<%BANNER%>

Structural Studies of the Catalytic Mechanisms of Two Superfast Metalloenzymes

Permanent Link: http://ufdc.ufl.edu/UFE0041215/00001

Material Information

Title: Structural Studies of the Catalytic Mechanisms of Two Superfast Metalloenzymes the Carbonic Anhydrases and Manganese Superoxide Dismutases
Physical Description: 1 online resource (130 p.)
Language: english
Creator: Domsic, John
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: ca, crystallography, kinetics, mnsod
Biochemistry and Molecular Biology (IDP) -- Dissertations, Academic -- UF
Genre: Medical Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Structural studies of the catalytic mechanisms of two metalloenzymes were performed to gain a better understanding of their catalytic mechanisms. Analysis of human manganese superoxide dismutase revealed that there are subtle variations compared to the bacterial MnSODs that account for the differences in catalytic efficiency between the two enzymes. For the first time, the binding of substrate to human carbonic anhydrase II was performed, showing how the enzyme is able to catalyze with a high turnover rate. Mutational analysis of two CAs showed how the active site environment is responsible for fine-tuning catalytic activity. These studies give insight into how metalloenzymes function and provide specific examples of how analysis must be performed outside of the active site to truly understand their function.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by John Domsic.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: McKenna, Robert.
Local: Co-adviser: Silverman, David N.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2010-06-30

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0041215:00001

Permanent Link: http://ufdc.ufl.edu/UFE0041215/00001

Material Information

Title: Structural Studies of the Catalytic Mechanisms of Two Superfast Metalloenzymes the Carbonic Anhydrases and Manganese Superoxide Dismutases
Physical Description: 1 online resource (130 p.)
Language: english
Creator: Domsic, John
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: ca, crystallography, kinetics, mnsod
Biochemistry and Molecular Biology (IDP) -- Dissertations, Academic -- UF
Genre: Medical Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Structural studies of the catalytic mechanisms of two metalloenzymes were performed to gain a better understanding of their catalytic mechanisms. Analysis of human manganese superoxide dismutase revealed that there are subtle variations compared to the bacterial MnSODs that account for the differences in catalytic efficiency between the two enzymes. For the first time, the binding of substrate to human carbonic anhydrase II was performed, showing how the enzyme is able to catalyze with a high turnover rate. Mutational analysis of two CAs showed how the active site environment is responsible for fine-tuning catalytic activity. These studies give insight into how metalloenzymes function and provide specific examples of how analysis must be performed outside of the active site to truly understand their function.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by John Domsic.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: McKenna, Robert.
Local: Co-adviser: Silverman, David N.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2010-06-30

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0041215:00001


This item has the following downloads:


Full Text

PAGE 1

1 STRUCTURAL STUDIES OF THE CATALYTIC MECHANISMS OF TWO SUPERFAST METALLOENZYMES: THE CARBONIC ANHYDRASES AND MANGANESE SUPEROXIDE DISMUTASES By JOHN FRANCIS DOMSIC 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 2009

PAGE 2

2 2009 John Francis Domsic

PAGE 3

3 To my parents, Kenneth and Gabrielle Domsic

PAGE 4

4 ACKNOWLEDGMENTS I would like to thank my ment or, Dr. Robert McKenna for his continued support and guidance. He has provided much insight into my research and has given myriad suggestions that have guided my research down the right path. His enthusiasm for research has been a key part of my motivati on and success. I would also like to thank Dr. David Silverman for his mentorship and his availability to answer any questions that I have had throughout the past few years. In addition, I would like to thank the other members of my committee, Dr. Art Ed ison and Dr. John Aris. My committee has provided me with insight and support for my research that has been a major part of keeping me on track. I would like to thank my fellow labmates and friends for all of their help throughout my time in the lab. Dr. Mavis AgbandjeMcKenna, Dr. Lakshmanan Govindasamy, Dr. Hyun Joo Nam, and Dr. Chingkuang Tu have provided me with assistance, without which none of this work would have been possible. My fellow graduate students in the McKenna Lab have provided an environment that has made working in the lab a wonderful experience. I would like to thank the undergraduates who I have had the pleasure to work with for all of their assistance with my research. Additionally I would like to acknowledge all of the researchers who I have had the pleasure to collaborate with over the past 4 years. Their assistance and insight was invaluable to this research. Finally, I would like to thank my family for their continued support and guidance during this important stage of my career.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................. 4 LIST OF FIGURES .......................................................................................................... 9 LIST OF ABBREVIATIONS ........................................................................................... 12 ABSTRACT ................................................................................................................... 15 CHAPTER 1 INTRODUCTION .................................................................................................... 17 Metalloproteins ....................................................................................................... 17 Superoxide Dismutases .......................................................................................... 18 The Superoxide Radical ................................................................................... 18 Superoxide Dismutases .................................................................................... 19 Phenotypes of SOD deficiency .................................................................. 20 Catalytic mechanism .................................................................................. 21 Structura l descriptions ............................................................................... 22 Carbonic Anhydrase ............................................................................................... 23 ..................................................................... 24 Catalytic mechanism .................................................................................. 24 Structure .................................................................................................... 25 class Isoforms ............................................................................................... 27 Carbonic Anhydrase Deficiencies ..................................................................... 27 2 THE ROLE OF THE DIMERIC INTERFACE IN CATALYSIS BY HUMAN MANGANESE SUPEROXIDE DISMUTASE ........................................................... 34 Introduction ............................................................................................................. 34 Materials and Methods ............................................................................................ 35 Enzymes ........................................................................................................... 35 Visible Absorption ............................................................................................. 36 Pulse Radiolysis ............................................................................................... 37 Determin ation of Manganese and Iron Content ................................................ 37 Crystallography ................................................................................................ 37 Differential Scanning Calorimetry ..................................................................... 38 Results .................................................................................................................... 39 Metal Content Analysis ..................................................................................... 39 pH Profile .......................................................................................................... 39 Catalysis ........................................................................................................... 40 Structural Analysis ............................................................................................ 41 Thermal Stability ............................................................................................... 42 Discussion .............................................................................................................. 42

PAGE 6

6 Structures ......................................................................................................... 43 Spectroscopic Properties .................................................................................. 44 Catalysis ........................................................................................................... 45 Conclusions ............................................................................................................ 47 3 COMPARATIVE STUDY OF HUMAN AND ESCHERICHIA COLI MANGANESE SUPEROXIDE DISMUTASE: KINETIC AND ST RUCTURAL INSIGHTS .............. 53 Introduction ............................................................................................................. 53 Materials and Methods ............................................................................................ 55 En zymes ........................................................................................................... 55 Pulse Radiolysis ............................................................................................... 56 Crystallization ................................................................................................... 56 Data Collection and Refinement ....................................................................... 56 Results .................................................................................................................... 57 Catalysis ........................................................................................................... 57 Structural Analysis ............................................................................................ 59 Discussion .............................................................................................................. 60 Catalysis ........................................................................................................... 61 Active Site Environment ................................................................................... 63 Conclusions ............................................................................................................ 65 4 SEQUESTRATION OF SUBSTRATE CARBON DIOXIDE IN THE ACTIVE SITE OF HUMAN CARBONIC ANHYDRASE II ............................................................... 71 Introduction ............................................................................................................. 71 Materials and Methods ............................................................................................ 72 Enzymes ........................................................................................................... 72 Crystallization ................................................................................................... 73 Carbon Dioxide Trapping .................................................................................. 73 Data Collection and Processing ....................................................................... 74 Structure Refinement ....................................................................................... 75 Results .................................................................................................................... 76 Carbon Dioxide in the Active Site ..................................................................... 76 A Second Carbon Dioxide Binding Site ............................................................ 77 Additional Protein Structural Features .............................................................. 78 Discussion .............................................................................................................. 78 Trapping Carbon Dioxide .................................................................................. 78 Physiological Relevance ................................................................................... 79 Implications For the Catalytic Mechanism ........................................................ 80 Conclusions ............................................................................................................ 81 5 THE ROLE OF SURFACE RESIDUES IN PROTON TRANSFER BY HUMAN CARBONIC ANHYDRASE II ................................................................................... 90

PAGE 7

7 Introduction ............................................................................................................. 90 Materials and Methods ............................................................................................ 92 Enzymes ........................................................................................................... 92 Crystallization and Data Collection ................................................................... 92 Data Processing and Refinement ..................................................................... 92 18O Exchange ................................................................................................... 93 Esterase Activity ............................................................................................... 94 Results .................................................................................................................... 95 Catalysis ........................................................................................................... 95 Structure ........................................................................................................... 96 Discussion .............................................................................................................. 97 Conclusions ............................................................................................................ 98 6 CA FROM METHANOSARCINA THERMOPHILA : A COMPARISON WITH THE CA, HUMAN CARBONIC ANHYDRASE II .................................................................................................... 105 Introduction ........................................................................................................... 105 Materials and Methods .......................................................................................... 107 Enzymes ......................................................................................................... 107 Crystallization and Data Collection ................................................................. 108 Refinement ..................................................................................................... 109 Results .................................................................................................................. 109 Discu ssion ............................................................................................................ 110 7 CONCLUSIONS AND FUTURE DIRECTIONS .................................................... 117 Summary and Conclusions ................................................................................... 117 Future Directions .................................................................................................. 120 LIST OF REFERENCES ............................................................................................. 122 BIOGRAPHICAL SKETCH .......................................................................................... 130

PAGE 8

8 LIST OF TA BLES Table page 1 1 The tissue specificities, structural topologies, and catalytic activities of the class carbonic anhydrases ................................................................... 28 2 1 Rate constants for catalysis by wildtype and mutant human MnSOD ............... 48 2 2 X ray diffraction data processing and structure refinement statistics for E162D and E162A human M nSOD .................................................................... 48 3 1 X ray diffraction data processing and structure refinement statistics for F66A and F66L human MnSOD ................................................................................... 65 3 2 Rate cons tants for catalysis by wild type and mutant human and E. coli MnSOD obtained by pulse radiolysis .................................................................. 66 3 3 Geometric distances with the active site of wildtype, F66A, and F66L human and wildtyp e E. coli MnSOD as illustrated in Figure 37. ................................... 66 4 1 Data collection and refinement statistics for the CO2bound holo and apo HCA II structures. ............................................................................................... 82 4 2 Distances between atoms in the structures of CO2bound holo and apo HCA II ......................................................................................................................... 83 5 1 Data collection and refinement statistics for the Lys170 variant structures of HCA II. .............................................................................................................. 100 5 2 Apparent values of pKa and maximal rate constants for kinetic measurements of catalysis by wild type and sitespecific mutants of HCA II at residue 170 obtained by 18O exchange at 25 oC.a ................................................................ 101 5 3 Values of the pKa for the zinc bound water and the maximal value of kcat/ KM for the catalysis of the hydration of 4nitrophenylacetate by variants of HCA II at 25 oC. ............................................................................................................ 101 6 1 Data processing and refinement statistics for the structures of the Cam mutants. ............................................................................................................ 112

PAGE 9

9 LIST OF FIGURES Figure page 1 1 The dismutation of superoxide as catalyzed by human MnSOD, monitored spectroscopically at 250 nm ............................................................................... 29 1 2 The four known classes of superoxide dismutase share very l ittle structural similarity ............................................................................................................. 30 1 3 A stereoview of the active site of human MnSOD illustrates the amino acid and solvent hydrogen bonded network ............................................................... 31 1 4 The tetrameric structure of human manganese superoxide dismutase contains 2 twofold axes of symmetry ................................................................. 31 1 5 The structures of three classes of carbonic anhydrase show that, despite catalyzing the same reaction, there is a large divergence in structure ............... 32 1 6 A stereoview of the active site of human CA II showing the hydrophilic side of the active site cavit y ........................................................................................... 33 2 1 The active site structure of wildtype human manganese superoxide dismutase shows an intricate interaction network ............................................... 49 2 2 The pH profile for molar abosorptivity at 480 nm for hMn3+SOD, wild type and mutants ............................................................................................................... 49 2 3 Change in absorbance at 420 and 480 nm over a 0.25 ms time scale after generation of superoxide by pulse radiolysis in a solution containing E162D human MnSOD ................................................................................................... 50 2 4 Normalized transitions for E162D (A), wildtype (B), and E162A (C) human MnSOD as determined by differential scanning calor imetry ............................... 51 2 5 Structures of the dimeric interface of wildtype (top), E162D (middle), and E162A (bottom) human MnSOD ......................................................................... 52 3 1 Superposition of the active sites of human (magenta) (Hearn et al. 2003) and E. coli MnSOD (green) (Borgstahl et al. 2000) shows the high degree of similarity in the active site ................................................................................... 67 3 2 Output fr om pulse radiolysis studies shows the decrease in absorbance at 260 nm over the course of the reaction demonstrating the zeroorder phase of catalysis seen in F66A and wildtype human MnSOD .................................... 67 3 3 The pH dependence of k4 in catalysis by F66A human MnSOD ......................... 68

PAGE 10

10 3 4 Spectroscopic evidence of the formation of an inhibited complex in wildtype human MnSOD ................................................................................................... 68 3 5 The effects of mutations at Phe66 in human MnSOD are clearly visible at position 119 when compared to wildtype (magenta) (Hearn et al. 2003) .......... 69 3 6 A ste reoscopic view of the active site environment of human MnSOD and its mutants F66A and F66L ..................................................................................... 69 3 7 A diagrammatic representation of the geometry of the active site of the MnSODs with the distances presented in Table 33 defined with dashed lines .. 70 3 8 A histogram showing the normalized B factors of the atoms located in the active site of MnSOD .......................................................................................... 70 4 1 The active site of human carbonic anhydrase II is located at the bottom of a conical cavity ...................................................................................................... 84 4 2 The electron density for the carbon dioxide molecule in both holo (top) and apo (bottom) HCA II is clearly observed in the hydrophobic patch (green mesh and sticks) ................................................................................................. 85 4 3 The non catalytic carbon dioxide binding site was found in a hydrophobic patch ( magenta surface) on the side of the enzyme opposite the hydrophobic patch ................................................................................................................... 86 4 4 Two glycerol molecules were found in the structures of CO2bound holo and apo HCA II .......................................................................................................... 86 4 5 A superposition of the CO2bound holo HCA II structure with that of V143Y HCA II (Alexander et al. 1991) clearly shows that the side chain of Tyr143 (white sticks) would directly interfere with CO2 binding (cyan sticks) .................. 87 4 6 The catalytic mechanism of CO2 hydration as catalyzed by HCA II, as proposed by Lindskog (1983) ............................................................................. 88 4 7 The stru cture of bicarbonatebound T200H HCA II reveals that the binding site of CO2 observed in this study is the catalytic site ......................................... 89 5 1 The protein environment around the proton shuttle residue His64 includes residues Trp5 and Lys170 ................................................................................ 101 5 2 A surface representation of HCA II shows the location of Lys170 relative to the active site core ............................................................................................ 102 5 3 The pH dependence of kcat/ KM is apparently unchanged in the four Lys170 mutants ............................................................................................................. 102

PAGE 11

11 5 4 Mutation of Lys170 results in a change in the rate and pH dependence of the proton t ransfer controlled value of RH2O/[E] ...................................................... 103 5 5 The electron density corresponding to His3 is visible in all four Lys170 HCA II mutants, although to varying degrees ............................................................... 103 5 6 The location of His3 in all four mutants is roughly the same, apparently pi stacking with the side chain of Trp5 .................................................................. 104 5 7 A novel interaction is observ ed between residues 64 and 170 in the structure of K170D HCA II ............................................................................................... 104 6 1 An alignment of Cam (from M. thermophila) and CamH (from M. acetivorans ) shows the lack of the acidic loop (black box) i n CamH ..................................... 113 6 2 The trimeric organization of Cam yields three active sites per enzyme, each at a monomer monomer interface .................................................................... 114 6 3 The structure of a Cam monomer shows a left helix topology with several N terminal (blue) surface loops and a C helix ............. 115 6 4 The 63 6465 loop in the mutants of Ca m occupies a different conformation then that observed in all reported Cam structures ............................................ 115 6 5 The electron density for the 626364 loop clearly defines its orientation in the structure of W19N Cam .................................................................................... 116

PAGE 12

12 LIST OF ABBREVIATION S Angstrom atm atmosphere (unit of pressure) BH+ protonated base C alpha carbon CA carbonic anhydrase Cam Methanosarcina thermophila class carbonic anhydrase CamH M. acetivorans class carbonic anhydrase CAPS N cyclohexyl 3 aminopropanesulfonic acid CNS Crystallography and NMR System CO2 carbon dioxide CTD C terminal domain Cu copper Cp change in eat capacity DNA deoxyribonucleic acid DSC differential scanning calorimetry E enzyme E. coli Escherichia coli EDTA ethylenediaminetetraacetic acid ETC electron transport chain Fe iron GPI glycosylphosphaidylinositol H+ prot on / hydrogen ion H2O2 hydrogen peroxide HCA human carbonic anhydrase

PAGE 13

13 HCO3 bicarbonate ion IPTG isopropyl D thiogalactopyranoside k rate constant kcal kilocalorie kcat turnover number kcat/ KM specificity constant KD dissociation constant kDa kilodalton kV kilovolt LB lysogeny broth M molar MES 2 (4 morpholino) ethane sulfonic acid Mn manganese MOPS 3 (N morpholino) propanesulfonic acid M micromolar mA milliampere mM millimolar mm millimeter mol mole NCOthiocyanate ion Ni nickel NTD N terminal domain nm nanometer O2 dioxygen O2 superoxide ion

PAGE 14

14 O2 2 superoxide dianion OD optical density OHhydroxide ion PDB Protein Data Bank pH negative log of proton concentration p Ka acid dissociation constant PSR proton shuttle residue rmsd root mean square deviat ion ROS reactive oxygen species RT room temperature SDS PAGE sodium dodecyl sulfate polyacrylamide gel electrophoresis SOD superoxide dismutase TAPS N tris(hydroxymethyl)methyl 3 aminopropanesulfonic acid TM transmembrane Tris tris(hydroxymethyl)aminomethane Zn zinc

PAGE 15

15 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 STRUCTURAL STUDIES OF THE CATALYTIC MECHANIMS OF TWO SUPE RFAST METALLOENZYMES: CARBONIC ANHYDRASES AND MANGANESE SUPEROXIDE DISMUTASES By John Francis Domsic December 2009 Chair: Robert McKenna Cochair: David Silverman Major: Medical Sciences Biochemistry and Molecular Biology The utilization of metals in biological enzymes is ubiquitous in the diverse kingdoms of life. These metals, on their own are inactive, but when incorporated into a protein allow for the enhancement a myriad of chemical reactions. To understand how an enzyme functions it is necessar y to gain detailed knowledge of its active site structure and how this correlates to catalytic efficiency. Additionally, a knowledge of the interactions between enzyme and substrates and products aids in the elucidation of catalytic pathways. The manganes e superoxide dismutases (MnSOD) neutralize naturally occurring toxic superoxide radicals. Mutational analysis of human MnSOD demonstrated that Glu162, a secondshell ligand of the Mn ion is necessary for efficient activity, due to tuning of the Mn. Addit ionally, the eukaryotic MnSODs are typically more product inhibited than their prokaryotic counterparts. Alteration of the active site mouth in human MnSOD resulted in a weakly product inhibited form. The carbonic anhydrases (CAs) are a family of structurally diverse enzymes that catalyze the reversible hydration of carbon dioxide to bicarbonate and a proton. Due to

PAGE 16

16 the high turnover rate, an understanding of enzymesubstrate interactions has been elusive. The use of a highpressure environment allowed for successful capture of carbon dioxide in the hydrophobic pocket in the active site of human CA II. The x ray crystal structure of both zinc bound and zinc free HCA II revealed that the active site remains relatively static, acting as a solvation site f or CO2, thus allowing for rapid turnover. A proton transfer step is also required in the catalytic cycle of CA to allow for the regeneration of the active zinc bound hydroxide. This is accomplished by proton transfer along a solvent mediated proton wire, leading to the final proton acceptor. Mutational analysis of the environment surrounding the proton shuttle residue, His64, in HCA II revealed that the enzyme finely tunes this region to allow for bidirectional proton transfer under physiological conditions. Mutational analysis of a CA revealed that residues located adjacent to the active site affect the proton transfer properties in this enzyme. These data suggest that one must carefully consider residues outside the active site environment when analy zing the catalytic activity of an enzyme.

PAGE 17

17 CHAPTER 1 INTRODUCTION Metalloproteins From the calcium ions required for calmodulin, to the zinc ion in zinc finger containing transcription factors, to the iron ion in the ironsulfur clusters of plant photosystems, metal ions are vital cofactors for the functions a myriad of proteins. The roles of metal ions in biological macromolecules are as varied as the functions of these molecules. Functions of metals include a role as conformational regulators that allow proteins, such as zinc finger transcription factors (zinc) and calmodulin (calcium) to fold and function properly. (Coleman, 1992; Chin & Means, 2000). The enzymatic functions of metals (metalloenzymes) involve s a modulation of the local environment to f avor more efficient reaction conditions One example is the formation of a redox active catalytic site as occurs in the superoxide dismutases. Another example is the modulation of the p Ka of a metal bound ligand, such as a solvent molecule, as is present in the carbonic anhydrases. Additionally, the protein itself can finely tune the metal to assist in the formation of a catalytically efficient active site. These effects can be both short and long range. Short range interactions involve the direct coor dination of the metal ion by amino acid side chains and ligands and the subsequent configuration of additional coordination sites Long range interactions are involved in finely tuning the environment surrounding the metal ion, thus modulating the kinetic properties of the enzyme. This work will focus on the effects of the protein on modulating the activity of two metalloenzymes, human manganese superoxide dismutase and human carbonic anhydrase II.

PAGE 18

18 Superoxide Dismutases The Superoxide Radical The formation of superoxide occurs during cellular respiration in the electron transport chain (ETC) of the mitochondria. Complexes I and III of the ETC have been implicated as the m ajor producers of superoxide due to a lack of efficiency in the transfer of electrons from two electron carriers to oneelectron carriers (St Pierre et al. 2002) The occasional leakage of the extra electron can result in the one electron reduction of molecular oxygen, thus forming the superoxide radical. This can further reduce to hydr ogen peroxide or a hydroxyl radical. S uperoxide cannot combine, spontaneously, with other nonradical molecules due to spin restriction. However, it does hav e the ability to interact with other radicals. Interaction with nitric oxide forms peroxynitrite, which can, in turn, cause lipid peroxidation and tyrosine nitration. Superoxide can also cause the release of iron from sources such as the ironbinding protein ferritin. The resulting elevated iron levels can result in interactions with hydrogen perox ide to produce hydroxyl radicals via the Fenton reaction (McCord, 2002) Ultimately hydroxyl radicals cause the formation of DNA lesions occurs due to 8oxoguanine and thymine glycol dimers, lipid peroxidation, and protein damage due to backbone breakage and the formation of noncanonical amino acid side chains (for a review see Davies, 2005). This oxidative stress caused by the formation of reactive oxygen species (ROS) is the cornerstone for the free radical theory of aging (Harman, 1956) This theory suggests that the accumulation of cellular abnormalities due to oxidative stress leads to a decline in cell processes and ultimately to cell death.

PAGE 19

19 Superoxide Dismutases The neutralization of superoxide is therefore necessary for the survival of cells that respire aerobically. This is accomplished by the superoxide dismutases (SODs), a family of enzymes (EC 1.15.1.1) of which t here are four classes that are categorized based on th e metal(s) in the active site: copper/zinc (Cu/Zn), iron (Fe), manganese (Mn ), and nickel (Ni). Cu/ZnSOD utilizes a two metal active site, with the copper ion being the catalytic metal. This class can be found in both prokaryotic organisms and eukaryotic organisms, from plants to mammals (Fink & Scandalios, 2002) T he iron and m anganese enzymes share very sim ilar properties structurally and, indeed, several cambialistic enzymes have been discovered that are active with iron or manganese active site ions (Sugio et al. 2000; Chen et al. 2002) The iron enzymes are found mainly in prokaryotes and the chloroplasts of plants (Muoz et al. 2005) The manganese enzymes are found ubiquitously in nature and occur as a mitochondrially targeted enzyme in eukaryotes (human SOD2 for example). NiSOD is the most recently discovered of these classes with existence confirmed in Streptomyces (Youn et al. 1996; Youn et al. 1996b; Barondeau et al. 2004) a nd a strain of cyanobacteria (Palenik et al. 2003). The physiological importance of the reaction catalyzed by the SODs is underscored by t he presence of multiple classes within the same organism. For example, Escherichia coli is known to express both an iron and a manganese SOD. Also, there are three human SODs, a cytosolic Cu/ZnSOD, an extracellular Cu/ZnSOD, and the mitochondrial MnSOD. Members of the Cu/Zn, Fe, and Mn SODs have been found in various plant species, though a given species doesnt necessarily contain all three (Gupta et al. 1993).

PAGE 20

20 Phenotypes of SOD deficiency The necessity o f the various forms of SODs has also been demons trated by numerous studies on SOD deficient models. Phillips et al. (1989) demonstrated that a mutational defect in Cu/ZnSOD in Drosophila leads to a drastic decrease in longevity and fertility and sensitivity to increased levels of superoxide. Superoxide sensitivity and increased rates of mutagenesis were observed in Cu/ZnSOD deficient yeast and, interestingly, normal function was rescued by the introduction of a bacteri al MnSOD (Bowler et al. 1990). In mice with defective Cu/ZnSOD a reproduct ive defi ciency was observed (Reaume et al. 1996) Probably the most well documented disease state associated with SOD is familial amyotrophic lateral sclerosis, otherwise known as Lou Gehrigs Disease (Noor et al. 2002; Strange et al. 2007) This disease is as sociated with defects in Cu/Zn SOD ( SOD1 ) that can cause both aggregation of misfolded protein and increase in the rate of production of hydroxyl radical, depending on the mutation(s) present Deficiencies in the manganese form of SOD have also been phenot ypically characterized. SOD doubleknockouts (Fe and Mn) in E. coli demonstrated an increased sensitivity to superoxide as well as an increase in the rate of spontaneous mutagenesis and an oxygendependent decrease in growth rate (Farr et al. 1986). In a very extreme case, mice lacking any MnSOD had a life span of 12 weeks as a result of faulty mitochondrial activity (Lebovitz et al. 1996; Li et al 1995). Disease states that propagate in the human population are not usually associated with a completely inactive enzyme, as one would expect a lack of MnSOD would lead to rapid oxidative damage and organism death. Rather, the most analyzed polymorphism results in the amino acid substitution V16A, located in the mitochondrial targeting peptide at the N te rminus of the

PAGE 21

21 immature protein. This sequence is recognized by transporters on the mitochondria that uptake the protein, followed by cleavage of the signal peptide. Thus, a mutation in the signal peptide results in a decreased level of targeting and, therefore, mitochondrial MnSOD, although the mature form of the enzyme has normal activity. This inadequate targeting has been suggested to cause an increased risk of nephropathy and retinopathy in certain diabetic patients (Mllsten et al. 2007; Hovnik et al. 2009). Also this polymorphism is associated with an increased risk of Parkinsons disease (Farin et al. 2001). Catalytic mechanism Even though there are differences in the catalytic metal, all families of SOD catalyze the disproportionation of two superoxide molecules to form hydrogen peroxide (H2O2) and dioxygen (O2). The detailed mechanism discussed here is specific to the MnSODs in its pathways, although all SODs share the general mechanism, the difference being the oxidation state of the metal and the occurrence of a product inhibited state. During the catalytic process, the active site metal cycles between an oxidized and reduced state. The catalytic cycle begins when superoxide binds at an unoccupied, 6th coordination site on the Mn. The fi rst step of catalysis is a first order reaction with respect to superoxide (Figure 1 1) In this step, one molecule of superoxide is oxidized to dioxygen with a concurrent reduction of the manganese and the protonation of the manganese bound hydroxide ( eq 1 1 ). There are two possible routes during the second step of catalysis by MnSOD. The first involves the uninhibited formation of H2O2 from one molecule of superoxide and two protons (one from Mnbound water and one from bulk buffer) as well as the concu rrent oxidation of the manganese ( eq 1 2 ). The second possibility involves a zeroorder

PAGE 22

22 decay that has been attributed to the presence of a reversibly product inhibited form of the enzyme (Figure 1 1) This is believed to involve the formation of a perox ide dianion in complex with oxidized manganese ( eq 1 3 ). Once the product inhibited state is released the catalytic cycle can complete ( eq 1 4 ). This r esults in the formation of H2O2 and a return of the enzyme to its initial state, a manganese bound hydr oxide and the oxidized manganese. Mn3+(OH-)SOD + O2 + H+ Mn2+(H2O)SOD + O2 [rate = k1] (1 1) Mn2+(H2O)SOD + O2 + H+ Mn3+(OH-)SOD + H2O2 [rate = k2] (1 2) Mn2+(H2O)SOD + O2 Mn3+(O2 2 -)SOD [rate = k3] (1 3) Mn3+(O2 2 -)SOD + H+ + H2O Mn3+(OH-)SOD + H2O2 [rate = k4] (1 4) The hydrogen peroxide i s then obviated directly by enzymes such as glutathione peroxidase, peroxiredoxin, and catalase (Smith et al. 2003). Structural descriptions Despite the similarities in the catalytic mechanisms across the SOD families, the structural features of the various forms of SODs generally vary greatly between families (Figure 1 2). The subunit organization of all known iron (Figure 11 A) and manganese SODs consists of a homodimer with several eukaryotic forms including the human manganese enzyme, further organized into a dimer of dimers (Figure 11 B). The nickel enzyme exists as a hexamer of monomers, which are structurally distinct from the other forms (Figure 1 1 C). The Cu/Zn SODs also exist as dimers but have a different monomeric structure than the iron and manganese SODs (Figure 11 D) (Strange et al. 2007). The general topology of all iron and manganese SOD monomers and dimers is virtually the same, and the majority of this work will focus on the human mitochondrial MnSOD unless otherwise stated.

PAGE 23

23 The human MnSOD monomer can be divided into two domains: the N terminal domain (NTD) (residues 1terminal domain (CTD) (residues 84active site metal ion is sandwiched between these two domains and is coordinated by two histidines from the NTD (His 26 and H is 74) and by an aspartate and histidine from the CTD (Asp 159 and H is 163). A fifth coordination site of the Mn is occupied by a sol vent molecule, either a hydroxide or water depending on the oxidation state of the manganese, thus completing a trigonal bipyramidal complex with the manganese (Figure 13). The active site may be further extended by considering the presence of an extensi ve hydrogenbonded network that may be utilized as a proton transfer pathway during the second stage of catalysis. Two monomers are organized to form the dimeric interface that is required for efficient catalytic activity in all known MnSODs (Figure 14). In the tetrameric MnSODs, two of these dimers are required to form the physiological tetramer. Carbonic Anhydrase As was mentioned at the beginning of this chapter, another role for metals in catalysis is altering the acid/base properties of metal bound ligands. One such class of enzymes that utilize this is the carbonic anhydrases (CA) (EC 4.2.1.1), which catalyze the reversible hydration of carbon dioxide to form bicarbonate and a proton. There are five known classes of CAs, categorized based on their overall fold and their amino acid sequence: , and The latter two classes were relatively recently discovered and have only been found in a few organisms. The class is found in diatoms and the lone member of the class is the only known cadmium containing enzyme and was discovered i n Thalassiosira weissflogii (Roberts et al. 1997; So et al. 2004; Lane et al. 2005). There was at one point an class CA, discovered in the carboxysomal shell of

PAGE 24

24 cyanobacteria. However, structural analysis revealed it to be a gene CA wit h one of the two domains having diverged to the point that sequence analysis no CA domain (Sawaya et al. CAs are g enerally mammalian enzymes, but have also been discovered in mosquitoes and green algae (Corena et al. 2002; Karlsson et al. 1995). The CAs are plant, bacterial, and fungal enzymes (Elleuche & Pggeler, 2009). The CAs have been found in archaea and plants (Alber & Ferry, 1994; Sunderhaus et al. 2006). and Carbonic Anhydrases Of the fiv e known classes, only three, and have been studied extensively. Structures of at least two representatives of each of these classes have been solved by x ray crystallography with an even larger number having been characterized kinetically. Zinc has been implicated as the metal of choice for these enzymes, though a recent study suggests that, when in an anaerobic environment, the CAs utilize iron as the active site metal. Catalytic m echanism The catalytic mechanism of the CAs has been the most extensively studied of all CAs and quite possibly the CAs, although the mechanics are slightly different (Rowlett, 2009; Ferry, 2009). As was shown with the superoxide dismutases, catalysis by the CAs involves a twostep mechanism. In the first step carbon dioxide binds to the active site and the zinc bound hydroxide nucleophilically attacks the carbon of the CO2 molecule, thus forming HCO3 (eq 1 5). A water molecule then diffuses freely i nto the active site, displaces the bicarbonate, and coordinates with the zinc. The zinc bound water is catalytically inactive, and therefore a hydroxide must be generated by proton removal. This occurs

PAGE 25

25 during the second step of catalysis in which a solvent mediated proton wire transfers a proton from the zinc bound water to the proton acceptor, amino acid His64 (eq 16) (Figure 1 6). This proton transfer is the ratelimiting step of catalysis, occurring at 106 s1 (Silverman & Lindskog, 1988). EZn OH+ CO2 EZn HCO3 EZn H2O + HCO3 (1 5) EZn H2O H+EZn OHEZn OH+ BH+ (1 6) Structure Despite catalyzing the same general reaction, there is an astounding lack of structural identity, thus providing an excellent example of convergent evolution. The class enzymes all share the same general structure: a catalytic core of a 10 stand, anti (Figure 1 5 A). The catalytic site is located in an active site cavity that is directly zinc ion. The zinc is covered by two flaps (residues 126 and 196207) and is directly coordinated by three histidines (HCA II numbering His94, His96, and His119) and a solvent molecule (H2O or OH-). CAs display the greatest intraclass structural dissimilarities, with a high degree of oligomeric heterogeneity across this family, with physiological molecular weight ranging from 45200 kDa. The l evel of oligomerization is most likely a result of surface structure variation, with the functionally active moiety being a dimer, and all structures having a number of monomers equal to 2n, where n is 1, 2, or 3 (Figure 1 5 B). Also, Porphyridium purpureum CA exists as pseudo dimer in which there appears to have been a gene duplication event leading to a functional monomer with two nearly identical domains (Mitsuhashi et al. 2000). There have been two

PAGE 26

26 subclasses identified based on metal coord ination (Rowlett, 2009). In type I enzymes, the zinc is coordinated by one histidine and two cysteines, with a fourth coordination site occupied either by a solvent molecule. In type II enzymes, the zinc is also coordinated by one histidine and two cysteines, with the fourth coordination site occupied by an aspartate. The zinc coordinating amino acids are located all within one monomer, with additional active site residues contributed by the neighboring monomer. CAs is also oligomeric with the physiological unit existing as a trimer with three active sites, each sandwiched at the three interfaces (Figure 1 5 C). The monomer exhibits an overall left strands) with several intervening loops a nd a C terminal helix that runs along base of helix. The intervening loops occur at three points in helix towards the N terminus of the protein. All three loops exist on the face opposite of the interfacial sides helix core. Three histidines and a solvent CAs, coordinate the active site metal. When synthesized aerobically in E. coli CA from Methanosarcina thermophila (Cam) contains zinc in its active site ( Alber et al. 1999). However, it has been demonstrated that Cam contains iron in its active site when synthesized in an archaeal expression system anaerobically (Macauley et al. CA, from Pyrococcus horikoshii reveals an identical fold, with the exception that one of the loops is absent (Jeyakanthan et al. 2008). class. In CA isolated as a carbon dioxide hydrating preparation from ox blood (Meldrum and Roughton, 1932). Since this initial discovery a

PAGE 27

27 plethora of studies have elucidated many physiological, catalytic, and structural properties of this enzyme. class Isoforms The human genome encodes 15 known CAs (I class. Three of these isoforms, VII, X, and XI, are not capable of typical CA activity as they lack one of the zinc binding histidine ligands. There is also an additional mammalian isoform, XV, which shows active expression in rats and m ice, but has not been observed in humans (Hilvo et al. 2005). The remaining human isoforms have all been characterized in terms of expression patterns (mRNA), while a large majority have also been characterized kinetically. Table 1 1 shows the localization, domain organization, and kinetic properties of all 14 human CAs. Like the considerable variations in tissue distribution, the human CAs also have a wide range of functions. CAs are believed to supply bicarbonate for the initial steps of ureagenesis and gluconeogenesis (Henry, 1996). They have also been demonstrated to be important in acid/base balance, including gastric acid production and renal and male reproductive tract acidification (Breton, 2001). There is also accumulat ing evidence for a role of CA in regulating the function of ion transport by the sodium/bicarbonate transporter, possibly by direct interaction between the two proteins (Vince & Reithmeier, 1998; Becker & Deitmer, 2007). Carbonic Anhydrase Deficiencies The variety of isoforms provides functional redundancy that allows for survival of organisms that may have a mutation leading to reduced or lost function of an isoform (Sly & Hu, 1995). For example, CA III, which makes up approximately 8% and 25% of the prot ein mass in skeletal muscle and adipocytes, respectively, can be knocked out in

PAGE 28

28 mice without any noticeable effects on fitness or appearance (Kim et al. 2004). Also, mutations in the gene encoding human CA I that have been shown to result in decreased or lost activity but do not cause any phenotypic differences (Venta, 2000). Presumably, human CA II (HCA II), which is also found in red blood cells, is able to make up for the loss of CA I, as CA II is a much more efficient enzyme. The importance of HCA I I is highlighted by mutations that cause a loss of functional HCA II resulting in a disease aptly known as HCA II deficiency syndrome (Hu et al. 1997). The symptoms of this disease are severe and illustrative of the myriad functions of HCA II. They incl ude renal tubular acidosis, hearing impairment, mental retardation, facial dysmorphism, growth failure, increased bone density, and intracerebral calcification (Venta et al. 1991; Borthwick et al. 2003). Table 11. The tissue specificities, structural t opologies, and catalytic activities of the class carbonic anhydrases. In the topology column, CA = CA domain, TM = transmembrane domain, and GPI = glycosylphosphatidylinositol lipid anchor. The domains shown as triangles are extracellular, circle s are intracellular, and rounded squares are transmembrane segments. Isozyme Localization Topology a k cat /K M (M1 s1 ) b kcat (s1) b I erythrocytes, epithelium of large intestine, adipose tissue, sweat glands, corneal epithelium 5.0x107 2.0x105 II v irtually all tissue types 1.5x108 1.4x106 III red skeletal muscle cells, adipocytes, several other tissues in low concentrations 3.0x105 1.0x104 IV membrane bound in lungs, kidneys, gastrointestinal tract 5.0x107 1.1x106 V liver mitochondria 3.0x107 3.0x105 VI saliva 1.6x107 7.0x104 VII cytosolic 7.6x107 9.4x105

PAGE 29

29 a derived from (Purkerson & Schwartz, 2007); b from (Duda & McKenna, 2004) Figure 11. The dismutation of superoxide as catalyzed by human MnSOD, monitored spectroscopically at 250 nm. The red shaded ar ea outlines the initial burst phase of first order convertsion of superoxide to dioxygen. The blue shaded VIII Purkinje cells N/A N/A N/A IX membrane bound, colorectal tumor cells 5.5x107 3.8x105 X mRNA found in adult brain N/A N/A N/A XI mRNA found in adult brain N/A N/A N/A XII membrane bound mRNA overexpressed in renal cell cancers 3.4x107 4.0x105 XIII salivary glands, kidney, small intestine, colon, uterus, testis active active XIV membrane bound in brain, heart, skeletal muscle, lung, liver 3.9 XV does not appear to be expressed in higher primates active active

PAGE 30

30 area outlines the zeroorder decay of superoxide that occurs when the enzyme enters a product inhibited state. (derived from Hsu et al. 1996). A B C D Figure 12. The four known classes of superoxide dismutase share very little structural similarity. The closest families, structurally, are the iron (A) and manganese (B) enzymes as shown here from E. coli (PDB ID 1isa, Lah et al. 1995) and human (P DB ID 1luv, Hearn et al. 2003) sources, respectively. Nickel SOD exists as a hexamer in Streptomyces coelicolor (PDB ID 1t6u, Barondeau et al. 2004). Another human SOD is copper/zinc SOD, the structure of which is a dimer with a dimetallic active site (PDB ID 2v0a, Strange et al. 2007 ).

PAGE 31

31 Figure 13. A stereoview of the active site of human MnSOD illustrates the amino acid and solvent hydrogen bonded network. The Mn is shown as a purple sphere, and dashed magenta lines represent distances that would allow for hydrogen bonding. The majority of the active site is made up of one monomer (orange) with some contribution from the neighboring monomer (blue). Figure 14. The tetrameric structure of human manganese superoxide dismutase contains 2 twofold axes of symmetry. Each dimeric interface forms two active sites. The monomers that constitute each dimer are colored similarly in this figure (dark and light blue or red). The dimeric interface can be observed by viewing the protein along the x axis (bottom right). The tetrameric interface, which is involved in protein stability, can be seen by viewing down the y axis (bottom left).

PAGE 32

32 A B C Figure 15. The structures of three classes of carbonic anhydrase show that, despite catalyzing the same reaction, there is a large divergence in structure. (A) The CAs (PDB ID 1tbt, Fisher et al. 2005 CAs are all believed to be oligomeric with either 2, 4, or 8 monomers. Shown here is the tetrameric structure of Porphyridium purpureum CA (PDB ID 1ddz, Mitsuhashi et al. 2000 CAs are all trimeric enzymes with a central left helix core, as illustrated by the CA from Methanosarcina thermophila (PDB ID 1thj, Kis ker et al. 1996 ).

PAGE 33

33 Figure 1 6. A stereoview of the active site of human CA II showing the hydrophilic side of the active site cavity. The residues in this half are responsible for forming the solvent mediated proton network, shown here as magenta dashes. It is along this network that a proton is transferred during the second catalytic step. The proton accepting residue, His64, is shown in both its in and out conformations. (PDB ID 1tbt, Fisher et al. 2005).

PAGE 34

34 CHAPTER 2 THE ROLE OF THE DIME RIC INTERFACE IN CATALYSIS BY HUMAN MANGANESE SUPEROXIDE DISMUTASE Introduction As discussed in Chapter 1, t he manganese superoxide dismutases (MnSOD) catalyze the disproportionation of superoxide to produce O2 and H2O2. The MnSODs are multimeric enzymes, with eac h monomer containing one active site manganese, coordinated by three histidines, an aspartate, and a solvent molecule. Human MnSOD is a mitochondrial enzyme that exists as a physiological tetramer, a dimer of dimers. The dimeric interface is quite conser ved, structurally, between the eukaryotic and prokaryotic forms of the enzyme (Borgstahl et al. 1992, Edwards et al 1998). The tetrameric interface, however, is unique to the eukaryotic forms of MnSOD and may act to stabilize the enzyme. Evidence for this is provided by thermal denaturation studies of human MnSOD and Escherichia coli MnSOD, a dimeric MnSOD. A major thermal unfolding transition occurs around 90 C for human MnSOD, while that for E. coli MnSOD is around 76 C (Mizuno et al. 2004, Greenleaf et al. 2004). There have been several studies that have examined the catalytic and structural role of dimeric and tetrameric interface residues in human MnSOD. In the I58T tetrameric interfacial mutant of h uman MnSOD, it was found that the enzyme p redominantly existed in the dimeric form and had a greatly reduced major unfolding temperature of 76 C, identical to that of E. coli MnSOD (Borgstahl et al. 1996 ). Also, there was only a slight reduction in the catalytic activity of the enzyme. Mutational analysis of a dimeric interfacial residue, Tyr166, revealed that the Y166F variant exhibited two major defects: a major unfolding temperature of 74 C and a 40fold red uction of activity (Hearn et al. 2004). The loss of activity can be attributed to the role

PAGE 35

35 of this residue in forming the putative proton transfer wire. An NMR study utilizing hydrogendeuterium exchange showed that in wild type human MnSOD the tetrameric interface was more dynamic than the dimeric interface (Quint et al. 2006). Tak en together these results support the hypothesis that the dimeric interface is vital for catalysis and overall stability, while the tetrameric interface only acts to slightly increase the stability of the enzyme. To further investigate the function of the dimeric interface the role of a glutamate residue, Glu162, was examined. Glu162 spans the dimeric interface and forms a hydrogen bond with one of the manganese coordinating histidines His163, in a neighboring monomer (Figure 2 1) (Quint et al. 2008) The importance of this interaction is highlighted by its conservation in all MnSODs, such as that from E. col i The mutation E170A in E. coli MnSOD r esulted in an enzyme that more readily dissociated into monomers and, intriguingly, preferentially bound to iron instead of manganese thus nullifying enzymatic activity ( Whittaker & Whittaker, 1998). This chapter will examine the corresponding mutation in human MnSOD, E162A, as well as a more conservative mutation, E162D. Despite the conservation of this g lutamic acid in all MnSODs the mutations did not exhibit all of the effects seen in the E. coli E170A mutant. T he mutant human enzyme s showed no appreciable change in metal specificity or stability However, major changes in catalytic efficiency were obs erved, with both mutants being catalytically deficient Materials and Methods Enzymes The gene encoding the wildtype human MnSOD gene was contained in the vector pTrc99A. Mutations were made using the QuikChange SiteDirected

PAGE 36

36 Mutagenesis kit from Stratagene (La Jolla, CA). Thermocycling was performed using oligonucleotides containing the desired mutation as primers The presence of the mutation was verified by DNA sequencing (ICBR, University of Florida). The plasmid containing the mutated gene was then transformed into QC774 strain of E. coli that lacks the genes that encode endogenous FeSOD ( SodB-) and MnSOD ( SodA-) (Carlioz and Touati, 1986). Cells were grown in LB broth supplemented with 6 mM MnCl2 and ampicillin for antiobiotic selection. Cultures were grown to OD580 of 0.8 and then induced with IPTG at 1 mM final concentration for 4 hours. Cells were then harvested via centrifugation and frozen at 20 C. Lysis was carried out using hen egg white lysozyme in 20 mM Tris, pH 8.2, 10% glycerol, an d 1 mM EDTA. Following lysis, the suspension was heated to 60 C for 20 minutes to remove the majority of contaminating proteins. The lysate was then centrifuged at 30000 x g for 30 minutes. Prior to further purification, the supernatant was dialyzed against three exchanges of 20 mM Tris, pH 8.2. Q sepharose HP (GE Healthcare) was used for ion exchange with elution by a 20% gradient of 500 mM sodium chloride. Purity was assessed via SDS PAGE analysis and final concentrations were determined using UV spe280 = 40500 M1 cm1 ( Greenleaf et al. 2004). Visible Absorption The visible spectrum of human MnSOD exhibits a broad absorption, with a maximum at 480 nm ( 4 80 = 610 M1 cm1) (Leveque et al. 2001). Enzyme samples were diluted 1:1 (final enzyme concentration ~500 M) in a buffer composed of 200 mM MES and 200 mM TAPS, with the pH adjusted by potassium hydroxide and the absorbance at 480 nm was recorded as a function of pH.

PAGE 37

37 Pulse Radiolysis Pulse radiolysis experiments were perf ormed by Dr. Diane E. Cabelli at Brookhaven National Laboratory using the 2 MeV van de Graaff accelerator to instantaneously produce superoxide in solution. The formation of superoxide radicals is driven by the exposure of air saturated solution to the hi gh dose electron pulse according to the methods of Schwarz (Schwartz, 1981). This results in the production of superoxide at final concentrations in the range of 45 M. Enzyme solutions were made in 2mM buffer (MOPS at pH 6.58.0, TAPS at pH 8.09.0, or CAPS at pH 9.010.0), 50 M EDTA, and 30 mM formate (a hydroxyl radical scavenger). Reactions were monitored using a Cary 210 spectrophotometer at 25 C, by followin g changes in either the absorbance of superoxide ( 260 = 2000 M1 cm1) (Rabani et al. 1969) or the absorbance of the enzyme (Cabelli et al ., 1999). Determination of Manganese and Iron Content Manganese concentrations for the enzymes were determined by Patrick Quint in Dr. David Silvermans lab using a PerkinElmer 308 flame atomic absorption spectrometer fitted with a multi ion lamp and a 3slit burner and the absorption at 279 nm was monitored. The typical occupancies for manganese content ranged from 5 4 90%, depending on the mutation. Iron content was measured by ABC Research Corp. (Gainesville, FL) and accounted for no more than 2% of the total metal content of the mutants. Therefore, the manganese concentrations were used as the active enzyme concen tration for catalytic measurements. Crystallography Crystals of E162D and E162A MnSOD were grown with a precipitant solution of 3 M ammonium sulfate, 50 mM imidazole, and 50 mM malate at pH 7.8 8.2 using the

PAGE 38

38 vapor diffusion hanging drop method. Hexagonal crystals with dimensions of approximately 0.2 mm 0.2 mm 0.3 mm grew at room temperature (RT) within 1 week and were magenta in color. Diffraction data were collected from visually selected single crystals wet mounted in quartz capillaries (Hampton Res earch) on an R AXIS IV++ image plate system with Osmic mirrors and a Rigaku RU H3R CU rotating anode operating at 50 kV and 100 mA. A 0.3 mm collimator was used with a crystal to detector using a 0.3 oscillation angle with an exposure time of 5 min/frame at RT. Both data sets were indexed using DENZO and scaled and reduced with SCALEPACK ( Otwinowski & Minor, 199 7). Useful diffraction data were collected to 2.3 and 2.5 resolution for t he E162D and E162A MnSOD crystals, respectively. To prevent model bias, the E162D and E162A MnSOD crystal data sets were phased using the human wildtype MnSOD structure (Quint et al. 2006) (PDB ID 2adq) in which the residue at position 162 was replaced with an alanine and all waters and the active site manganese had been removed. The structures were phased and refi ned using CNS ( Brunger et al. 1998). Refi nement cycling (using rigid body, simulated annealing for the fi rst cycle, and energy minimizatio n and individual B facto r refi nement for all subsequent cycles ) was done in conjunction with rounds of manual model building using COOT for molecular modeling (Emsley & Cowtan, 2004). The refined model statistics are given in Table 22 The refi n ed model s and structure factor fi les have been deposited with the Protein Data Bank as entries 3c3t and 3c3s for E162D and E162A MnSOD, respectively. Differential Scanning Calorimetry Samples of each mutant were buffered with 20 mM potassium phosphate at pH 7.8, w ith a final enzyme concentration of 1 mg/mL. The samples and references were

PAGE 39

39 then degassed for 10 minutes prior to data collection. Scans were performed by Patrick Quint in the Silverman lab, using a temperature range from 25110 C at a rate of 1 C/min (Microcal VP DSC). A buffer blank was subtracted from the final protein scan and a cubic baseline was fit to the profile. Changes in heat capaci ty ( Cp) for the unfolding peaks were corrected by fitting a reversible, nontwo state model with two compone nts. Baseline correction and peak fitting were performed using Origin (Microcal Software, Northampton, MA). Results Metal Content Analysis The replacement of Glu162 with either aspartate (E162D) or alanine (E162A) had no effect on the preferential binding of manganese to the active site, with both enzymes having an iron content of only 2 %. However, there was a large change in the amount of active enzyme present as determined by manganese analysis. The Mn occupancy of E162D was determined to be 88% and t hat for E162A was 54%, indicating a weaker metal binding site. pH Profile Wild type human MnSOD with Mn3+ exhibits a characteristic absorbance at 480 nm. The reduced form, Mn2 +, on the other hand shows no significant visible absorbance peaks. The pH prof ile for wild type human MnSOD was fit with a single ionization that had a p Ka of 9. 2 0 1 (Figure 22). Two different effects wer e seen with mutation of Glu162: E162D had a lower p Ka of 8.7 0.2 while E162A had a higher p Ka of 10.1 0.1.

PAGE 40

40 Catalysis The reaction catalyzed by human MnSOD is the disproportionation of superoxide via a twostep mechanism in which the active site metal cycles between Mn2+ and Mn3+. Concurrent with this cycling is the uptake or release of a proton by the manganese bound solvent molecule, respectively (see Introduction for details). Additionally, during the second step of catalysis, a product inhibited form of the enzyme occurs. In total, four rate constants for catalysis can be measured (eq 11 1 4): k1 is the rate of rea ction for the first step, the reduction of the Mn, the protonation of the Mnbound hydroxide, and the release of dioxygen; k2 is the rate of the second step along an uninhibited pathway, the oxidation of the Mn, deprotonation of the Mnbound water, and the formation of H2O2; k3 is the rate of formation of the product inhibited state; k4 is the rate of release from the product inhibited state and subsequent formation of product, H2O2. The values of k1k4 for wild type, E162A, and E162D are provided in Table 1 1. The rate constant k1 was measured using two different methods, the disappearance of superoxide ( =260 nm) and the rate of change in the visible absorption of human MnSOD ( =480 nm) under single turnover conditions ([E] >> [S]) (Cabelli et al. 1999) The resulting rates as measured by both methods were in agreement. The rate constants k2 and k4 were measured by first reducing the active site with H2O2 followed by superoxide generation via pulse radiolysis. This caused an increase in absorption at 480 nm, indicating the oxidation of the Mn and a return to the initial state of the enzyme. The initial part of the curve gave k2 and the later part gave k4. Finally, k3 was measured by monitoring the increase in absorption at 420 nm, the characteristic absorbance peak for the product inhibited complex (Bull et al. 1991, Hearn et al. 2001), and the decrease in absorption at 480 nm (Figure 23).

PAGE 41

41 There was no pH dependence observed for the values of k1k4 of wild type human MnSOD in the pH range of 7.09. 5. However, the value for k1 in the E162D mutant was pH dependent and was fit to a single ionization with a maximum at 355 33 1 s1 and a p Ka of 8.7 0.2. The data for k2 and k3 for E162D decreased to approximately 50 1 s1 at 1 s1, respectively) (Table 11). Additionally the value for k4 of E162D showed no pH d ependence. Conversely, the values of k1k3 for the E162A mutant showed no pH dependence, but k4 exhibited a 10fold decrease as pH changed from 8 (30 3 s1) to 10 (3 1 s1) and was unable to be fit by a single ionization. Another rate constant, k0/[E ], describing the product inhibited, zeroorder region of catalysis at steady state is near 500 s1 for wild type, with no pH dependence (Hearn et al. 2001, Hsu et al. 1996). For E162D, the rate was decreased to a value of 270 s1, and showed no pH dependence. The value of k0/[E] for E162A showed greater pH dependence with the rate decreasing from 190 s1 at pH 7.7 to 18 s1 at pH 8.4. Structural Analysis Both mutants of human MnSOD crystallized in the hexagonal space group P 6122 under similar crystall ization conditions. The asymmetric unit consisted of one dimer with the tetramer formed by a crystallographic 2fold symmetry operator. The structures of the two mutants did not show any large, overall structural deviations compared to wild type with a r msd of 0.2 for both mutants. The most significant difference between the mutant and wildtype structures was, not surprisingly an alteration of the interaction between residue 162 and His163 at the dimeric interface (Figure 25). In the wildtype enzym e, there is a direct interaction between Glu162B (where B indicates monomer B) and His163A (where A indicates

PAGE 42

42 monomer A). In E162D, the interaction between the two residues is mediated by a solvent molecule that apparently interacts with the carboxylate o f Asp162B and an imidazole nitrogen of His163A. This interaction is completely lost in the E162A mutant due to the distance between the two residues and the hydrophobic nature of Ala162B. In addition to this change in interaction, a putative sulfate is s een to penetrate into the channel near the 162BHis163A interaction. The sulfate is seen to enter further into the channel ( and closer to the 162163 interaction) as the side chain of residue 162 is shortened (Figure 25). Thermal Stability The major thermal transition temperatures for the unfolding of E162A and E162D were determined by differential scanning calorimetry The data for E162D showed two peaks with melting temperatures that were very similar to that of wildtype human MnSOD. The thermal inactivation temperature for E162D was 72 C with an unfolding temperature of 88 C (Figure 24). For comparison, the values for wildtype enzyme were reported as 68 and 90 C, respectively (Borgstahl et al. 1996) with the values independently determined by our group to be 72 and 94 C, respectively. The inactivation temperature for E162A was increased to 77 C, though the denaturation temperature was decreased to 81 C. Discussion Mutagenic analysis of residue 162 in human manganese superoxide dismutase has provided insights into the role of this second shell Mn ligand both catalytically and structurally. The sight specific mutants E162D and E162A create catalytic deficiencies that have been correlated to the accompanying structural changes.

PAGE 43

43 Structures Th e manganese ion in wildtype human MnSOD is coordinated by three histidines, one aspartate, and one solvent molecule, either a water or hydroxide depending on the stage of catalysis. Additionally there are several secondshell ligands that possibly play r oles in fine tuning the redox state of the Mn and the pKas of surrounding solvent molecules and amino acids. Glu162 spans the dimeric interface and interacts directly with the adjacent monomer with His163, a coordinating ligand of the Mn (Figure 21, Fig ure 25 top). Mutation of glutamate to aspartate effectively s hortens the length of the side chain by one carboncarbon bond, and thus lengthens the interaction distance between the two amino acids. Keeping the carboxylate group, however, allows for a wa ter molecule to intervene between the aspartate and histidine. Mutation to alanine completely abolishes the interaction due to the much shorter, hydrophobic side chain. Interestingly, a sulfate molecule was observed adjacent to this portion of the dimeric interface, and represents a novel sulfate binding site. This indicates the widening of a solvent accessible channel that exists along the dimeric interface. In E162D, this sulfate displaces a water molecule ( S2 in Figure 21) that is believed to be part of a putative proton wire. Further inward movement of the sulfate in the E162A mutant allows it to interact directly with the side chains of His30 and Tyr34, both of which are partially solvent exposed in wildtype. This suggests that part of the cataly tic deficiency of the mutant enzymes is due in part to a weakening of the proton wire, thus providing further support for a role of proton transfer in the catalytic mechanism. It should be noted that the catalytic effects of sulfate were not examined, as no sulfate was present in the kinetic or spectroscopic assays.

PAGE 44

44 Thermal denaturation analysis, as measured by differential scanning calorimetry, indicates that the Glu162BHis163A interaction plays a role in stabilizing the protein. The maintenance of the interaction by a water molecule in E162D results in only a slight, possibly insignificant, decrease in the major unfolding transition (88 C for E162D, 9094 C for wild type). The E162A mutant in which this interaction is lost and the major unfolding transition occurs at 81 C further supports this stabilizing role. The loss of this interaction did not affect the ability of the protein to form tetramers, indicating that there were no major changes in overall protein mobility. Spectroscopic Properties Wi ld type human Mn3+SOD exhibits a broad absorption spectrum with a peak at 480 nm (Bull et al. 1991, Hsu et al. 1996, Maliekal et al. 2002). The pH profile at this maximum titrates with a pKa value of 9.2. It is believed that the source of this ionizat ion is Tyr34 with evidence provided by a study done on E. coli MnSOD (Maliekal et al. 2002) and Y34F human MnSOD, which has a significantly altered p Ka of 11 (Hsu et al. 1996, Guan et al. 1998). There is a shift of this critical pKa in E162D and E162A human MnSOD, though the shifts occur in opposite directions (wildtype = 9.2 0.1, E162D = 8.7 0.2, E162A = 10.1 0.1 ) (Figure 2 2). This effect is likely due to the indirect interaction between Glu162B and Tyr34A (Figure 2 1). The side chain of Tyr34A is located 6.2 away from the carboxylate of Glu162B in wild type human MnSOD. The interaction traces a path from Glu162B to His163A, and then to solvent molecule S2, which interacts with Tyr34A. The preservation of the Asp162BHis163A interaction vi a a mediating water in E162D results in an enzyme in which the pKa is only slightly altered. The loss of this indirect interaction in E162A results in a much larger (1 pH unit) pKa shift with a value similar to that of tyrosine in solution.

PAGE 45

45 Catalysis A d iminished level of catalysis (48 fold) was seen in E162D MnSOD as measured by k1 and k2 (Table 11). Mutation to alanine results in an enzyme with a 2224 fold lower rate for k1 and k2. The observed changes in catalytic rates are a result of the weakened or lost dimeric interfacial interaction between 162B and His163A and the subsequent effects on the properties of the active site metal and the atoms in the active site. Similar effects have been observed in previous studies that have shown that mutations of second shell ligands and also at the dimeric interface can have substantial effects on the catalytic properties of human MnSOD. Mutation of the second shell ligand (via the Mnbound solvent) Gln143 to asparagine resulted in a 100fold reduction in ca talytic activity. Also, there was evidence of an increase in the redox potential of the active site, as Q143N enzyme did not have an absorbance peak at 480 nm (Leveque et al. 2000, Hsieh et al. 1998). Another mutation, the dimeric interfacial mutant Y166F, resulted in a breaking of the proton wire and, not surprisingly, caused a 10fold decrease in catalysis (Hearn et al. 2004). The catalytic rate, k1, for the E162D mutant exhibited an interesting property. The pH dependence of k1 gave a pKa near 8.7 that is nearly in agreement with the pKa from the spectroscopically monitored titration of the active site (Figure 22). This indicates that there is a catalytic dependence on the protonated state of a single group with a pKa near 8.7. It is possible that the group of interest is the Mnbound hydroxide/water, as one would expect the most noticeable effects to occur closer to the metal due to the second shell nature of the mutated residue. It is also possible that this group is an amino acid that lies near the active site, as this rate is dependent on protontransfer to the Mnbound hydroxide. However, pinpointing this location would require concurrent

PAGE 46

46 mutational analysis at other positions in the active site, such as Tyr34 and Gln143, for example. This property was not observed in E162A, which had a spectroscopically titrated p Ka of around 10.1. It may very well be that there is an equivalent kinetic ( k1) p Ka, though this pH value is above the range utilized in the kinetic measurements. As was previousl y mentioned, the mutant E170A in E. coli MnSOD is equivalent to the E162A mutant of human MnSOD. Interestingly, the properties of these two mutants showed some dramatic differences. The E170A E. coli MnSOD mutant resulted in an enzyme that preferentially bound iron, was catalytically dead, and had greatly weakened dimeric stability in solution. In contrast, E162A human MnSOD retained apparent dimeric stability and its preference for manganese though with substantially reduced activity relative to wild ty pe. Not surprisingly, the enzyme appeared to retain its tetrameric structure in solution and this tetramerization may keep the protein stable despite the weakened dimeric interface. The reason for the differences in the properties of these two mutants is not immediately apparent upon structural alignment of their structures. The dimeric interfaces of the two enzymes are nearly identical, as are all other amino acids (not including position 162) throughout the protein. However, analyzing the area at the mouth of the active site reveals that there exists an interaction between a phenylalanine and glutamine that is flipflopped between the two enzymes. The role of this interaction will be discussed in the next chapter. Another feature of catalysis that is prominent in human MnSOD is the large extent of product inhibition relative to the prokaryotic enzymes (Hearn et al. 2001, Hsu et al. 1996). The value of k0/[E], a measure of the zeroorder rate constant for superoxide turnover in the product inhibited pathway at steady state was lower for E162D (270 s1)

PAGE 47

47 and E162A (190 s1) than that for wildtype (500 s1). Furthermore, the rate of formation ( k3) and release ( k4) of the product inhibited state were diminished in both mutants (Table 21). Yet another measure of the extent of product inhibition is obtained by comparing the values of k2 and k3, known as the gating ratio ( k2/ k3). The value of this ratio obtained for both mutants is roughly 0.6, indicating enhanced flux through the pathway of formation of the product inhibited complex (eq 13). These data indicate that the extent of product inhibition was greater for the two mutants due to a preferential entrance into the product inhibited state ( k3) and a slow rate of release of the product inhibited state ( k4). These results suggest that Glu162 plays a role not only in the tuning of the pKa and redox state of the active site environment, but also in the release of the product inhibited state. Conclusions The dimeric interface of human MnSOD plays a rol e in several aspects of the enzymes structure and function. Glu162 at the dimeric interface interacts directly with His163 on a neighboring monomer, a direct ligand of the Mn. Mutational analysis (E162D and E162A) showed that this interaction is importa nt in determining the stability of the tetrameric complex as evidenced by a decrease in the temperature of the major unfolding transition that was dependent on the degree of interaction between residue 162B and His163A. Kinetic analysis showed that there was also a substantial loss of catalytic activity in the mutants and that there was a greater degree of product inhibition exhibited. These results illustrate the need to understand the role of the dimeric interface as the development of novel antiproliferative forms of MnSOD is undertaken (Davis et al. 2004), as mutations may cause unexpected, undesirable effects. The

PAGE 48

48 dimeric interface must also be examined to further understand its role in altering the properties of the eukaryotic MnSODs as compared to their prokaryotic counterparts. Table 21. Rate constants for catalysis by wildtype and mutant human MnSOD Enzyme k 1 ( 1 s 1 ) k 2 ( 1 s 1 ) k 3 ( 1 s 1 ) k 4 ( s 1 ) wild type a 1500 1100 1100 120 E162D b 355 33 133 16 215 20 40 4 E162A b 63 4 50 4 87 8 30 3 a From Quint et al. 2006. b In 2 mM TAPS (pH 7.7), 50 mM EDTA, and 30 mM formate at 25 C Table 22. X ray diffraction data processing and structure refinement statistics for E162D and E162A human MnSOD Parameter E162D E162A S pace group P 6 1 22 P 6 1 22 Unit cell parameters ( ) a = 81.3, c = 241.7 a = 81.3, c = 242.5 Resolution ( ) 20 2.2 (2.28 2.20) a 20 2.5 (2.59 2.50) No. of unique reflections 23358 15609 Completeness (%) 93.5 (90.7) 90.1 (92.7) R sym b (%) 11.2 (19.6) 11.4 (21.0) I/ (I) 28.2 22.2 R cryst c (%) 17.58 17.7 R free d (%) 20.09 22.2 No. of protein atoms 3106 3100 No. of water molecules 175 103 rmsd for bond lengths ( ) 0.006 0.006 rmsd for bond angles () 1.248 1.325 Average B factors (main/side/solvent) 21. 7/24.4/35.3 26.8/28.2/33.4 Ramachandran Plot (%) (favored/additional/generous) 91.2/1.2/7.6 91.8/7.0/1.2 a Data in parentheses are for the highest resolution shell. b Rsym = ( I I |/ I ) x 100. c Rcryst = ( o| |Fco|) x 100. d Rfree i s calculated the same as Rcryst, except with 5% of the data omitted from refinement.

PAGE 49

49 Figure 21. The active site structure of wildtype human manganese superoxide dismutase shows an intricate interaction network. The active site is made mostly of amino acids from one monomer (blue) coordinating the manganese (purple sphere) with two solvent waters (red spheres). Second shell ligands of the Mn are contributed by a neighboring monomer (green). Figure 22. The pH profile for molar abosorptivity at 480 nm for hMn3+SOD, wild type and mutants. Data for wildtype ( ), E162D ( ), and E162A ( ) were fit to a single ionization with pKa values of 9.2 0.1, 8.7 0.2, and 10.1 0.1, respectively. Measurements were made in solutions buffered with 200 mM MES a nd TAPS at 25 C.

PAGE 50

50 Figure 23. Change in absorbance at 420 and 480 nm over a 0.25 ms time scale after generation of superoxide by pulse radiolysis in a solution containing E162D human MnSOD. The decrease in the 480 nm reading is due to the reduction of the active site manganese after the first round of catalysis. The increase in the 420 nm reading is due to the formation of the product inhibited complex. This is achieved by first reducing the Mn with H2O2, followed by generation of superoxide by pulse radiolysis. Both changes in absorbance were fit to first order processes to give k3.

PAGE 51

51 A B C Figure 24. Normalized transitions for E162D (A), wildtype (B), and E162A (C) human MnSOD as determined by differential scanning calorimetry. Enzyme concentrations for each enzyme were 1.0 mg/mL. The data was fit with a nontwo state model. The normalized calorimetric trace is in black and the red line is a fit to the model, assuming noncooperative transitions. The average temperatures for inactivation and melting are listed in the text. Each experiment is the average of three scans with the subtraction of the reference buffer.

PAGE 52

52 Figure 25. Structures of the dimeric interface of wildtype (top), E162D (middle), and E162A (bottom) human MnSOD. The active site manganese ions are shown as a purple spheres, waters as small, red spheres, and putative hydrogen bonds as magenta dashes. The black ellipsoids illustrate the twofold axis of symmetry that separates one chain from the other. A solvent molecule was observed to bridge the interaction between Asp162 and His163 in E162D, while no interaction was observed in E162A.

PAGE 53

53 CHAPTER 3 COMPARATIVE STUDY OF HUMAN AND ESCHERICHIA COLI MANGANESE SUPEROXIDE DISMUTASE: KINETIC AND STRUCTURAL INSIGHTS As was discusse d in the previous chapter, the dimeric interface of the manganese superoxide dismutases plays an important role in the catalytic and structural properties of the enzyme. Mutagenic analysis of a second shell ligand, Glu162, in human MnSOD resulted in an enzyme, E162A, which had very different properties from its E. coli MnSOD counterpart, E170A. In chapter 3, further mutagenic analysis at the dimeric interface of human MnSOD is performed in order to create an enzyme that is more E. colilike. That is, an enzyme with a lesser degree of product inhibition. Introduction Manganese superoxide dismutases catalyzes a twostep reaction in which the active site Mn cycles between +2 and +3 oxidation states and the manganesebound solvent is protonated and subsequent ly deprotonated (eq 11 1 4). As is measured by pulse radiolysis the catalytic cycle begins with an ultrafast neutralization of superoxide anion. However, during the second step of the catalytic cycle, a product inhibited form of the enzyme appears. T his complex involves the binding of oxidized manganese (Mn3 +) to superoxide in the absence of a proton, forming an Mn3+O2 2 intermediate. The structure of this inhibitedstate has not been visualized directly, but there is evidence for two conformations. The first theory is based on spectroscopic evidence and suggests that the inhibited state is a sideon peroxo complex with the manganese (Bull et al. 1991; McAdam et al. 1977; Abreu et al. 2007). Computational calculations, on the other hand, suggest an endon complex in which the peroxo compound is coordinated by the manganese and the hydroxyl of Tyr34 (Abreu et al. 2005). In order for the enzyme to continue, a proton must be transferred into the active site. Evidence

PAGE 54

54 for this proton transfer event is demonstrated by an observed solvent hydrogen isotope effect of 3.1. It has further been suggested that this proton transfer is the limiting step in release of the product inhibited state, and therefore catalysis (Hsu et al. 1996). Analysis of the human MnSOD active site hints at how the proton would be transferred into (or out of) the active site. At the core, the manganese ion is coordinated by three histidines, an aspartate, and a solvent molecule (water or hydroxide). Extending out of the active site is an apparent hydrogen bonded network of amino acid side chains and waters (Figure 21). This network is present in both human and E. coli MnSODs as well as in iron SODs (Miller et al. 2004; Borgstahl et al. 1992; Lah et al. 1995; Smith & Doolit tle, 1992) (Figure 31). Despite the similarities at the core of the active site, the human and E. coli MnSODs exhib it one major catalytic difference: the E. coli enzyme, as well as mos t other bacterial MnSODs, has a reduced level of product inhibition ( Bull et al. 1991; McAdam et al. 1977; Hsu et al. 1996; Abreu et al. 2007). Prior efforts have attempted to lower the magnitude of product inhibition of human MnSOD (Hearn et al. 2001). This work was motivated by the observation that the lower produc t inhibition of an H30N mutant of human MnSOD is anti proliferative when over expressed in human cancer cells (Davis et al. 2004). The theory is that the there is an increase in H2O2 production in the lesser product inhibited forms. It has been shown pr eviously that H2O2 is a potent regulator of various cellular processes and that increased H2O2 levels lead to an arrest of cellular growth (Sundaresan et al. 1995; Rodriguez et al. 2000; Davis et al. 2004). The problem with identifying lesser product i nhibited forms of human MnSOD is that mutation typically results in enzymes that have lower catalytic efficiencies, as demonstrated in Chapter 2 (Quint et al. 2008).

PAGE 55

55 Therefore, rational design of an enzyme with wildtype efficiency and lower product inhi bition requires an understanding of the mechanism of product inhibition. A large amount of research has been devoted to understand the role of Tyr34 in the catalytic mechanism of human MnSOD (Maliekal et al., 2002; Whittaker & Whittaker, 1997; Guan et al. 1998; Edwards et al. 2001). Mutation of the tyrosine to phenylalanine (Y34F) resulted in an enzyme that had a significantly increased level of product inhibition (Guan et al. 1998). Thus, Tyr34 is indispensible for catalysis, and it may be that its environment is an important determinant of this role. In fact, this environment differs between the human and E. coli enzymes and represents one of the very few active site differences between the two enzymes. This region exists at the dimeric interface an d involves the interaction between a phenylalanine and a glutamine (human) or asparagine ( E. coli) (Figure 3 1). In the human enzyme, Phe66A lies near Gln119B. In E. coli MnSOD, this interaction is flip flopped, with Gln73A neighboring Phe124B. To better understand the catalytic differences between E. coli and human MnSODs, mutagenic analysis was performed at residue Phe66 of human MnSOD (Zheng et al. 2007). Two mutants were created, F66A and F66L, with the F66L enzyme closely resembling the E. coli en zyme in terms of catalytic activity (reduced product inhibition), thus identifying a region of the active site that affects product inhibition. Materials and Methods Enzymes Mutant enzymes were created by sitedirected mutagenesis of the cDNA encoding h uman MnSOD in a pTrc99A vector using the QuikChange kit (Stratagene, La Jolla,

PAGE 56

56 CA). Successful mutagenesis was confirmed with DNA sequencing (ICBR, University of Florida). Mutationcontaining plasmids were then transformed into a strain of E. coli that l acked the genes encoding FeSOD ( SodB-) and MnSOD ( SodA-) (Carlioz et al. 1986). The protein was then expressed and purified as discussed in Chapter 2. Manganese concentrations were checked by inductively coupled plasma mass spectrometry with the active enzyme concentration taken as the manganese concentration. Pulse Radiolysis Kinetic rate constants were determined via pulse radiolysis performed by Diane Cabelli at Brookhaven National Lab. A more detailed version of the reaction conditions can be found in Chapter 2. Crystallization The enzymes both crystallized in normal human MnSOD conditions (2.5 M ammonium sulfate, 100 mM imidazole, 100 mM malic acid) using the hanging drop vapor diffusion technique. Data Collection and Refinement X ray diffraction d ata were collected using a Rigaku RU H3R CU rotating anode generator running at 50 kV and 100 mA, with Osmic mirrors, a 0.3 mm collimator and R AXIS IV++ image plates. Due to the large unit cell vector along the c axis, the crystal to detector distance was set at 190 mm with an oscillation of 0.3 per image over 45 total collection. One crystal was used for the F66L structure whereas two crystal data sets were merged for F66A. X ray data processing was performed using DENZO and the data were scaled (and merged) with SCALEPACK (Otwinowski & Minor, 1997). Initial attempts at direct phasing using the structure of wildtype human MnSOD (PDB ID 1luv, Hearn et al. 2003) were unsuccessful. Molecular replacement was then

PAGE 57

57 performed with the program MOLREP from the CCP4 suite of software (Collaborative Computational Project, Number 4, 1994; Vagin & Teplyakov, 1997). This showed that the tetrameric interfacial 2fold coincided with one of the crystallographic 2folds. Normally the dimeric interfacial 2fold is c oincident for this space group. To avoid phase bias, prior to refinement the side chains of residue 66 and Gln119 were mutated to alanines and all solvent atoms and the Mn were removed. Both structures were refined using the CNS suite of programs with ini tial rounds of rigid body refinement and simulated annealing to 3000 K (Brunger et al. 1998). Iterative rounds of energy minimization and B factor were then performed with an intervening round of manual model building and automated water picking in the g raphics program C OOT (Emsley & Cowtan, 2004). The F66A structure was refined to 2.2 resolution with a final Rcryst of 19.5%. Similarly the F66L structure was refined to 2.3 resolution with a final Rcryst of 19.9%. Complete data processing and model refinement statistics are given in Table 31. The atomic coordinates and structure factors for F66A and F66L were deposited in the Protein Data Bank with PDB IDs 2qka and 2qkc, respectively. Results Catalysis Rate constants for the catalytic pathway descr ibed in Chapter 1 (eq 11 1 4) were determined using pulse radiolysis to generate superoxide and monitoring the decrease in absorbance and 260 nm, the peak absorbance for superoxide ( 260 = 2000 M1 cm1) (Rabani et al. 1969). For a detailed explanation of the procedures used to determine the rate constants, see Results in Chapter 2. The F66L mutant showed a more rapid initial catalytic rate than F66A, however both enzymes were impaired as compared to wildtype (complete kinetic data are given in Table 32). F6 6A, but not

PAGE 58

58 F66L, showed a region of zeroorder catalysis corresponding to the product inhibited form of the enzyme that is predominant in catalysis by human MnSOD (Figure 3 2 ). The rate of superoxide decay in this zeroorder region, k0/[E], for F66A was 1740 s1. This value is about three times greater than that for wildtype human MnSOD (500 s1), indicating a lower degree of product inhibition in F66A. Due to the lack of a zero order region in catalysis by F66L, k2k4 were determined using the Numerical Integration of Chemical Kinetics program in PRWIN (H. Schwartz, Brookhaven National Laboratory), fitting to the data shown in Figure 32. The rate constants for k1k4 were virtually pH independent over the pH range 6.58.5 for both F66L and wildtype human MnSOD. There was a small decrease in rate at pH greater than pH 8.5, however the data could not be fit with a single ionization. F66A exhibited similar traits for k1k3, however k4 showed an increasing value with increasing pH (Figure 33) (from 65 s1 to 150 s1). It should be noted, though, that there is an increase in experimental error as pH increases. The product inhibition of F66L as well as wild type human MnSOD a nd wild type E. coli MnSOD were examined in greater detail by examining the extinction coefficients of the enzymes over a range of wavelengths, on short (0.21.0 ms, all) and long (>50 ms, wild type human) time scales (Figure 34). Wild type human MnSOD s howed signs of the formation of a reversible intermediate on short time scales, as evidenced by the drop in absorbance at 480 nm. On longer time scales, the spectrum returned to that of Mn3+SOD, with a peak at 480 nm. Examination of both time scales for F66L human MnSOD and wild type E. coli MnSOD did not reveal the formation of this reversible state (data for short time scale shown in Figure 34).

PAGE 59

59 Structural Analysis Mutation of Phe66 to alanine and leucine showed no global effects on the structure of human MnSOD. There were effects on the positions of two amino acids near residue 66A, Tyr34A and Gln119B. In the F66L structure, Leu66 was positioned similar to Phe66 in the wildtype enzyme, with the and 1 carbons of Leu66 nearly superimposing on the s ame atoms of Phe66. This positioning caused the 2 carbon of the leucine to extend towards the surface of the enzyme and Gln119B, thereby shifting the side chain of the glutamine towards the surface of the enzyme to maintain the distance between the two r esidues (Figure 35). The loss of the phenyl group in the mutation F66A allowed for the side chain of Gln119B to face into the active site with a 5). In both mutants, the position of Tyr34 was shifted towards t he surface of the enzyme by about 0.5 This position of Tyr34 is more similar to that seen in E. coli MnSOD. In addition to the effects on amino acids, the positions of two water molecules were affected (Figure 36). W2A connects the side chains of Ty r34 and His30 and W2B is positioned between residues 66 and 119 and the hydroxyl of Tyr34. The distances between the two waters and Tyr34 remain relatively unchanged in the mutant structures. However, the two waters are further apart from each other in t he mutants relative to wildtype human MnSOD (Figure 37, Table 33). Further structural analysis was undertaken to examine the changes in the mobility in the residues surrounding the mutation site. This is accomplished by looking at the B factors of thes e atoms. A B factor is a crystallographic measure of the thermal motion associated with a given atom. To make this comparison, the B factors had to be normalized across all structures (F66A, F66L, wildtype human MnSOD, and wildtype E. coli MnSOD). Thi s was especially important becaus e the E. coli MnSOD structure

PAGE 60

60 was solved at 100 K, while all of the human structures were determined at room temperature (Borgstahl et al. 2000). Normalization was performed for each atom individually the equation: Bnorm = Batom x [/] (3 1) Where Batom is the calculated B factor for the atom and and are the average B factors for the given protein and all four proteins over a conserved region, respectively. This conserved region is the main chain atoms (C, N, O, and C) of residues in regions of conserved tertiary structure between the human and E. coli MnSODs (human residues 2134, 120126, 157175; E. coli residues 2134, 125 131, and 165178). This selection of normalization residues (rmsd 0.2 ) was chosen due to the high degree of structural variation between human and E. coli MnSODs, outside of the core of the enzyme. The error bars shown in Figure 38 were calculated by comparing the normalized B factor s for identical residues in the A and B monomers, where A and B were solved independently (i.e. without the use of a noncrystallographic symmetry operator). This B factor analysis showed that the manganese ion, solvent W1, Tyr34 and His30 were well ordered in all four enzymes (Figure 38). Solvent molecule W2B was relatively more disordered in F66A, F66L, and E. coli MnSOD compared to wildtype human MnSOD. Interestingly, solvent molecule W2A was more disordered in the two enzymes that exhibited a lower level of product inhibition, F66L and E. coli MnSOD, while it was relatively equally ordered in wildtype and F66A human MnSOD. Discussion Despite the high degree of structural similarity between the monomers of eukaryotic and prokaryotic MnSODs, there ar e substantial catalytic differences between

PAGE 61

61 the classes, most notably in the level of observed product inhibition. A comparison of human and E. coli MnSODs reveals one significant difference at the mouth of the active site. This is the interaction between a phenylalanine on one monomer and a glutamine (human) or asparagine ( E. coli) on a neighboring monomer. Phe66 in human MnSOD lies near the catalytically important residue Tyr34, a residue that has been extensively studied in human (Greenleaf et al. 200 4, Guan et al. 1998) and E. coli MnSODs (Maliekal et al. 2002; Whittaker & Whittaker, 1997; Edwards et al. 2001). It therefore can be hypothesized that Phe66 exerts effects, indirectly, on Tyr34. To understand the role of this region, sitespecific mu tants of human MnSOD were made at residue 66 (F66A and F66L). Catalysis Initial kinetic analysis of F66A and F66L human MnSODs demonstrated that Phe66 is not vital for catalytic activity of the enzyme, as both mutants exhibited only minor decreases in catalytic rates when compared to wildtype (compare to the large decreases seen in Table 21 for E162D and E162A human MnSOD). Despite the small change in rate constants, it is very obvious from Figure 32 that there are differences in the progress curves of the enzymes. The zeroorder catalysis seen in F66A is similar to that of wild type human MnSOD, indicating that this mutant retains a similar level of product inhibition (Bull et al. 1991; McAdam et al. 1977). Interestingly, there is virtually no zeroorder catalysis seen in F66L human MnSOD, which is similar to the properties of E. coli MnSOD, indicative of a low level of product inhibition. Similarly, spectroscopic evidence shows that wildtype human MnSOD forms a short lived intermediate that is not seen in E. coli or F66L human MnSOD (Figure34).

PAGE 62

62 Further analysis of the degree of product inhibition is provided by analysis of the gating ratio, k2/ k3. A value of 1, as seen in wildtype human MnSOD, indicates that there is an equal transition of this enzyme into an uninhibited second step as well as into a stage of product inhibition. In E. coli MnSOD this ratio is 4, which means that this enzyme favors an uninhibited route of catalysis during the second step (Table 32). The mutant F66A has a gating ratio that is nearly equal to that of wildtype human MnSOD, whereas F66L has a gating ratio that is very similar to that of E. coli MnSOD, thus providing additional evidence that F66L has a low level of product inhibition. One may note that k4 is also p art of the inhibited step of catalysis and therefore could be included in analysis of product inhibition. In fact, analysis of k4 of these mutants shows that k4 is decreased in both enzymes, making each more similar to the E. coli enzyme. However, it sho uld to be noted that the lower rate of k3 in these enzymes lowers the level of inhibited state and, therefore, could directly lower the rate of release from the inhibited state. In other words, the number of cycles of release from the product inhibition i s lower in F66L, as there is a lower concentration of the product inhibited state available (increase k2, decreased k3). This is what is seen in F66L, as this enzyme exhibits the lowest rate for k4 (Table 32), but also demonstrates a nearly undetectable level of product inhibition (Figure 32). Previous mutagenic analysis of human MnSOD has provided another mutant that exhibits a lower level of product inhibition, H30N (Hearn et al. 2001; Greenleaf et al. 2004). The reason for the decreased inhibition is not the same in H30N as it is in F66L. In H30N, there is a greatly increased rate, k4 (480 s1), and no change in the gating ratio k2/ k3, as compared to wildtype. Therefore, the product inhibited state forms just as

PAGE 63

63 readily, possibly due to impaired proton transfer, however it is released much more rapidly (Ramilo et al. 1999). This is in contrast to the lower rate of entry into the product inhibited state ( k3) by F66L. Active Site Environment An examination and comparison of the crystal structures of these enzymes provides some rationalization of these results. Among the most obvious structural differences is the conformation of the amino acids surrounding the site of mutation. In wild type human MnSOD Phe66A neighbors Gln119B at the dimeric interface. Mutation of phenylalanine to alanine results in a conformational change in Gln119. The loss of the bulk of the side chain at residue 66 allows the side chain of Gln119 to bury itself into the active site, occupying space that would normally steric ally clash with Phe66 (Figure 3 5). This rotation allows the polar side chain of residue 119A to interact with solvent molecule W2B, which normally only interacts directly with Tyr34A. This novel interaction does not alter the interaction distance between W2B and Tyr34A, however the position of W2B moves outward, away from the active site Mn (Figure 36, Table 33). Associated with this movement is an increase in the inherent mobility of W2B as measured by the normalized atomic B factor (Figure 3 8). Thi s increased mobility is also associated with E. coli and F66L MnSOD when compared with wildtype human MnSOD. Even more notable is the change in mobility of solvent molecule W2A, which mediates an interaction between Tyr34 and His30. There is no apparent change in the normalized B factor of W2B in F66A compared to wild type human MnSOD (Figure 38). However both F66L and E. coli MnSOD show an increased level of mobility for W2B relative to wild type human MnSOD. Additionally, solvent molecule W2A is m ore dynamic in both mutants as well as the E. coli enzyme.

PAGE 64

64 This increased mobility of the solvent in the active site correlates to the level of product inhibition; that is, the higher the mobility, the lower the level of product inhibition. It is possible that the high degree of product inhibition seen in human MnSOD is due to the rigidity of the active site environment One can envision a state of catalysis in which the product inhibiting moiety is bound in the active site, possibly in an endon state as is observed for the MnSOD inhibitor azide (Lah et al. 1995). In this position the Mn and the hydroxyl of Tyr34 would coordinate this moiety. A more static environment may lead to the stabilization of this intermediate, which kinetically would be observ ed as an decrease in the gating ratio ( k3 > k2) and/or a decrease in k4. The converse, therefore, would be true for a more dynamic active site, as is observed in the lesser product inhibited enzymes, E. coli and F66L human MnSOD. Recall that the major di fference between these two wildtype enzymes is the level of quaternary structure. It has been shown that the tetrameric organization of human MnSOD acts to stabilize the enzyme. This stabilization can be correlated to a more rigid protein, suggesting that part of the reason for the increased product inhibition of human MnSOD is due to the higher level of quaternary organization. Additional evidence for the effects of changes in the solvent network on catalysis is provided by previous studies examining so lvent hydrogen isotopes effects (SHIE). In a study by Hearn et al. (2001), it was shown that the rate of formation of H2O2 along the inhibited pathway ( k4) has a very large SHIE, indicating a role for proton transfer during this step. Compared to wildty pe human MnSOD, the two mutants and E. coli enzyme have a significantly decreased rate k4 (Table 32). The higher degree of disorder in the

PAGE 65

65 active sites of these three enzymes agrees with this, in that there would be a loss of support for proton transfer in a disordered system. Conclusions The active site environment of the MnSODs has been shown to be vital for efficient catalysis. Mutational analysis of Phe66 in human MnSOD has demonstrated a possibility to alter the kinetic properties of the enzyme, particularly in the formation and release of the product inhibited state. Both F66A and F66L mutants demonstrated lower rates of activity at all stages of catalysis. Additionally, F66L provided an enzyme with a significant decrease in the extent of produc t inhibition with only a minor decrease in the rate of the first stage of the reaction. Thus, this enzyme has characteristics more similar to the E. coli enzyme. Through crystallographic B factor analysis, it was also shown that both enzymes, as well as E. coli MnSOD, have a more dynamic solvent structure in the active site pocket. Proton transfer in these enzymes may be impeded as a result of the increased mobility of solvent molecules along the proposed proton wire. Also, this dynamic nature leads to t he conclusion that the mobility of the active site solvent correlates directly with the extent of product inhibition, with less dynamic enzymes stabilizing the product inhibited intermediate state. Table 31. X ray diffraction data processing and structur e refinement statistics for F66A and F66L human MnSOD Parameter F66A F66L Space group P 6 1 22 P 6 1 22 Unit cell parameters ( ) a = 81.2, c = 242.8 a = 81.1, c = 242.5 Resolution ( ) 20 2.2 (2.28 2.20) a 20 2.5 (2.59 2.50) Redundancy 5.2 (3.7) 3.1 (2.8) Completeness (%) 93.5 (90.7) 92.6 (90.4) R sym b (%) 8.0 (17.0) 7.3 (13.1) I/ (I) 31.2 38.4 R cryst c (%) 19.5 19.9 R free d (%) 21.7 20.6

PAGE 66

66 No. of protein atoms 3097 3103 No. of water molecules 136 142 rmsd for bond lengths ( ) 0.008 0.007 rmsd for bond angles () 1.4 1.3 Average B factors (main/side/solvent) 24.5/27.7/33.1 22.8/25.5/24.1 a Data in parentheses are for the highest resolution shell. b Rsym = ( I I |/ I ) x 100. c Rcryst = ( o| |Fco|) x 100. d Rfree is calculated the s ame as Rcryst, except with 5% of the data omitted from refinement. Table 32. Rate constants for catalysis by wildtype and mutant human and E. coli MnSOD obtained by pulse radiolysis Enzyme k 1 ( 1 s 1 ) k 2 ( 1 s 1 ) k 3 ( 1 s 1 ) k 4 ( s 1 ) Wild type h uman a 1500 1100 1100 120 F66A b 600 500 700 82 F66L b 700 800 200 40 Wild type E. coli c 1100 900 200 60 a From Hearn et al. 2001 b In 2 mM HEPES (pH 8.0), 50 M EDTA, and 30 mM formate at 25 C Experimental uncertainties in rate constants are no greater than 15%. c From Abreu et al. 2007. Table 33. Geometric distances with the active site of wildtype, F66A, and F66L human and wildtype E. coli MnSOD as illustrated in Figure 37. distance wild type human F66A F66L wild type E. coli d 2A 2.8 a 3.0 3.0 3.2 d 2B 2.9 2.7 3.3 2.6 d 2AB 2.9 4.0 3.8 2.7 d 30 2.6 2.9 2.9 2.8 d 166 2.7 2.6 2.6 2.6 a Distances are in angstroms.

PAGE 67

67 Figure 31. Superposition of the active sites of human (magenta) ( Hearn et al. 2003) and E. coli MnSOD (g reen) (Borgst ahl et al. 2000) shows the high degree of similarity in the active site. The manganese is shown as a purple sphere with the ordered solvent molecules shown as small spheres. A (B) following the residue label indicates that the residue belongs to the neighboring monomer. Text color correspond to the enzyme color, black text indicates that the residue type and number is the same in both enzymes. Figure 32. Output from pulse radiolysis studies shows the decrease in absorbance at 260 nm over the cours e of the reaction demonstrating the zeroorder phase of catalysis seen in F66A and wildtype human MnSOD. In the foreground is shown F66A (gray) and F66L (black) catalytic progressions. The inset shows the same reaction for wildtype human (gray) and E. coli (black). The zeroorder (linear) phase is clearly seen in F66A and wildtype human MnSOD after about 1 ms. The initial superoxide concentration was 10 M and enzyme was at 1 M.

PAGE 68

68 Figure 33. The pH dependence of k4 in catalysis by F66A human MnSOD. The data shown are the average of between 3 and 5 experiments. The reaction conditions were 30 mM sodium formate, 50 M EDTA, and buffer at 2 mM (HEPES at pH 6.4 7.9, TAPS at pH 8.08.8, CHES at pH 8.910.9. Note that the error in measurement increases with pH. Figure 34. Spectroscopic evidence of the formation of an inhibited complex in wildtype human MnSOD. The formation of the product inhibited st ate occurs over short time scales (0.21.0 ms) in human MnSOD ( ) as illustrated by the decrease in extinction coefficient at 480 nm. This inhibited state was not seen on longer time scales in the same enzyme (>50 ms) ( ) indicating a return to the Mn3+ s tate. There was no distinguishable formation of the product inhibited state observed in either E. coli ( ) or F66L ( ) human MnSOD on the short time scale.

PAGE 69

69 Figure 35. The effects of mutations at Phe66 in human MnSOD are clearly visible at position 119 when compared to wildtype (magenta) ( Hearn et al. 2003). The glutamine at this position shifts outward in the F66L mutant (blue), due to a steric clash with the 1 carbon of the leucine. Removal of the phenyl ring by mutation to alanine (F66A, orange) allows the side chain of Gln119 to rotate towards the active site, bending by about 90 at the carbon. Also notice the slight outward movement of Tyr34 observed in both mutant structures. Figure 36. A stereoscopic view of the active site environment of human MnSOD and its mutants F66A and F66L. The manganese ion is shown as a purple sphere and the water molecules as small spheres colored according to the structure to which they belong (wildtype = magenta, F66A = orange, F66L = blue). Notice the change in the positions of the water molecules W2A and W2B.

PAGE 70

70 Figure 37. A diagrammatic representation of the geometry of the active site of the MnSODs with the dis tances presented in Table 33 defined with dashed lines. The finely dashed line for d2AB indicates that this is most likely not a hydrogen bonding interaction, as the distance is generally too long. Figure 38. A histogram showing the normalized B fact ors of the atoms located in the active site of MnSOD. Shown here are the B factors for wild type human (magenta) ( Hearn et al. 2003), wild type E. coli (green) (Borgstahl et al. 2000), F66A (orange), and F66L (blue). The values given for Tyr34 and His3 0 are those of the average normalized B factor for the entire side chain. Normalization was performed against the average B factor for a selection of main chain atoms as described in the Results section. Error bars represent the normalized standard deviations across independently refined monomers from the same structure.

PAGE 71

71 CHAPTER 4 SEQUESTRATION OF SUBSTRATE CARBON DIOXIDE IN THE ACTIVE SITE OF HUMAN CARBONIC ANHYDRASE II The previous two chapters have focused on the enzyme manganese superoxide dismutase and the role of amino acids near the active site in modulating the enzymes activity. It was shown that the manganese ion at the active site center is finely tuned by second shell ligands and that product inhibition may be lessened by a more mobile active site. The following chapters will extend active site analysis to another class of enzymes, the carbonic anhydrases (CA). The CAs utilize a mechanism that does not involve a redox reaction, as the metal at the center is typically a zinc. This chapter exa mines the binding of substrate in the active site of human CA II (HCA II), and it is shown that the active site residues are very static and that this is beneficial for rapid catalysis (Domsic et al ., 2008; Domsic & McKenna, 2009). Introduction HCA II is a very efficient enzyme, with rates approaching the diffusioncontrolled limit ( kcat ~106 s1, kcat/KM ~108 M1 s1). The reaction catalyzed is the reversible hydration of carbon dioxide to form bicarbonate and a proton. As discussed in Chapter 1, the reaction is two step with the hydration occurring in the first step, resulting in a zinc bound water. The transfer of a proton in the second step regenerates a catalytically ready zinc bound hydroxide. It is the active site make up of HCA II that allows it to accept a hydrophobic substrate on one side and form a solvent mediated proton wire on the other side. The entrance to the active site is about 15 wide at the surface of the protein, tapering down in a roughly conical fashion to the zinc ion. It is t he cavity formed by this cone that provides both hydrophobic and hydrophilic environments, divided nearly in half at the zinc (Figure 41). The atoms of the

PAGE 72

72 hydrophilic half are responsible for forming the proton wire and include Tyr7, Asn62, His64, Asn67, Thr199O and Thr200O (Figure 41) ( Tu et al. 1989; Jackman et al. 1996; Fisher et al. 2005). Direct structural analysis of hydrophobic substitutions at positions 62 and 67 (N62L and N67L) have shown that these mutations affect the ability of the protein to form an active proton wire, resulting in drastically reduced rates of proton transfer (Fisher et al. 2007). The hydrophobic half is made up of atoms including Val121, Val143, Leu198, Thr199CH3, Val207, and Trp209. It has been suggested by molecular dynamics studies that this half is important for binding carbon dioxide (Liang & Lipscomb, 1990; Merz, 1991). Also, mutational analysis has shown that increasing the bulk of the side chain at position 143 has a deleterious effect on catalytic activity ( Fierke et al ., 1991; Alexander et al. 1991). However, there has been little direct structural evidence for the mechanics of the first part of the reaction, specifically the binding of substrate. It has been stated that the low solubility of CO2 in aqueous solution and the extremely rapid turnover of HCA II catalyzed CO2 hydration preclude any structural studies of substrate binding in HCA II (Liang & Lipscomb, 1990). These limitations were overcome by utilizing a high pressure environment, however, and the structure of CO2bound HCA II was solved. The results suggest that the CO2 passively binds to the hydrophobic half of the active site cavity and that the zinc plays no role in substrate binding, allowing for rapid catalytic turnover (Domsic et al ., 20 08). Materials and Methods Enzymes Human CA II was expressed in E. coli BL21(DE3)pLysS cells by induction with 0.1 mM IPTG for four hours. Following induction, cells were harvested via centrifugation,

PAGE 73

73 frozen overnight, and then lysed in 100 mM Tris, 200 mM sodium sulfate, pH 9.0 with ~1 mg/mL lysozyme. The lysate was clarified and the protein was purified using affinity chromatography on agarose resin coupled with p (aminomethyl) benzenesulfonamide (Sigma Aldrich, St. Louis, MO). Elution was performed with 100 mM Tris, 400 mM sodium azide, pH 7.0. The azide was then removed by extensive dialysis against 10 mM Tris, pH 8.0 and the protein concentrated using centrifugal ultrafiltration. To test the effects of the zinc on CO2 binding, a zinc free, or apo, enzyme was prepared by Balendu Avvaru in the McKenna lab. Briefly, the enzyme was incubated at 20 C in the presence of the strong zinc chelator, pyridine 2,6 dicarboxylic acid, (100 mM in 25 mM MOPS, pH 7.0). The chelator was then removed by dialysis against 50 mM Tris, pH 8.0 and removal of the zinc was verified by a complete loss of catalytic activity of the sample. Crystallization Crystals of the apo and holo enzymes were both obtained using the hanging drop vapor diffusion technique. The reservoir contained 1.3 M sodium citrate and 100 mM Tris pH 8, and trays were placed at room temperature. Crystals appeared within 7 days and grew to approximately .2 mm3. Carbon Dioxide Trapping Several considerations had to be made with regards to trapping the CO2 in the crystals. First, it was necessary that the CO2 remain in the crystal after treatment and that the catalytic reaction did not take place. Secondly, the low solubility of CO2 in aqueous solutions precluded the use of CO2 saturated solutions, as the molarity of CO2 would be too low to saturate the binding sites in all molecules of the crystal. These issues were solved by utilizing a technique developed in Sol Gruners laboratory in

PAGE 74

74 which protein crystals were placed under highpressure helium to freeze the crystals without the use of cryoprotectant (Kim et al ., 2005). This method involves putting the crystals on loops, ready for data collection, and putting them in highpressure tubes. The tubes are then filled with helium gas at pressures >10 0 atm. Once the pressure is reached, the crystals are dropped into a liquid nitrogen bath to prevent boiling of the trapped helium gas. All of this is done remotely with the entire device placed in carbon steel box. This method was modified and performed at much lower pressures for these experiments. Protein crystals have a very high solvent content (~50% on average), so prior to exposure to CO2, the crystals were coated with mineral oil. As with the helium method, the crystals were mounted onto nylon l oops and placed in the highpressure tubing. Previous attempts at rapidly plunging the crystals into liquid nitrogen were unsuccessful. Therefore, the crystals were left under 15 atm of CO2 for 25 minutes at room temperature. After this incubation, the end of the metal tubing was dipped slowly into a liquid nitrogen bath over a 2 min. time course. (The slow cooling required that the crystals be treated with 20% glycerol + reservoir solution prior to treatment.) This allowed for the solidification of the CO2 gas, as evidenced by a drop to 1 atm in the internal pressure of the tubing. The presence of CO2 in the crystal on the loop was confirmed by placement of the crystals at room temperature (by blocking the 100 K cryostream) resulting in a bubbling o f the crystal. Data Collection and Processing The crystals were transferred under liquid nitrogen to the goniometer and data was collected at a wavelength of 0.9772 on the A1 beamline at Cornell High Energy Synchrotron Source. The crystal to detector di stance was set at 65 mm with an

PAGE 75

75 oscillation angle of 1. To ensure completeness and data validi ty, a large number of images were collected for both holo and apo enzyme (624 and 360 images, respectively). Diffraction data were then indexed, integrated, and scaled using the program HKL2000 (Otwinowski & Minor, 1997). Crystals of both enzymes diffracted to 1.1 resolution with completeness = 99.9% and Rsym = 8.8% for holo and completeness = 93.1% and Rsym 8.0% for apo. Complete data reduction statistics are given in Table 41. Structure Refinement To ensure acceptable structure comparison, both the apo and holo enzymeCO2 co mplex structures were solved in the same manner using the program SHELXL (Sheldrick, 2008). Prior to refinement a random 5% of data were flagged for Rfree analysis (Brunger, 1992). A previously determined HCA II structure was used as the initial model (PDB ID: 2cba, Hakansson et al ., 1992). Before any fitting though, all heteroatoms and alternate conformations were deleted from the structure. The initial round of refinement was least squares, rigidbody fitting at 2.5 resolution. This resulted in Rfactor/Rfree of 33.3/33.2% for holo and 28.0/28.6% for apo. The resolution was then extended to 1.5 and refinement was switched to the conjugant gradient least squares (CGLS) method. After 20 cycles, the protein model and weighted electron density maps were read into the graphics program Coot (Emsley & Cowtan, 2004). All side chain positions were verified manually and the zinc was built into the appropriate density in the holo structure. In the next cycle of CGLS refinement, water molecules were added and then checked manually against the generated electron density maps. Data was then extended to 1.1 resolution and more water molecules were added. After all waters were satisfactorily added the carbon

PAGE 76

76 dioxide molecule(s) were added in the appropriate density. Leaving the CO2 molecules out ensured that any density observed was not an artifact of model bias, but rather was due to the unaccounted presence of substrate. All alternate conformations were then modeled in followed by the addition of riding hydrogens (at all positions except histidines). The final round involved increasing the weighting factor to 0.2. The final Rfactor/Rfree was 10.9/12.9% for holo and 10.4/13.9% for apo. Complete refinement statistics can be found in Table 41. Geometries of the structures were then analyzed in PROCHECK (Laskowski et al. 1993). Results Carbon Dioxide in the Active Site Carbon dioxide was found bound in both the holo and apo HCA II structures, located on the hydrophobic half of the active site (Figure 42). The highpressure environment and binding of CO2 had negligible effects on the overall structure of the enzyme as compared to wildtype (PDB ID: 2ili; Fisher et al ., 2007), with an overall C rmsd of 0.21 for holo and 0.15 for apo. It was assumed that the CO2 molecules in the holo and apo structures were at full occupancy and, as such, refined with average B factors of 10.90 and 10.35 2, respectively (comparable with neighboring protein atoms) (Table 41). The CO2 molecule lies within 4 of residues Val121, Val143, Leu198, and Trp209 (Figure 4 2, Table 42). The location of CO2 displaces an ordered water molecule termed the deep water (WDW) that has been observed in many other HCA II structures. In the holo st ructure, one of the CO2 oxygens, O(1) interacts with the amide of Thr199 with a distance of 3.5 between the atoms. The other CO2 oxygen, O(2), is positioned between the zinc ion and the side chain of Val121. Due to these interactions, the CO2

PAGE 77

77 molecule is placed in a sideon orientation with respect to the zinc, with both CO2 oxygens nearly equidistant from the zinc bound solvent (~3.1 ). This positions the carbon 2.8 from the zinc bound solvent. There is also a new, well ordered water molecule obs erved in a position that has not been observed in crystal structures of HCA II, which was termed the intermediate water (WI). WI is located between Thr200O and the O(2) oxygen of CO2. A list of interactions and the associated distances can be found in Table 42. Despite the absence of zinc, the geometry of the CO2 molecule in the apo structure is nearly identical to that of CO2 in the holo structure. The position that would normally be occupied by zinc is occupied by a water molecule, with the atom ce nter shifted ~0.6 closer to the histidine ligands (Figure 42 B). The distance between this water and the CO2 oxygens is still ~3.1 Therefore the CO2 has shifted position slightly, pivoting about its O(2) oxygen, into a closer interaction with the a mide nitrogen of Thr199. Its interaction distance is now 3.15 versus a distance of 3.5 observed in the holo structure (Figure 42). A Second Carbon Dioxide Binding Site Electron density corresponding to a carbon dioxide molecule was also found in anot her portion of the protein. This location is a hydrophobic pocket situated roughly 11 away from the active site (Figure 43). The binding of CO2 in this location displaces the phenyl ring of Phe226, causing the ring to tilt approximately 30. An addit ional residue in this pocket, Trp97, has been shown to act as a nucleation site for protein folding (Jonasson et al ., 1997).

PAGE 78

78 Additional Protein Structural Features The tilting of the side chain of Phe226 was the largest change observed in the CO2bound st ructures, when compared to wildtype HCA II. A large number of alternate conformations were observed in the holo and apo structures, a common feature of highresolution crystal structures. In the holo structure alternate conformations were seen for Ile22, Leu47, Ser50, Asp52, His64, Ser152, Ser217, and Val223. For the apo structure, alternate conformations were seen at Ile22, His64, Gln103, Asp162, Lys172, Glu214, Ser217, and Val223. The two conformations seen for His64 have been observed before and are termed the in and out conformations, with the in conformation being the typically preferred orientation (Silverman & McKenna, 2007; Fisher et al ., 2007). In the CO2bound holo and apo structures, the out conformation was favored. The presence of a glycerol molecule adjacent to His64 caused this preference, with the oxygens of the glycerol positioned where the ordered proton wire would normally be observed (Figure 4 4 A). Glycerol was also observed in both structures near residues 243245 on the surface of the protein (Figure 44 B). Discussion Trapping Carbon Dioxide The rate of CO2 hydration catalyzed by HCA II (108 M1 s1) approaches the diffusion controlled limit, with a ratelimiting proton transfer step that brings the maximal turnover to 106 s1. This high level of catalysis dictates that the substrate be bound loosely and is reflected by the KM of HCA II, 10 mM. Infrared spectroscopic measurements also suggest that HCA II has a weak affinity for CO2, with a calculated KD of 100 mM (Kreb s et al ., 1993). Overall, the weak binding of CO2 coupled to the high turnover rate of HCA II suggests that trapping substrate would be nearly impossible.

PAGE 79

79 The success of these experiments can be explained by the physical and chemical properties of this s ubstrate and enzyme. Firstly the large hydrophobic pocket in HCA II provides an excellent environment for CO2 solvation, allowing for the selective solubilization of CO2. Also, the constant of Henrys Law, for CO2 solubility, confirms that the concentrat ion of CO2 in the highpressure experiments (15 atm) is 450 mM, nearly 5 times the KD (Butler, 1982). Additionally, the catalytic activity of the enzyme was hindered by two factors. Firstly, by lowering the temperature to 100 K, catalysis was negligible because the energy barrier for HCA II driven catalysis is 10 kcal mol1. Secondly, the CO2 pressurization at room temperature likely acidified the crystal, thereby protonating the zinc bound hydroxide, effectively negating any enzymatic activity towards C O2. Taken together, this information suggests that, under these conditions, complete occupancy in the holo and apo HCA II crystals would be expected. Physiological Relevance Previous biochemical studies suggest that the CO2 binding site described in these crystal structures is physiologically relevant, and not an artifact of the pressurization used for substrate trapping. Mutational analysis of the hydrophobic pocket, specifically at Val143, indicated that bulky side chain substitutions at this position l ed to drastically reduced catalytic activities (Fierke et al ., 1991, Alexander et al ., 1991). For example, mutation to tyrosine (V143Y) resulted in an enzyme that had only 0.02% of the activity of wild type HCA II (Fierke et al ., 1991). The crystal structure of V143Y HCA II showed that the side chain of Tyr143 juts into the active site, effectively removing the hydrophobic pocket. Structural superposition of CO2bound holoHCA II with the V143Y structure (overall C rmsd = 0.26 ) clearly shows that the side chain of Tyr143 would

PAGE 80

80 directly block the observed CO2 binding site, thereby negating any catalytic activity (Figure 4 5). Further evidence is provided by structural analysis of CO2analog inhibitors of HCA II. In one study, the binding of the potent, isoelectronic HCA II inhibitor thiocyanate (NCO-) was observed crystallographically (Lindahl et al ., 1993). In this case, NCObinds in the hydrophobic pocket, and does not displace the zinc bound solvent. The inhibitor also displaces the solvent molecule WDW, forming a hydrogen bond with the amide nitrogen of Thr199. The distances between atoms in the active site and the inhibitor are very similar to those observed in the CO2bound structures. The carbon of NCOis 2.4 from the zinc bound solvent, s imilar to the CO2 carbons distance from the zinc bound solvent (2.8 ). NMR and IR studies of CO2 binding have, previously suggested the lack of a role for the zinc ion in CO2 binding, as suggested by a comparison of the CO2bound holo and apo structures (Bertini et al ., 1987; Williams & Henkens, 1985). It should be noted that, in these studies, the zinc was replaced with either copper or cobalt, since zinc is not paramagenetic. As will be discussed later, these studies also suggested that the product, bicarbonate, is coordinated directly to the metal. Implications For the Catalytic Mechanism The position of the CO2 molecule in the hydrophobic pocket and its direct interaction with the amide nitrogen of Thr199 optimally orient the CO2 carbon for nucleophilic attack by the zinc bound hydroxide. The subsequent formation of bicarbonate results in the formation of a bidentate ZnHCO3 complex as was proposed by Lindskog (1983) (Figure 4 6). The structure of bicarbonatebound HCA II was obtained previously by creating a mutation, T200H, which resulted in a greater affinity

PAGE 81

81 for bicarbonate (Xue et al ., 1993). A comparison of this structure with that of CO2bound holo HCA II shows that the CO2 lies nearly in plane with the bicarbonate (Figure 4 7). This suggests that these structures represent the beginning and end points of CO2 hydration. T he release of bicarbonate is facilitated by the diffusion of a water molecule into the active site. The presence of a never before seen water, WI, in close proximity to t he zinc suggests that WI may be a bicarbonatedisplacing water. WI may not be a new water, but rather the location of WDW when substrate is bound. The binding of CO2 may have local electrostatic effects, therefore allowing the displaced WDW to occupy a p reviously unfavorable position. However, the presence of a glycerol at the proton wire obfuscates this analysis. Conclusions The challenges of trapping substrate carbon dioxide in the active sites of one of the fastest enzymes known were overcome by a co mbination of pressurization and low temperatures. X ray crystal structures of both holo (with zinc) and apo (without zinc) HCA II in complex with CO2 revealed that the hydrophobic pocket on one side of the active site is responsible for substrate capture due to a more favorable CO2 solvating environment. The binding of substrate occurs in a passive manner, with no changes in local protein conformation, and the rearrangement of only one water molecule. The lack of change illustrates the ephemeral nature of substrate binding, which allows for rapid turnover (108 s1). Despite the apparent static nature of the hydrophobic pocket, it still plays a role in placing the substrate for efficient catalysis. Optimal orientation of CO2 is provided by the amide nitr ogen of Thr199, allowing for subsequent nucleophilic attack by the zinc bound hydroxide. Furthermore, the observation of a second, nonproductive

PAGE 82

82 CO2 binding site in another hydrophobic pocket suggests that this method may be used to probe hydrophobic str uctures in other proteins, providing insights into protein folding and catalytic mechanisms. Table 41. Data collection and refinement statistics for the CO2bound holo and apo HCA II structures. Parameter Holo Apo Space group P 2 1 P 2 1 Cell dimensions a, b, c () 42.4, 41.5, 72.4 42.2, 41.5, 72.3 90.0, 104.1, 90.0 90.0, 104.2, 90.0 Resolution () 20 1.1 (1.12 1.10) a 20 1.1 (1.12 1.10) R sym b (%) 8.8 (51.9) 8.0 (50.6) I/( )I 21.0 (4.1) 35.6 (4.3) Completeness 99.9 (100.0) 93 .1 (89.7) Redundancy 11.4 (10.8) 7.0 (5.8) No. reflections 98,494 86,919 R factor c /R free d (%) 10.90/12.89 10.35/13.87 No. atoms Protein 2096 2121 Zinc/CO 2 /glycerol 1/6/12 0/6/6 Water 404 352 B factors Protein (main/side) 10.4/ 15.2 11.1/15.6 Zinc/CO 2 e /glycerol 5.1/14.0, 39.5/20.1 NA/15.7, 59.7/19.1 Water 31.7 29.4 Ramachandran plot (%) Allowed 89.4 88.9 Additionally allowed 10.2 10.6 Generously allowed 0.5 0.5 rmsd f 0.192 0.144 a Values in parentheses are for the highest resolution shell. b Rsym = ( I I |/ I ) x 100. c Rcryst = ( o| |Fco|) x 100. d Rfree is calculated the same as Rcryst, except with 5% of the data omitted from refinement. d The first number is for the active si te CO2, the second is for the CO2 in the noncatalytic site. e The root mean square deviation when the respective structure is superimposed on the 1.1 structure of wild type HCA II (Fisher et al. 2007) using only Cpositions

PAGE 83

83 Table 42. Distances between atoms in the structures of CO2bound holo and apo HCA II. The distances shown are only for those atoms that lie within 3.9 of the CO2 molecule. The second atom in each row corresponds to the CO2 molecule. For information on CO2 atom naming, re fer to Figure 4 7 Interaction Holo Apo Zn bound H 2 O C 2.8 N/A H 2 O C N/A 2.9 Zn bound H 2 O O(1) 3 N/A H 2 O O(1) N/A 3.1 Zn bound H 2 O O(2) 3.1 N/A H 2 O O(2) N/A 3.1 Zn 2+ O(1) 3.2 N/A His94(C 1 ) O(1) 3.3 3.2 Leu198(C ) O(2) 3.4 3.5 His119(N ) O(1) 3.4 3.5 Thr199(N) O(2) 3.5 3.2 W I O(2) 3.5 3.2 His119(C ) O(1) 3.5 3.6 Val121(C ) O(1) 3.5 3.4 Trp209(C ) O(2) 3.5 3.3 W I C 3.6 3.2 His94(N 2 ) O(1) 3.6 3.5 His119(C ) O(1) 3.7 3.8 Leu198(C ) O(2) 3.7 4.1 Zn 2+ C 3.7 N/A Val143(C ) O(1) 3.7 3.8 Trp209(C ) C 3.9 3.9 Trp209(C ) C 3.9 3.9 H2O is the water molecule that occupies the site of the zinc ion in CO2bound apo HCA II

PAGE 84

84 Figur e 41. The active site of human carbonic anhydrase II is located at the bottom of a conical cavity. On one side is a hydrophilic patch (blue surface, top; blue sticks, bottom) that forms the solvent mediated proton wire (red spheres, bottom). The other side is lined with hydrophobic residues (magenta surface, top; magenta sticks, bottom). Three histidines (cyan sticks, bottom) and a solvent molecule coordinate the zinc.

PAGE 85

85 A B Figure 42. The electron density for the carbon dioxide molecule in both holo (top) and apo (bottom) HCA II is clearly observed in the hydrophobic patch (green mesh and sticks). The electron density shown for the protein and waters is a weighted 2FOFC Fourier map. The density for the CO2 is a weighted 2 FOFC Fourier map made prior to the modeling of the CO2 coordinates. Both maps are contoured to 2.25

PAGE 86

86 A B Figure 43. The noncatalytic carbon dioxide binding site was found in a hydrophobic patch (magenta surface) on the side of the enzyme opposite the hydrophobic patc h (green surface) (A). (B) A close up of this site clearly shows the CO2 density ( weighted 2FOFC at 1.5 ) and the tilting of the side chain of Phe226 (green) compared to its position in wildtype HCA II (red). A B Figure 44. Two glycerol molecul es were found in the structures of CO2bound holo and apo HCA II. (A) A glycerol molecule in the active site of CO2bound holo enzyme displaced the proton wire waters. (B) A second glycerol binding site was observed on the surface of CO2bound holo enzym e. Both electron density maps are weighted 2FOFC at 1.0 The structures shown are of the bound holo enzyme, however glycerol molecules were found in the same locations in the bound apo enzyme structure.

PAGE 87

87 Figure 45. A superposition of the CO2bound holo HCA II structure with that o f V143Y HCA II (Alexander et al. 1991) clearly shows that the side chain of Tyr143 (white sticks) would directly interfere with CO2 binding (cyan sticks). The hydrophobic patch is shown as a green surface.

PAGE 88

88 Figure 46. The catalytic mechanism of CO2 hy dration as catalyzed by HCA II, as proposed by Lindskog (1983). The enzyme begins with a zinc bound hydroxide, with an optimally oriented lone electron pair (top left). The binding of CO2 is followed by nucleophilic attack of the CO2 carbon by the zinc b ound hydroxide (top middle). This leads to formation of a monodentate bicarbonatezinc transition state (right). A bidentate zinc bicarbonate state then forms (bottom right) with the complex dissociating when a water molecule diffuses into the active sit e and coordinates with the zinc (bottom left). The regeneration of the active zinc bound hydroxide occurs with the transfer of a proton out of the active site (left).

PAGE 89

89 Figure 47. The structure of bicarbonatebound T200H HCA II reveals that the binding site of CO2 observed in this study is the catalytic site. The CO2 (cyan sticks) lies nearly in plane with the bicarbonate in the T200H structure (magenta sticks) (Xue et al. 1993). Additionally, it can be seen that the transition of WI, a water molecule never before seen in HCA2, may in fact be WDW displaced upon CO2 binding.

PAGE 90

90 CHAPTER 5 THE ROLE OF SURFACE RESIDUES IN PROTON TRANSFER BY HUMAN CARBONIC ANHYDRASE II The previous chapter discussed the mechanism by which human carbonic anhydrase II (HCA II) captures substrate, carbon dioxide (CO2). The crystal structures of CO2bound holo and apo HCA II demonstrated that the hydrophobic patch on one half of the active site is responsible for this process and that binding of substrate is passive, allowing for rapid turnover. This chapter will discuss the second stage of the reaction, the regeneration of zinc bound hydroxide by proton transfer. Specifically, the roles of residues forming acidic and basic patches on the surface of HCA II will be examined in t erms of their effects on tuning the p Ka of the proton acceptor, His64. Introduction As was discussed previously, the reaction catalyzed by HCA II is a two step cycle. The first step, in the hydration direction, is the hydration of CO2 to form bicarbonate (eq 1 5). The bicarbonate is then displaced from the zinc when a water molecule diffuses into the active site. A zinc bound water is inactive for catalysis, so a proton must be transferred off of this water molecule to regenerate a zinc bound hydroxide, a step that is rate limiting for the overall catalytic cycle (eq 16). This is accomplished by an ordered network of solvent molecules that form a hydrogenbonded wire (Figure 15). The end point of this proton wire is His64, which transfers the proton t o buffer in bulk solvent. The importance of His64 was revealed by mutational analysis, which showed that an H64A m utation results in a dramatically reduced rate of proton transfer (Tu et al. 1989). As was discussed in Chapter 1, His64 is known to occupy two conformations: an in conformation in which the side chain points toward the zinc and is flanked by two

PAGE 91

91 water molecules of the proton wire (W2A and W2B) with N 2 7.5 from the zinc, and an out conformation with the side chain facing bulk solvent and pi stacking with the indole ring of Trp5, with N 2 12.0 from the zinc (Figure 51). An earlier xray crystallographic study examined the effect of pH on the conformation of His64 (Nair & Christianson, 1991). It was observed that the transition from in to out occurs as the pH is decreased, indicating that the conformation of His64 is pH sensitive. The sensitivity of His64 to its environment has also been demonstrated by mutational analysis of residues located in other areas of the active site. O ne study showed that mutation of Asn62 to leucine caused His64 to be in the in conformation at pH 6.0 and 8.2 (Fisher et al. 2007). However, in the mutant N67L, His64 was observed in the out conformation, independent of pH. Another mutant, Y7F, resulted in an enzyme with a greatly increased rate of proton transfer, and His64 always in the in conformation. Interestingly, the pKa of His64 was determined to be lower when His64 was in the in conformation (pKa = 6.0 in Y7F) and was higher when in the out conformation (p Ka = 7.5 in N67L). Mutation of Thr200, a residue on the opposite end of the hydrophilic pocket, to serine resulted in the observation that His64 was solely in the out conformation (Krebs et al. 1991). Interestingly, there were no observed kinetic changes in this enzyme variant as compared to wildtype. This chapter will d iscuss the role of an amino acid on the surface of the protein, Lys170, which is situated 4.4 from the N 2 of His64 and 15 from the zinc (Figure 52). Intere stingly, Lys170 shows a high level of conservation across CAs from various species, including rat, chicken, bovine, and turtle. Mutations at Lys170 are shown to

PAGE 92

92 effect the side chain orientation and pKa of His64. An increased rate of proton transfer is o bserved in the dehydration direction because of these effects. Materials and Methods Enzymes Mutations at residue 170 (to alanine, aspartate, glutamate, and histidine) were made by sitedirected mutagenesis of a plasmid containing the cDNA encoding the human CA II gene using the QuikChange mutagenesis kit (Stratagene, La Jolla, CA). Successful mutagenesis was verified by DNA sequencing (ICBR, University of Florida). Mutated plasmids were then transformed into BL21(DE3)pLysS cells and protein expression w as induced by addition of IPTG and zinc sulfate (1 mM final concentration each). The protein was then purified as described in Chapter 4. Purity analysis was performed using SDS PAGE (data not shown). Crystallization and Data Collection All four mutants were successfully purified and crystallized. Crystallization was accomplished using the hanging drop vapor diffusion with a reservoir solution of 1.3 M sodium citrate, 100 mM Tris pH 8. Crystals grew to 0.2 mm3 within one week. The crystals were then sealed in quartz capillaries and x ray diffraction data was collected at room temperature using the system discussed in Chapter 2. An oscillation angle of 1 was used with a crystal to detector distance of 100 mm. Data Processing and Refinement Diffraction data were indexed, integrated and scaled using program HKL2000 (Otwinowski & Mino r, 1997). All crystals diffracted to ~1.7 with isomorphous unit cells (Table 51). Model refinement was performed using the prog ram REFMAC5 in the CCP4 suite of programs (Collaborative Computational Project, Number 4, 1994;

PAGE 93

93 Murshudov et al. 1997) with the starting model derived from the 1.54 structure of wild type HCA II (PDB ID 2cba, Hakansson et al. 1992), with all hetero atoms removed and Lys170 and His64 mutated to alanine. Refinement was carried out using an initial round of rigidbody fitting, followed by placement of the zinc using the molecular graphics program COOT (Emsley & Cowtan, 2004). After a round of restrained refinement, the side chains of residues 170 and 64 as well as waters were placed manually COOT Any difference map density present was modeled appropriately and the structure was refined with REFMAC5 followed by manual model building in COOT This iterative process was performed until Rfactor rea ched convergence. The final Rfactor for each structure was around 17%. Complete data processing and refinement statistics can be found in Table 51. 18O Exchange This technique measures the depletion of 18O from species of CO2 measured via membrane inlet mass spectrometry (Silverman, 1982). The CO2 species passing across the membrane enter a mass spectrometer (Extrel EXM 200) thus providing a measure of the isotopic content of the CO2 species. The dehydration of the labeled bicarbonate has a probabilit y of transiently labeling the active site zinc with 18O (eq 5 1). The subsequent protonation the zinc bound hydroxide produces H2 18O, which is then released into bulk solvent (eq 52). HCOO18O+ EZnH2O EZnHCOO18O COO + EZn18OH(5 1) H+His64 EZn18OHHis64 EZnH2 18O His64EZnH2O + H2 18O (5 2) This method provides two rates for CA catalyzed 18O exchange. The first is R1, the rate of exchange of CO2 and HCO3 at chemical equilibrium. R1/[E] = kex cat[S]/( KS eff + [S]) (5 3)

PAGE 94

94 Here, kex cat is the rate constant for the maximal interconversion of substrate and product, KS eff is the apparent binding constant for substrate to enzyme and [S] is the concentration of substrate, either carbon dioxide or bicarbonate. The ratio kex cat / KS eff is, in theory and in practice, equal to kcat/ KM obtained by steady state methods. The second rate obtained, RH2O, is the rate of release of water that bears 18O from the enzyme. It is this component of 18O exchange that is dependent on the donation of protons to the 1 8O labeled zinc bound hydroxide. RH2O/[E] = kB obs [B]/( Keff B + [B]) + RH2O 0/[E] (5 4) The value of this rate can be interpreted in terms of the rate constant for proton transfer from the proton donor to the zinc bound hydroxide according to eq 7. In this equation, kB is the rate constant for proton transfer and ( Ka)donor and ( Ka)ZnH2O are the ionization constants of the proton donor and zinc bound water. To determine the kinetic constant kB and the ionization constants, nonlinear least squares methods were used in the program Enzfitter (Elsevier Biosoft, Cambridge, U.K.) kB obs = kB/{[1 + ( Ka)donor/[H+]][1 + [H+]/( Ka)ZnH2O]} (5 5) The measurement for CA catalyzed and uncatalyzed 18O exchange were measured at 25C in the presence of a total substrat e concentration of 25mM by Dr. Chingkuang Tu in the Silverman lab. Additionally, the total ionic strength of the solution was kept at 0.2 M by the addition of sodium sulfate. Esterase Activity To correctly assign p Kas obtained from 18O exchange, another kinetic assay was performed, which allows assigment of the zinc bound solvent pKaCAs also possess the ability to hydrolyze ester linkages, with no known physiological relevance for this activity (Gould & Tawfik, 2005). To measure this activity, t he hydrolysis of 4-

PAGE 95

95 nitrophenol acetate was monitored by UV spectroscopy at 348 nm, the isosbestic point of nitrophenol and its conjugate, nitrophenolate ( = 5000 M1 cm1) (Verpoorte et al. 1967). Results Catalysis Mutations at residue 170 in HCA II had little effect on kcat/ KM of the enzyme, as the pH profiles of all four mutants were nearly superimposable on that of wildtype enzyme (Figure 5 3). It was also found that the pKa of the zinc bound water was approximately 7 in each of the mutants, thus unchanged with respect to wildtype (Table 52). This value was also confirmed by independent measurement using the esterase activity assay (Table 53). The catalytic effects of mutation at residue 170 were much more notable for the proton transfer stage of catalysis. As can be seen from the superimposed pH profiles of RH2O/[E] in Figure 5 4, there is a clear change in this proton transfer dependent rate constant. In this figure, the solid lines represent fits of the data to eq 54, with previous knowledg e of the pKa of the zinc bound water available from the determination of kcat/ KM discussed above. This fit provides values for the pKa of the proton donor, His64, and the rate constant for proton transfer, kB. One apparent effect of mutation and residue 170 is that the pKa of the proton donor, His64, is lowered relative to wildtype. This change is not very dramatic, with values between 6.3 and 6.7 as compared to a value of 7.2 for wildtype. Interestingly, the rate of proton transfer, kB, was increased in all four mutants with values ranging from 1.5 to 5 times greater than wildtype (800 ms1) (Fisher et al. 2007b; Duda et al. 2001). The final values for the four mutants and wildtype are presented in table Table 52.

PAGE 96

96 Structure The structures of all four mutants were solved to ~1.7 with final Rfactor/Rfree of ~17/20 %. Overall, no major structural perturbations were observed with C rmsd of approximately 0.09 for all four mutants when compared to wildtype. The solvent network that constitutes the proton wire was conserved in all four structures, indicating that the enzyme was structurally capable of proton transfer. A volume of density corresponding to an atom with more electron density than oxygen was observed near the side chains of His64, A sn62, and Asn67 and was modeled as a sodium ion, though the exact identity of this atom is not clear from crystallographic analysis. However, the resolution of the structures and the possibility of partial occupancy of this atom convolute an absolute determination of the atom type. The proton shuttle residue, His64, which normally occupies two conformations, was observed only in the in conformation in all four mutants. Additionally, there was a change in the position of residues His3 and His4. Typica lly at this resolution, electron density for His3 and His4 are not observed due to the inherent disorder of the N terminus of HCA II. In these four mutants, distinct electron density is observed for His3 and His4, showing that this residue occupies a conf ormation that is distinct from structures that have reported coordinates for this residue (Fisher et al. 2005; Fisher et al. 2007a) (Figure 55). The position of these residues is such that the side chain of His3 is located near the out conformation o f His64 (Figure 56). A small amount of additional density can be observed extending off of the backbone of His3, however it was not modeled into, as it does not cover an entire residue.

PAGE 97

97 Discussion Despite its position near the surface of the protein, Lys1 70 may play a role in finetuning the properties of His64, the proton shuttle residue. Mutation of Lys170 to alanine, aspartate, glutamate, or histidine resulted in alterations of the conformation and p Ka of His64 as well as the rate of proton transfer. The crystal structures of the four mutants reveal that His64 is always located in the in conformation. Also, it was observed that two N terminal histidines, His3 and His4 have distinct electron density, indicating an ordering of the N terminus. The loc ation of His3 is such that the out conformation of His64 is blocked, with the side chain of His3 located in an apparent pi stacking interaction with Trp5 (Figure 56). It cannot be said for certain whether the ordering of these two residues is a result of or a cause of the conformational locking of His64. Although support for the former is provided by the structure of K170D, in which the side chain of Asp170 interacts with the carbonyl oxygen of Asn62 (distance = 3.2 ). In turn, the carbonyl oxygen appears to interact with the Natom of His64 (Figure 5 7). Interestingly, the density for His3 appears to be the strongest for this mutant. Regardless of this cause/effect relationship, distinct changes in the kinetic properties of proton transfer were observed that are attributable to the orientation of His64. The p Ka of His64, as described by 18O exchange, was decreased in all four mutants, with values ranging from 6.3 6.7, compared to 7.2 for wildtype. Such a decreases is likely due to the hydrophobic environment occupied by His64 in this conformation. In fact, a decrease in pKa to a value of 6 was observed in the structure of Y7F HCA II, a mutant in which His64 was also seen to occupy only the in conformation (Fisher et al. 2007b).

PAGE 98

98 The decre ase in p Ka allows His64 to become a better proton donor, an effect of which is an increase in the rate of proton transfer, kB, in the dehydration direction (Table 5 2). This effect is reversed, however, when considering the opposite direction, the hydrati on of CO2. As an example, the value of kcat in the hydration direction is 1 x 106 s1 for wild type HCA II (Khalifah, 1971), but is decreased to about 0.5 x 106 s1 in K170A. This difference reflects evolution of the enzyme to have no preference for one direction of the catalytic cycle, thus allowing for a tight regulation of intracellular pH. This is due to a nearly equal pKa for both the zinc bound water and His64. However, this work does suggest that Lys170 plays a role in tuning the pKa of His64, al lowing for equal enzymatic efficiency in both directions. Conclusions These results highlight the role of a residue located very distant from the active site core in finetuning the acidbase properties of HCA II. Lys170 is located approximately 15 away from the zinc, near the location of His64. The results presented in this chapter show that mutation at Lys170 results in a decrease in the pKa of His64, a residue that is vital for efficient proton transfer. This decrease allows for His64 to be a more e fficient proton donor, which is its role in the dehydration direction of catalysis. Therefore, in the dehydration direction, the mutant enzymes, particularly K170E, offer an enzyme that has a higher rate of proton transfer. The cost of this, however, is a decrease in the catalytic efficiency of the enzyme in the hydration direction. Taken together with previous mutagenic analyses, it is apparent that the active site environment is finely tuned by several residues near the proton wire to allow for efficie nt proton transfer in both the hydration and dehydration directions (Fisher et al. 2007b). It can also be concluded from these studies that certain areas of the protein

PAGE 99

99 are intolerant to mutation, even though, at first glance, these areas may be hypothes ized to have only minute effects. It is thus not surprising that these mutations in HCA II have not been observed in nature, as the enzyme would be less effective in maintaining pH homeostasis. These studies also suggest that further analysis should be c arried out in these regions to identify other residues that may be vital for proper catalysis.

PAGE 100

100 Table 51. Data collection and refinement statistics for the Lys170 variant structures of HCA II. Parameter K170A K170D K170E K170H Space group P 2 1 P 2 1 P 2 1 P 2 1 Cell dimensions a, b, c () 42.6, 41.6, 72.8 42.6, 41.6, 72.8 42.6, 41.5, 72.7 42.7, 41.6, 72.8 104.5 104.4 104.4 104.5 Resolution () 50 1.65 (1.71 1.65) a 50 1.75 (1.81 1.75) 50 1.75 (1.81 1.75) 50 1.75 (1.81 1.75) R sym b (%) 6.9 (41.4) 6.2 (38.7) 6.7 (38.8) 5.4 (32.3) I/( )I 12.5 (2.1) 20.7 (2.4) 16.4 (2.3) 19.0 (3.3) Completeness 98.4 (98.1) 92.2 (80.9) 93.1 (80.4) 92.7 (80.9) Redundancy 2.7 (2.6) 3.2 (3.0) 3.1 (2.8) 3.2 (3.1) R factor c /R free d (%) 16.8/19.7 16.6 (19.8) 16.9 (20.3) 16.3 (19.5) No. atoms Protein 2076 2071 2075 2081 Water 139 101 84 102 B factors Protein (main/side) 14.9/17.5 14.8/17.4 15.7/17.1 15.3/18.1 Water 29.2 26.7 26.4 27.3 rmsd (bond/angle) 0.011/1.299 0.012/1.361 0.01 3/1.388 0.012/1.334 Ramachandran plot (%) Allowed 89.4 88.9 88.9 88.9 Additionally allowed 10.1 10.6 10.6 10.6 Generously allowed 0.5 0.5 0.5 0.5 a Values in parentheses are for the highest resolution shell. b Rsym = ( I I |/ I ) x 100. c Rcryst = ( o| |Fco|) x 100. d Rfree is calculated the same as Rcryst, except with 5% of the data omitted from refinement.

PAGE 101

101 Table 52. Apparent values of pKa and maximal rate constants for kinetic measurements of c atalysis by wild type and sitespecific mutants of HCA II at residue 170 obtained by 18O exchange at 25 oC.a Enzyme p K a (His64)b k B ( s 1 ) b p K a (ZnH2O) k cat / K M ( M1 s1) Wild type 7.2 0.1 0.8 0.1 6.9 100 K170A 6.7 0.1 1.2 0.1 7.0 110 K170D 6.7 0.1 1.5 0.2 7.0 120 K170E 6.3 0.1 4.0 0.4 7.1 150 K170H 6.7 0.1 1.6 0.1 6.9 120 a Solutions contained 25 mM of all species of CO2 at sufficient sodium sulfate to maintain ionic strength at 0.2 M, without any added buffers. b These values were obtained by a least squares fit of the data of Figure 53 to eq 54 in which the value of pKa (ZnH2O) was fixed at the value determined from R1 shown in the fourth column of this Table. Table 5 3. Values of the p Ka for the zinc bound water and the maximal value of kcat/ KM for the catalysis of the hydration of 4nitrophenylacetate by variants of HCA II at 25 oC. Enzyme p K a (ZnH 2 O) k cat / K M (M 1 s 1 ) Wild type 7.0 2800 K170A 7.0 1600 K170D 7.1 1850 K170E 6.9 1890 K170H 6.9 1760 Figure 51. The protein environment around the proton shuttle residue His64 includes residues Trp5 and Lys170. The out conformation of His64 is in a near pi stacking interaction with Trp5. On the opposite side, His64 is neighbored by the solvent mediated proton w ire (red spheres), which connects it to the zinc (purple sphere).

PAGE 102

102 Figure 52. A surface representation of HCA II shows the location of Lys170 relative to the active site core. Shown in green is the base of the active site pocket, located at the center of the enzyme. Adjacent to this is the proton shuttle residue His64, shown in orange. Directly adjacent to His64 is a basic patch, composed of Lys168, Lys170 and Lys172 (blue). Of these three, Lys170 lies the closest to His64 (4.4 ), as indicated by th e red arrow. Figure 53. The pH dependence of kcat/ KM is apparently unchanged in the four Lys170 mutants. The data shown are measured for the dehydration reaction using the 18O exchange method. All measurements were taken at 25 C with a total CO2 con centration of 25 mM and an ionic strength of 200 mM and no buffer.

PAGE 103

103 Figure 54. Mutation of Lys170 results in a change in the rate and pH dependence of the proton transfer controlled value of RH2O/[E]. Measurements were made using the same conditions described in Figure 53 A B C D Figure 55. The electron density corresponding to His3 is visible in all four Lys170 HCA II mutants, although to varying degrees. The density shown for the four mutants [(A) K170A, (B) K170D, (C) K170E, (D) K170H] is a 2FOFC Fourier map contoured to 1.5 The density is also shown for residues His4, Trp5 and His64. The density around Trp5 serves as a comparison for the four panels.

PAGE 104

104 Figure 56. The location of His3 in all four mutants is roughly the same, apparently pi stacking with the side chain of Trp5. In this location, the side chain of His3 directly interferes with His64 in its out conformation, with less than 3 between His3 and the in conformation of His64 (magenta dashes). Figure 57. A novel interaction is observed between residues 64 and 170 in the structure of K170D HCA II. This is an indirect interaction mediated by the carbonyl oxygen of Asn62. The distances of the interactions Asp170Asn62 and Asn62His64 are approximately 3.4 and 3.6 respectively (magenta dashes). As a reference, the proton wire solvent molecules are shown (red sphere.

PAGE 105

105 CHAPTER 6 CA FROM METHANOSARCINA THERMOPHILA : A COMPARISON WITH THE CA, HUMAN CARBONIC ANHYDRASE II As was discussed in Chapter 1, despite the catalysis of the same reaction, there is virtually no structural homology among the classes of CA, an indication of the physiological importance of this enzyme in all forms of life. The proposed mechanism for all classes involves a metal bound hydroxide nucleophil ically attacking the carbon of CO2, resulting in bicarbonate. Chapters 4 and 5 discussed catalysis from a structural point of view for the CA, HCA II. Myriad other studies have been done on the CAs, with relatively few studies dealing with any of the other classes. Not surprisingly, a large majority of the crystal structures available are of the CAs, with only a handful of CA struc tures and only two unique structures of CAs (Rowlett, 2009; Ferry, 2009). This chapter will look at the active site of the archetypal CA from Methanosarcina thermophila, particularly as compared to the active site organization of HCA II, in order to h elp elucidate the structural basis for catalysis. Introduction The reaction catalyzed by the carbonic anhydrases is vital for the survival of nearly every known organism, as evidenced by the occurrence of at least one CA in nearly every organism, with the exception of Mycoplasma genitalium (Fraser et al. 1995). Despite the wide occurrence of CAs and the population size of the Archaea domain of life, only two unique CAs have been characterized from organisms in this domain. The first was a CA from Methanobacterium thermoautotrophicum (Smith & Ferry, 1999; Strop et al. 2001). The other was the CA from M. thermophila (Cam), the first structurally characterized enzyme from this class (Kisker et al. 1996).

PAGE 106

106 Sequencebased database searches have revealed the existence of CA like genes in plants, green algae, proteobacteria, archaea, and cyanobacteria (Parisi et al. 2004). Alignment of these sequences suggests that there are two distinct classes of CA, based on the presence (Cam) or abs ence (CamH) ( Pyrococcus horikoshii CA, for example) of an acidic loop that lies directly adjacent to the active site (Figure 6 1). The metal coordinating histidines appear to occur in nearly all of these sequences, indicating metal binding competence. Interestingly, previous biochemical studies suggest that this acidic loop and another residue, Cam Glu62 (not present in the CamH class), are important for the presence of CA activity (Tripp & Ferry, 2000). The lack of these two features in CamH may indic ate a different function, although two CamH enzymes have been shown to have some carbonic anhydrase activity (Ferry, 2009). Cam is a trimeric enzyme with three active sites, each situated at the interface between two monomers (Figure 62). Unless purpos ely replaced with cobalt, the active site contains zinc coordinated by three histidines and a solvent molecule. Two of the histidines are from one monomer while one is contributed by the neighboring monomer. The physiological relevance of zinc is debatab le, however, as the preparation of Cam anaerobically in either E. coli or M. acetivorans produces an iron enzyme that is more active than ZnCam (MacAuley et al. 2009). The aquatic environment of these archaea suggests that iron is the relevant metal, as zinc is limiting in these environments. In any case, the active site is flanked by two surface loops, one of which is the previously mentioned acidic loop (Figure 6 3). This chapter presents structural data on mutations at two sites in Cam, Trp19 and Ty r200. A comparison of Cam with HCA II shows that these two residues may have

PAGE 107

107 catalytic implications that were previously deduced for similar residues in HCA II. Trp19 in Cam is positioned adjacen t to the proton shuttle residue Glu84 (Tu et al ., 2002). S imilarly, in HCA II, Trp5 apparently stacks with the proton shuttle residue His64. The importance of this tryptophan is highlight by the interaction of the proton rescue agent 4methylimidazole (4MI) with HCA II in the proton transfer deficient mutant H64A (Duda et al. 2003). 4MI sits in a location similar to the His64 out conformation, stacking with the side chain of Trp5. Another HCA II residue that has been shown to effect catalysis is Tyr7. A mutation to phenylalanine resulted in a mutant tha t had a greatly increased rate of proton transfer, as discussed in the previous chapter (Fisher et al. 2007b). Tyr200 in C am appears to be positioned in an equivalent location and is therefore hypothesized to play a similar role. The structures of four mutants, W19A, W19F, W19N, and Y200A are presented. Materials and Methods Enzymes Site directed mutagenesis of the Cam genecontaining plasmid was performed to create mutations at residues 19 and 200. The proteins were over expressed in E. coli and the c ells were harvested by centrifugation. The pellet was suspended in 50 mM potassium phosphate, pH 6.8, supplemented with zinc sulfate and passed twice through a French pressure cell at 20000 lb/in2. DNase I and RNase A were added after the first pass to r educe lysate viscosity. The lysate was clarified by centrifugation at 20000 x g and the protein was purified as follows by Sabrina Zimmerman in James G. Ferrys laboratory at Penn St. Univeristy. First, the supernatant was applied to a Q Sepharose column (GE Healthcare) and the protein was eluted by a linear gradient of 0 1 M sodium chloride with elution occurring at about 0.5 M NaCl. Further purification was

PAGE 108

108 accomplished by adding ammonium sulfate to the fractions (1.5 M final) and running the sample on a phenyl Sepharose column (GE Healthcare) equilibrated with 100 mM potassium phosphate, pH 7 and 1.5 M ammonium sulfate. The protein was eluted with a 1.5 0 M ammonium sulfate gradient with a peak around 0.75 M. Protein purity was assessed by SDS PAGE an alysis and the protein concentration was determined by Bradford assay. Crystallization and Data Collection All mutants were crystallized using the hanging drop vapor diffusion technique with a reservoir solution of 5% PEG 8000 and 250 mM ammonium sulfate. After sitting overnight, a heavy amount of precipitation was observed. Despite this, crystals were observed after approximately 4 weeks, with a size of ~1 mm3. Two of the mutants, Y200S and Y200F crystallized, however there was no observed diffraction, despite multiple attempts. Diffraction data were collected using the inhouse set up described in Chapter 2. The crystal detector distance was set at 100 mm with an oscillation angle of 1 per image. A total of 90 of data were collected for each crystal The data were then indexed, integrated, and scaled using HKL2000 (Otwinowski & Minor, 1997) All four mutants were solved to comparable resolution (1.61.8 ) in the space group P 213 with isomorphous unit cells ( a ~83.5). The Rsym for each mutant wer e: 4.6 for W19A, 6.4 for W19F, 3.4 for W19N, and 5.7 for Y200A. The high symmetry of the unit cell allowed for nearly 100% completeness in all cases with high redundancy, though only about 50 images were used for each structure. Complete data processing statistics are presented in Table 61.

PAGE 109

109 Refinement All structures were refined using the program REFMAC5 in the CCP4 suite of programs (Collaborative Computational Project, Number 4, 1994; Murshudov et al. 1997) Structures were refined until the Rfactor reached convergence, with a final Rfactor/Rfree of approximately 17/20% for all mutants. One residue, Met65, refined to a position with unfavorable and angles. The electron density was clear for this residue, so it was left in this conformation. Final refinement statistics are given in Table 6 1. Results Overall, the structures of the four mutants were identical to that of wildtype Cam (PDB ID 1qrg, Iverson et al. 2000). The largest difference between the mutant and wild type structures was the conformation of a loop composed of resides Glu62, Gly63, and Met64 (6263 64 loop) (Figure 6 4). This conformation was observed in all four mutants an d does not occur in any of the other Cam structures currently available in the PDB. To ensure a lack of model bias in the electron density maps, the loop was removed prior to refinement. The FOFC density map clearly shows the position of the loop, indic ating that this is the correct conformation, despite the unfavorable nature of 5). There was some evidence that this loop occupied both conformations in the W19F structure, so both conformations were built in this str ucture. There was a large electron density peak observed adjacent to the zinc ion, indicating that something electron rich was coordinated to the metal. Two previous structures of Zn and CoCam have been solved with a sulfate molecule in this exact posit ion. Therefore, this density was modeled as a sulfate in all four structures, though the occupancy of the sulfate was adjusted so that there were no large, negative FOFC

PAGE 110

110 density peaks at the sulfate. Additionally a metal was found at the 3 fold symmetry axis in Y200A Cam. This metal was modeled as an iron, which left no residual FOFC electron density (negative or positive). The iron is coordinated by three methionines, Met55, one from all three monomers. It should be noted that the analysis of this density is convoluted due its overlap with the 3fold crystallographic symmetry axis. Discussion The structures of four mutants of Cam have been solved using x ray crystallography and have revealed a previously unseen structural feature. The loop composed of residues 6264 occupies a distinctly different conformation then that reported in previous Cam structures. In this conformation the side chain of Met64 is shifted by nearly 11 as measured at the S atoms In wild type Cam the Met64 side chain buries itself into the core of the protein. The observed rotation of the loop seen in the mutant Cam structures results in the side chain of Met64 pointing out towards the surface of the protein, almost completely solvent exposed. No role for this has been pos tulated and kinetic profiles will need to be completed before a complete understanding is reached. One effect of this conformational change is that the proton shuttle residue Glu84 is observed almost solely in one conformation. In the previously reported Cam structures this residue occupies at least two distinct positions, analogous to the in and out conformations of His64 in HCA II. The side chain of Glu84 points towards the zinc in the structure of ZnCam with a water bound at the zinc (PDB ID 1qrg ) (Iverson et al. 2000). In the structures of ZnCam bound with either bicarbonate or sulfate, the side chain of Glu84 points away from the zinc ion. In its inward orientation, Glu84 appears to hydrogen bond to a water molecule that is within hydrogen bonding distance of the zinc -

PAGE 111

111 bound solvent or compound. This indicates that there may be a pH effect on the orientation of Glu84, as has been observed many times for His64 (for example, Nair & Christianson, 1991). Preliminary kinetic analysis suggests that the Trp19 mutants are catalytically deficient, while the Tyr200 mutants are more catalytically efficient (J. G. Ferry, personal communication). However, a complete kinetic work up will be required before any conclusions are drawn about the mutational ef fects on catalysis.

PAGE 112

112 Table 61. Data processing and refinement statistics for the structures of the Cam mutants. Parameter W19A W19F W19N Y200A Space group P 2 1 3 P 2 1 3 P 2 1 3 P 2 1 3 Cell dimensions ( a ) ( ) 83.492 83.621 83.580 83.498 Resolution ( ) 20 1 .6 (1.66 1.6) a 20 1.8 (1.86 1.8) 50 1.65 (1.71 1.65) 25 1.8 (1.86 1.8) R sym b (%) 4.6 (46.0) 6.4 (48.4) 3.4 (24.9) 5.7 (34.9) I/( )I 34.2 (3.6) 21.8 (2.8) 47.0 (4.7) 26.5 (4.3) Completeness 99.8 (100.0) 99.9 (99.9) 99.6 (99.1) 100.0 (100.0) Redundancy 6.8 (6.5) 5.6 (5.4) 7.2 (3.5) 6.5 (6.2) R factor c / R free d (%) 17.5/20.8 16.5 (20.3) 16.9 (20.6) 16.0 (19.4) No. atoms Pr otein 1612 1587 1603 1573 Water 71 62 80 59 B factors Protein (main/side) 22.1/25.0 23.0/26.3 21.1/24.4 18.0/21.1 Water 36.4 37.1 35.4 30.0 RMSD (bond/angle) 0.013/1.456 0.015/1.597 0.013/1.459 0.015/1.491 Ramachandran plot (%) Allowed 88.4 90.7 90.2 90.7 Additionally allowed 11.0 8.7 9.2 8.7 Disallowed e 0.6 0.6 0.6 0.6 a Values in parentheses are for the highest resolution shell. b Rsym = ( I I |/ I ) x 100. c Rcryst = ( o| |Fco|) x 100. d Rfree is calculated the same as Rcryst, except with 5% of the data omitted from refinement e The orientation of Met64 is disallowed, however the electron density for this residue clearly shows its position.

PAGE 113

113 Figure 61. An alignment of Cam (from M. thermophila) and CamH (from M. acetivorans ) shows the lack of the acidic loop (black box) in CamH. Residues are numbered according to each sequence, with green indicating identical resi dues, cyan showing conservative mutation, and red showing insertions/deletions.

PAGE 114

114 Figure 62. The trimeric organization of Cam yields three active sites per enzyme, each at a monomer monomer interface. The active sites are indicated by the zinc ion at th eir center (purple spheres). The 3fold axis of symmetry is readily apparent in this view (black triangle).

PAGE 115

115 Figure 63. The structure of a Cam monomer shows a left handed helix topology with several N terminal (blue) surface loops and a C terminal (r ed) helix. The acidic loop that differentiates the Cam and CamH classes is identified in green and the loop 636465 loop is inidicated by a red arrowhead. Figure 64. The 6364 65 loop in the mutants of Cam occupies a different conformation then tha t observed in all reported Cam structures. The orientation of the side chain of Met64 is clearly different in the two structures (magenta = wildtype, green = mutants) with a maximal distance of 11 between identical atoms. The loop of W19N is shown here, however the conformation of the loop is the same for all mutants, with W19F displaying both conformations. Side chains for the W19 mutants are also shown here.

PAGE 116

116 Figure 65. The electron density for the 6263 64 loop clearly defines its orientation in the structure of W19N Cam. Shown as blue mesh is FOFC electron density calculated with the loop removed. The side chains are built into this density (green sticks) and are in a novel conformation as compared to wildtype (magenta backbone) (PDB ID 1qrg, Iverson et al. 2000). Also shown is Pro65, a residue which remains more or less unchanged between the mutant and wildtype structures.

PAGE 117

117 CHAPTER 7 CONCLUSIONS AND FUTURE DIRECTIONS Summary and Conclusions X ray crystallographic studies provide insight i nto an enzymes structural basis of catalysis. In order to truly understand catalysis, however, one must acquire knowledge of the structural changes that underlie kinetic defects in mutant and inhibited enzymes. The trapping of substrates, products, and transition states in the enzymes active site will provide additional knowledge. This analysis can be complicated by many factors including the ability to crystallize enzymes, the resolution of the structure, as well as the lifetime of the enzymesubstra te/product/transition state complex. Chapters 2 and 3 presented mutational analyses of human manganese superoxide dismutase in order to understand the proteins role in catalysis. It was shown that secondary metal coordinating ligands play a major role in finely tuning the catalytic properties of the active site metal. This was illustrated by the E162D and E162A mutants of human MnSOD, which had a fraction of the activity of the wildtype enzyme. Crystal structures revealed that this was due to a loss or weakening of the interaction between residue 162 and His163 of a neighboring monomer. His163 acts as a manganese ligand, and mutation at residue 162 is hypothesized to affect the tuning of the redox properties of the manganese. These effects were not as drastic as those observed in the equivalent mutant of E. coli MnSOD, E170A, which lost metal specificity and, therefore, activity. This difference highlights the importance of the entire protein in modulating the active site environment, as human and E. coli MnSODs are nearly identical at all atomic positions within 10 of the manganese.

PAGE 118

118 Chapter 3 discussed mutational analysis at residue Phe66, located at the mouth of the active site, away from the Mn. The results showed that even the slightest change in the enzymes structure could lead to drastic catalytic deficiencies. Both mutants, F66A and F66L, exhibited slightly deficient catalytic activity as compared to wild type. Analysis was extended to the role of the active site cavity was described with respect to the product inhibited state of MnSOD. This site represents one of the only differences between E. coli and human MnSOD, with the human form being the more strongly product inhibited of the two. It was shown that F66L human MnSOD has a very low level of product inhibition, with a level nearly identical to that of E. coli MnSOD. Crystallographic analysis suggested that the mobility of the solvent in the active site pocket correlated to the level of product inhibition, with more mobility resulting in a lower level of product inhibition. These results suggest that the unique features of this product inhibited state may be controlled by the flexibility of these active site amino acids. In Chapter 4 the capturing of the substrate CO2 was shown to be a rather static process in HCA II. The fast turnover of this complex was previously thought to be a severe limitation to ever achieving this complex. However, the use of high pressures and low temperatures allowed for a stoppage of catalytic activity. T his demonstrated that the hydrophobic portion of the HCA II active site is responsible for providing a solvating environment for the hydrophobic CO2 molecule. Additionally, it was shown that the zinc has virtually no role in substrate binding. All of the presented results were in agreement with previous biochemical studies, therefore supporting the role of the observed binding site as the physiologically relevant site.

PAGE 119

119 The role of the protein environment in catalysis with regards to HCA II was discussed i n Chapter 5, as elucidated by mutational analysis of Lys170. It was shown that despite being a long distance from the active site, Lys170 plays a role in finetuning the environment of the proton shuttle residue His64. Kinetic analysis indicated that the p Ka of His64 was decreased in all four mutants examined. The result is an increase in the rate of proton transfer in the dehydration direction. These variants dont occur in nature, however, as the lowering of the pKa of the proton shuttle residue would decrease the catalytic efficiency when this residue acts as a proton acceptor, in the hydration direction. This suggests that when one is examining the structure of an enzyme that one must look beyond the obvious active site residues and consider how nei ghboring amino acids may effect the local electrostatic environment. An extension of this analysis to the CA from M. thermophila was discussed in Chapter 6. Despite a lack of overall structural homology, there may be localized similarities between the and CAs. Mutational analysis of residues Trp19 and Tyr200 in Cam illustrated the structural role of t hese residues and provided insight into a unique conformation of a loop lying near the active site metal. Preliminary kinetic analysis suggests that these mutations also cause changes in the catalytic efficiency of the enzyme. Taken together, the results presented here suggest that there is an underappreciation of the role of the total protein in tuning the catalytic efficiency of an enzyme. It is acceptable to begin analysis of the structurebased catalytic mechanism of an enzyme. However, one must extend analysis beyond the defined active site in order to truly understand how an enzyme functions. The body of the protein acts in a

PAGE 120

120 way beyond simply precise placement of catalytic residues and cofactors, even in apparently static catalytic mechanisms. Fut ure Directions The antiproliferative effects of weakly product inhibited forms of human MnSOD provide and means of countering the growth of cancerous cells. The use of nonhuman SODs for this purpose is complicated by the possible immune response to these enzymes. Therefore further study of the product inhibited state is required to generate efficient forms of the enzyme for this purpose. Additional mutational analyses at Phe66 offer a starting point. Combining single mutants that exhibit lower product inhibition may also yield variants with therapeutic potential. One such double mutant would be H30N/F66L, where both single mutants offer decreased product inhibition for different reasons. The double mutant may provide a weakly product inhibited enzyme with otherwise wildtype activity. It will also be necessary to directly visualize the product inhibited state of the enzyme. Crystallographic analysis of this state will require high resolution as the substrate will only appear as a dioxygen species (the protons are nearly invisible to all but the highest resolution crystal structures). One method may be to initially reduce the Mn with hydrogen peroxide, as it is known that product inhibition occurs during the second stage of catalysis in which the Mn ex ists in the +2 state. Subsequent addition of superoxide (i.e. potassium superoxide) may allow for the trapping of the substrate if cryopreservation occurs rapidly after its addition. The elucidation of the substrate binding site in HCA II, as presented in Chapter 4, provided a major leap forward in understanding the catalytic mechanism during the first stage of catalysis. The extension of this technique to and class enzymes, as well as

PAGE 121

121 other CO2binding proteins, will facilitate the elucidation of how substrate binds in these enzymes, and, therefore, how the catalytic cycle proceeds. Additionally for HCA II, the events that occur during the proton transf er step are not completely understood. The importance of this stage of catalysis in the CAs is underscored by the presumed presence of hydrogenbonded networks across all classes of CA. Little knowledge is available on the protonation states of the atoms involved in this step. One solution is to use neutron diffraction to visualize the locations of protons. Neutrons scatter off of atomic nuclei, and the replacement of protons with deuterons provides signal levels on the same order as carbons, nitrogens, and oxygens. There has been great progress on the neutron crystal structure of HCA II (Fisher et al. 2009), though a complete analysis has not yet been finished. The initial structure will only scratch the surface, though, and neutron structures at var ious pHs and in the presence of transition states/products (bicarbonate in T200H, for example) will provide atomic level detail of the catalytic process. The research presented in the Chapter 6 will also need to be completed in order to help determine the structural basis for catalysis in Cam. Kinetic studies will be completed to determine the effects of the mutations on catalysis. Additionally a complete analysis will only be possible if the crystal structures of the Tyr200 mutants are solved. Also, the crystal of ironbound Cam will be required to visualize the positions of active site atoms and to gain an understanding of what is bound to the metal in more physiologically relevant conditions.

PAGE 122

122 LIST OF REFERENCES Abreu, I. A., Hearn, A., An, H., Nick H. S., Silverman, D. N. & Cabelli, D. E. (2007). Biochemistry 47, 2350 2356. Abreu, I. A., Rodriguez, J. A. & Cabelli, D. E. (2005). J. Phys. Chem. B 109 2450224509. Alber, B. E., Colangelo, C. M., Dong, J., Stalhandske, C. M., Baird, T. T., Tu, C., Fierke, C. A., Silverman, D. N., Scott, R. A. & Ferry, J. G. (1999). Biochemistry 38, 1311913128. Alexander, R. S., Nair, S. K. & Christianson, D. W. (1991) Biochemistry 30, 1106411072. Barondeau, D. P., Kassmann, C. J., Bruns, C. K., Tainer, J. A. & Getzoff, E. D. (2004). Biochemistry 43, 8038 8047. Becker, H. M. & Deitmer, J. W. (2007). J. Biol. Chem. 282, 13508 13521. Bertini, I., Luchinat, C., Monnanni, R., Roelens, S. & Moratal, J. M. (1987) J. Am. Chem. Soc. 109 7855 7856. Borgstahl, G. E., Pa rge, H. E., Hickey, M. J., Beyer, W. F., Jr., Hallewell, R. A. & Tainer, J. A. (1992). Cell 71 107 118. Borgstahl, G. E., Pokross, M., Chehab, R., Sekher, A. & Snell, E. H., (2000). J. Mol. Biol. 296 951959. Borthwick, K. J., Kandemir, N., Topaloglu, R ., Kornak, U., Bakkaloglu, A., Yordam, N., Ozen, S., Mocan, H., Shah, G. N., Sly, W. S. & Karet, F. E. (2003). J. Med. Genet. 40 115121. Bowler, C., Van Kaer, L., Van Camp, W., Montagu, M. V., Inz, D. & Dhaese, P. (1990). J. Bacteriol. 172 1539 1546. B reton, S. (2001). J. Panc. 2 159 164. Brunger, A. T. (1992) Nature, 355 472 475 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. & Warren, G. L. (1998). Acta Crystallogr. D54 905921. Bull, C., Niederhoffer, E. C., Yoshida, T. & Fee, J. A. (1991). J. Am. Chem. Soc. 113 40694076. Butler, J. N. (1982). Carbon Dioxide Equilibria and Their Applications Reading, MA: A ddisonWesley.

PAGE 123

123 Cabelli, D. E., Guan, Y., Leveque, V., Hearn, A. S., Tainer, J. A., Nick, H. S. & Silverman, D. N. (1999). Biochemistry 38, 11686 11692. Carlioz, A. & Touati, D. (1986). EMBO J. 5 623 630. Chen, H. Y., Hu, R. G., Wang, B. Z., Chen, W. F., Liu, W. Y., Schrder, W., Frank, P. & Ulbrich, N. (2002). Arch. Biochem. Biophys. 404 218226. Chin D. & Means A. R. (2000). Trends Cell Biol. 19 322 328. Coleman, J. E. (1992). Annu. Rev. Biochem. 61, 897946. Collaborative Computational Project, Number 4 (1994). Acta Cryst. D 50, 760 763. Corena, M. P., Seron, T. J., Lehman, H. K., Ochrietor, J. D., Kohn, A., Tu, C. & Linser, P. J. (2002). J. Exp. Biol. 205 591602. Davies, M. J. (2005). Biochim. Biophys. Acta, 170, 93 109. 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. & Nick, H. S. (2004). J. Biol. Chem. 279 12769 12776. Domsic, J. F., Avvaru, B. S., Kim, C. U., Gruner, S. M., Agbandje McKenna, M., Silver man, D. N. & McKenna, R. (2008). J. Biol. Chem. 283 30766 30771. Domsic, J. F. & McKenna, R. (2009). Biochim. Biophys. Acta, In the press. Duda, D., Govindasamy, L., AgbandjeMcKenna, M., Tu, C, Silverman, D. N., McKenna, R. (2003). Acta Cryst. D 59, 93 10 4. Duda, D. & McKenna, R. (2004). Handbook of Metalloproteins edited by A. Messerschmidt, W. Bode & M. Cygler, pp. 249263. West Sussex, England: John Wiley & Sons, Ltd. Duda, D., Tu, C. K., Qian, M. Z., Laipis, P., AgbandjeMcKenna, M., Silverman, D. N. & McKenna, R. (2001). Biochemistry 40, 17411748. Edwards, R. A., Baker, H. M., Whittaker, M. M., Whittaker, J. W., Jameson, G. B. & Baker, E. N. (1998). J. Biol. Inorg. Chem. 3 161 171. Edwards, R. A., Whittaker, M. M., Whittaker, J. W., Baker, E. N. & Jameson, G. B. (2001). Biochemistry 40, 15 27. Elleuche, S. & Pggeler, S. (2009). Curr. Genet. 55 211222. Emsley, P. & Cowtan, K. (2004). Acta Crystallogr. D60 21262132.

PAGE 124

124 Farin, F. M., Hitosis, Y., Hallagan, S. E., Kushleika, J., Woods, J. S., Jansse n, P. S., Smith Weller, T., Franklin, G. M., Swanson, P. D., Checkoway, H. (2001). Mov. Disord. 16 707 707. Farr, S. B., DAri, R. & Touati, D. (1986). Proc. Natl. Acad. Sci. USA 83, 82688272. Ferry, J. G. (2009). Biochim. Biophys. Acta In the press. Fierke, C. A., Calderone, T. L. & Krebs, J. F. (1991) Biochemistry 30, 11054 11063 Fink, R. C. & Scandalios, J. G. (2002). Arch. Biochem. Biophys. 399, 19 36. Fisher, S. Z., Kovalevsky, A. Y., Domsic, J. F., Mustyakimov, M., Silverman, D. N., McKenna, R & Langan, P. (2009). Acta Cryst. F 65 495 498. Fisher, S. Z., Maupin, C. M., BudayovaSpano, M., Govindasamy, L., Tu, C. K., AgbandjeMcKenna, M., Silverman, D. N., Voth, G. A. & McKenna, R. (2007a ) Biochemistry 46, 2930 2937. Fisher, S. Z., Prada, J. A. H., Tu, C., Duda, D., Yoshioka, C., An, H. Q., Govindasamy, L., Silverman, D. N. & McKenna, R. (2005). Biochemistry 44, 10971105. Fisher, S. Z., Tu, C., Bhatt, D., Govindasamy, L., AgbandjeMcKenna, M, McKenna, R. & Silverman, D. N. (2007b). Biochemis try 46, 38033813. Fraser, C. M., Gocayne, J. D., White, O., Adams, M. D., Clayton, R. A., Fleischmann, R. D., Bult, C. J., Kerlavage, A. R., Sutton, G. G. & Kelley, J. M. et al. (1995). Science 270 397403. Gould, S. M. & Tawfis, D. S. (2005). Biochemi sty, 44 54445452 Greenleaf, W. B., Perry, J. J., Hearn, A. S., Cabelli, D. E., Lepock, J. R., Stroupe, M. E., Tainer, J. A., Nick, H. S. & Silverman, D. N. (2004). Biochemistry 43, 70387045. Guan, Y., Hickey, M. J., Borgstahl, G. E., Hallewell, R. A., Lepock, J. R., OConnor, D., Hsieh, Y., Nick, H. S., Silverman, D. N. & Tainer, J. A. (1998). Biochemistry 37, 47224730. Gupta, A. S., Webb, R. P., Holaday, A. S. & Allen, R. D. (1993). Plant Physiol. 103 10671073. Hakansson, K., Carlsson, M., Svensson L. A. & Liljas, A. (1992) J. Mol. Biol. 227 11921204. Harman, D., (1956). J. Gerontol. 11 298 300. Hearn, A. S., Fan, L., Lepock, J. R., Luba, J. P., Greenleaf, W. B., Cabelli, D. E., Tainer, J. A., Nick, H. S. & Silverman, D. N. (2004). J. Biol. Ch em. 279, 58615866.

PAGE 125

125 Hearn, A. S., Stroupe, M. E., Cabelli, D. E., Lepock, J. R., Tainer, J. A., Nick, H. S. & Silverman, D. N. (2001). Biochemistry 40, 1205112058. Hearn, A. S., Stroupe, M. E., Cabelli, D. E., Ramilo, C. A., Luba, J. P., Tainer, J. A., N ick, H. S. & Silverman, D. N. (2003). Biochemistry 42, 2781 2789. Henry, R. P. (1996). Annu. Rev. Physiol. 58, 523538. Hovnik, T., Dolzan, V., Ursic, B. N., Podkrajsek, K. T. & Battelino, T. (2009). Diab. Care In the press. Hsieh, Y., Guan, Y., Tu, C., Bratt, P. J., Angerhofer, A., Lepock, J. R., Hickey, M. J., Tainer, J. A., Nick, H. S. & Silverman, D. N. (1998). Biochemistry 37, 47314739. Hsu, J. L., Hsieh, Y., Tu, C., OConnor, D., Nick, H. S. & Silverman, D. N. (1996). J. Biol. Chem. 271 17687 176 91. Hu, P. Y., Lim, E. J., Ciccolella, J., Strisciuglio, P. & Sly, W. S. (1997). Human Mut. 9 383387. Iverson, T. M., Alber, B. E., Kisker, C., Ferry, J. G. & Rees, D. C. (2000). Biochemistry 39, 92229231. Jackman, J. E., Merz, K. M., Jr. & Fierke, C. A. (1996). Biochemistry 35, 16421 16428. Jeyakanthan, J., Rangarajan, S., Mridula, P., Kanaujia, S. P., Shiro, Y., Kuramitsu, S., Yokoyama, S. & Sekar, K. (2008). Acta Cryst. D 64, 10121029. Jonasson, P., Aronsson, G., Carlsson, U. & Jonsson, B. H. (1997) Biochemistry 36, 51425148. Karlsson, J., Hitonen, T., Husic, H.D., Ramazanov, Z. & Samuelsson, G. (1995). Plant Physiol. 109 533 539. Khalifah, R. G. (1971). J. Biol. Chem. 246, 25612573. Kim, C. U., Kapfer, R. & Gruner, S. M. (2005) Acta Crystallog r. D 61, 881 890 Kim, G., Lee, T.H., Wetzel, P., Geers, C., Robinson, M. A., Myers, T. G., Owens, J. W., Wehr, N. B., Eckhaus, M. W., Gros, G., Wynshaw Boris, A. & Levine, R. L. (2004). Mol. Cell. Biol. 24, 99429947. Kisker, C., Schindelin, H., Alber, B E., Ferry, J. G. & Rees, D. C. (1996). EMBO J. 15 23232330. Krebs, J. F., Fierke, C. A., Alexander, R. S. & Christianson, D. W. (1991). Biochemistry 30, 91539160.

PAGE 126

126 Krebs, J. F., Ippolito, J. A., Christianson, D. W. & Fierke, C. A. (1993) Biochemistry 32, 44964505. Lah, M. S., Dixon, M. M., Pattridge, K. A., Stallings, W. C., Fee, J. A. & Ludwig, M. L. (1995). Biochemistry 34, 16461660. Lane, T. W., Saito, M. A., George, G. N., Pickering, I. J., Prince, R. C. & Morel, F. M. (2005). Nature, 435 42. Laskowski R. A., MacArthur, M. W., Moss, D. S. & Thornton, J. M. (1993). J. App. Cryst. 26, 283291. Lebovitz, R. M., Zhang, H., Vogel, H., Cartwright, J., Jr., Dionne, L., Lu, N., Huang, S. & Matzuk, M. M. (1996). Proc. Natl. Acad. Sci. USA 93 9782 978 7. Leveque, V. J., Stroupe, M. E., Lepock, J. R., Cabelli, D. E., Tainer, J. A., Nick, H. S. & Silverman, D. N. (2000). Biochemistry 39, 71317137. Leveque, V. J., Vance, C. K., Nick, H. S. & Silverman, D. N. (2001). Biochemistry 40, 1058610591. 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. & Epstein, C. J. (1995). Nat. Gene. 11 376381. Liang, J. Y. & Lipscomb, W. N. (1990). Proc. Natl. Acad. Sci. US A 87, 36753679. Lindahl, L., Svensson, L. A. & Liljas, A. (1993) Proteins Struct. Funct. Bioinformat. 15, 177182 Lindskog, S. (1983). Zinc Enzymes edited by T. G. Spiro, pp. 78121. West Sussex, England: John Wiley and Sons. Macauley, S. R., Zimmerma n, S. A., Apolinario, E. E., Evilia, C., Hou, Y. M., Ferry, J. G. & Sowers, K. R. (2009). Biochemistry 48, 817819. Maliekal, J., Karapetian, A., Vance, C., Yikilmaz, E., Wu, Q., Jackson, T., Brunold, T. C., Spiro, T. G. & Miller, A. F. (2002). J. Am. Che m. Soc. 124 15064 15075. McAdam, M. E., Fox, R. A., Lavelle, F. & Fielden, E. M. (1977). Biochem. J. 165 71 79. McCord, J. M. (2002). Methods Enzymol. 349 331 341. Meldrum, N. U. & Roughton, F. J. W. (1932). J. Physiol. (London) 75 15 16. Merz, K. M., Jr. (1991). J. Am. Chem. Soc. 113 406 411. Miller, A. F. (2004). Curr. Opin. Chem. Biol. 8 162168.

PAGE 127

127 Mitsuhasi, S., Mizushima, T., Yamashita, E., Yamamoto, M., Kumasaka, T., Moriyama, H., Ueki, T., Miyachi, S. & Tsukihara, T. (2000). J. Biol. Chem. 275, 55215526. Mllsten A., Marklund, S. L., Wessman, M., Svensson, M., Forsblom, C., Parkkonen, M., Brismar, K., Groop, P. H. & Dahlquist, G. (2007). Diabetes 56, 265 269. Mizuno, K., Whittaker, M. M., Bachinger, H. P. & Whittaker, J. W. (2004). J. Biol. Che m. 279, 2733927344. Muoz, I. G., Moran, J. F., Becana, M. & Montoya, G. (2005). Protein Sci. 14 387 394. Murshudov, G. N., Vagin, A. A. & Dodson, E. J. (1997). Acta. Cryst. D53 240 255. Nair, S. K. & Christianson, D. W. (1991). J. Am. Chem. Soc. 113 9 455 9458. Noor, R., Mittal, S. & Iqbal, J. (2002). Med. Sci. Monit. 8 RA210 RA215. Otwinowski, Z. & Minor, W. (1997). Methods Enzymol. 276, 307 326. Palenik, B., Brahamsha, B., Larimer, F. W., Land, M., Hauser, L., Chain, P., Lamerdin, J., Regala, W., All en, E. E., McCarren, J., Paulsen, I., Dufresne, A., Partensky, F., Webb, E. A. & Waterbury, J. (2003). Nature 424 1037 1042. Parisi, G., Perales, M., Fornasari, M., Colaneri, A., Schain, N., Casati, D., Zimmermann, S., Brennicke, A., Araya, A., Ferry, J. G., Echave, J. & Zabaleta, E. (2004). Plant Mol. Biol. 55 193207. Phillips, J. P., Campbell, S. D., Michaud, D., Charbonneau, M. & Hilliker, A. J. (1989). Proc. Natl. Acad. Sci. USA 86 2761 2765. Purkerson, J. M. & Schwartz, G. J. (2007). Kidney Int. 71, 103 115. Quint, P., Ayala, I., Busby, S. A., Chalmers, M. J., Griffin, P. R., Rocca, J., Nick, H. S. & Silverman, D. N. (2006). Biochemistry 45, 82098215. Quint, P. S., Domsic, J. F., Cabelli, D. E., McKenna, R., Silverman, D. N. (2008). Biochemistry 47, 4621 4628. Quint, P., Reutzel, R., Mikulski, R., McKenna, R. & Silverman, D. N. (2006). Free Radical Biol. Med. 40, 453458. Rabani, J. & Nielsen, S. O. (1969). J. Phys. Chem. 73, 37363744. Ramilo, C. A., Leveque, V., Guan, Y., Lepock, J. R., Tainer J. A., Nick, H. S. & Silverman, D. N. (1999). J. Biol. Chem. 274, 27711 27716. Reaume, A. G., Elliott, J. L., Hoffman, E. K., Kowall, N. W., Ferrante, R. J., Siwek, D. R., Wilcox, H. M., Flood, D. G., Beal, M. F., Brown, R. H., Jr., Scott, R. W. & Snider W. D. (1996). Nat. Gene., 13 43 47.

PAGE 128

128 Roberts, S. B., Lane, T. W. & Morel, F. M. M. (1997). J. Phycol 33 845850. Rodriguez, A. M., Carr ico, P. M., Mazurkiewicz, J. E. & Melendez, J. A. (2000), Free Radic. Biol. Med 29, 801 813 Rowlett, R. S. (2009). Biochim. Biophys. Acta In the press. Sawaya, M. R., Cannon, G. C., Heinhorst, S., Tanaka, S., Williams, E. B., Yeates, T. O. & Kerfeld, C. A. (2006). J. Biol. Chem. 281 75467555. Schwartz, H. (1981). J. Chem. Educ. 58, 101 105. Sheldrick, G. M. (2008) Acta Crystallogr. A64 112 122 Silverman, D. N. (1982). Methods Enz. 87, 732 752. Silverman, D. N. & Lindskog, S. (1988). Acc. Chem. Res. 21, 30 36. Silverman, D. N. & McKenna, R. (2007) Acc. Chem. Res. 40, 669 675 Sly, W. S. & Hu, P. Y. (1995). Annu. R ev. Biochem. 64, 375 401. Smith, M. W. & Doolittle, R. F. (1992). J. Mol. Evol. 34, 175184. Smith, K. S. & Ferry, J. G. (1999). J. Bacteriol. 181 6247 6253. Smith, R. A., Kelso, G. F., Blaikie, F. H., Proteous, C. M., Ledgerwood, E. C., Hughes, G., Jame s, A. M., Ross, M. F., Asin Cayuela, J., Cochem, H. M., Filipovska, A. & Murphy, M. P. (2003). Biochem. Soc. Trans. 31, 12951299. So, A. K., Espie, G. S., Williams, E. B., Shively, J. M., Heinhorst, S., Cannon, G. C. (2004). J. Bacteriol. 186 623 630. S t Pierre, J., Buckingham, J. A., Roebuck, S. J. & Brand, M. D. (2002). J. Biol. Chem. 277, 4478444790. Strange, R. W., Yong, C. W., Smith, W. & Hasnain, S. S. (2007). Proc. Natl. Acad. Sci. USA 104 1004010044. Strop, P., Smith, K. S., Iverson, T. M., F erry, J. G. & Rees, D. C. (2001). J. Biol. Chem 276, 1029910305. Sugio, S., Hiraoka, B. Y. & Yamakura, F. (2000). Eur. J. Biochem. 267 34873495. Sundaresan, M., Yu, Z., Ferrans, V. J., Irani, K. & Finkel, T. (1995) Science 270 296299.

PAGE 129

129 Sunderhaus, S ., Dudkina, N. V., Jansch, L., Klodmann, J., Heinemeyer, J., Perales, M., Zabaleta, E., Boekema, E. J. & Braun, H. P. (2006). J. Biol. Chem. 281 64826488. Tripp, B. C. & Ferry, J. G. (2000). Biochemistry 39, 92329240. Tu, C., Rowlett, R. S., Tripp, B. C., Ferry, J. G. & Silverman, D. N. (2002). Biochemistry 41, 1542915435 Tu, C. K., Silverman, D. N., Forsman, C., Jonsson, B. H. & Lindskog, S. (1989). Biochemistry 28, 7913 7918. Vagin, A. & Teplaykov, A. (1997). J. Appl. Crystallogr. 30 10221025. Ve nta, P. J. (2000). The Carbonic Anhydrases, New Horizons edited by W. R. Chegwidden, N. D. Carter & Y. H. Edwards, pp. 403412. Switzerland: Birkhauser Verlag Basel. Venta, P. J., Welty, R. J., Johnson, T. M., Sly, W. S. & Tashian, R. E. (1991). Am. J. Hu m. Genet. 49, 10821090. Verpoorte, J. A., Mehta, S. & Edsall, J. T. (1967). J. Biol. Chem. 242 42214229. Vince, J. W. & Reithmeier, R. A. F. (1998). J. Biol. Chem. 273 28430 28437. Whittaker, M. M. & Whittaker, J. W. (1997). Biochemistry 36, 8923893 1. Whittaker, M. M. & Whittaker, J. W. (1998). J. Biol. Chem. 273 22188 22193. Williams, T. J. & Henkens, R. W. (1985) Biochemistry 24, 24592462 Xue, Y., Liljas, A., Jonsson, B. H. & Lindskog, S. (1993) Proteins 15 80 87. Youn, H. D., Kim, E. J., R oe, J. H., Hah, Y. C. & Kang, S. O. (1996). Biochem. J. 318 889896. Youn, H. D., Youn, H., Lee, J. W., Yim, Y. I., Lee, J. K., Hah, Y. C. & Kang, S. O. (1996b). Arch. Biochem. Biophys 334 341348. Zheng, J., Domsic, J. F., Cabelli, D., McKenna, R. & Si lverman, D. N. (2007). Biochemisty 46, 14830 14837.

PAGE 130

130 BIOGRAPHICAL SKETCH John Francis Domsic was born in Erie, PA in 1983. He attended Fort LeBoeuf High School, graduating in 2001. He immediately entered the undergraduate program at Allegheny College in Meadville, PA and completed his B.S. in Biology in 2005. He spent his senior year studying the capsid protein of bacteriophage HK97 under the guidance of Dr. Brandi Baros. In the fall of 2005, John began his graduate school training in the University of Floridas Interdisciplinary Program in the Biomedical Sciences. He joined the lab of Dr. Robert McKenna in the spring of 2006 to begin his research on the structural basis of catalysis in the metalloenzymes human manganese superoxide dismutase and the ca rbonic anhydrases.