceeded over a 3-day period against 5 mM potassium acetate, with 2 changes of buffer. Solution containing FeSOD was first passed over a CM52 column; eluate was kept. After buffer exchange to 5 mM potassium phosphate, pH 7.4, solution was passed over a DE52 resin. Protein was eluted by a gradient of 5-50 mM potassium phosphate. Pure protein, as shown by gel eletrophoresis, was obtained at this point; the hydroxylapatite column was omitted. The absorbance shoulder at 350 nm, which corresponds to Fe3+ (350 = 1850 M-1 cm-1) was used to determine total active enzyme concentration spectroscopically. Stopped-Flow Spectrophotometry Activation of catalysis by E. coli FeSOD was measured by stopped-flow according to the procedures of McClune and Fee (1978). Potassium superoxide was dissolved in anhydrous dimethyl sulfoxide (DMSO), with solubility aided by addition of an approximately equimolar amount of 18-crown-6 ether. Changes in the absorbance of superoxide at 250 nm (250 = 2000 M-1 cm-1; Rabani and Nielson, 1969) were observed using a dual drive sequential mixing stopped-flow spectrophotometer (Applied Photophysics, SX.18MV). All experiments were maintained at 25 C, and a pathlength of 2 mm or 10 mm was used. Four to eight progress curves were averaged, and the initial 5-10% of these curves were used to measure initial velocities, which were calculated using a least-squares fitting process (Enzfitter; Biosoft). All calculations accounted for the uncatalyzed dismutation of superoxide. Instrumental dead times of 4 ms on average

PAGE 25

16 were obtained by use of sequential dilutions in which superoxide in DMSO is first mixed from a 250 L syringe in a 1:10 ratio with an aqueous solution containing 2 mM CAPS and 1 mM EDTA at pH 11. After a 500 ms ageing time, this solution was rapidly mixed in a 1:1 ratio with a solution containing enzyme, 100 mM buffer (sodium borate, TAPS, CHES), primary amines as activators, and 1 mM EDTA. Addition of EDTA was done to reduce superoxide dismutase activity of free metals such as iron and copper. By volume, final DMSO concentration was 4.5%. Maximal catalytic activity of E. coli FeSOD was not enhanced by addition of increasing concentrations of borate or TAPS between 5 and 180 mM. High ionic strength was maintained by addition of 100 mM Na2SO4. Above concentrations of 200 mM, borate buffer decreased maximal catalytic activity of FeSOD by as much as 40% (Benovic et al., 1983). Solvent hydrogen isotope effects were measured using D2O (99.9 atom %) obtained from Isotec (Miamisburg, Ohio). Before use D2O was charcoal filtered and redistilled. Results Saturable activation by exogenous primary amines of catalysis by E. coli FeSOD at steady-state was observed via stopped-flow spectrophotometry. Figure 2.1 shows the results for ethanolamine, in which concentrations of superoxide were maintained near saturating levels of 780 M. Ethanolamine concentration is plotted as the amount of ethanolammonium ion. The Km for superoxide was calculated to be 100 20 M under the conditions of Figure 2.1. The maximal turnover rate kcat was calculated to be 25 ms-1, which along with KmO2 agreed well with previously reported values (Bull and Fee, 1985). In the absence of ethanolamine, appreciable catalytic activity was present (Fig. 2.1).

PAGE 26

17 In previous activation studies of E. coli FeSOD, saturation was not observed even at > 100 mM concentrations of primary amine (Bull and Fee, 1985). For the activating amines of Table 2.1 except 3-aminopropionic acid, activation was saturable in this work. A likely explanation for this behavior is the different experimental conditions used in each study. Bull and Fee (1985) used ionic strength > 0.2 M, while here ionic strength was 0.1 M. Further, the experiments in this work were carried out at 25 C, whereas Bull and Fee (1985) performed their experiments at 4 C. Further investigation of activation by primary amines of catalysis by FeSOD was carried out by altering both activating amine concentration and superoxide concentration. Figure 2.2 shows the activation of FeSOD by glycine, in a Hanes-Woolf plot. All lines share a common intercept on the y-axis, which indicated kcat/KmO2 is not altered by varying the concentration of glycine. This is consistent with the activation being uncompetitive (Segel, 1975). The curve resulting from the replot of the slope of each line in the Hanes-Woolf plot against activator concentration was hyperbolic. Similar results to Fig. 2.2 were observed for ethanolamine as activator, in which again lines on the Hanes-Woolf plot shared a common intercept on the ordinate. Activation by exogenous amines of catalysis by FeSOD was therefore found to be nonessential and uncompetitive. Shown below in Eq. 2.1 is expression for the nonessential activation consistent in describing the data shown in Figs. 2.1 and 2.2: (2.1) [] EkOKOBHK1+BHKcatm O+mB+mBbtbt[]2221

PAGE 27

18 In Eq. 2.1 BH+ represents the primary amine, while represents the factor by which kcat is increased by the activating amine. Experiments similar to those of Fig. 2.1 used near saturating concentrations of superoxide. When this is taken into account for Eq. 2.1, the following expression results: (2.2) vEkk1BHKBHKcatobscat+mB+mBbb1 Under conditions without activator present, Eq. 2.2 reduces to v/[E] = kcat. When activating amine is at saturating concentrations, Eq. 2.2 reduces to v/[E] = kcat where > 1.0, which shows the degree of activation at saturation. Equation 2.2 may be rearranged to give the following expression (Segel, 1975): 11111catobscatkkBHbbbtbbmBcatcatKkknnn[] 1 (2.3) Equation 2.3 indicates that plotting 1/(kcatobs – kcat) against 1/[BH+] should give a line which intercepts the ordinate at the maximal velocity of activation, kcat(. The slope of this line gives the reciprocal of the apparent second-order rate constant of activation, ()kcat/KmB, which for clarity is designated (kcat/Km)donor. The inset in Fig. 2.1 shows this replot for activation by ethanolamine. Other activating amines were investigated; their and (kcat/Km)donor values are listed in Table 2.1. General acids that are not primary amines, such as morpholine and 1,2-dimethylimidazole, were tested but did not activate superoxide decay catalyzed by E. coli FeSOD under the conditions of Fig 2.1.

PAGE 28

19 Values of the apparent second-order rate constant of activation, (kcat/Km)donor, calculated using Eq. 2.2 were plotted against the solution pKa of the activating amine (Fig. 2.3). The most efficient activator of catalysis, glycine methyl ester (Table 2.1), had a (kcat/Km)donor = 4.7 1.4 M-1 s-1, well below the diffusion controlled limit. This is in sharp contrast to the apparent second-order rate constant for superoxide decay, 250 M-1 s-1, closer to diffusion controlled. The Brnsted plot of Fig. 2.3 is adequately fit by a straight line with slope of 0.50 0.07. However, similar plots of and kcat(-1) versus pKa of activating amine did not display a linear correlation. Solvent hydrogen isotope effect (SHIE) measurements on the second-order rate constant of activation, D[(kcat/Km)donor], were near unity for all amines investigated, except for ethanolamine, which displayed a large standard deviation (Table 2.1). Without being able to estimate the pKa of bound amines, we have instead used solution pKa values for all plots and calculations. However, as indicated by KmB values, amine binding to E. coli FeSOD was weak which would indicate little change in pKa of bound activator. This was the case for the activation of catalysis by carbonic anhydrase by imidazole and pyridine derivatives (An et al., 2002). Discussion General Acid Catalysis and FeSOD For enzymatic reactions that require or generate protons, the role of exogenous proton donors or acceptors has been well established. This includes the hydration and dehydration of CO2/HCO3 as catalyzed by carbonic anhydrase (An et al., 2002; Lindskog, 1997), and activation of a site-directed mutant of aspartate aminotransferase (Toney and Kirsch, 1989). For both of these examples, however, removal of activator

PAGE 29

20 Table 2.1: Steady-state constants (Eqs. 2.1, 2.2) and solvent hydrogen isotope effects for activation of E. coli FeSOD catalysis by various primary amines. The entries are the exogenous proton donors as number in Fig. 2.3 with conditions described in the legend to Fig. 2.1. (kcat/Km)donor is (-1)kcat/KmB of Eq. 2.3 and is the contribution to kcat/KmB due to the exogenous proton donors. Proton Donor Solution pKa (kcat/Km)donor (M-1 s-1) D[ (kcat/Km)donor ] glycine methyl ester (1) 7.8 3.7 0.1 4.7 1.4 –a 2-bromoethylamine (2) 8.5 2.0 0.4 1.2 0.2 1.1 0.2 2-fluoroethylamine (3) 8.6 2.4 0.3 0.89 0.25 1.0 0.4 2-aminoethylsulfonic acid (4) 8.9 3.5 0.1 1.65 0.03 1.1 0.1 benzylamine (5) 9.3 1.1 0.1 0.70 0.07 1.2 0.5 ethanolamine (6) 9.5 1.8 0.1 0.48 0.15 1.9 0.5 glycine (7) 9.8 2.6 0.4 0.51 0.15 1.0 0.3 3-aminopropionic acid (8) 10.2 3.3 0.6 0.31 0.11 1.4 0.5 ethylamine (9) 10.6 1.1 0.1 0.08 0.03 –a a , SHIE not measured.

PAGE 30

21 Figure 2.1: The dependence on ethanolammonium ion concentration of the turnover number kcatobs for catalysis of superoxide decay by 0.5 M E. coli FeSOD. Upon final dilution solution contained 100 mM borate, 1.0 mM CAPS, and 1.0 mM EDTA at pH 8.3. Experiments were carried out at 25 C, with initial concentration of superoxide 780 M. Solid line is a fit to Eq. 2.2 with kcat = 21 1 ms-1, KmB = 13 2 mM, and = 1.8 0.1. Inset: Data for activation of FeSOD by ethanolammonium ion replotted based upon Eq. 2.3.

PAGE 31

22 Figure 2.2: Activation of glycine of catalysis of superoxide decay by 0.5 M E. coli FeSOD. Experiment was carried out at pH 9.0 maintained by 100 mM CHES. Other conditions listed in the legend of Figure 2.1. Total glycine concentration was 0 (), 15 (), 30 (), and 60 mM ().

PAGE 32

23 Figure 2.3: Variation with the pKa of proton donors of log[(kcat/Km)donor] for activation of E. coli FeSOD. (kcat/Km)donor is the contribution to kcat/KmB due to the exogenous proton donors. Conditions were as described in the legend to Figure 2.1. The exogenous proton donors were: glycine methyl ester (1); 2-bromoethylamine (2); 2-fluoroethylamine (3); 2-aminoethylsulfonic acid (4); benzylamine (5); ethanolamine (6); glycine (7); 3-aminopropionic acid (8); ethylamine (9).

PAGE 33

24 causes catalytic activity to approach zero, unlike what occurs in the case of activation of E. coli FeSOD. Specifically for carbonic anhydrase, the steady-state turnover for the dehydration of HCO3 nears zero when concentrations of exogenous proton donors such as 1,2-dimethylimidazole are less than 1 mM (Jonsson et al., 1976; Pocker et al., 1985). Histidine 64 acts as a proton shuttle in several carbonic anhydrase isozymes, and transfers protons from buffers in solution to zinc-bound hydroxide (Tu et al., 1989). Reprotonation of the zinc-bound hydroxide must occur from solution in order to complete a catalytic cycle. Interestingly, water cannot serve as a sufficiently efficient proton donor to support catalysis faster than 103 s-1. Thus, activity is very small in the absence of exogenous proton donors. In contrast to the carbonic anhydrase system, catalysis by E. coli FeSOD does not approach zero as the concentration of exogenous proton donor is lowered to zero, but instead maintains the appreciable rate of 20 ms-1 (Fig. 2.1). This therefore excludes a mechanism similar to carbonic anhydrase in which a residue (i.e. His64) on the enzyme serves to transfer protons to the product, and then the exogenous proton donor reprotonates the residue. This mechanism is excluded because reprotonation of the proton shuttle would necessarily be rapid, and in the absence of proton donor catalysis would decrease beyond what is seen in Fig. 2.1. Further, this suggests that solvent is the major proton donor in the case of FeSOD. Two other features of FeSOD support this conclusion: the pH independence of kcat (Bull and Fee, 1985) and crystal structures of FeSOD and azide inhibited complexes (Lah et al., 1995). The iron-bound solvent molecule will have a depressed pKa, and thus is a likely candidate for the proton transfer group in the active site. When in the oxidized state of FeSOD, the solvent ligand has a

PAGE 34

25 pKa near 5. While in the reduced state, this pKa is 9-11 (Fee et al., 1981). Analysis of activation data from this work (discussed later) also suggests the metal-bound solvent serves as the proton donor to the bound peroxide dianion, and is replaced in this role by exogenous primary amines. Free Energy Plot for General Acid Catalysis The Brnsted plot (Fig. 2.3) has several features that yield information about the catalysis of superoxide decay by FeSOD. The logarithm of the apparent second order rate constants of activation, (kcat/Km)donor, depended linearly upon the pKa of the activating amine, thus confirming activation by primary amines occurs by proton transfer. No correlation of any type was observed between kcatobs or and pKa of the proton donor. Two explanations for this behavior are that these components of activation do not require a proton transfer step, or that the parameter is inaccurately measured due to the technical difficulty of supplying large concentrations of primary amines and superoxide. The slope of the free energy plot in Fig. 2.3 gives information about the position of the transition state in comparison to reactants and products (Kresge, 1975a). For this case, b, the Brnsted coefficient, is 0.50 0.07, suggesting the transition state position (position of the proton) is midway between the reactant and product. If the Brnsted plot were extensively curved, as in the case of carbonic anhydrase (Rowlett and Silverman, 1982), it becomes possible to estimate the pKa of the proton donor group on the enzyme (Eigen and Hammes, 1963). With the case of FeSOD, however, it is not possible to make this estimate. In contrast to the linear free energy diagram of this work, Brnsted plots for the nonenzymatic bimolecular proton transfer between nitrogen and oxygen acids and bases

PAGE 35

26 in solution show a high degree of curvature (Kresge, 1975b). Also displaying a high degree of curvature, and consequently a low kinetic barrier, is general acid catalysis for site-specific mutants of carbonic anhydrase (Silverman et al., 1993). Thus, the lack of curvature seen in the free energy diagram of Fig. 2.3, despite being over a narrow range of pKa values, suggests a large intrinsic kinetic barrier for proton transfer from primary amines to FeSOD when compared to the non-enzymatic proton transfer between electronegative atoms, near 2 kcal/mol (Kresge, 1975a). Within a narrow range of 2-6 pKa units near the pKa of the proton acceptor, SHIE values increase to a maximum at [pKa(donor) – pKa(acceptor)] = 0 for the transfer of protons between electronegative atoms. Outside of this narrow range, SHIE values are near unity, an observation made for both nonenzymatic (Bergman et al., 1978; Cox and Jencks, 1978) and enzymatic proton transfer in carbonic anhydrase (Silverman et al., 1993; Taoka et al., 1994). Westheimer (1961) explains this behavior of a maximal SHIE near tpKa = 0 by comparison of symmetric and asymmetric transition states. For the case of FeSOD, the SHIE on (kcat/Km)donor is near unity for the range of pKa values of the activating amines used (Table 2.1), indicating the amines used for this study do not have pKa values near the pKa of the proton acceptor on the enzyme. Based on this, it is possible to eliminate Tyr34 (pKa = 8.5 for E. coli Fe(II)SOD; Sorkin and Miller, 1997) as the proton acceptor. Further, if the iron-bound solvent molecule in Fe(II)SOD has a pKa in the range of 9-11 (Fee et al., 1981), it is not likely to be the proton acceptor either. Again, these arguments are consistent with the pH independent nature of kcat for E. coli FeSOD over the pH range 7-11 (Bull and Fee, 1985), since it is kcat that displays a SHIE of 3 (Fee et al., 1986). The proton acceptor on the enzyme is therefore proposed to be

PAGE 36

27 bound peroxide dianion, which would be expected to have a pKa well above the range investigated in this work. Uncompetitive Activation Activation of E. coli FeSOD by glycine (Fig. 2.2) and ethanolamine shows that activation is uncompetitive with respect to superoxide. Despite the complexity of the steady-state rate equation derived from Eqs. 1.1 and 1.2, several conclusions may be drawn from the uncompetitive nature of this activation. The simplest explanation taking into account general catalytic features developed by Lah and coworkers (1995) is that proton donation occurs in a ternary complex including protonated primary amine, enzyme, and bound peroxide dianion, as shown in Scheme 2.1 below: Fe(III)OOHHO-Fe(III)OOH2O2-BH+BH+-Fe(III)OOH2O2-Fe(II)OH2 + O2-B-Fe(III)OOH2+BH++HO-Fe(II)OH2 + O2-BH+Fe(III)OHFe(III)OHHO2-HO2-B, H+kcatkcat Scheme 2.1 In Scheme 2.1, proton transfer from primary amine to enzyme assists in product release. Scheme 2.1 may reflect both uncompetitive activation and inhibition, depending on the value of in equations 2.1-2.3 ( > 1 is activation, < 1 is inhibition). Taken into account in scheme 2.1 is the expansion of coordination about the iron to octahedral upon binding of substrate (Lah et al., 1995), but not shown are the four protein ligands. This scheme provides a model for the nonessential activation of catalysis by FeSOD as observed in Fig. 2.1, with the enzymatic proton donor being the iron-bound solvent

PAGE 37

28 molecule, which is H2O in the Fe(II)SOD form. No data indicates binding positions of primary amines. The active-site channel of FeSOD is narrow and may not permit binding of exogenous proton donors to a specific site on the enzyme, which may also be consistent with (kcat/Km)donor values being far from diffusion controlled (Table 2.1). Another possibility is the exogenous proton donors may transfer their protons from a distance through intervening water molecules in a hydrogen-bonded network. Such a network is present in Feand MnSODs, and in human MnSOD mutagenesis of residues in this chain reduce kcat levels by at least an order of magnitude (Silverman and Nick, 2002).

PAGE 38

CHAPTER 3 HYDROGEN BONDING IN THE ACTIVE SITE OF HUMAN MANGANESE SUPEROXIDE DISMUTASE Introduction An extensive hydrogen-bonded network of residues is present in the active site of MnSOD (Figure 1.3); these residues are critical in maintaining catalytic activity (Silverman and Nick, 2002). The manganese-bound solvent molecule forms a hydrogen bond with the critical active-site residue Gln143. Mutagenesis of Gln143 to Asn reduces kcat/Km by 3 orders of magnitude and changes the resting state of the enzyme from Mn(III) to Mn(II) (Hsieh et al., 1998). Glutamine 143 forms hydrogen bonds with two adjacent residues, Trp123 and Tyr34. This work focuses on Trp123, which has not yet been investigated. Previous studies have shown that Tyr34 is a functional residue that serves to maintain levels of product inhibition (Guan et al, 1998). Residue His30 forms a hydrogen bond extension from Tyr34 through an intervening water molecule. Interestingly, when mutated to Asn, product inhibition was greatly alleviated (Ramilo et al., 1999) again indicating that this non-liganded residue is not critical in maintaining catalytic activity. Continuing this network from His30 is residue Tyr166, which originates from the adjacent subunit and is approximately 8 from the manganese. Similar to His30Asn, the Tyr166Phe mutant displays decreased product inhibition, and is not critical to maintaining catalytic activity (Ramilo et al., 1999). This work estimates the role of the hydrogen bonds between Gln143, Tyr34 and Trp123. A crystal structure of the double mutant Y34F-W123F Mn(III)SOD reveals little 29

PAGE 39

30 change in active-site structure other than removal of hydrogen bonds between Gln143, Trp123 and Tyr34. Differential scanning calorimetry (DSC) experiments show little change in the main thermal unfolding transition near 90 C. Replacements of Tyr34 and Trp123 with Phe decreased catalytic activity and increased the degree of product inhibition compared to wild-type MnSOD. Analysis of the double mutant Y34F-W123F MnSOD indicated that Tyr34 and Trp123 act cooperatively in their effects on catalysis. Materials and Methods Site-Directed Mutagenesis The cDNA from human MnSOD (cDNA sequence published in Beck et al., 1987) was amplified via polymerase chain reaction using two oligonucleotides, 5’-GCATATGAAGCACAGCCTCC-3’ and 5’-GGAGATCTC-AGCATAACGATC-3’. The plasmid pHMNSOD4 (ATCC 59947) containing human MnSOD was subcloned into the TA cloning vector pCRII (Invitrogen Corp.). Four primers were used to generate the mutant W123F in two separate reactions. The first pair of primers was used to regenerate the coding region: 5’-CGCTAGTAATCATTTCATGAAGCACAGCCT-3’ (primer 1) and 5’-CGCCAAAACAGCCAAGCTTGCATGCCTGCA-3’ (primer 2). The second pair of primers (primers 3 and 4) differed for each mutant made, and served as complimentary internal primers encoding the desired mutation (underlined). For the site-directed mutant W123F, the second pair of primers used was 5’-AAGGCTCAGGTTTT GGTTGGCTTG-3’ (primer 3) and 5’-CAAGCCAACCAAA ACCTGAGCCTT-3’ (primer 4). The Y34F mutant human MnSOD cDNA (Guan et al., 1998) was the template used to generate the Y34F-W123F site-directed mutant, using the same primers as the W123F site-directed mutant. Since each site is distant from the other, no modifications in primers were necessary. Four

PAGE 40

31 other site-directed mutants at position 123 were made including Ala, Val, His, and Tyr. Three separate reactions were used to generate each mutant MnSOD. In the first PCR reaction, primers 1 and 4 were used to generate the 5’ half of the coding sequence, while in the second PCR reaction primers 2 and 3 were used to generate the 3’ half of the coding sequence. After the products of reactions 1 and 2 were gel purified using the QIAgen gel extraction kit (QIAgen Corp.) they were used as templates along with primers 1 and 2 in a third PCR reaction to generate the entire coding sequence containing the mutation of interest. The final PCR product was then cut with BspHI and PstI to make the ends compatible with the pTRC99a (Pharmacia Corp.) cloning vector. The BspHI site, located at the N-terminal end of the protein, was annealed to the compatible NcoI site in the pTRC99a vector, while the PstI site at the C-terminal end was annealed to the PstI site in pTRC99a. The sequence of all mutants generated was verified by DNA sequence analysis. Protein Expression and Purification The expression vector pTRC99a containing the mutant MnSOD of interest was transformed into SodA/SodB E. coli (Carlioz and Touati, 1986), which lacks the bacterial Mnand FeSOD. Final protein product was tagged with an N-terminal methionine. Cells were grown to an OD600 of 0.4, then supplemented with 660 M MnCl2 and induced to express protein by addition of IPTG. After 4 hours of expression, cells were pelleted by centrifugation, lysed, and then heat treated at 60 C for 10 minutes. Precipitate was removed by centrifugation. Lysate was then extensively dialyzed against appropriate buffer for 3 days with 2 changes of buffer. Enzyme was purified via FPLC on a Q-Sepharose anion exchange resin (Amersham Biosciences), and when necessary by

PAGE 41

32 size exclusion chromatography on a Sephacryl S-300 column. Each stage of purification was monitored by SDS-PAGE, until one band at 22 kDa corresponding to monomeric MnSOD was observed. Total protein concentration was measured by UV-spetrophotometry (280 = 40,500 M-1 cm-1). The total manganese concentration in each purified sample, as determined by flame atomic absorption spectroscopy, was used as the active enzyme concentration. Typically, manganese content was 60-80% of the total protein. Pulse Radiolysis Pulse radiolysis experiments were carried out at Brookhaven National Laboratory with the assistance of Dr. Diane Cabelli, using a 2 MeV van de Graaff accelerator. The path length for all experiments was either 2.0 or 6.1 cm, and all experiments were performed at 25 C. Spectra were recorded on a Cary 210 spectrophotometer. Dosimetry was determined by the methods of Hearn et al. (2001). Solutions were air saturated, and upon pulsing up to 30 M superoxide was generated according to the mechanisms of Schwarz (1981). Solutions contained enzyme, 30 mM sodium formate, 50 M EDTA, and 2 mM buffer depending on pH of the experiment: MOPS (pH 6.5-7.5), TAPS (8.0-8.5), CHES (9.0-9.5) or CAPS (10.0-10.5). Measurements of the pH dependence of rate constants showed complete reversibility, giving reproducible progress curves upon changing the pH of a solution at low pH to high pH, and then returning to low pH. Rate constants were determined by measuring the absorbance of superoxide (Rabani and Nielson, 1969) or of the enzyme (Cabelli et al., 1999). Experiments determining SHIEs on rate constants were performed in 98% D2O.

PAGE 42

33 Differential Scanning Calorimetry Differential scanning calorimetry (DSC) experiments were carried out in collaboration with Drs. James Lepock and Harold Frey at the University of Toronto. Studies were performed using a Nano DSC (Calorimetry Sciences, American Forks, UT, USA). The concentration of mutants and wild type MnSOD protein samples was 1 mg/ml in 20 mM potassium phosphate, pH 7.8. A solution containing only 20 mM potassium phosphate pH 7.8 was used for the reference. Sample and reference solutions were degassed under mild vacuum at 4 C before scanning from 10 to110 C at a rate of 1 C/min. The baseline and change in specific heat (tCp) upon thermal denaturation were corrected as described previously (Lepock et al., 1990; McRee et al., 1990). Transitions in the DSC profile were fit by a non-two-state, single component model (Sturtevant, 1987) using the software package Origin (Microcal Software, Northampton, MA, USA). All ttG values are in kcal/mol tetramer and were calculated using the van’t Hoff enthalpy and Tm from the best fits of the transitions. The estimated error of the transition temperature is 0.3 C. Stopped-Flow Spectrophotometry For wild-type and mutant MnSODs of this study, the zero-order rate constant k0/[E] was measured as described in the Materials and Methods section of Chapter 2, with several modifications. Solutions upon final dilution contained 2 M enzyme, 100 mM TAPS, and 500 M EDTA at pH 8.0, and 4.5% DMSO by volume. Dead time was 4 ms on average. Enzfitter (Biosoft) was used to calculate zero-order rate constants by fitting the average of 6-8 reaction traces to the sum of a linear (catalyzed) and second-order

PAGE 43

34 (uncatalyzed) process. Fits of second-order rates were confirmed by similar analysis of curves measured in the absence of enzyme. X-ray Crystallography Crystallographic studies were carried out in collaboration with Drs. M. Elizabeth Stroupe and John A. Tainer at the Scripps Research Institute. Crystals of mutant Y34F-W123F human MnSOD were grown from 2-3 M (NH4)2SO4 buffered by 50 mM imidazole/malate at pH 8.0; the mother liquor was supplemented with 20% ethylene glycol to facilitate freezing. X-ray diffraction data were collected to 2.20 -resolution for a single crystal of the Y34F-W123F variant at 100 K on beamline 9-1 at the Stanford Synchrotron Radiation laboratory with a Q315 ADSC CCD detector. Data were indexed and reduced with Denzo and scaled with Scalepack (Otwinowski et al., 1997). Phases were obtained by molecular replacement against the wild-type human MnSOD structure in Crystallography and NMR Systems, v. 1.1 (Brunger et al., 1998) (CNS). The structure was fit against CNS-calculated 2Fo-Fc, Fo-Fc, and composite omit electron density maps in the Xfit module of XtalView (McRee, 1999) and refined in CNS. Results Visible Spectrophotometry The maximal absorbance of the visible spectrum for wild-type human MnSOD is at 480 nm (Hsu et al., 1996). The extinction coefficient at the maximal absorbance (480) shows a dependence upon pH which can be fit by a single ionization with pKa = 9.2 0.1 (Figure 3.1; Hsu et al., 1996; Guan et al., 1998). The visible absorption spectrum for the W123F mutant again showed a maximum at 480 nm; however, compared with wild-type, 480 was larger below pH 10, though the change in 480 could also be fit to a single ionization with pKa = 9.2 0.1. The pH profile of 480 for Y34F-W123F MnSOD was

PAGE 44

35 featureless over the pH range measured, while for Y34F MnSOD it showed a decrease above pH 10.5 (Figure 3.1, Guan et al., 1998). Metal specificity was unchanged in the W123F substitution. Similarly, the manganese content of Y34F-W123F was 80% per monomer, while iron content was undetectable. Catalysis Using stopped-flow spectrophotometry, catalytic decay of superoxide was measured for wild-type MnSOD and the mutants of this work. These curves were adequately fit to the sum of zero-order and second-order processes, though the catalytic (product inhibited) portion of the curve dominated (Bull et al., 1991; Guan et al., 1998). Figure 3.2 shows representative data for W123F MnSOD, which is predominantly zero-order. Compared to wild-type, k0/[E] for the mutants of this work were reduced from 3to 8-fold (Table 3.1), indicating an increase in product inhibition in these variants. McAdam and coworkers (1977) described a simplified kinetic scheme to model catalysis by MnSOD (Eqs. 3.1-3.4) (Hearn et al., 2001). This scheme has been modified to include the observation that MnSOD takes up a proton upon reduction of the metal, the site of which is suggested to be the metal bound solvent (Miller et al., 2003). Mn(III)SOD + O2•– 1k H Mn(II)SOD(H+) + O2 (3.1) Mn(II)SOD(H+) + O2•– 2k H Mn(III)SOD + H2O2 (3.2) Mn(II)SOD(H+) + O2•– Mn(III)(O2-)SOD(H+) (3.3) 3k Mn(III)(O2-)SOD(H+) 4k H Mn(III)SOD + H2O2 (3.4) The product-inhibited form of the enzyme is represented here by Mn(III)(O2-)SOD(H+). Each stage of catalyis is treated irreversibly, an assumption justified by the favorable

PAGE 45

36 equilibrium constants for formation of products. Equations 3.1 and 3.2 represent a full catalytic cycle; equations 3.3 and 3.4 represent formation and dissociation of the product-inhibited form. Rate constants k1-k4 were measured via pulse radiolysis in single turnover experiments by observing changes in absorbance of the enzyme (Hearn et al., 2001). Rate constant k1 is calculated from the decrease in 480 nm of the Mn(III) enzymatic form; k2 is difficult to measure in highly product inhibited mutants such as those in the present study; from reduced enzyme k3 is estimated from the increase at 420 nm which corresponds to the maximal absorbance of the product-inhibited form (McAdam et al., 1977); k4 is calculated from the decay of the 420 nm peak and subsequent increase of the 480 nm peak which is absent during product inhibition. Representative single turnover experiments are shown in Figure 3.3. Starting with the featureless spectrum of Mn(II)SOD, addition of micromolar amounts of superoxide resulted in a first-order increase of absorbance at 420 nm on a microsecond time scale (Eq. 3.3) (Figure 3.3, top). Over millisecond time scales the product-inhibited form decays into Mn(III)SOD the absorbance of which appears at 480 nm (Eq. 3.4) (Figure 3.3, bottom). Replacements of Tyr34 and Trp123 with Phe in single and double mutants resulted in a 2to 3-fold decrease in rate constants k1 and k3 (Tables 3.1, 3.2). The decrease in k4 caused by these mutants was by as much as 4-fold in the case of the double mutant. Direct measurement of k2 is not possible in highly product inhibited mutants since significant changes at 480 nm are not measurable over microsecond time scales. The spectrum of the product inhibited form for Y34F-W123F MnSOD is shown in Fig. 3.4; this demonstrates that little or no Mn(III)SOD is being directly formed from Mn(II)SOD (Eq. 3.2 of the catalytic cycle) upon addition of superoxide. Using the kinetic simulation

PAGE 46

37 program KINSIM (Barshop et al., 1983), upper estimates of k2 were made by fitting superoxide decay traces for the mutant MnSODs of this work with k1, k3, and k4 as determined experimentally (Tables 3.1, 3.2), while varying k2 until the simulated catalyzed superoxide decay curve closely resembled the experimentally observed trace. For the mutants of Table 3.1, it was possible to set an upper limit of 0.02 nM-1 s-1 for k2 (Table 3.1). The rate constant k1 (Eq. 3.1) for wild-type human MnSOD was independent of pH up to pH 9.5. Beyond 9.5, k1 decreased, the pKa of which was estimated at greater than 10.0 (Figure 3.5). The pH profiles of k1 for Y34F, W123F, and Y34F-W123F MnSOD were all similar in magnitude and in the pKa of 9.5 0.2 (shown for Y34F-W123F in Figure 3.5 for clarity). Similarly, the rate constant k3 (Eq. 3.3) resembled k1 for both wild-type and mutant MnSODs in both magnitudes and pKas (shown for Y34F-W123F in Fig. 3.6 for clarity). In contrast to k1 and k3, the rate constant k4 showed no dependence on pH for wild-type and Y34F MnSOD over the range of pH values measured (Figure 3.7). Both mutants containing W123F show the appearance of a pKa of approximately 9.2 for k4 (Figure 3.7). At pH > 10.0 in the presence of 78 M hydrogen peroxide, k4 is near 45 s-1 for Y34F-W123F MnSOD possibly indicating an effect on this rate due to hydrogen peroxide. This point has been omitted in the fitting of the data for Y34F-W123F. Experiments in H2O and D2O (98%) measured SHIEs on rate constants k1, k3 and k4. For wild-type and the three mutant MnSODs of this work, no SHIE was observed for k1 and k3. However, a SHIE near 2 was observed for k4 for these enzymes (Tables 3.1, 3.2).

PAGE 47

38 Calorimetry The thermal stability of human wild-type and mutant MnSODs was measured by differential scanning calorimetry (Table 3.3). Only small changes in stability as assessed by Tm and ttG were observed for substitutions at positions 34 and 123 singly and in tandem. Values of ttG were calculated assuming tCp=0 since it is difficult to measure tCp accurately and incorporating tCp into the calculation for ttG has only a small effect when changes in stability are small. The main unfolding transition (Tm) of wild type enzyme with Gln143 hydrogen bonded to both Tyr34 and Trp123 was 90.7nC. Replacing Trp123 with Phe (W123F) had no effect on stability while replacing Tyr34 with Phe (Y34F) increased stability (tTm=3.5nC). The substitution of phenylalanine at both positions in the double mutant (Y34F-W123F) was destabilizing (tTm= -2.0 nC). Substitutions at position 123 that could not be purified included Ala, Val, His, and Tyr. Typically this means these variants are unstable and do not fold properly. Crystallography The overall structure of human Y34F-W123F Mn(III)SOD crystallized at pH 8.0 was nearly identical to the wild-type structure (active-site region in Figure 3.8). There was no difference in bond lengths to the metal from any of the ligands. In addition, the conformations of the catalytically significant residues Gln143 and His30 appeared unchanged in the double mutant. The phenyl rings of Phe34 and Phe123 were superimposed on the phenol and indole rings of wild-type at these positions (Figure 3.8). As in the wild-type crystals, the asymmetric unit contains a homodimer and the biologically-relevant tetramer is formed by one of the crystallographic two-fold symmetry axes in the p6122 space group. The unit cell for the Y34F-W123F double

PAGE 48

39 mutant was 79.5 x 79.5 x 242.8 . The dataset is 98.1% complete (99.5% for the highest resolution shell, 2.28-2.20 ; 23,666 unique reflections measured) with an Rsym of 7.6% (37.8% for the highest resolution shell). The data have an overall I/ of 30 with an I/ in the highest resolution shell of 8; crystals had a mosaicity of 0.4o. The structure was refined to an Rworking of 25.2% and an Rfree (Adams et al., 1997) of 27.1% (with 10% of the reflections flagged for the test set). Discussion The side-chain of the active-site residue Gln143 forms a hydrogen bond to the manganese-bound solvent molecule (Figure 3.8; Borgstahl et al., 1992). Mutagenesis of this residue to Asn indicated Gln143 is critical to maintaining efficient catalytic activity (Hsieh et al., 1998; Edwards et al., 2001), likely because it influences the pKa of the metal-bound water and thus the redox potential at the manganese (Maliekal et al., 2002). The side-chains of Tyr34 and Trp123 form hydrogen bonds with the carboxamide side-chain of Gln143, and thus may serve to adjust the position of Gln143. In another metalloenzyme, carbonic anhydrase, the third-shell ligand Glu106 orients second-shell ligands for catalysis, and influences the pKa of the zinc-bound water (Liang et al., 1993). In this study, replacements were made of Tyr34 and Trp123 through site-directed mutagenesis, and examined for their effects on structure and function. The crystal structure of Y34F-W123F Mn(III)SOD at 2.2 resolution was nearly superimposable over the wild-type structure (Figure 3.8). Despite the loss of hydrogen bonds between the side-chain of Gln143 and its neighbors Tyr34 and Trp123, the phenyl rings of residues 34 and 123 are superimposable on the corresponding phenol and indole rings in the wild-type structure. Perhaps surprisingly, there is no change in the crystal

PAGE 49

40 Table 3.1: The zero-order rate constant k0/[E] describing a product-inhibited region of catalyzed decay of superoxide and the second-order rate constants k1 and k2 for steps in the catalytic mechanism (Eqs 3.1, 3.2) for human wild-type MnSOD and site-specific mutants. Enzyme k0/[E] (s-1)a k1 (nM-1 s-1)b Dk1 k2 (nM-1 s-1)b Wild-type 220 1.5 0.1c –d 1.1 0.1c Y34F 26 0.55 0.03 1.2 0.1 <0.02 W123F 75 0.76 0.05 1.2 0.2 <0.02 Y34F-W123F 58 0.66 0.06 1.3 0.1 <0.02 W161F 50 0.30 0.08c –d <0.01c a Data taken from stopped flow at pH 8.0, 100 mM TAPS, 500 M EDTA, 2 M enzyme, 20 C. Error is less than 10 percent for each value; b From pulse radiolysis with conditions as in Figure 3.3, pH 7.5. Values are the average and standard deviations of at least three measurements; c Rate constant for wild-type MnSOD and W161F MnSOD at pH 8.2 and 25 oC taken from Hearn et al. (2001); d Not measured.

PAGE 50

41 Table 3.2: Rate constants k3 and k4 and their solvent hydrogen isotope effects for steps in the formation and dissociation of the product inhibited complex (Eqs 3.3, 3.4) for human wild-type MnSOD and site-specific mutants.a Enzyme k3 (nM-1 s-1) Dk3b k4 (s-1) Dk4b Wild-typec 1.1 0.1 1.2 0.2 120 10 1.7 0.2 Y34F 0.46 0.05 1.1 0.2 52 3 1.8 0.2 W123F 0.64 0.08 1.0 0.2 79 4 1.8 0.2 Y34F-W123F 0.70 0.09 1.0 0.2 27 2 1.9 0.2 W161F c 0.46 0.04 1.0 0.1 33 3 2.2 0.2 a Conditions as in Figure 3.3; pH 7.5. Values are the average and standard deviations of at least three measurements; b Solvent hydrogen isotope effects are indicated by superscript D, as in Dk3 = (k3)H2O/(k3)D2O; c Rate constants for wild-type MnSOD and W161F MnSOD at pH 8.2 and 25 C taken from Hearn et al. (2001).

PAGE 51

42 Table 3.3: Unfolding transition temperatures of wild-type human MnSOD and mutants at positions 34 and 123. MnSOD Tm (C)a ttG (kcal/mol-tetramer) Wild-type 90.7 — Y34F 94.2 1.2 W123F 90.6 0 Y34F-W123F 88.7 –0.7 aStandard deviation of Tm was estimated at 0.3 C.

PAGE 52

43 3004005006007008007.58.59.510.511.5pH480 (M-1 cm-1) Figure 3.1: Change in molar absorptivity at 480 nm as a function of pH for wild-type (), W123F (),Y34F (), and Y34F-W123F () human MnSOD. A single ionization of pKa = 9.2 0.1 was used to fit the data for W123F and wild-type human MnSOD.

PAGE 53

44 00.20.40.60.800.30.60.91.21.5time (s)A250 Figure 3.2: Decrease in A250 (250 = 2000 M-1 cm-1) due to decay of superoxide catalyzed by 2 M W123F human MnSOD measured by stopped-flow spectrophotometry. Solution contained 100 mM CAPS, 500 M EDTA at pH 10.5, with initial concentration of superoxide 390 M. The solid line is a fit to the sum of a zero-order process with k0/[E] = 76 s-1 and an uncatalyzed, second-order process of rate constant of 440 M-1 s-1.

PAGE 54

45 Figure 3.3: Single turnover experiments describing rate constants k3 and k4 for Y34F-W123F MnSOD. Top: Increase in the absorbance at 420 nm of 120 M Y34F-W123F Mn(II)SOD over microsecond time scales after pulse radiolysis generated 6.7 M O2 in solution. Enzyme was reduced with H2O2 prior to pulsing. Solutions contained 30 mM formate, 50 M EDTA, and 2 mM MOPS at pH 7.5 and 25 C. The solid line is a fit to first-order process corresponding to k3 of 0.64 0.06 nM-1 s-1. Bottom: Increase in the absorbance at 480 nm of 120 M Y34F-W123F MnSOD over millisecond time scales after pulse radiolysis generated 3.2 M O2. Solution conditions were identical to those described in the legend for the top figure. The solid line is a fit to a first-order process corresponding to a k4 of 25 1 s-1.

PAGE 55

46 01000200030004000300350400450wavelength, nm480, M-1 cm-1 Figure 3.4: Spectrum of the product inhibited form of Y34F-W123F human MnSOD measured by pulse radiolysis. Plotted is the maximum absorbance of Fig. 3.3 (top) versus wavelength. Conditions are identical to those in the legend of Fig. 3.3 (top).

PAGE 56

47 0369121567891011pHk1 x 108 (M-1 s-1) Figure 3.5: Dependence of k1 with pH for Y34F-W123F () and wild-type human MnSOD (). Solutions at 25 C contained 120 M Y34F-W123F, 30 mM formate, 50 M EDTA, and 2 mM of one of the following buffers: MOPS (6.5-7.5), TAPS (8.0-8.5), CHES (9.0-9.5), or CAPS (10.0-10.5). Data for WT MnSOD taken from Hearn et al. (2001). Each data point is the mean and standard deviation of 6-12 measurements. The data for Y34F-W123F was fit to a single ionization with pKa of 9.5 0.2; wild-type data was consistent with pKa > 10.0.

PAGE 57

48 048121667891011pHk3 x 108, M-1 s-1 Figure 3.6: Dependence of k3 with pH for Y34F-W123F () and wild-type human MnSOD (). Solution conditions were identical to those in the legend of Figure 3.5. Data for wild-type MnSOD taken from Hearn et al. (2001). Each data point is the mean and standard deviation of 6-12 measurements. The data for Y34F-W123F was fit to a single ionization with pKa of 9.5 0.2; wild-type data was consistent with an ionization of pKa > 10.0.

PAGE 58

49 040801201606810pHk4 (s-1) Figure 3.7: Dependence of k4 on pH for Y34F-W123F () and W123F () MnSOD. Solution conditions were identical to those described in the legend of Figure 3.5. Each data point is the mean and standard deviation of 6-12 measurements. A single ionization was fit to the data for Y34F-W123F with a pKa of 9.3 0.1 and to W123F with a pKa of 9.1 0.2. The values of k4 for wild-type (solid line) and Y34F (dashed line) human MnSOD are invariant over this pH range and are given in Table 1.

PAGE 59

50 Figure 3.8: Crystal structure of the active site of human Y34F-W123F MnSOD (black) super-imposed upon wild-type human MnSOD (blue) (Borgstahl et al., 1992). Hydrogen bonds connecting Tyr34, Gln143, and Trp123 are shown in orange for wild type but cannot form in the mutant.

PAGE 60

51 WT Y34F W123F Y34F-W123F 2.5 kJ mol-1 0.5 kJ mol-1 0.3 kJ mol-1 1.7 kJ mol-1 WT Y34F W123F Y34F-W123F 2.2 kJ mol-1 1.0 kJ mol-1 0.2 kJ mol-1 1.3 kJ m ol-1 Figure 3.9: Double mutant cycles for the rate constants k1 and k3 in Tables 3.1 and 3.2. Top: Double mutant cycle for k1 of Eq. 3.1, showing cooperativity between residues 34 and 123 for this step. Bottom: Double mutant cycle for k3 of Eq. 3.3, showing cooperativity between residues 34 and 123 for this step.

PAGE 61

52 structure in the double mutant of the position of Gln143 compared to wild-type. Further, little change in geometry about the manganese is observed in Y34F-W123F MnSOD, with ligand-metal bond distances unchanged (Figure 3.8). The most significant alteration in structure of the double mutant then is the lack of hydrogen bonds to the critical residue Gln143. The pH profiles of the visible absorbance of wild-type and mutant Mn(III)SODs show that Tyr34 is the likely source of the pKa near 9.2 (Figure 3.1), due to the absence of this ionization in mutants containing Y34F. This is consistent with previously published reports (Guan et al., 1998; Maliekal et al., 2002). As shown in the single mutant W123F MnSOD, Trp123 does not appear to influence the pKa of Tyr34, which indicates no interaction between Tyr34 and Trp123 for this property (Figure 3.1). Yamakura et al. (2003) have suggested that subtle changes in the position of Trp123 may account for the changes in metal selectivity seen in mutants of P. gingivalis cambialistic SOD. The mutant W123F human MnSOD did not display any changes in manganese selectivity, and iron occupancy was less than 1% per monomer. This may indicate that in W123F MnSOD the replacement of Trp with Phe is not conservative enough to allow for changes in metal specificity, or that this residue does not play a role in metal specificity in human MnSOD. Further, human MnSOD is not a cambialistic enzyme, and thus other determinants of metal selectivity may be playing a role. The effects on metal specificity due to Trp123 may therefore be masked. Though not essential to activty, Trp123 helps to maintain efficient catalysis. Replacement of Trp123 with Phe resulted in a 2-3 fold decrease in rate constants k1 and k3, and at least a 50-fold decrease in the rate constant k2 for the simplified catalytic

PAGE 62

53 mechanism of Eqs. 3.1-3.4. A significant observation is that k2 is much less than k3 for the mutants Y34F, W123F, and Y34F-W123F. This accounts for the large increase in product inhibition in these mutants compared to wild-type. The k3/k2 ratio determines gating into the product-inhibited form, and a significant result is that partitioning between the catalytic and product-inhibited portions of catalysis has been significantly altered by substitutions at these positions (Table 3.1). Because of the high k3/k2 ratio, the W123F MnSOD catalyzed dismutation of superoxide shown in Figure 3.2 can be explained by catalysis proceeding entirely through the product-inhibited form (and the uncatalyzed dismutation). As shown in Table 3.1, the increase in product inhibition for the Y34F, W123F, and Y34F-W123F mutants was seen as a 3to 8-fold decrease in k0/[E] compared to wild-type. In a previous study, replacement of Trp161 with Phe displayed similar results for levels of product inhibition (Hearn et al., 2001). Residues Tyr34, Trp123, and Trp161 are all active-site residues that have their side chains within 6 of the manganese, and mutations at each site remove support for the efficient reoxidation of the metal (Eq. 3.2). Remarkably, the rate constants k1 and k3 have very similar magnitudes and pH profiles for wild-type MnSOD and the mutants of Figures 3.5 and 3.6. This similarity suggests that the stages of catalysis represented by k1 and k3 (Eqs. 3.1, 3.3) are insensitive to the formal charge on the manganese, since the step defined by k1 proceeds from Mn(III)SOD and the step defined by k3 starts with Mn(II)SOD. For E. coli MnSOD it has been observed that a proton is taken up by the enzyme upon reduction of the metal, the site of which is suggested to be the solvent ligand of the manganese (Miller et al., 2003). Therefore, the oxidized form of the active site may be SOD-Mn+3OH while the reduced

PAGE 63

54 form of the active site may be SOD-Mn+2H2O. The delocalization of charge into the ligands as well as the formal neutralization of charge on the oxidized metal by hydroxide indicates the electrostatics of the active site maybe be similar for k1 and k3. Interestingly, k2 appears to behave quite differently from k1 and k3 in these mutants, despite sharing the same electrostatics at the active site. As suggested by Cabelli et al. (2000), the processes represented by k1 and k3 may be inner sphere reactions, while k2 is an outer sphere process. Because of this, the reaction defined by k2 may responding to a different region of the active site cavity than reactions defined by k1 and k3. Since mutations at residues Tyr34, Trp123, and Trp161 decrease the rate constant k2, they may be the region of the active site responsible for nonspecific binding of superoxide in the outer-sphere mechanism of Eq. 3.2. In contrast, residues His30 and Tyr166 are farther away from the metal, and mutations of these amino acids result in an enzyme which gates more through the catalytic cycle (Hearn et al., 2004), thus indicating by this hypothesis that neither His30 nor Tyr166 are in the region of superoxide binding for the outer-sphere process of Eq. 3.2. In contrast to k1 and k3 which decrease with increasing pH for wild-type and mutant MnSODs of this work, k4 for both single and double mutants containing W123F showed a pH dependence with pKa = 9.2 with maximal values at high pH (Figure 3.7). Wild-type and Y34F MnSOD showed little variance in pH over the range measured. Similar behavior of k4 was observed for the W161A human MnSOD, although the pKa measured was about 2 pH units lower (Hearn et al., 2001). Therefore, the deprotonated form of an ionizable group made available by substitution of Phe for Trp123 promotes dissociation of the product-inhibited complex. Due to the lack of changes observed in the crystal

PAGE 64

55 structure of Y34F-W123F MnSOD, it is not possible to pinpoint the source of the ionization. Figure 3.7 indicates that the ionization at pKa 9.2 is not due to Tyr34. Proton transfer is rate limiting for k4, as shown by the SHIE of 2 (Table 3.2), and it is possible that this ionization modifies the proton transfer step in some manner. No SHIE were measured at high pH to clarify this point. The double mutant Y34F-W123F MnSOD was created in order to determine any interactions between residues Tyr34 and Trp123. A striking feature is that the rate constants k1 and k3 (Eqs. 3.1, 3.3) were nearly identical for the double mutant as compared to the single mutants Y34F and W123F (Tables 3.1, 3.2). Therefore, the kinetic effects due to either mutation represent a threshold and mutation at the other site may not enhance the catalytic effect further. Another manifestation of this threshold effect is seen in the kinetic pKa values for k1 and k3, which are not changed in the double mutant from their values at 9.5 0.2 in the single mutants (Figures 3.5, 3.6). This indicates an interaction between Tyr34 and Trp123 for reduction of the enzyme and formation of the product-inhibited complex. These data may also be interpreted by converting rate constants for each step in the catalysis to a change in free energy using the expression –RTln(k) where R is the universal gas constant, T is the temperature at which experiments were carried out, and k is the rate constant. By subtracting the resulting tG value from that of wild-type, a double mutant cycle may be created (Mildvan et al., 1992). For the present case of k1 and k3, tG1+2 = tG1 = tG2 indicates a cooperative interaction, where tG1+2 represents the free energy change for the double mutant and tG1 and tG2 for single mutants (Figure 3.9). The data therefore suggest that loss of a hydrogen bond from Trp123 to Gln143 is as detrimental to activity as loss of the

PAGE 65

56 Gln143-Tyr34 hydrogen bond. Considering the minor structural changes seen in the double mutant compared to wild-type (Figure 3.8), a loss of a hydrogen bond appears to be the cause of the decrease in k1 and k3, rather than a conformational change. The thermal stability of the single mutants of Table 3.3 was either increased (Y34F) or unchanged (W123F) compared to wild-type. Therefore, neither Tyr34 nor Trp123 stabilize the protein compared with Phe at these positions. The small changes in thermal stability upon removal of hydrogen bonds to the carboxamide side chain of Gln143 imply that the hydrogen bonds themselves are not structural in natural, but rather functional. This conclusion agrees with the catalytic data of this study, which indicate reduction of rate constants for these mutants compared to wild-type MnSOD. However, the W123F mutant was the only mutant at the Trp123 site that could be purified. Other substitutions attempted were Ala, Val, His, and Tyr; none of these were stable. This suggests that while the hydrogen bonds around Gln143 may not serve a structural role, Trp123 itself does have a significant structural role. In summary, Trp123 plays a major role in maintaining the gating between the catalytic and product-inhibited phases. When Phe was substituted at the Trp123 position catalysis was overwhelmed by appearance of a peroxide-inhibited form at the metal, as was also found for the Y34F (Guan et al., 1998) and the W161F (Cabelli et al., 1999) mutants. Davis et al. (in press) have suggested that product inhibition exists to prevent mitochondrial overload of H2O2; compared with phenylalanine, Trp123 and Tyr34 in wild-type MnSOD serve to maintain catalytic activity and reduce levels of product inhibition.

PAGE 66

CHAPTER 4 KINETIC PROPERTIES OF Porphyromonas gingivalis CAMBIALISTIC SUPEROXIDE DISMUTASE Introduction Superoxide dismutase from P. gingivalis is a homodimer composed of 22 kDa subunits, each of which contains a single active site that binds manganese or iron depending on aerobic conditions of the bacterial culture (Amano et al., 1990). The SOD from P. gingivalis may be reconstituted exclusively with manganese or iron, with full activity retained regardless of the metal present, a property known as cambialism (Yamakura et al., 1998). The active site is nearly identical to the active site of E. coli FeSOD, with the metal liganded by three histidine residues, one aspartic acid, and a solvent molecule in a trigonal bipyramidal geometry (Figure 4.1; Sugio et al., 2000). The hydrogen-bond structure in the active site is also similar to E. coli FeSOD in that glutamine originates from position 70 (numbering according to P. gingivalis SOD sequence) rather than position 142. Switching Gln70 to the more manganese-like position of Glu142 increased the Mn-substituted activity relative to the Fe-substituted SOD (Hiraoka et al., 2000). Further, the iron-like mutant Gly155Thr in P. gingivalis SOD increases the Fe-specific activity (Yamakura et al., 2003). Obligate Feand MnSODs have several distinct mechanistic properties. Although both catalyze the dismutation of superoxide through a cyclic redox process as described in Eqs. 1.1 and 1.2, only MnSODs are complicated by rapid appearance of a product-inhibited form (Eqs. 1.3, 1.4; Silverman and Nick, 2002). On the other hand, FeSODs 57

PAGE 67

58 can be inactivated by addition of H2O2, while MnSODs show no such inactivation (Beyer et al., 1989). Finally, catalysis by FeSOD is activated by exogenous proton donors such as primary amines (Bull and Fee, 1985); activation of MnSOD in a similar manner has not been observed. The SOD from P. gingivalis provides a very similar (though not necessarily the same) protein structure for both the Mnand Fe-substituted enzymes. This enables experimentation to determine if the metal is responsible for the mechanistic properties of activation and product inhibition. This work investigates the catalytic properties of Feand Mn-substituted P. gingivalis SOD. Activation of catalysis via exogenous proton donors was observed for P. gingivalis Feand potentially for MnSOD. The mechanism of activation for P. gingivalis FeSOD and MnSOD was consistent with the uncompetitive activation mechanism of E. coli FeSOD (Greenleaf and Silverman, 2002). The resting state of P. gingivalis MnSOD appeared to be mixed, with the Mn(II) form predominant. Product-inhibition of P. gingivalis MnSOD was less than that of human MnSOD, suggesting a connection between observation of activation and levels of product inhibition. Materials and Methods Enzymes Enzyme samples of P. gingivalis MnSOD and P. gingivalis FeSOD for stopped-flow and pulse radiolysis experiments were graciously provided by Dr. Fumiyuki Yamakura. Metal occupancy of the wild-type P. gingivalis MnSOD was 63.4% manganese and 1.1% iron. Metal occupancy of the wild-type P. gingivalis FeSOD was 59.0% iron and 0.6% manganese. The slight activity of the incorrectly substituted metal was not significant enough to affect data.

PAGE 68

59 Stopped-Flow Spectrophotometry The stopped-flow methods as described in Chapter 2 were used for determining kcat and kcat/Km and activation properties of P. gingivalis Fe-reconstituted SOD. Solution conditions upon final mixing were 0.5 M enzyme, 100 mM TAPS, 1 mM EDTA, up to 100 mM activating amine, and 4.5% DMSO by volume at pH 9.0. All experiments were carried out at 25 C with a 10 mm pathlength. Initial rates were determined by fitting the averaged data of 4-8 progress curves after subtraction of the uncatalyzed rate of dismutation. Pulse Radiolysis Pulse radiolysis experiments were carried out at Brookhaven National Laboratory in collaboration with Dr. Diane Cabelli using the 2 MeV van de Graaff accelerator. Dosimetry was determined using the KSCN dosimeter as described in Chapter 3. All UV and visible spectra were recorded on a Cary 210 spectrophotometer with a path length of 2.0 or 6.1 cm. Solutions contained enzyme, 30 mM formate and 0.5 M ethanol (as hydroxyl radical scavengers), 50 M EDTA, and 2 mM TAPS at pH 8.2. All solutions were air saturated. Superoxide radicals were formed upon pulsing by the mechanisms of Schwarz (1981). Superoxide decay was observed spectrophotometrically at 260 nm, while changes in the enzyme spectra were observed at 420 nm or 480 nm, as described in Chapter 3. Initial rates were calculated from the first 5% of progress curves; steady state parameters were determined by a least-squares fitting process (Enzfitter; Biosoft). Results Steady-state rate constants were measured for P. gingivalis Fe-substituted SOD by stopped-flow spectrophotometry and pulse radiolysis. Under stopped-flow conditions, P. gingivalis FeSOD was found to saturate with increasing superoxide concentration, with

PAGE 69

60 kcat = 11 1 ms-1 (Table 4.1). Pulse radiolysis, which cannot generate high enough concentrations of superoxide to saturate, was used to measure a kcat/Km of 190 10 M-1 s-1, in close agreement with the value obtained from stopped-flow. Pulse radiolysis was also used to measure the steady-state rate constants of P. gingivalis Mn-substituted SOD (Table 4.1). Saturation of P. gingivalis MnSOD was not observed at the low concentrations of superoxide generated by pulse radiolysis; kcat and kcat/Km values were similar to those of P. gingivalis FeSOD (Table 4.1). Under stopped-flow conditions of 800 M superoxide, O2•– decay catalyzed by 1 M P. gingivalis MnSOD displayed the zero-order kinetics characteristic of MnSODs (McAdam et al., 1977; Bull et al., 1991), with k0/[E] calculated at 2500 500 s-1 (Table 4.1). Catalytic activity of P. gingivalis Feand Mn-substituted SOD was enhanced by addition of primary amines. Using stopped-flow, activation of P. gingivalis FeSOD by 2-aminopropionate appeared to approach saturation, as shown in Figure 4.2, with (kcat/Km)donor = 120 mM-1 s-1, calculated using Eq. 2.3. Pulse radiolysis was used to measure the activation of P. gingivalis MnSOD by 40 mM glycine (Figure 4.3). Although modest, as superoxide concentration was increased, the degree of activation also appeared to increase, although saturating superoxide concentrations could not be reached. In the presence of 40 mM glycine, the value for kcat/Km for superoxide was 120 10 M-1 s-1, unchanged from the value in the absence of glycine. Rate constants in the simplified kinetic scheme described in Eqs. 3.1-3.4 were measured for P. gingivalis MnSOD. Single turnover experiments showed an increase in absorbance at both 420 (Figure 4.4) and 480 nm (Figure 4.4, inset) when starting in the resting state of the enzyme. Reduction of MnSOD by incubation with H2O2 caused a

PAGE 70

61 one-third increase in the maximal absorbance at 420 (Figure 4.5) and 480 nm measured after 2 ms (Figure 4.5, inset) upon pulsing with superoxide compared with Figure 4.4. The rate of increase at 420 and 480 nm was the same for both resting state and reduced enzyme, 80 7 M-1 s-1. Both the 420 and 480 peaks show similar rates of decay over longer time scales, near 60 s-1. Discussion For most MnSODs catalysis can be broken down into four steps (Eqs. 3.1-3.4), in which the enzyme rests as Mn+3SOD (Silverman and Nick, 2002). The Mn+3SOD form shows a distinct visible spectrum with a maximum absorbance at 480 nm; Mn+2SOD has a featureless visible spectrum. The product-inhibited form of the enzyme absorbs at 420 nm, but only very weakly at 480 nm, allowing separate measurement of these rate constants k1-k4 in single turnover experiments (Hearn et al., 2001). The MnSOD from P. gingivalis shows an increase in absorbance at both 480 and 420 nm upon pulsing the resting form of the enzyme with superoxide (Figure 4.4, inset). The rate of formation of these species is the same, suggesting that these two processes represent the same catalytic step. The increase in absorbance at both 420 and 480 nm suggests the kinetic process being observed is k2 of Eq. 3.2, and therefore implies P. gingivalis MnSOD rests in the Mn+2 form. Addition of H2O2 to P. gingivalis MnSOD would reduce any remaining Mn3+SOD to Mn2+SOD. The total absorbance at 480 nm corresponding to Mn+3SOD was greater by about 50% when starting with fully reduced enzyme (Figure 4.5). This is consistent with a mixed resting state in which two-thirds of P. gingivalis MnSOD rests as Mn+2, while one-third rests as Mn+3. A slow decay process followed the rapid formation

PAGE 71

62 Table 4.1: Steady-state constants for the decay of superoxide catalyzed by FeSOD and MnSOD and for the activation of catalysis by 3-aminopropionatea Enzyme kcat (ms-1) kcat/Km (M-1 s-1) (kcat/Km)donor (mM-1 s-1) k0/[E] (s-1) E. coli FeSOD 25 250 310 –c P. gingivalis FeSOD 11 1 190 10 120b –c Human MnSOD 40 800 –d 500 P. gingivalis MnSOD –e 110 10 –e 2500 500 a kcat for E. coli and P. gingivalis FeSOD determined by stopped flow with conditions as described in Figure 4.2 (pH 8.3, 25 oC); kcat/Km for P. gingivalis FeSOD and MnSOD determined by pulse radiolysis with solution conditions described in Materials and Methods (pH 8.2, 25 oC). Data for human MnSOD (pH 9.4, 20 oC) determined by pulse radiolysis was reported by Guan et al. (1996). Standard errors are typically 15% or less; b This experiment at pH 9.0, 100 mM TAPS. Other conditions as listed in footnote a; c No product inhibition observed; d No activation by buffers observed; e Not measured.

PAGE 72

63 Trp123 Tyr35 Gln70 Figure 4.1: The active site region of P. gingivalis FeSOD. The active site metal is shown as an orange sphere. The four protein ligands, consisting of three histidines and an aspartic acid, are shown in blue. The fifth ligand, a solvent molecule, is represented by a red sphere. Hydrogen bonds are shown in yellow.

PAGE 73

64 024681012020406[2-aminopropioniate], mMkcatobs 0 Figure 4.2: The dependence on 2-aminopropionate concentration of the turnover number kcatobs for catalysis of superoxide decay by 0.3 M P. gingivalis FeSOD. A pH of 9.0 was maintained by 100 mM TAPS. Initial superoxide concentration was 325 M; other final solution contents were as listed in the legend to Figure 2.1. The solid line is a fit to Eq. 2.2 with kcat = 6500 100 s-1, KmB = 30 8 mM, and = 2.3 0.3.

PAGE 74

65 00.511.522.530102030[O2•], Mv/[E], ms-1 Figure 4.3: Rate of superoxide decay catalysed by 0.5 M P. gingivalis MnSOD in the presence () and absence () of 40 mM glycine, measured by pulse radiolysis. Final solution conditions at 25 C included 30 mM formate, 0.5 M ethanol, 50 M EDTA, and 0.5 M perchlorate at pH 8.2 maintained by 2 mM TAPS. The solid line for superoxide decay in the absence of glycine was fit to the Michaelis-Menten equation, with (kcat/Km)O2 = 110 M-1 s-1; the solid line for superoxide decay in the presence of 40 mM glycine was fit to Eq. 2.3 with (kcat/Km)donor = 120 mM-1 s-1. Fitting of the data in the presence of 40 mM glycine to the Michaelis-Menten equation showed (kcat/Km)O2=120 10 M-1 s-1.

PAGE 75

66 02468101201020304050time, msA420 x 103 02468101201020304050time, msA480 x 103 Figure 4.4: Increase and subsequent decrease of absorbance at 420 nm of 25 M P. gingivalis MnSOD after pulsing with 2.7 M O2•. Solid line is the fit to the sum of two exponential equations with the increase at 420 nm described by a rate of 2.2 ms-1 and the subsequent decrease described by a rate of 64 s-1. Solution at 25 C and pH 8.2 contained 30 mM formate, 2 mM TAPS, 0.5 M ethanol, 50 M EDTA, and 0.5 M perchlorate. Inset: Increase and subsequent decrease at 480 nm of 25 M P. gingivalis MnSOD. Increase was described by a rate of 1.8 ms-1 and decrease was described by a rate 58 s-1. Conditions identical to those listed above.

PAGE 76

67 048121601020304050time, msA420 x 103 0510152001020304050time, msA480 x 103 Figure 4.5: Increase and subsequent decrease at 420 nm of 25 M P. gingivalis MnSOD that was incubated with 250 M H2O2 then pulsed with 2.7 M O2•. Increase was described by a rate of 2.0 ms-1 and decrease was described by a rate of 62 s-1. Conditions were as described in the legend of Figure 4.4. Inset: Increase and subsequent decrease at 480 nm of 25 mM P. gingivalis MnSOD that was incubated with 250 M H2O2 then pulsed with 2.7 M O2•. Increase was described by a rate of 1.9 ms-1 and decrease was described by a rate of 57 s-1.

PAGE 77

68 of Mn+3SOD from Mn+2SOD. One possibility is that this decay represents a return to the equilibrium of mixed oxidation states of P. gingivalis MnSOD. Vance and Miller (1998) proposed that superoxide dismutases provide an environment that reduces the potential of the free iron or manganese by a fixed amount. In order to achieve efficient catalysis, the midpoint potential of the complexed metal must be midway between the redox potentials of the two half reactions of superoxide, mV for O2/O2 and +870 mV for O2/H2O2. Free manganese has a much higher redox potential (1510 mV) than does free iron (770 mV; Sawyer et al., 1995); therefore, it was suggested that obligate MnSODs must more strongly depress E0 than obligate FeSODs. Cambialistic SODs would therefore need to lower the potentials of iron and manganese by approximately 700 mV, an amount that would leave the E0 of both metals between the E0 of the two half-reactions of superoxide oxidation and reduction. For P. gingivalis MnSOD, the redox potential of the metal would be anticipated to be near 700-800 mV. Manganese-substituted FeSODs with redox potentials greatly elevated above that of MnSOD (E0 near 400 mV) rest in the Mn+2 state (Vance and Miller, 2001). Therefore, P. gingvalis MnSOD with a potentially slightly elevated redox potential could be expected to rest in a mixed Mn+2/Mn+3 state. If P. gingivalis MnSOD does not have an optimized redox potential then the steps in the catalysis for oxidation (k1 of Eq. 3.1) and reduction (k2 of Eq. 3.2) of superoxide would be expected to occur with varying efficiency. The step in the catalysis converting O2 to H2O2 (Eq. 3.2) would be the slow step for a cambialistic MnSOD; for P. gingivalis MnSOD k2 is 80 7 M (Figure 4.3), significantly slower than k2 for wild-type human MnSOD which is 1100 M-1 s-1 (Hearn et al., 2001). Further, compared with the

PAGE 78

69 steady-state rate constants of human MnSOD, Mn-substituted SOD from P. gingivalis shows decreases in both kcat and kcat/Km (Table 4.1). For P. gingivalis MnSOD kcat/Km is of a similar magnitude as k2, which is to be expected. Iron-substituted SOD from P. gingivalis would also be anticipated to have less efficient catalytic steps as compared to obligate FeSODs; kcat/Km is depressed roughly 3-fold compared to E. coli FeSOD (Table 4.1). Activation of the proton transfer pathway leading to escape from product inhibition has not been observed in human MnSOD (Ramilo et al., 1999). It is not clear if this effect is due to protein influence or to the metal. The cambialistic MnSOD from P. gingivalis would likely provide a similar protein structure for both the Feand Mn-substituted forms, thus supplying an internal control. Properties of activation and product inhibition were measured in P. gingivalis SOD to test if both the Feand MnSOD could be activated in a similar manner to E. coli FeSOD. Figure 4.2 demonstrates that activation of P. gingivalis FeSOD by a primary amine, 2-aminoproprionic acid, was saturable and nonessential, since significant activity existed in the absence of primary amine. The apparent-second order rate constant of activation for P. gingivalis FeSOD, (kcat/Km)donor = 120 mM-1 s-1, is consistent with that of E. coli FeSOD at 310 mM-1 s-1, indicating that activation might be a general property of all FeSODs. Although inconclusive, the data are suggestive that P. gingivalis MnSOD was activated by glycine under pulse radiolysis conditions (Figure 4.2). Activation by glycine did not appear to change kcat/Km for superoxide, which is consistent with the proposed uncompetitive activation mechanism for FeSOD (Greenleaf and Silverman, 2002). This implies that at least the proton transfer pathway supporting kcat in MnSOD is capable of

PAGE 79

70 being activated by exogenous proton donors. Further, the catalytic steps for P. gingivalis MnSOD (Eqs. 3.1, 3.2) are possibly very similar to those of FeSOD. These studies do not rule out the possibility that the more constricted active site of human MnSOD compared to E. coli and P. gingivalis FeSODs prevents access of the activating amine.

PAGE 80

CHAPTER 5 CONCLUSIONS AND FUTURE DIRECTIONS Conclusions The work in this dissertation investigated the similarities between the proton transfer pathways of Feand MnSOD. The proton transfer pathway of E. coli FeSOD was studied by examining the activation of catalysis by exogenous primary amines. As was suggested by Bull and Fee (1985), this study showed activation to be proton-transfer dependent by correlating the apparent second-order rate constant of activation with the pKa of the activating amine. The mechanism of activation was found to be nonessential and uncompetitive, consistent with proton transfer in a ternary complex of enzyme, amine, and the proposed proton acceptor peroxide dianion. The bound peroxide dianion would be anticipated to have a high pKa, consistent with the pH independent nature of kcat, which contains the proton-transfer dependent step(s) of catalysis (Bull and Fee, 1985; Fee et al., 1986) (Chapter 2). The potential proton transfer pathway in human MnSOD was examined by investigating the effects of site-specific substitutions at Trp123 and Tyr34, which form hydrogen bonds with the critical active-site residue Gln143. Mutants at positions 123 and 34 (Y34F, W123F, Y34F-W123F) were shown to have a rate limiting proton transfer occurring for the release of the product-inhibited form (k4 of Eq. 3.4). For k1 and k3 (Eqs. 3.1, 3.3), all of these mutants displayed similar rates and dependence on pH, suggesting superoxide interacts with the enzyme in a similar manner for both catalytic processes. However, k2 (Eq. 3.2) values were greatly decreased in these mutants, implying 71

PAGE 81

72 superoxide interacts in a different manner than k1 and k3 for this step of catalysis. Because of the similar results seen in mutants W123F, Y34F, and the double mutant Y34F-W123F, these residues are thought to interact cooperatively in their effects on catalysis. Interestingly, substitutions at Trp161 reproduce similar effects on k2, k3 and k4 (Hearn et al., 2001), suggesting the general role for the tryptophans in the active site of MnSOD is in maintaining efficient reoxidation of the enzyme, possibly by providing a nonspecific binding site for superoxide (Chapter 3). This dissertation also addressed the reason why activation of the proton transfer pathway of MnSOD in a manner analogous to FeSOD has not yet been observed (Ramilo et al., 1999). Possible activation of the less product-inhibited cambialistic P. gingivalis MnSOD suggested that rapid appearance of product inhibition in human MnSOD does not allow for observation of activation (Chapter 4). This work suggests that overall the catalytic pathways of MnSOD and FeSOD are quite similar, with the exception that FeSOD does not proceed through a product inhibited complex, the reasons for which are not clear. Further, the proton transfer dependent processes defined by k2 and k4 are likely quite different, suggested by the lack of activation of the proton transfer dependent step k4. Future Directions Activation by Exogenous Proton Donors Further activation studies on both P. gingivalis Feand MnSOD would be useful to determine several subtle differences between Feand MnSODs. Construction of a free energy diagram similar to Figure 2.3 for both P. gingivalis Feand MnSOD may give different indications of the pKa of the proton donor and acceptors on the enzymes, since both of these enzymes would be anticipated to have a different redox potential than E.

PAGE 82

73 coli FeSOD, and thus different pKa’s of the metal bound solvent molecule. Further, it would be useful to compare the Feand MnSODs from P. gingivalis to give some indication of the role the protein structure plays during activation. Finally, confirmation of the activation mechanism of P. gingivalis MnSOD is necessary to establish the similarities of the proton transfer pathways of Feand MnSODs, which is suggested by the preliminary studies of P. gingivalis MnSOD in Chapter 4. Site-Directed Mutagenesis Yamakura and coworkers have identified several mutations in P. gingivalis SOD that shift the metal specific activity of the enzyme. First, switching the position of the active site Gln from the wild-type Gln70 which resembles most FeSODs to Gln142 which resembles most MnSODs, confers a greater degree of Mn-specific activity upon the mutant enzyme (Hiraoka et al., 2000). Second, replacement of Gly155 with Thr, similar to most FeSODs, increases the enzymatic activity of Fe-substituted SOD compared to Mn-substituted SOD (Yamakura et al., 2003). These mutations have not been attempted in human MnSOD, although the G77Q-Q146A E. coli MnSOD mutant was constructed, and showed an increase from 0% to 7% of MnSOD activity when substituted with iron (Schwartz et al., 2000). Possible mutants that would favor iron specificity in human MnSOD would be Gly70Gln-Gln143Ala, Gly157Thr, and a double mutant containing both substitutions, made in an effort to convert MnSOD to FeSOD. Based on the metal-specific catalytic properties of P. gingivalis MnSOD shown in chapter 4, a G70Q-Q143A human MnSOD might be anticipated to have a significantly lower degree of product inhibition compared to wild-type, and could be therapeutically useful in treatment of cancer or radiation damage.

PAGE 83

74 In Y34F-W123F human MnSOD two hydrogen bond partners of Gln143 were removed, with little change in structure about the active-site (Figure 3.8). In wild-type human MnSOD, Asn73 is 2.8 from Gln143 and is also capable of forming a hydrogen bond with Gln143. Similarly in Y34F-W123F MnSOD, Asn73 is 3.2 from Gln143 and could form a weak hydrogen bond, stabilizing the position of Gln143. Therefore, the N73A mutant was created to remove this hydrogen bond partner to Gln143 (Greenleaf et al., unpublished). However, N73A MnSOD was not able to be purified, and was likely unstable. Several other substitutions could be attempted at this site in an effort to find a stable variant at the Asn73 position. Once created this mutation could be introduced in a triple mutant containing Y34F-N73X-W123F in which all the hydrogen-bond partners of Gln143 have been removed. This triple mutant would be useful to determine the protein components necessary to establish a catalytically competent MnSOD. Several tryptophan residues in the active site of MnSOD could be mutated in tandem to investigate the role these residues play in maintaining efficient reoxidation of the manganese. These residues include Trp78, Trp123, Trp125, and Trp161. Neither Trp78 nor Trp125 have been investigated in any detail, and deserve further study. Residue Trp78 is more distant from the active-site metal than residues 123 and 161, and emanates from a point opposite the metal compared to the other tryptophan residues. Dr. Deepa Bhatt in the Silverman laboratory mutagenized Trp78 to Phe, and found that unlike the Phe mutants at positions 123 and 161, levels of product inhibition were not increased compared to wild-type. Since product inhibition in W123F and W161F results from a reduction in k2 (Chapter 3; Hearn et al., 2001), Trp78 may not be in the region of superoxide interaction suggested in Chapter 3 to be present for W123 and W161 for k2.

PAGE 84

75 Further investigation of this site would confirm its role in catalysis. Residues Trp123 and Trp125 form a f-stacking interaction in wild-type MnSOD. Changing the more distant Trp125 may alter the position of Trp123, which due to its proximity to the metal and to Gln143 could alter catalytic properties. This residue is also a good candidate for establishing a favorable reoxidation (k2) step. Both k1 and k3 for the mutant MnSODs examined in this work display a similar pH dependence of pKa = 9.5, the source of which is unclear (Chapter 3). Because these steps are measured from different states of the enzyme, it is possible that it is their common interaction with superoxide that is the source of the pKa. That is, deprotonation of a group on the enzyme may cause a charge-charge interaction with superoxide to occur thus decreasing the amount of superoxide that reaches the active site. At least two residues would be capable of contributing to this ionization, His30 and Tyr166. While free tyrosine has a pKa near that of the ionizations of k1 and k3, shifts in pKa values of up to 3 units are not uncommon, thus His30 is also a reasonable candidate. Since Tyr166 would be negatively charged at high pH, it would be more likely to have a charge-charge interaction with superoxide. Double mutants W123F-H30N and W123F-Y166F would be useful in testing this hypothesis. Further, the W123F-H30N mutant would be interesting because of the contrast in single mutants, with W123F being highly product inhibited and H30N displaying a low degree of product inhibition. Redox Potential Measurement of the redox potential of a cambialistic SOD would help to confirm the mechanism of action of these enzymes, and how they maintain activity with either iron or manganese present in the active site. Based upon kinetic studies of Chapter 4 and

PAGE 85

76 the proposal made by Vance and Miller (1998), P. gingivalis Feand MnSODs would be expected to have redox potentials lower than E. coli FeSOD and greater than human MnSOD, respectively. Because of the buried nature of the active sites of Mnand FeSODs, mediators are required to carry charge between the metal and the electrode. Typically, mediators must have a similar E0 to that of the metal in order to transfer charge. Since differences in E0 > 200 mV could exist in P. gingivalis SODs compared to obligate SODs, new mediators must be used. It is likely that a cambialistic SOD protein would lower the redox potential of free iron more than the E. coli FeSOD protein would, thus for P. gingivalis FeSOD, analogs with lower E0 values than p-benzoquinone and dichloroindophenol (near 200 mV) and higher than benzylviologen (near –200 mV) would be necessary to use. Similarly, a cambialistic SOD protein would lower the redox potential of free manganese less than the human MnSOD protein would, resulting in a redox potential greater than the 400 mV measured for human MnSOD (Lvque et al., 2001), possibly near 700 mV. The mediators used for human MnSOD include ferrocene, which has a limited solubility in phosphate buffer at pH 7.8, and ferricyanide. Therefore, other conditions and mediators would need to be explored in order to measure the redox potential of MnSOD with a higher E0. Mediators capable of measuring elevated MnSOD redox potentials could also be useful in measuring E0 values for W123F and Y34F-W123F MnSOD, as well as mutants at the W161 position. Based on the kinetic studies of Chapter 3 in this dissertation and work done by Hearn et al. (2001), the step involving reoxidation of the manganese has been slowed, reflected by k2 values reduced by at least 50-fold. It may then be possible that the redox potential of these mutants is elevated in a manner analogous to P.

PAGE 86

77 gingivalis MnSOD, which also shows a slowed k2 step (Chapter 4). Measurement of redox potentials of these mutants could confirm if this hypothesis is true, or if the earlier suggestion of a reduction in the inner sphere binding affinity of superoxide for the k2 step would support the data. X-Ray Crystallography Activation of E. coli FeSOD by addition of primary amines as exogenous proton donors was found to be saturable, indicating binding of the primary amine to the enzyme (Chapter 2). Thus, it may be possible to determine the location of the bound amine by diffusing in a primary amine to an FeSOD crystal and then determining the crystal structure with activator bound. This approach could improve the understanding of the proton transfer pathway in FeSOD, as these amines do not have to directly transfer their proton to substrate, but instead could do so through an intervening water structure. Similar experiments have been performed in carbonic anhydrase II, in which activation by 4-methylimidazole of a mutant lacking the proton shuttle His64 was saturable and a binding site for the activator was determined by X-ray crystallography (Duda et al., 2001). While a crystal structure of Y34F-W123F MnSOD revealed few differences between mutant and wild-type, no crystal structure of the single mutant W123F MnSOD was obtained. A crystal structure of this mutant may be useful in explaining the large difference observed for the resting extinction coefficient at 480 nm between wild-type and this mutant (Chapter 3). Another unexplained phenomenon observed in W123F containing mutants is the appearance of a pH dependence in the proton transfer dependent decay of the product inhibited form (Chapter 3). Clearly this is not due to titration of the proton donor group, since the rate is increasing as protons become less

PAGE 87

78 available. One possibility is deprotonation of a group on the enzyme removes a steric hindrance and allows more rapid release from product inhibition. Most crystal structures of MnSOD, including the one in this work, have been at pH 8.0 or less. This work suggests that crystal structures at high pH (10.0 or greater) of wild-type and W123F containing mutant MnSODs could provide interesting structural data to explain these catalytic properties, and may give insight into the differences in the proton transfer pathways for the main catalytic cycles of Feand MnSOD, and the proton transfer pathway necessary for the decay of the product inhibited form. Identification of the structure of the product-inhibited form of MnSOD would provide the most information in determining the differences in proton transfer between Feand MnSODs. Because the product-inhibited form of the enzyme decays relatively quickly, it must be trapped for a sufficient time for a structural determination to be made. The most promising approach would be to use the H30V human MnSOD mutant for crystallography, which has a very slow decay of the product-inhibited form compared to wild-type and other mutant MnSODs (Hearn et al., 2003). Upon obtaining a crystal of H30V MnSOD, hydrogen peroxide could be diffused in, and subsequent rapid freezing could trap the product-inhibited form.

PAGE 88

LIST OF REFERENCES Adams, P. D., Pannu, N. S., Read, R. J., and Brunger, A. T. (1997) Proc. Natl. Acad. Sci. U. S. A. 94, 5018-5023. Amano, A., Shizukuishi, S., Tamagawa, H., Iwakura, K., Tsunasawa, S., and Tsunemitsu, A. (1990) J. Bacteriol. 257, 1457-1463. An, H., Tu, C. K., Duda, D., Montanez-Clemente, I., Math, K., Laipis, P. J., McKenna, R., and Silverman, D. N. (2002) Biochemistry 41, 3235-3242. Bannister, J. V., Bannister, W. H., and Rotilio, G. (1987) CRC Crit. Rev. Biochem. 22, 111-180. Barshop, B. A., Wrenn, R. F., and Frieden, C. (1983) Anal. Biochem. 130, 134-145. Beck, Y., Oren, R., Amit, B., Levanon, A., Gorecki, M., and Hartman, A. J. (1987) Nucleic Acids Res. 15, 9076. Benovic, J., Tillman, T., Cudd, A., and Fridovich, I. (1983) Arch. Biochem. Biophys. 221, 329-332. Bergman, N.-A., Chiang, Y., and Kresge, A. J. (1978) J. Am. Chem. Soc. 100, 5954-5956. Beyer, W. F., Jr., and Fridovich, I. (1987) Biochemistry 26, 1251-1257. Beyer, W. F., Jr., Reynolds, J. A., and Fridovich, I. (1989) Biochemistry 28, 4403-4409. Borgstahl, G. E. O., Parge, H. E., Hickey, M. J., Beyer, W. F., Hallewell, R. A., and Tainer, J. A. (1992) Cell 71, 107-118. Brunger, A. T., Adams, P. D., Clore, G. M., DeLano, W. L., Gros, P., Grosse-Kunstleve, R. W., Jiang, J. S., Kuszewski J., Nilges, M., Pannu, N. S., Read, R. J., Rice, L. M., Simonson, T., and Warren, G. L. (1998) Acta Crystallogr D Biol Crystallogr 54 (Pt 5), 905-921. Bull, C., and Fee, J. A. (1985) J. Amer. Chem. Soc. 107, 3295-3304. Bull, C., Niederhoffer, E. C., Yoshida, T., and Fee, J. A. (1991) J. Am. Chem. Soc. 113, 4069-4076. 79

PAGE 89

80 Cabelli, D. E., Guan, Y., Leveque, V., Hearn, A. S., Tainer, J. A., Nick, H. S., and Silverman, D. N. (1999) Biochemistry 38, 11686-11692. Cabelli, D.E., Riley, D., Rodriguez, J.A.A., Valentine, J.S., and Zhu, H. (2000) in Biomimetic Oxidations Catalyzed by Transition Metal Complexes (Meunier, B., Ed.) Chapter 10, Imperial College Press, London, pp 461-508. Carlioz, A., and Touati, D. (1986) EMBO J. 5, 623-630. Choudhury, S. B., Lee, J.-W., Davidson, G., Yim, Y.-I., Bose, K., Sharma, M. L., Kang, S.-O., Cabelli, D. E., and Maroney, M. J. (1999) Biochemistry 38, 3744-3752. Cox, M. M., and Jencks, W. P. (1978) J. Am. Chem. Soc 100, 5956-5957. Cullen, J. J., Weydert, C., Hinkhouse, M. M., Ritchie, J., Domann, F. E., Spitz, D., and Oberley, L. W. (2003) Cancer Res. 63, 1297-1303. Davis, C. A., Hearn, A. S., Fletcher, B., Bickford, J., Garcia, J. E., Leveque, V., Melendez, J. A., Silverman, D. N., Zucali, J., Agarwal, A., and Nick, H. S. (In press) J. Biol. Chem. DiDonato, M., Craig, L., Huff, M. E., Thayer, M. M, Cardoso, R. M., Kassman, C. J., Lo, T. P., Bruns, C. K., Powers, E. T., Kelly, J. W., Getzoff, E. D., and Tainer, J. A. (2003) J. Mol. Biol. 332, 601-615. Duda, D., Tu, C., Qian, M., Laipis, P., Agbandje-McKenna, M., Silverman, D. N., and McKenna, R. (2001) Biochemistry 40, 1741-1748. Edwards, R. A., Whittaker, M. M., Whittaker, J. H., Baker, E. N., and Jameson, G. B. (2001) Biochemistry 40, 15-27. Eigen, M., and Hammes, G. G. (1963) Adv. Enzymol. 25, 1-38. Epperly, M. W., Epstein, C. J., Travis, E. L., and Greenberger, J. S. (2000) Radiat. Res. 154, 365-374. Farr, S. B., D’Ari, R., and Touati, D. (1986) Proc. Natl. Acad. Sci. U. S. A. 83, 8268-8272. Fee, J. A., McClune, G. J., Lees, A. C., Zidovetzki, R., and Pecht, I. (1981) Isr. J. Chem. 21, 54-58. Fee, J. A., Yoshida, T., Bull, C., O’Neill, P., and Fielden, E. M. (1986) in Superoxide and Superoxide Dismutase in Chemistry, Biology and Medicine (Rotilio, G., ed), pp. 205-211, Elsevier Science Publishing Co., Inc., New York Fridovich, I. (1997) J. Biol. Chem. 272, 18515-18517. Greenleaf, W. B., and Silverman, D. N. (2002) J. Biol. Chem. 277, 49282-49286.

PAGE 90

81 Guan, Y., Hickey, M. J., Borgstahl, G. E. O., Hallewell, R. A., Lepock, J. R., O’Connor, D., Hsieh, Y., Nick, H. S., Silverman, D. N., and Tainer, J. A. (1998) Biochemistry 37, 4722-4730. Haber, F. and Weiss, J. (1934) Proc. R. Soc. London Ser. A 147, 332-351. Hearn, A. S., Fan, L., Lepock, J. R., Luba, J. P., Greenleaf, W. B., Cabelli, D. E., Tainer, J. A., Nick, H. S., and Silverman, D. N.(2004) J. Biol. Chem. 279, 5861-5866. Hearn, A. S., Stroupe, M. E., Cabelli, D. E., Lepock, J. R., Tainer, J. A., Nick, H. S., and Silverman, D. N. (2001) Biochemistry, 40, 12051-12058. Hearn, A. S., Stroupe, M. E., Cabelli, D. E., Ramilo, C. A., Luba, J. P., Tainer, J. A., Nick, H. S., and Silverman, D. N. (2003) Biochemistry 42, 2781-2789. Hearn, A. S., Tu, C. K., Nick, H. S., and Silverman, D. N., (1999) J. Biol. Chem. 274, 24457-24460. Hiraoka, B. Y., Yamakura, F., Sugio, S., and Nakayama, K. (2000) Biochem J. 345, 345-350. Hsieh, Y., Guan, Y., Tu, C., Bratt, P. J., Angerhofer, A., Lepock, J. R., Hickey, M. J., Tainer, J. A., Nick, H. S., and Silverman, D. N. (1998) Biochemistry 37, 4731-4739. Hsu, J.-L., Hsieh, Y., Tu, C. K., O’Connor, D., Nick, H. S., and Silverman, D. N. (1996) J. Biol. Chem. 271, 17687-17691. Jonsson, B.-H., Steiner, H., and Lindskog, S. (1976) FEBS Lett. 64, 310-314. Kresge, A. J. (1975a) Acc. Chem. Res. 8, 354-360. Kresge, A. J. (1975b) in Proton-Transfer Reactions (Caldin, E. F., and Gold, V., eds), pp. 179-199, John Wiley & Sons, Inc., New York. Kresge, A. J., and Silverman, D. N. (1999) Methods Enzymol. 349, 61-74. Lah, M. S., Dixon, M. M., Pattridge, K. A., Stallings, W. C., Fee, J. A., and Ludwig, M. L. (1995) Biochemistry 34, 1646-1660. Lepock, J. R., Frey, H. E., and Hallewell, R. A. (1990) J. Biol. Chem. 265, 21612-21618. Lvque, V. J.-P., Vance, C. K., Nick, H. S., and Silverman, D. N. (2001) Biochemistry 40, 10586-10591. Li, Y., Huang, T. T., Carlson, E. J., Melov, S., Ursell, P. C., Olson, J. L., Noble, L. J., Yoshimura, M. P., Berger, C., Chan, P. H., Wallace, D. C., and Epstein, C. J. (1995) Nature Genetics 11, 376-381.

PAGE 91

82 Liang, Z., Xue, Y., Behravan, G., Jonsson, B. H., and Lindskog, S. (1993) Eur. J. Biochem. 211, 821-827. Lindskog, S. (1997) Pharmacol. Ther. 74, 1-20. Maliekal, J., Karapetian, A., Vance, C., Yikilmaz, E., Wu, Q., Jackson, T., Brunold, T. C., Spiro, T. G., and Miller, A.-F. (2002) J. Am. Chem. Soc. 124, 15064-15075. McAdam, M. E., Fox, R. A., Lavelle, F., and Fielden, E. M. (1977) Biochem. J. 165, 71-79. McCord, J. M. and Fridovich, I. (1968) J. Biol. Chem. 243, 5753-5760. McCord, J. M. and Fridovich, I. (1969) J. Biol. Chem. 244, 6049-6055. McRee, D. E. (1999) J Struct Biol 125, 156-165. McRee, D. E., Redford, S. M., Getzoff, E. D., Lepock, J. R., Hallewell, R. A., and Tainer, J. A. (1990) J. Biol. Chem. 265, 14234-14241. Mildvan, A. S., Weber, D. J., and Kuliopulos, A. (1992) Arch. Biochem. Biophys. 294, 327-340. Miller, A.-F., Padmakumar, K., Sorkin, D. L., Karapetian, A., and Vance, C. K. (2003) J. Inorg. Biochem. 93, 71-83. Misra, H., and Fridovich, I. (1972) J. Biol. Chem. 247, 6960-6962. Misra, H. P., and Fridovich, I. (1978) Arch. Biochem. Biophys. 189, 317. Orr, W. C., and Sohal, R. S. (1994) Science 263, 1128-1130. Otwinowski, Z. and Minor, W. (1997) in Methods in Enzymology (Carter, C. W., Jr., and Sweet, R. M., Eds.) Vol. 276, pp 307-326 Academic Press, New York. Phillips, J. D., Campbell, S. D., Michaud, D., Charbonneau, M., and Hilliker, A. J. (1989) Proc. Natl. Acad. Sci. U. S. A. 86, 2761-2765. Pocker, Y., Janjic, N., and Miao, C. H. (1985) in Zinc Enzymes (Bertini, I., Luchinat, C., Maret, W., Zeppezauer, M., eds.) pp 341-356, Birkhauser, Boston. Rabani, J., and Nielson, S. O. (1969) J. Phys. Chem 73, 3736-3744. Ramilo, C. A., Leveque, V., Guan, Y., Lepock, J. R., Tainer, J. A., Nick, H. S., and Silverman, D. N. (1999) J. Biol. Chem. 274, 27711-27716. Rosenblum, J. S., Gilula, N. B., and Lerner, R. A. (1996) Proc. Natl. Acad. Sci. U. S. A. 93, 4471-4473.

PAGE 92

83 Rowlett, R. S., and Silverman, D. N. (1982) J. Am. Chem. Soc. 101, 6734-6741. Sawyer, D. T, Sobkowiak, A., and Roberts, J. L., Jr. (1995) Electrochemistry for Chemists, John Wiley and Sons, Inc., New York. Schmidt, M., Meier, B., and Parak, F. (1996) J. Biol. Inorg. Chem. 1, 532-541. Schwartz, A. L., Yikilmaz, E., Vance, C. K., Vathyam, S., Koder, R. L., and Miller, A.-F. (2000) J. Inorg. Biochem. 80, 247-256. Schwarz, H. A. (1981) J. Chem. Educ. 58, 101-105. Segel, I. H. (1975) Enzyme Kinetics, pp 183-188, 227-231, John Wiley & Sons, Inc., New York. Silverman, D. N. and Nick, H. S. (2002) in Methods in Enzymology (Packer, L., Ed.) Vol. 349, pp 61-74 Academic Press, New York. Silverman, D. N., Tu, C. K., Chen, X., Tanhauser, S. M., Kresge, A. J., and Laipis, P. J. (1993) Biochemistry 32, 10757-10762. Slykhouse, T. O., and Fee, J. A. (1976) J. Biol. Chem. 251, 5472-5477. Stallings, W. C., Pattridge, K. A., Strong, R. K., and Ludwig, M. L. (1985) J. Biol. Chem. 260, 16424-16432. Stallings, W. C., Powers, T. B., Pattridge, K. A., Fee, J. A., and Ludwig, M. A. (1983) Proc. Natl. Acad. Sci. U. S. A. 80, 3884-3888. Sturtevant, J. M. (1987) Annu. Rev. Phys. Chem. 265, 463-488. Sugio, S., Hiraoka, B. Y., and Yamakura, F. (2000) Eur. J. Biochem. 267, 3487-3495. Taoka, S., Tu, C. K., Kistler, K. A., and Silverman, D. N. (1994) J. Biol. Chem. 269, 17988-17992. Toney, M. D., and Kirsch, J. F. (1989) Science 243, 1485-1488. Tsan, M.-F., White, J. E., Caska, B., Epstein, C. J., and Lee, C. J. (1998) Am. J. Respir. Mol. Cell. Biol. 19, 114-120. Tu, C. K., Silverman, D. N., Forsman, C., Jonsson, B. H., and Lindskog, S. (1989) Biochemistry 28, 7913-7918. Vance, C. K. and Miller, A.-F. (1998) J. Amer. Chem. Soc. 120, 461-467. Vance, C. K. and Miller, A.-F. (2001) Biochemistry 40, 13079-13087. Westheimer, F. H. (1961) Chem. Rev. 61, 265-273.

PAGE 93

84 Whittaker, M. M., and Whittaker, J. W. (1996) Biochemistry 35, 6762-6770. Wilshire, J., and Sawyer, D. T. (1979) Acc. Chem. Res. 12, 105-110. Yamakura, F., Rardin, R. L., Petsko, G. A., Ringe, D., Hiraoka, B. Y., Nakayama, K., Fujimura, T., Taka, H., and Murayama, K. (1998) Eur. J. Biochem. 253, 49-56. Yamakura, F., Sugio, S., Hiraoka, B. Y., Ohmori, D., and Yokota, T. (2003) Biochemistry 42, 10790-10799.

PAGE 94

BIOGRAPHICAL SKETCH William B. Greenleaf was born in Opelika, Alabama, on July 30, 1977, and grew up in nearby Auburn, Alabama. Upon graduating from Auburn High School in 1995, he enrolled as a National Merit Scholar at Florida State University. After graduating in the fall of 1998 with a degree in biochemistry, he worked as a research technician until starting graduate studies at the University of Florida in the fall of 1999. He joined the lab of Dr. David N. Silverman in the Department of Pharmacology and Therapeutics, where he focused his dissertation on the role of proton transfer in the catalytic mechanisms of iron and manganese superoxide dismutase. After graduation he plans on pursuing a post-doctoral position at the University of Colorado Health Science Center in the lab of Dr. Xiaojiang Chen, and would like to eventually direct his own research group in an academic setting. 85