<%BANNER%>

Structural Insights into the Active Site of Alpha-Carbonic Anhydrases

HIDE
 Title Page
 Dedication
 Acknowledgement
 Table of Contents
 List of Tables
 List of Figures
 Abbreviations
 Abstract
 Introduction
 Structural and kinetic analysis...
 Structural and kinetic effect of...
 Working towards a neutron structure...
 Novel inhibitors and their binding...
 Expression, purification, kinetic,...
 Conclusions and future directi...
 References
 Biographical sketch
 

PAGE 1

STRUCTURAL INSIGHTS INTO THE ACTIVE SITE OF ALPHA-CARBONIC ANHYDRASES By SUZANNE ZO FISHER A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2006

PAGE 2

Copyright 2006 By Suzanne Zo Fisher

PAGE 3

This document is dedicated to my parents, Alet and Peter Fisher.

PAGE 4

ACKNOWLEDGMENTS I would like to thank my advisor and me ntor, Dr. Robert McKenna. Throughout my stay in his lab, he was an inspiration and a constant source of enthusiasm and encouragement. He was always available for que stions and very patient, to say the least. Without him, none of this work would have been possible and I thank him deeply and sincerely for his outstanding guidance over the y ears. Another key person to thank is Dr. David Silverman. I have worked closely with him and his lab and appr eciate the excellent advice and insightful discussions that we have had. I would also like to thank my other committee members, Dr. James Flanegan and Dr. Maurice Swanson. Together, my committee has shown nothing but support and re spect for me and my work and their supervision has kept me on tr ack and focused on my goals. I would like to thank some of the other labmembers and friends. Mavis AgbandjeMcKenna, Lakshmanan Govindasamy, Deepa Bh att, and Huyn-Joo Nam have all given me a lot of assistance, genera l advice, and technical expertis e. The graduate students John Domsic, Brittney Whitaker, and Nicolette Case have made a lot of tough times manageable and provided many a good laugh. I would also like to thank the undergraduates Edward Miller, Mike DiMattia, and Caroli Genis for great conversations and being good company. The McKenna lab r eally is a special and unique place and I will miss it a great deal. I would also like to thank Kathy Conture, Susan Gardner, and Wayne McCormack who are part of the administrative staff for the IDP and have helped me with all things iv

PAGE 5

practical over the years. Al so, Pat Jones and Terry Rickey from the Biochemistry and Molecular Biology Department have been help ful with everyday things, from registering for class to purchasing chemicals for the lab. Lastly, I would like to thank my family. Despite living far away in South Africa, they have always supported and encouraged me to stay in the United States and continue my studies. A special thank you also goes to my American fa mily, the Hunts. They have been generous and kind by welcoming me into their home and making me a part of their family. v

PAGE 6

TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................iv LIST OF TABLES .............................................................................................................ix LIST OF FIGURES .............................................................................................................x ABBREVIATIONS ..........................................................................................................xii ABSTRACT .....................................................................................................................xiv CHAPTER 1 INTRODUCTION........................................................................................................1 The Discovery of Carbonic Anhydrase........................................................................1 A Fine Example of Convergent Evolution ...................................................................2 Human Carbonic Anhydrases (HCAs) .........................................................................5 Structure of HCA II ......................................................................................................9 Activity and Catalytic Mechanism of CAs .................................................................10 Substrate Binding in the CA Active Site ....................................................................11 Measuring CA Activity ...............................................................................................13 HCA and Human Disease ...........................................................................................14 CA Inhibitors ..............................................................................................................16 2 STRUCTURAL AND KINETIC ANALYSIS OF PROTON SHUTTLING IN HUMAN CARBONIC ANHYDRASE II..................................................................28 Introduction .................................................................................................................28 Materials and Methods ...............................................................................................32 Enzymes ..............................................................................................................32 Crystallography ...................................................................................................33 Activity Analysis by 18O Exchange .....................................................................35 Results and Discussion ...............................................................................................36 Crystallography ...................................................................................................36 Effect of pH on the Wild Type HCA II Active Site ............................................37 Effect of pH on the Mutant HCA II Active Site ..................................................39 Catalysis ..............................................................................................................41 Conclusion ...........................................................................................................42 vi

PAGE 7

3 STRUCTURAL AND KINETIC EFFE CTS OF HYDROPHOBIC MUTATIONS IN THE ACTIVE SITE OF HU MAN CARBONIC AN HYDRASE II.....................51 Introduction .................................................................................................................51 Materials and Methods ...............................................................................................52 Enzymes ..............................................................................................................52 Crystallography ...................................................................................................53 Kinetics and Activity Analysis ............................................................................54 Results and Discussion ...............................................................................................55 Structural Effects of Hydrophobic Mutations .....................................................55 Kinetic Effects of Hydrophobic Mutations .........................................................58 Solvent Structure and Implic ations for Proton Transfer ......................................60 Conclusion ...........................................................................................................63 4 WORKING TOWARDS A NEUTRON STRUCTURE OF PERDEUTERATED HUMAN CARBONIC ANHYDRASE II..................................................................75 Introduction .................................................................................................................75 Materials and Methods ...............................................................................................82 Production and Crystallizati on of Perdeuterated HCA II ....................................82 Crystallography ...................................................................................................83 Activity Analysis by 18O Exchange Methods ......................................................84 Results and Discussion ...............................................................................................84 Structural Effects of Perdeuteration ....................................................................85 Kinetic Effects of Perdeuteration ........................................................................87 Conclusion ..................................................................................................................88 5 NOVEL INHIBITORS AND THEIR BINDING MODES TO HCA II....................99 Introduction .................................................................................................................99 Materials and Methods .............................................................................................101 Expression, Purification, and Crystallization ....................................................101 Synchrotron X-ray Data Collection ...................................................................101 Structure Determination and Model Refinement ..............................................102 Results and Discussion .............................................................................................103 Conclusion ................................................................................................................105 6 EXPRESSION, PURIFICATION, KINETIC, AND STRUCTURAL CHARACTERIZATION OF AN ALPHA-CLASS CA FROM AEDES AEGYPTI (AACA1)................................................................................................114 Introduction ...............................................................................................................114 Materials and Methods .............................................................................................116 Expression and Purification ...............................................................................116 Activity Analysis by 18O Exchange ...................................................................117 Determination of Inhibition Constants ..............................................................118 Model Building ..................................................................................................118 vii

PAGE 8

Results and Discussion .............................................................................................119 Conclusion ................................................................................................................122 7 CONCLUSIONS AND FUTURE DIRECTIONS...................................................130 Summary and Conclusions .......................................................................................130 Future Directions ......................................................................................................134 LIST OF REFERENCES .................................................................................................136 BIOGRAPHICAL SKETCH ...........................................................................................149 viii

PAGE 9

LIST OF TABLES Table page 1-1 Catalytic constants and subcellular locations of active -CA isozymes ..................19 2-1 Data set and model statistics for wild type HCA II from pH 5.1 to 10.0. ................44 2-2 Data set and model statistics for H64A/N62H and H64A/N67H HCA II at pH 6.0 and 7.8. ...............................................................................................................45 2-3 pH-Independent rate constants for proton transfer and pKa for proton donor and acceptors in wild type and mutant HCA II. ..............................................................46 3-1 Data set and final model statisti cs for N62L HCA II and N67L HCA II. ................65 3-2 Data set and final model statistics for Y7F HCA II. ................................................66 3-3 Maximal values of rate constants for hydration of CO2, proton transfer, and pKa of the zinc-bound water. ...........................................................................................67 4-1 Scattering amplitudes and cross secti ons of atoms by X-rays and neutrons. ...........90 4-2 Data collection and model refinement statistics. ......................................................91 4-3 Distances between solvent molecules and active site residues for hydrogenated and perdeuterated HCA II. .......................................................................................92 5-1 Data collection, refinement and final model statistics. ..........................................107 6-1 Maximal values of kcat/Km for the hydration of CO2 and kB for the proton transfer dependent release of H2 18O from isoforms of carbonic anhydrase. ..........124 ix

PAGE 10

LIST OF FIGURES Figure page 1-1 Ribbon diagram of CA from three classes ...............................................................20 1-2 Multiple sequence alignment of fourteen HCA domains......................................... 21 1-3 Conservation of HCAs. ............................................................................................22 1-4 Active sites of HCA I, II, and III.............................................................................. 23 1-5 Molecular and electrostatic surface potentials of HCAs. .........................................24 1-6 Ribbon diagram of HCA II with secondary structure elements ...............................25 1-7 Active site of wild type HCA II ...............................................................................26 1-8 HCA II in complex with inhibitors.......................................................................... 27 2-1 Crystal structures of wild type HCA II active site ...................................................47 2-2 Crystal structures of mutant HCA II active site .......................................................48 2-3 The pH profiles for kcat/KM catalyzed by wild type and mutant HCA II. ................49 2-4 The pH profiles for RH2O/[E] catalyzed by wild type and mutant HCA II .............50 3-1 Active site of wild type human carbonic anhydrase II .............................................68 3-2 Active site of N62L a nd N67L at pH 8.2 and pH 6.0 ..............................................69 3-3 Active site of Y7 F HCA II at various pH .................................................................70 3-4 The pH profiles for kcat/KM catalyzed by wild type and mutant HCA II .................71 3-5 The pH profiles for RH2O/[E] catalyzed by wild type and mutant HCA II .............72 3-6 Activation of RH2O/[E] catalyzed by wild t ype and mutant HCA II by the addition of 4-methylimidazole .................................................................................73 3-7 Active sites of wild type and Y7F HCA II. ..............................................................74 x

PAGE 11

4-1 The LANSCE position-sensitive 3He-filled detector. ..............................................93 4-2 Optical photograph of perd euterated wild type HCA II. ..........................................94 4-3 Backbone superposition of hydrogenate d and perdeuterated wild type HCA II ......95 4-4 Active site comparison of hydrogenate d and perdeuterated wild type HCA II .......96 4-5 The pH profiles for kcat/KM catalyzed by hydrogenated and perdeuterated HCA II. ..............................................................................................................................97 4-6 The pH profiles for RH2O/[E] catalyzed by hydrogenated and perdeuterated HCA II. ..............................................................................................................................98 5-1 Classical carbonic anhydra se inhibitors bound to HCA II .....................................108 5-2 Chemical structure of novel inhibitors...................................................................109 5-3 Active site structures of HCA II in complex with two novel inhibitors ................110 5-4 HCA II active site with inhibitors superimposed ...................................................111 5-5 Interactions of BB3 and TDM with HCA II ..........................................................112 5-6 HCA II with BB3 and TDM superimposed ...........................................................113 6-1 Sequence alignment of AaCA1 with HCA I and II ................................................125 6-2 The 12% Coomassie stained polyacr ylamide gel of AaCA1 expression and purification .............................................................................................................126 6-3 The pH profile of rate co nstants for catalysis by AaCA1 ......................................127 6-4 Molecular model of AaCA1. ..................................................................................128 6-5 Active sites and surface charge dist ribution of AaCA1, HCA I, and HCA II .......129 xi

PAGE 12

ABBREVIATIONS Angstrom A. aegypti Aedes aegypti CA carbonic anhydrase CAPS N-cyclohexyl-3-aminopropanesulfonic acid CHES N-cyclohexyl-2-aminoethanesulfonic acid CO 2 carbon dioxide cm centimeter E. coli Eschericia coli g gravitational force gm gram H + proton/ hydrogen ion HCA human carbonic anhydrase HCl hydrochloric acid HCO 3 bicarbonate ion HEPES N-(2-hydroxyethyl)-piper azine-N-2-ethanesulfonic acid IPTG isopropyl-D-thiogalactopyranoside k cat turnover number k cat /K M specifity constant kD kilo Daltons K i inhibition constant xii

PAGE 13

K M Michaelis-Menten constant kV kilovolt LB luria broth M molar MES 2-(4-morpholino)-ethane sulfonic acid MOPS 3-(N-morpholino)-propanesulfonic acid g microgram l microliter M micromolar mA milliampere mg milligram ml milliliter mm millimeter mM millimolar nm nanometer nM nanomolar pAMBS para-aminomethylbenzenesulfonamide pH negative log of the hydrogen ion concentration rmsd root mean square deviation TAPS N-Tris(hydroxymet hyl)methyl-3-aminopropanesulfonic acid Tris tris(hydroxymethyl)aminomethane Zn zinc xiii

PAGE 14

Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy STRUCTURAL INSIGHTS INTO THE ACTIVE SITE OF ALPHA-CARBONIC ANHYDRASES By Suzanne Zo Fisher December 2006 Chair: Robert McKenna Major Department: Medical Sciences--Biochemi stry and Molecular Biology Carbonic anhydrases (CA) are ubiquitously expressed metalloenzymes that are found in all organisms, ranging from bacteria to humans. Human CA II (HCA II) is the most well-studied and utilizes a zinc-hydroxide mechanism to catalyze the reversible hydration of carbon dioxide to produce bicarbonate and a proton. Catalysis involves an intramolecular proton transfer event that de livers an excess prot on from the zinc-bound water to an internal proton acceptor, His64. His64 then shuttles this proton to the bulk solvent, thus regenerating the active site for the next round of catalysis. An extensive analysis of the structural and kinetic stability of w ild type and several mutants of HCA II was conducted over a broa d pH range. The results show that the enzyme, and the water network in the active site, is extremely stable. It is also the first observation of sulfate ion binding in the active site of wild type HCA II. Attempts to disrupt not only the proton shuttle His64, but other residues involved in stabilizing the water network were also successful as refl ected in changes of the measured proton transfer rates. Overall, the results give in sights into the structural requirements for efficient proton transfer as cat alyzed by CA. To directly obs erve the active site waters xiv

PAGE 15

and protonation state of His64, perdeute rated wild type HCA II was produced, crystallized and the X-ray structure determin ed. This work lays the foundation for future proposed neutron diffraction experiments. Classical, clinically used CA inhibitors (CAI) are not very water-soluble and this feature has implications for bi oavailability of these drugs. The X-ray structures of two novel, water-soluble CAIs bound to HCA II were determined. They reveal that incorporation of spacer groups and fluorines can change the binding modes of CAIs. This work has implications for the clinical use and bioavailability of systemically applied CAIs and for targeting different isozymes of HCA. A CA from mosquito larvae (AaCA1) was also expressed, purified, and structurally and kinetically characterized. AaCA1 is a high activity CA that shows inhibition with all the classical sulfonamide-based CAIs. This enzyme represents an interesting new drug target for the control of mos quito populations and further understanding of CA function in other organisms. xv

PAGE 16

CHAPTER 1 INTRODUCTION The Discovery of Carbonic Anhydrase Prior to the discovery of carbonic anhydrase (CA) there were two main theories as to the physiological mechanism of car bon dioxide transport in blood. The first (bicarbonate theory) stated that carbon dioxi de was transported as bicarbonate and upon reaching the lung, was converted by blood protei ns to carbon dioxide and expelled. The second (direct combination theory) declared that carbon dioxide is carried directly by blood proteins, primarily hemoglobin, and can reversibly dissociate from them, upon reaching the lung (Forster, 2000). From 1917 to 1921 the bicarbonate theory gained a lot of support through the efforts of many physio logists. However, early experiments by Thiel in 1913 studied the uncatalyzed rate of carbon dioxide hydration and found it slow at near 0.02 s -1 This low rate was obviously insu fficient as red blood cells (RBCs) only have about 1 second in the lung to exchange carbon dioxide. In 1926 Henriques calculated the hydration rate under physiologi cal conditions using velocity constants obtained by others and concluded that there must be a catalyst in blood that speeds up the hydration reaction. He also pred icted that there must be an other mechanism that does not purely rely on the bicarbonate theory (H enriques, 1928). Later experiments involved determination of the rate of carbon dioxide production from hemoglobin solutions. These showed that there was a dramatic increase in the rate compared to just buffer alone. These observations gave support to the direct combination theory, which led to the suggestion that carbamate binds to hemoglobin givi ng rise to a hypotheti cal complex called 1

PAGE 17

2 carbhemoglobin (Henriques, 1928). As it later turned out, the hem oglobin purification technique was not perfect and actually contai ned a contaminant, carbonic anhydrase, that was mediating the observed catalysis. It was not until 1932 that Meldrum and Roughton isolated a non-hemoglobin protein from ox blood and showed that it was catalyzing the carbon dioxide hydration reaction. They named this protein carbonic anhydrase (Meldrum and Roughton, 1932). Then in 1933 the same authors published a paper that described the details of CA pr eparation and its catalytic properties. Their rather crude initial experiments tested an impressive ar ray of CA properties in cluding temperature and pH stability. Their results indica ted that pure preparations of CA that were heated to 65 C for 30 minutes still retained 40% activity and the protein was stable and active from pH 4.0 to 12.0. In the same paper they also showed that CA was potently inhibited by cyanide and azide (Meldrum and Roughton, 1933). Since the early discoveries and observations in the 1930s the CA field has grown enormously and a lot is known about the various isoforms, physiological functions, catalytic activity, and crystal structures. A Fine Example of Convergent Evolution From humans to plants, the ubiquitous zinc metalloenzyme CA catalyzes the reversible hydration of carbon dioxide (CO 2 ) to form bicarbonate (HCO 3 ) and a proton (H + ). There are five evolutionary distinct classes of CAs: , , and The -class was discovered first and is found primarily in mammals but has also been identified in such diverse organisms as the mosquito and plant green algae, Chlamydomonas reinhardtii The -class is found in plants but ther e are also examples found in Escherichia coli ( E. coli ) and Synechococcus (Hewett-Emmett and Tashian, 1996). The -class is found mainly in ar chaebacteria and was initiall y discovered in the archeon

PAGE 18

3 Methanosarcina thermophila (Alber and Ferry, 1996). Arabidopsis thaliana and Synechococcus seem to be the exceptions by having se quences that have similarity to all three classes (Hewett-Emmett and Tash ian, 1996). The recently discovered and classes have not been extensively studied and are found in diatoms and cyanobacteria, respectively (Tripp et al. 2001; So et al. 2004). Most living organi sms have genes that encode for CAs, except for Mycoplasma genitalium that appears to lack any CAencoding gene (Fraser et al. 1995). There are numerous crystal structures of -, -, and -class CAs (Figure 1-1) that are available in the Protein Data Bank (www.pdb.org). Currently, there are well over 200 CA structures; most of these are class CAs and are, more sp ecifically, represented by 5 human CA (HCA) isozyme structures, HCA I, II, as well as extracellular domains of HCA IV, XII, and XIV (Liljas et al. 1972; Kannan et al. 1975; Stams et al. 1996; Whittington et al. 2001; Whittington et al. 2004). Crystal st ructures of the -class (Figure 1-1(a)) include various mutants of diffe rent isoforms as well as complexes of the protein with inhibitors and/or activators. The first CA crys tal structure of human isozyme II was determined in 1972 (Liljas et al. 1972). More recently, the first and -CA crystal structures have b een reported (Mitsuhashi et al. 2000; Kisker et al. 1996). Visual inspection of the crystal st ructures of the three classes ( -, -, and -classes) reveals a dramatic picture of the variation in topology of these zinc-containing enzymes (Figure 1-1 (a)-(c)) and it is no surprise th at the different class members are found in phylogenetically diverse organisms (Strop et al. 2001). The CAs act mainly as monomers except in the case of HCA XII th at was shown to be a dimer (Whittington et al. 2001). The structures of and -class CA reveal that they oligomerize to form

PAGE 19

4 pseudo-dimers and trimers, respectively (Mitsuhashi et al. 2000; Iverson et al., 2000). Despite the obvious overall struct ural differences between thes e classes, closer inspection of the active sites reveal a remarkably simila r architecture of the catalytic zinc center. The carbonic anhydrases are an excellent example of convergent evolution where distinctly varied life forms ha ve found structurally altern ative ways to construct an enzyme that has the same catalytic mechanism (Figure 1-1; Lindskog, 1997). The many physiological functions of CA ha ve been most extensively characterized in plants and mammals. In mammals CA is involved in many physiological processes such as acid/base homeostasis, renal acidifi cation, bone resorption, cellular re spiration, gluconeogenesis, formation of gastric aci d, cerebrospinal fluid and aqueous humour production, tumor metastasis, a nd the interconversion of CO 2 /HCO 3 in red blood cells (RBCs) during respiration (Henry, 1996; Breton, 2001; Chegwidden and Carter, 2000). Due to all these functions that are vita l to all life proce sses CAs are found in virtually all tissue types. In plants CA is found in both the cytosol and the chloroplast where it is primarily involved with providi ng carbon for the fixation of inorganic carbon into sugar. Carbonic anhydr ase provides either HCO 3 or CO 2 as a source of inorganic carbon for either PEP carboxylase or RuBisCO (ribulose-1,5-bisphosphate carboxylase/oxygenase), respectively (Burnell, 20 00). The functions of CA in bacteria are less well understood except for two well-docum ented cases. The first is the role of CA in E. coli where, as a part of the cyn operon, it prevents depletion of CO 2 during cyanate breakdown (Guilloton et al. 1992). The second is in cyanobacteria where CA is part of the carboxysome shell. Carboxysom es are polyhedral microcompartents that consist of several shell proteins that p ackage RuBisCO for carbon fixation. RuBisCO is

PAGE 20

5 a notoriously slow enzyme and it is thought th at CA, present in the carboxysomal shell, can convert HCO 3 into CO 2 and deliver this substrate dire ctly to RuBisCO in high local concentrations (So et al. 2004). Human Carbonic Anhydrases (HCAs) In the human -class there are fourteen identified expressed CA isoforms (HCA I XIV) and there are examples of cyto solic (HCA I, II, III, and VII), transmembrane/membrane anchored (HCA IV, IX, XII and XIV), secretory (HCA VI and XI) and mitohondrial isoforms (HCA VA and VB) (Chegwidden and Carter, 2000; Duda and McKenna, 2004). The percentage sequence id entity varies from 26 to 61% among all fourteen isoforms and they differ in catalytic efficiency and subcellular location (Table 11, Figure 1-2). Table 1-1 summarizes some of the properties for re presentative active CAs. Some of the isozymes, such as HCA II and IV, are high activity forms that work near the diffusion limit. Most of the ones liste d in Table 1-1 fall in the medium activity level range while HCA III represents the isozyme with the lowest catalytic rate constants. Figure 1-2 is a multiple sequence alignment of the CA domains of all fourteen HCAs, and the conserved and similar residue s are marked as described in the figure legend. Figure 1-3 indicates the location, on the backbone of a CA domain, of all the conserved residues found in all f ourteen HCAs. It is striking that these residues are not located directly in the active site but seem to cluster around it. This distribution implies that the variation seen in the active sites is necessary for the wide range of catalytic rates for different isozymes (Table 1-1). The conser ved residues overall appear to have a role in maintaining the distinctive CA fold. Ve ry few of these residues are found on the

PAGE 21

6 surface and the resulting surface heterogeneit y lends support for the different cellular locations and functions of the variou s HCAs (Table 1-1; Figure 1-3) HCA I is found predominantly in RBCs wh ile HCA II, also abundant in RBCs, is also found in cells of all tissue types (Tas hian, 1992). HCA I and HCA IIs presence in the RBCs are very important for converting HCO 3 into CO 2 during respiration. Figure 14 shows the active sites of cy toplasmic HCA I, HCA II, and HCA III. The active sites of these three isozymes are highly conserved a nd the differences indicated in Figure 1-4 could account for the observed differences in their respective activities (Table 1-1). HCA III is a major part of the soluble protein in adipose and muscle tissue, but its function in these tissues remains elusive. Despite is a bundance, recent studies with mouse knock-out models of HCA III showed no phenotype (Sly and Hu, 1995; Kim et al. 2004). HCA IV is the only glycosylphosphatidyli nositol(GPI)-linked isoform and, in contrast to CA IV in other species, not glycosylated. The CA domain of HCA IV is located on the extracellular face of cells in many tissue types that include kidney, lung, and the eye (Sly and Hu, 1995; Chegwidden and Carter, 2000). Of the mitochondrial CAs there are tw o HCA V variants, CA VA and CA VB. HCA VA is expressed only in the liver while HCA VB expression is everywhere except the liver. HCA V is mainly involved with providing HCO 3 to metabolic enzymes in the gluconeogenesis and ureagene sis pathways (Dodgson, 1991). HCA VI is heavily glycosylated and the only secreted isoform in humans. It is found predominantly in saliva where it is involved with pH control of the mouth (Murakami and Sly,1987). HCA VII is another cytoplasmic, soluble isoform that is expressed mainly in the brain, salivary gla nd, and lung. It is the most highly conserved

PAGE 22

7 (compared to the consensus CA domain sequenc e) of all the active CA isoforms, but its function remains unknown (Chegwidden and Carter, 2000). HCA VIII, X, and XI have no catalytic activity but contain a CA domain. These inactive isoforms are subsequently known as the CA-Related Proteins, or CA-RPs (Khalifah, 1971; Jewell et al. 1991; Kato, 1990; Skaggs et al. 1993). The lack of catalytic activity in the HCA-RPs is due to deleterious mutations of the zinc ligand histidines (Figure 1-2). CA-RP VIII is highl y conserved across species and shares about 98% sequence identity between humans and mice (Bergenhem et al. 1998). HCA-RP is widely expressed throughout the brain and its expression is devel opmentally controlled (Taniuchi et al. 2002). HCA IX was first identified in HeLa cel ls and is a glycosylated extracellular enzyme with a membrane-spanning region (Pastorekov et al. 1992). It is normally expressed only in the gastrointestinal epithelial lining but is usua lly absent in normal tissues. HCA IX displays constitutive expres sion in tumors such as clear cell renal carcinoma and has potential as a biomarker for certain tumors (Murakami et al. 1999; Ortova Gut, 2002). HCA XII is another transm embrane glycoprotein with its CA domain located on the extracellular side of cells. It is expressed mainly in colon, kidney, and prostrate, but it was discovered due to overe xpression of its mRNA in renal and lung cancer cells (Treci et al. 1998; Ulmasov et al. 2000). The crystal stru cture revealed that this membrane protein exists as a dimer (Whittington et al. 2001). HCA XIII is the most recent addition to the HCA family and is predicted to be a cytosolic isoform, similar to HCA I and II. Not much is known about its function but it has been identified in thymus spleen, and colon (Lehtonen et al., 2004).

PAGE 23

8 HCA XIV is the last of the human isofor ms and shows 45% sequence identity to HCA XII (Fujikawa-Adachi et al. 1999). It is also a single-membrane spanning glycoprotein with an extracellular catalytic domain but, unlike HCA XII, the crystal structure of murine CA XIV show s it to be monomeric (Whittington et al. 2004). In humans, it is widely expresse d but is found mainly in hear t and kidney (Chegwidden and Carter, 2000). Figure 1-5 shows the molecular surface and electrostatic charge distributions of several representative HCAs. HCA II and III are the most and least efficient of the catalytically active cytosolic HCAs, respectiv ely. HCA II appears to be more negatively charged around the active site region compar ed to HCA III. HCA VI (Figure 1-5 (c)) is the only secreted isoform and is more hydrophobic around the active site compared to the other soluble forms shown in Figure 1-5 (a) and (b). HCA-RP VIII is one of the acatalytic isoforms and its subcellular location is unknown. Compared to the other isoforms shown in Figure 1-5, HCA-RP VIII has a much higher overall negativ e charge distribution. HCA IX is a membrane protein and the surface shown is for the catalytic CA domain (Figure 1-5 (e)). It has an asymmetric distribution of negative charge and hydrophobicity at the region adjacent to the ac tive site. HCA XII is dimeric membrane protein (Figure 1-5 (f)) with the two active sites located on th e same side of the dimer. These surface features of the HCAs reflect the variation in am ino acid sequence among them. As is shown in Figure 1-3, the most c onservation of sequence is not in the active site or the surface, but in the regions that form the scaffold of the metal binding center.

PAGE 24

9 Structure of HCA II The first crystal structure of HCA II was determined over thirty years ago by Liljas and colleagues (Liljas et al. 1972). The overall fold of HCA II can be described as a single-domain, mixed / globular protein that is almo st spherical with approximate dimensions of 5 x 4 x 4 nm 3 (Figure 1-6; Lindskog, 1997). These enzymes are overall small and compact proteins with the possible exception of the N-terminal region (residues 1 to 24) that is more loosely connected to th e rest of the molecule. N-terminal residues 14 are usually disordered in the crystal struct ures. A cluster of arom atic residues at the Nterminus of the enzyme consisting of Trp5, Tyr7, Trp16, and Phe20 has been suggested to assist in the anchoring of this region to the rest of the enzyme (Lindskog, 1997). It has been shown that the removal of the N-termin al region does not result in a major loss of protein stability or enzyme activity. Another cluster of aromatic residues located under the sheet consists of Phe66, Phe70, Ph e93, Phe95, Trp97, Phe176, Phe179, and Phe226 (Lindskog, 1997). The active site folds first and inde pendently of the N-terminal region (Aronsson et al. 1995). The central feature of the HCA II structure can be described as a 10-stranded ( AJ) twisted -sheet, which is decorate d on the surface by seven helices (Figure 1-6; AG). The strands of the -sheet are mainly antiparallel, with the exception of two pairs of pa rallel strands (Figure 1-6; F and G, I and J). There is a conserved loop region extending towards th e active site that contains the proton shuttling residue, His64. The activ e site consists of a conical cleft that is ~ 15 deep with the catalytic Zn 2+ placed at the bottom of the cleft (Figure 1-6). The Zn 2+ is tetrahedrally coordi nated by four direct liga nds: the imidazole groups of three conserved His residues (His94, His96, and His119) and a H 2 O/ OH molecule

PAGE 25

10 (Figure 1-7). The direct metal ligand histidin es, in turn, are held in position by hydrogen bonding interactions with other residues, the i ndirect or second shell metal ligands. The side chains of Gln92 and Glu117 interact with His94 and His119, respectively, while the carbonyl oxygen of Asn244 coordinates Hi s96. Thr199 makes a hydrogen bond with the solvent metal ligand and is optimally oriented for this by interacting with Glu106 (Christianson and Fierke, 1996). Activity and Catalytic Mechanism of CAs HCA II kinetics and catalytic mechanism ha ve been studied extensively; however, it is thought that all CAs perform the same general act ivity, which is known as the zinc-hydroxide mechanism. The interconversion between CO 2 and HCO 3 is a two-step reaction that can be described as a ping-pong mechanism (eq 1-1 and 1-2; Silverman and Lindskog, 1988; Lindskog, 1997; Christ ianson and Fierke, 1996). H 2 O EZnOH + CO 2 EZn(OH )CO 2 EZnHCO 3 EZnH 2 O + HCO 3 (1-1) EZnH 2 O H + EZnOH + B EZnOH + BH + (1-2) The first step (eq 1-1) in the hydration direction is the binding of CO 2 in a hydrophobic region adjacent to the zinc atom. Substrate binding is then followed by a nucleophilic attack on the car bon by the zinc-bound hydroxide leading to the formation of HCO 3 Water can freely diffuse into the active site and displace HCO 3 leaving a water molecule bound at the zinc atom. The second pa rt of the reaction (eq 1-2) occurs at 10 6 s -1 and is the rate-limiting step of the overall reaction and invol ves the removal of an excess proton from the zinc-bound water to regene rate the hydroxide needed for catalysis (Khalifah, 1971; Steiner et al. 1975; Silverman and Lindskog, 1988). The transfer of a proton out of the active site involves an intramolecular and intermolecular proton transfer

PAGE 26

11 event. The intramolecular proton transport occurs between the Zn-bound solvent and the side chain of His64 through an interven ing chain of hydrogen-bonded water molecules, W1, W2, W3a, and W3b (Figure 1-6; Lindskog and Silverman, 2000; Fisher et al. 2005). Mutation of His64 to an alanine results in a 10-50 fold reduction in the proton transfer rate (Tu et al., 1989). From His64 the intermolecular tr ansfer event delivers the proton to bulk solvent/buffer. In eq 1-2, B signifies ei ther an acceptor on the pr otein (His64) or an exogenous acceptor that becomes protonated, BH + (Silverman and Lindskog, 1988). For CO 2 hydration in HCA II, both the k cat and k cat /K m have pH profiles that appear as simple titration curves with a pK a ~7 and with maximal activ ity at high pH (Silverman and Lindskog, 1988). HCA II is the most efficient isozyme (k cat = 1.4 x10 6 s at 25 C) of this class while HCA III is the slowest (k cat = 8 x 10 3 s at 25 C) (Khalifah, 1971; Jewell et al. 1991). HCA III has similar catalytic features to HCA II in that catalysis also occurs in the same two, separate steps (eq 1-1 and 1-2) but the enzyme is resistant to the classic HCA II inhibitors, the sulfonamides (e.g., acetazolamide and HCA III yield a K i of 40 M vs. 0.06 M for HCA II; LoGrasso et al. 1991). In fact, there seems to be a consistent, inverse relationshi p between the turnover number (k cat ) and level of inhibition by sulfonamide of CAs across species (Tufts et al. 2003). Substrate Binding in the CA Active Site The precise location of CO 2 binding in the CA active site remains elusive. The binding site has been narrowed down to a hydrophobic region behind the Zn, but the exact interactions that me diate substrate binding are still unknown. The binding of CO 2 is weak and the binding of HCO 3 is even weaker with approximate K d of 100 and 500 mM, respectively (Krebs et al. 1993; Lindskog and Silverman, 200 0). It is thought that CO 2 binding in the hydrophobic pocket displaces th e deep water that is normally hydrogen

PAGE 27

12 bonded to the amide group of Thr199 (Hkansson et al. 1992). CO 2 binding causes it to be polarized by the interaction with the b ackbone amide of Thr199 in addition to the electrostatic effects exerted by the zi nc. This polarization causes the CO 2 to become susceptible to nucleophilic attack by the Zn-bound solvent, as described above. The hydrophobic substrate binding pocket in HCA II is defined by four residues: Val121, Val143, Leu198, and Trp209 (Lipscomb, 1990; Merz, 1991). The crystal structure of a HCA II mutant (Thr200 His) in complex with HCO 3 (PDB accession code: 1BIC, Xue et al. 1993). The structure shows one of the HCO 3 oxygen atoms acting as a metal ligand, replacing the Znbound solvent while the hydrogen is involved in an H-bond with the side chain of Th r199. The second oxygen atom acts as a fifth ligand to the Zn 2+ and the third seems to be in an H-bond with the backbone amide of His199 (Earnhardt and Silverman, 1998). The st ructure of this mutant enzyme:product complex implicates Thr199 as a very important residue for orienting water molecules in the active site as well as the direct interaction with HCO 3 A similar complex with the native enzyme has not been obtained, as th e native form probably does not exhibit the same extent of binding to HCO 3 The use of competitive inhibitors to elucid ate the substrate-binding site has yielded interesting but conflicting results. Imidazole is a competitive inhibito r of HCA I with a K i of 20 mM and phenol is a competitive inhibitor of HCA II with a K i of 10 mM (Khalifah, 1971; Tibell et al. 1985). Crystal structures of HCA I in complex with imidazole and HCA II in complex with phenol show different binding modes of these inhibitors in the active site, making it hard to elucid ate the possible binding mode of CO 2 (Kannan et al. 1977; Nair et al. 1994; Earnhardt and Silverman, 1998).

PAGE 28

13 Measuring CA Activity One of the techniques used to m easure CA activity is stopped-flow spectrophotometry under steady stat e conditions using pH-indicat or pairs. This technique is used to determine k cat and K M by measuring the initial rates of CO 2 hydration. The turn-over number k cat reflects the part of catalysis that involves rate-limiting proton transfer, while k cat / K M reflects the steps involved in CO 2 / HCO 3 interconversion (Khalifah, 1971; Steiner et al. 1975). Another technique used to measure CA activity is 18 O-exchange and this is done at chemical equilibrium. This me thod is based on the exchange of 18 O between 12 C and 13 C-containing species of CO 2 and water that occurs because of the hydrationdehydration reaction of CA. Two rates can be determined by this method, R 1 and R H2O The first rate, R 1 is a measure of the exchange between CO 2 and HCO 3 at chemical equilibrium. The second rate determined by these methods, R H2O indicates the rate of release from the enzyme of water carrying substrate oxygen. The rate of water release from the active site depends on the rate of proton transfer, thus R H2O is used to measure proton transfer activity (Silverman et al. 1979; Silverman 1982). Analysis of the catalyzed reaction by CA at steady state and chemical equilibrium has led to a model of its mechanism of action that implicates two ionizing groups in the active site that have pK a values near 7 (Steiner et al. 1975; Silverman and Lindskog, 1988; Lindskog 1997). One of these groups corresponds to the zinc-bound water, which ionizes to a hydroxyl ion, a nd is responsible for th e interconversion of CO 2 / HCO 3 while the other group, His64, is involved in proton transfer (Tu and Silverman, 1989).

PAGE 29

14 HCA and Human Disease Due to their wide-spread distribution and various physiological functions, the HCAs support many systemic and cellular HCO 3 / CO 2 transport processes as well as biosynthetic pathways. There ar e not many examples of CA-associated diseases and this reflects its crucial role in so many fundamental life pr ocesses. In general, HCA deficiencies are rare and a possible reason is that HCAs are ubiqu itously expressed and, in some cases, a different HCA can possibly compensate for the loss of a particular isoform (Sly and Hu, 1995). A number of genetic variants of HCA I have been described and they generally differ only in one amino acid due to nucleotid e substitutions. HCA I variants are less heat stable compared to the wild type version but the catalytic differen ces are very small or even insignificant. Individuals that either ha ve a variant or lack HCA I completely, show no phenotype (Osborne and Tashian, 1974; Vent a, 2000). Despite there being a lot more HCA I than HCA II in RBCs, maybe a loss of HCA I shows no phenotype because HCA II, a very active isoform compared to HCA I (Table 1-1), can functionally compensate for this loss. HCA II is the most well-studied isoform of all the HCAs and part of the reason is that there are some severe dise ases associated with mutant versions of this isoform. One of these is an inherited diseases called HCA II deficiency syndrome. It manifests as a lack of erythrocyte HCA II and is associated with osteopetrosis, renal tubu lar acidosis and, in extreme cases, cerebral calcificat ion that leads to mental re tardation (Sly and Hu, 1995). A number of mutations that include nonsense, frameshift, a nd splicing mutations can lead to a lack of HCA II. Most of these changes cause HCA II deficiency syndrome, but a few other variants have been identified that do not lead to a change in activity or amount of

PAGE 30

15 enzyme (Venta, 2000). The first mutation identified that is associated with the disease results in His107 Tyr substitution and leads to an unstable enzyme that shows a threefold lower catalytic rate (Venta et al. 1991; Tu et al. 1993). HCA II deficiency syndrome is quite rare and is mostly f ound in homozygous individuals from families where some level of inbreeding ha s occurred (Sly and Hu, 1995). No variants or deficiencies have been described for other HCAs. There is a polymorphism in HCA III where Ile31 Val and variation in activ ity is predicted based on the location of this substitution. Genetic knock-out studies in mice of murine CA III did not show any phenotype under all the standa rd muscle-stress tests and longevity of these animals was also not affected (Kim et al. 2004). Recent studies with CA IV and CA IX single and double knock-out mice implicated these isoforms in buffering and pH regulation of the extracellular space in the hippocampus. The CA IV knock-out mice offs pring were produced in lower than expected numbers and females seemed to die more frequently than males during gestation and immediately after birth. CA IX knock-out mice had normal fertility and viability. The double knock-out mice were smaller than wild type and most of the females died before ten months of age. Electrophysiological measur ements on brain slices from these animals showed that either one of these membrane-bound CAs can buffer the hippocampus after synaptic firing. A loss of both in the double knock-out mice completely ablated this buffering effect (Shah et al. 2005). CA IX is a highly active isoform with an extracellu lar CA domain that has functionally been implicated with acid/ba se balance and interc ellular communication. Aberrant expression of CA IX is associated with various tumors and has become of

PAGE 31

16 significant clinical inte rest. In 2002 Ortova Gut et al. constructed a CA IX knock-out mouse and investigated the effects on gastro intestinal epithelia cells. Although these mice had normal stomach pH and acid secretion, they developed gastric hyperplasias and several cysts. These studies highlight the importa nt role of CA IX in cell proliferation and differentiation (Ortova Gut et al. 2002). CA Inhibitors Not long after the discovery of CA by Meldrum and Roughton (Meldrum and Roughton, 1932) these authors also investigated inhibition of CA by small molecule inhibitors such as azide and cyanate (Figure 1-8 (a) and (b); Meldrum and Roughton, 1933). In the 1940s, Mann and Keilin found that sulfonamide-based compounds are specific and strong inhibitors of CA (Mann and Keilin, 1940). Figure 1-8 shows several crystal structures of HCA II with small molecule inhibitors as well as the clinically used sulfonamide-based drugs. Since the early findings, many other strong and selective inhibitors have been investigated and these are the aromatic and heterocyclic sulfonamides of the R-SO 2 NH 2 or R-SO 2 NH(OH) form (Maren, 1967; Maren 1974). All the sulfonamide-based inhibitors interact with CA by the sa me mechanism: they bind to the metal ion and interfere with the ZnOH coordination by either displacing or replacing the hydroxide, thus disrupti ng the interconversion of CO 2 and HCO 3 (Figure 1-8; Bertini and Luchinat, 1983). Several crystal structures of complexes of various sulfonamide -based inhibitors with CA show similar inte ractions: the -NH group of th e sulfonamide moiety binds directly to the metal and simultaneously donates a hydrogen bond to hydroxyl of Thr199. An oxygen of the sulfonamide also interacts with the amide backbone of Thr199 and thus displaces the deep water (Figure 1-8; Li ndskog, 1997). The key group in determining this

PAGE 32

17 displacement/replacement is the hydroxyl of Thr199 and this residue is sometimes referred to as the gate keeper. Examples of clinically important dr ugs include acetazolamide (Diamox) and brinzolamide (Azopt) that have applications in the treatment of congestive heart failure, altitude sickness and epilepsy (Figure 1-8 (c) and (d); Mansoor et al. 2000). The most common sulfonamide inhibitor in clinical use is acetazolamide, which is a strong inhibitor of HCA II with a K i value near 0.01 M (Maren and Conroy 1993). Analysis of acetazolamide bound to HCA II revealed the bi nding interactions of this compound. The thiadiazole ring is in van der Waals contact with Val121, Leu198, and Thr200 and the carbonyl oxygen of the amido group shares a hydrogen bond interact ion with the side chain amide of Gln92. The methyl group was shown to interact with the side chain of Phe 131 (Vidgren et al. 1990). CA inhibitors are commonly prescribed to treat a major sympto m of glaucoma, i.e. increased intraocular pressure. The inhibition of CA in the eye by t opical application of the drug suppresses the secretion of Na + HCO 3 and subsequently production of aqueous humor thus lowering intraocu lar pressure. (Maren, 1987). In the following chapters the detailed active si te structures of wild type and mutant CAs, and how it relates to proton transfer pr ocesses, will be presented. Chapter 2 will deal with pH stability of wild type and site -specific mutants as well as the position of a proton shuttling residue in the active site. Thes e structural features will be correlated with kinetic measurements. In Chapter 3 a different approach will be disc ussed where, instead of moving the proton shuttle to different positions, mutations were made in the active site of HCA II in order to disrupt solvent networks that mediate proton tr ansfer. That data will

PAGE 33

18 also be correlated with kinetic measurements and implications for proton transfer will be discussed. Chapter 4 will be a discussion of initial experiments performed for the eventual determination of a neutron diffracti on structure of perdeuterated wild type HCA II. A detailed structural comparison of overall and active site features between hydrogenated and perdeuterated HCA II will be presented. This work shows proof-ofprinciple for using HCA II crystals for neutron diffraction experiments. In Chapter 5, the unique binding modes of two novel CA inhib itors, as revealed by X-ray crystallography, will be discussed. These are new compounds and these inhibitors target different residues compared to other canonical CA inhibitors. Chapter 6 will deal with the characterization of a CA from the mosquito, Aedes aegypti. This work includes kinetic characterization, inhibition studies, and a homology model of the enzyme. Work presented in Chapters 5 and 6 have implicati ons for the search and design of novel CA inhibitors with possible applications for controlling mosquito populations and treating certain cancers that have associated CA overexpression. Finally, Ch apter 7 will contain a summary and concluding remarks on the work pr esented elsewhere in this thesis, as well as possible future directions for the topics discussed.

PAGE 34

Table 1-1. Catalytic constants and subcellular locati ons of active -CA isozymes.* Isoform k cat (s -1 ) k cat /K M (M -1 s -1 ) Subcellular Location Activity Level HCA I 2.0 x 10 5 5.0 x 10 7 Cytoplasm Medium HCA II 1.4 x 10 6 1.5 x 10 8 Cytoplasm High HCA III 1.0 x 10 4 3.0 x 10 5 Cytoplasm Low HCA IV 1.1 x 10 6 5.0 x 10 7 Membrane-bound High Murine CA V 3.0 x 10 5 3.0 x 10 7 Mitochondrial Medium Rat CA VI 7.0 x 10 4 1.6 x 10 7 Secreted Medium Murine CA VII 9.4 x 10 5 7.6 x 10 7 Cytoplasm High HCA IX 3.8 x 10 5 5.5 x 10 7 Transmembrane Medium-high HCA XII 4.0 x 10 5 7.4 x 10 7 Transmembrane Medium-high 19 Adapted from Chegwidden and Carter (2000) HCA XIII and XIV are not included as defi nitive rate constants have not been determined at the time of this writing.

PAGE 35

20 Figure 1-1. Ribbon diagram of CA from three classes. (a) -class CA; (b) -class CA; (c) -class CA. Panels on the right of (b) a nd (c) represent the biological assembly as a dimer and trimer, respectively. Colo ring is from blue for the N-terminus to red for the C-terminus, gray spheres are Zn 2+ Figures were generated and rendered with Bobscript and Raster3D (Esnouf 1997; Merritt and Bacon, 1997). PDB accession codes for (a), (b), and (c) are 1MOO, 1DDZ and 1QRG, respectively (Duda et al. 2003; Mitsuhashi et al. 2000; Iverson et al. 2000).

PAGE 36

21 Figure 1-2. Multiple sequence alignment of fourteen HCA domains. Alignment was performed using ClustalW and residue numbering is according to HCA II (Thompson et al. 1994). Conserved and similar substituted residues are indicated by and #, respectively. Regi ons in bold that are red and blue signify strand and helical regions of HCA II.

PAGE 37

22 Figure 1-3. Conservation of HCAs. Gray ribbon backbone repr esentation of HCA II with completely conserved residues of fourt een HCAs indicated as red spheres. Active site residues are shown as yell ow ball-and-stick (HCA II numbering), the zinc atom is a black sphere and the Nand C-terminus as labeled.

PAGE 38

23 Figure 1-4. Active sites of HCA I, II, and III. (a) HCA I, (b) HCA II, and (c) HCA III. Active site residues are in yellow ball-and-stick and are as labeled, zinc atom = black sphere. Figure was generated w ith Bobscript and Raster3D (Esnouf, 1997; Merritt and Bacon, 1997).

PAGE 39

24 Figure 1-5. Molecular and electr ostatic surface potentials of HCAs. (a) HCA II; (b) HCA III; (c) HCA VI; (d) HCA-RP VIII; (e ) HCA IX; (f) HCA XII dimer. Al models are shown in the sa me orientation with the active site in the center. Negative and positive charge is represented by red and blue, respectively. Figures was generated with GRASP (Nicholls et al. 1991).

PAGE 40

25 Figure 1-6. Ribbon diagram of HCA II with s econdary structure elements. Red regions are strands ( A J), blue regions are helices ( A G). Zinc atom is shown as a black sphere and the Nand C-termini are labeled. Figure generated with Bobscript and Raster 3D (Esnouf, 1997; Merritt and Bacon, 1997).

PAGE 41

26 Figure 1-7. Active site of wild type HCA II. Ca talytic residues are shown in yellow balland-stick, solvent molecules = red spheres, zinc atom = black sphere. Residues are as labeled and inferred hydrogen bonds are indicated by the dashed orange lines. Figure was genera ted with Bobscript and rendered with Raster3D (Esnouf, 1997; Merritt and Bacon, 1997).

PAGE 42

27 Figure 1-8. HCA II in complex with inhibitors. (a) Sulfate, (b) Azide, (c) acetazolamide, and (d) brinzolamide. HCA II active site residues are shown as yellow balland-stick and the black sphere is the zinc atom. PDB accession codes used (a)(d): 1T9N, 1RAY, 1YDA, and 1A42 (Fisher et al. 2005; Jonsson et al. 1993; Nair et al. 1995; Stams et al. 1998). Figure was genera ted with Bobscript and Raster3D (Esnouf, 1997; Merritt and Bacon, 1997).

PAGE 43

CHAPTER 2 STRUCTURAL AND KINETIC ANALYSIS OF PROTON SHUTTLING IN HUMAN CARBONIC ANHYDRASE II Introduction Human carbonic anhydrase II (HCA II) is one of the most efficient enzymes with a turnover number near 10 6 s -1 and k cat /K M of 1 x 10 8 M -1 s -1 This fast rate indicates that the catalysis is limited by the rate of diffusion of substrate into the active site of HCA II (Khalifah, 1971). The overall catalytic mechanism was presented in some detail in Chapter 1 and will only be discussed here briefly. The first step in the interconversion of CO 2 and HCO 3 is the hydration of CO 2 by the zinc-bound OH which is then followed by the displacement of HCO 3 by water. The second step is the deprotonation of the zincbound water to regenerate the zinc-bound OH and involves both intramolecular and intermolecular proton transf er steps (Silverman, 1982; Silverman and Lindskog, 1988). Imidazole and derivatives act as nucleophilic and gene ral base catalysts. The structure of imidazole, with two almost identic al N atoms, allow it to pick up a proton off one of its N atoms forming a cation and then delivering it to the second N atom. As the functional group of histidine, it is commonly a ssociated with proton transport in proteins (Scheiner and Yi, 1996). His64 in HCA II acts as a proton shuttle be tween the zinc-bound solvent and buffer in solution and mutation of His64 to an alanine reduces enzymatic activity 10-50 fold. This mutation does not affect the hydration/ dehydration part of catalysis, but the observed decrease in prot on transfer can be re scued by supplying free imidazole in the reaction buffer. It has been postulated that the pr oton transfer between 28

PAGE 44

29 the zinc-bound water and His64 occurs through intervening water molecules (Venkatasubban and Silverman, 1980; Tu et al. 1989). In other protein systems, such as cy tochrome c oxidase and the bacterial photosynthetic reaction center, hydrogen-bonded water chains have been observed in several crystal structures and the location and geometry of these chains suggest that they participate in proton transfer reactions (P oms and Roux, 1996). These chains are thought to be effective in long-range proton transloc ation and such hydrogen-bonded water chains have been called proton wires (Nagle a nd Morowitz, 1978). Protons do not move by diffusion as a hydrated proton (hydronium ion, H 3 O + ), instead the high observed mobility of protons is thought to occur from successi ve protonation-deprotonation events. In these cases, protons hop from oxygen to oxygen along a pre-existing hydrogen-bonded water chain in a Grotthus-type m echanism (Poms and Roux, 1996). Similar proton wires are observed in the active site of HCA II and th ese water molecules bridge the distance between the zinc-bound solvent and the proton sh uttle, His64. The water molecules that span the 8 distance between zinc and His64 appears to be hydrogen bonded to each other but not to the imidazo le group of His64 (Eriksson et al. 1988). The proton transfer path in HCA II was firs t described by Steiner et al. based on assays that showed significant solvent hydrogen isot ope effects. These experiments indicated that maximum velocity, k cat was limited in rate by the intermolecular proton transfer (Steiner et al. 1975). The Brnsted relation, as applied to pr oton transfer, is a linear free-energy relationship that correlates the rate constant (k B ) for proton transfer with the difference in

PAGE 45

30 acid/base strength ( pK a ) of the proton acceptor and donor, is shown in eq 2-1 (Silverman et al. 1993; Silverman, 2000) log(k B ) = [pK a (acceptor) pK a (donor)] + constant (2-1) The slope ( ) of a Brnsted plot of log(k B ) versus pK a is used to investigate a reaction mechanism and can be used as an es timate of the extent of proton transfer between reactants and products in the transiti on state. Observed variation in the slope over a range of pK a values has been interpreted through Marcus theory and yields 1) an intrinsic energy barrier for proton tr ansfer, and 2) two work terms, w r and w p that quantitate the energy required for aligning the reactants in the reaction complex in both the forward and reverse direc tions, respectively (Silverman et al. 1993; Silverman, 2000; Kresge and Silverman, 1999). Application of Marcus rate theory to Brnsted plot s of the proton transfer processes in carbonic anhydrase has provided a way to experimentally determine the intrinsic energy barrier as well as separati ng different thermodynamic contributions from the observed activation energy (Silverman, 2000). The rate constant fo r proton transfer is maximal when the difference between the pK a values of the proton acceptor and donor as near zero (Rowlett and Silverman, 1982; Silverman et al. 1993). Further an alysis of free energy plots has shown there is a significan t energetically unfavorable pre-equilibrium that exists before a very ra pid proton transfer event occurs. It was determined that catalysis proceeds with an intrinsic free ener gy of activation that is very small (~ 1.5 kcal/mol) but with a large work function w r (~ 10 kcal/mol) (Kresge and Silverman, 1999; Silverman, 2000). The magnitude of this work function accounts for the relative low proton transfer rate in HCA II, at most 10 6 s -1 compared to maximal rates of 10 11 s -1

PAGE 46

31 measured for proton transfer between naphtol-related photo acids to acetate in solution (Pines et al. 1997). It is not well under stood what processes add to the work functions, but in HCA II this may involve the rearra ngement and formation of a hydrogen bonded water chain as well as orienta tion of His64. These factors im ply that setting up the active site environment for proton transfer takes far more energy than the actual proton transfer itself (Silverman et al., 1993). The shuttling of protons between bul k solvent and the zinc-bound solvent molecule may require some conformational mobility of His64 as the proton donor/acceptor. Considerable support for this comes from various crystal structures in which the proton shuttle residue His64 in HCA II and Glu84 in the archaeal carbonic anhydrase from Methanosarcina thermophila show two conformational rotamers in the active-site cavity (Nair and Christianson, 1991; Tripp and Ferry, 2000; Iverson et al. 2000). A chemically modified cysteine residue acting as a proton shuttle in a mutant of CA V also shows evidence of multiple orientations (Jude et al. 2002). Structural studies by Nair and Christianson showed that His64 occupies different positions depending on the pH. At pH 5.7 th e side chain of His64 rotates around 1 by 64 to occupy what is termed the out position (~ 12 away from the active site) compared to the structure at pH 8.5 that s hows the side chain of His64 pointing towards the active site (~ 8 away from the activ e site) in what is known as the in conformation (Nair and Christianson, 1991). A similar observation was made for the Thr200Ser mutant at pH 8.0 where His64 was observed to occupy an even further out position due to a rotation about 1 of 105 (Krebs and Fierke, 1991). In both cases,

PAGE 47

32 the conformational mobility of His64 as a function of pH or the mutation at position 200 did not seem to adversely affect proton transfer kinetics (Nai r and Christianson, 1991). The current understanding of the requirements for rapid proton transfer, as measured for HCA II, involves several aspects that could be investigated by comparisons of kinetics and structure. The kinetic effect of the location of a histidine proton shuttle has been studied by introducing histidine residues at different positions in the active-site cavity of HCA II. These results show that a hi stidine residue at sites other than position 64 is able to participate in proton transfer; specifically, His67 app ears capable of more efficient proton transfer than His62 (Liang et al. 1993). To further characterize and understand th e relationship between efficient proton transfer and the existence of proton wires in HCA II, detailed comparisons of kinetics and active site water structure we re conducted. These included st ructure determinations of wild-type HCA II and HCA II mutants (H 64A/N62H and H64A/N67H) from pH 5.1 to 10.0. These data were correlated with catalytic activities that were measured by the exchange of 18 O between CO 2 and water from pH 5.0 to 9.0 as catalyzed by these enzymes to determine what common structural features are important for proton transfer. Materials and Methods Enzymes Plasmids with the appropriate mutations in the cDNA of HCA II were provided by Professor Sven Lindskog, Ume University, Swed en. Expression of wild type and mutant HCA II was performed in E. coli BL21(DE3) pLysS cells that were grown to an optical density of ~ 0.60 as measured at a waveleng th of 600 nm. Protein expression was induced by the addition of 1 mM isopropyl-D-1-thiogalactopyranoside (IPTG) and 1 mM zinc sulfate was also added for uptake in the expressed protein. At 4 hours post-induction,

PAGE 48

33 cells were harvested by centrifugation a nd the cell pellets frozen at -20 C overnight. Cell pellets were lysed by freeze/ thawing and solubilized in 0.2 M sodium sulfate, 50 mM Tris-Cl (pH 9.0). The soluble cell fraction wa s obtained by centrifuging the lysates at 100,000 x g for 1 hour at 4 C. Enzymes were further purified from the supernatant by affinity chromatography using p-amino-met hyl-benzenesulfonamide (pAMBS; a specific binder to the active site of -CAs) coupled to agarose b eads as described elsewhere (Khalifah, 1977). Purity of the protein was verified by electrophoresis on a 12% polyacrylamide gel stained with Cooma ssie. The concentration of HCA II was determined by measuring the absorbance at 280 nm and using a molar absorptivity of 5.4 x 10 4 M -1 cm -1 (Coleman, 1967). Crystallography Crystals of wild type and mutant HC A II were obtained using the hanging drop method (McPherson, 1982). The crystalliza tion drops were prepared by mixing 5 l of protein (10.5 mg/ml concentration in 50 mM Tris-HCl, pH 7.8) with 5 l of the precipitant solution (50 mM Tris-HCl, pH 7.8, 2.5-2.9 M ammonium sulfate) at 4 C against 600 l of the precipitant soluti on. Useful crystals were obs erved within five days after crystallization setup. The pH of the wild type crystals were obtained by equilibrating crystals in appropriate buffers (50 mM sodi um acetate, pH 5.1; 50 mM MES, pH 6.1; 50 mM Tris-HCl, pH 7.0, 7.8 and 9.3; 50 mM CAPS, pH 10.0) and 3.0 M ammonium sulfate. The pH of the double mu tant crystals were obtained by using the same approach as above using the buffers (50 mM Tris -HCl, pH 6.0 and 7.8) and 3.0 M ammonium sulfate. Crystals were allowed to equilibrate for 4-12 hours at 4 C before data collection

PAGE 49

34 to ensure that complete solvent exchange in th e crystal lattice occurred (the pH stated was that as measured at the start of the experiment). X-ray-diffraction data sets were obtained using an R-AXIS IV++ image plate system with Osmic mirrors and a Rigaku HU-H3R Cu rotating anode operating at 50 kV and 100 mA. The detector to crystal distance was set to 100 mm for H64A/N67H, and 120 mm for H64A/N62H and wild type HCA II X-ray data collection. Each data set was collected at room temperature from a single crystal mount ed in a quartz capillary. The oscillation steps were 1 with a 3 minute exposure per image. X-ray data processing was performed using DENZO and scaled and reduced with SCALEPACK software (Otwinowski and Minor, 1997). Data set statistics for the wild type and muta nt structures at different pHs are given in Tables 21 and 2-2, respectively. All models were built using the program O, version 7 (Jones et al. 1991). Refinement was carried out with the software package CNS, version 1.1 (Brnger et al. 1998). The wild type HCA II structure (Protein Data Bank accession number 2CBA; Hkansson, 1992), which was isomorphous with all the data se ts collected, was used to phase the data sets. To avoid phase bias of the model, th e zinc ion, mutated side chains and water molecules were removed. After one cycle of rigid body refinement, annealing by heating to 3000 K with gradual cooling, geometry-restrained position refinement, and temperature factor refinement, the 2F o F c Fourier maps were generated. These density maps clearly showed the position of the zi nc and the mutated residues, which were subsequently built into the respective models. After several cycles of refinement, solvent molecules were incorporated into the models using the automatic water-picking program in CNS until no more water molecules were found at a 2.0 level. Refinement of the

PAGE 50

35 models continued until convergence of the R work and R free was reached. Final model statistics for wild type and mutant stru ctures are given in Tables 2-1 and 2-2, respectively. Activity Analysis by 18 O Exchange These assays were performed by Dr. Chingkuang Tu in the Silverman lab and the details of determinating the ra te constants for catalysis by HCA II are discussed in detail below. The 18 O exchange method is based on the m easurement, using membrane-inlet mass spectrometry, of the exchange of 18 O between CO 2 and H 2 O at chemical equilibrium (Silverman, 1982) (eqs 2-2 and 2-3). HCOO 18 O + EZnH 2 O EZnHCOO 18 O COO + EZn 18 OH (2-2) H 2 O EZn 18 OH + BH + EZnH 2 18 O + B EZnH 2 O + H 2 18 O + B (2-3) An Extrel EXM-200 mass spectrometer with a membrane inlet probe was used to measure the isotopic content of CO 2 Solutions contained a to tal concentration of all species of CO 2 of 25 mM and the ionic strength was maintained by the addition of 0.2 M sodium sulfate. This appro ach yields two rates for the 18 O exchange catalyzed by carbonic anhydrase. The first is R 1 the rate of exchange of CO 2 and HCO 3 at chemical equilibrium, as shown in eq 2-4. R 1 /[E] = k cat ex [S]/(K eff S + [S]) (2-4) Here k cat ex is a rate constant for maximal inte rconversion of substrate and product, K eff S is an apparent binding constant for substr ate to enzyme, and [S] is the concentration of substrate, either CO 2 or HCO 3 The k cat ex /K eff S ratio is, in theory and in practice, equal to k cat /K M obtained by steady state methods. The binding of CO 2 and HCO 3 to the active site of HCA II is weak (Krebs et al. 1993), and in this work too it is assumed that [CO 2 ]

PAGE 51

36 << K eff S The pH dependence of k cat /K M depends on the ionization state of the zinc-bound water, as shown in eq 2-5. k cat /K M = (k cat /K M ) max (1 + [H + ]/(K a ) ZnH2O ) -1 (2-5) A second rate determined by the 18 O exchange method is R H2O the rate of release from the enzyme of water bearing substrate oxyg en (eq 2-3). This is the component of the 18 O exchange that is enhanced by exoge nous proton donors (Silverman, 1982). The pH dependence of R H2O /[E] is often bell-shaped, consistent with the transfer of a proton from a single predominant donor to the zinc-bound hydroxide. In thes e cases, the pH profile is adequately fit by eq 2-6 in which k B is a pH-independent rate c onstant for proton transfer, and (K a ) donor and (K a ) ZnH2O are the noninteracti ng ionization constants of the proton donor BH + of eq 2-3 and the zinc-bound water. k B obs = k B /{[1 + (K a ) donor /[H + ]][1 + [H + ]/(K a ) ZnH2O ]} (2-6) Results and Discussion Crystallography All crystals of wild type H64A/N62H, and H64A/N67H HCA II were isomorphous in the P2 1 space group with mean unit cell dimensions: a = 42.7 2.0 b = 41.6 1.0 c = 72.9 2.0 and = 104.6 2.0. Wild type crystals from pH 5.1 to 10.0 all diffracted to 2.0 while crystals of H64A/N62H a nd H64A/N67H diffracted to between 1.63 and 1.90 Tables 2-1 and 2-2 contain summaries of the diffraction data set and model statistics from wild type and mu tant HCA II crystals, respectively. A least squares superposition of wild type HCA II structures (at pH 5.1, 6.1, 7.0, 9.0, and 10.0) onto the structure at pH 7.8 indi cated no significant differences between them and had an average root mean square deviation (rmsd) of less than 0.1 for all atoms. Superposition of H64A/N62H HCA II at pH 6.0 and 7.8, as well as H64A/N67H

PAGE 52

37 at pH 6.0 and 7.8, onto the pH 7.8 wild type HCA II structure showed an average rmsd near 0.15 for all atoms. When the mutants were compared to each other, the rmsd was only 0.07 The rmsd of bond lengths and angles for all structures were between 0.038 and 0.005 and between 1.3 and 2.1, respectively. Ramachandran statistics for all amino acids were ~ 90% in most favored region and ~ 10% in the additionally allowed region with no residues in the gener ously and disallowed regions. Effect of pH on the Wild Type HCA II Active Site The tetrahedral coordination of the zi nc atom by His94, His96, His199, and a nonprotein atom (either H 2 O/OH or sulfate) was intact over the broad pH range from 5.1 10.0 used in this study. At pH 5.1 a sulfate ion bound directly to the zinc atom was observed (Figure 2-1). The sulfat e binding displaced the zinc-bound H 2 O/OH but had no effect on the coordination of the zinc atom. In all the wild type structures determined from pH 5.1 to 10.0, there was a highly cons erved, well-ordered water network extending from the zinc-bound solvent (or sulfate) to the imidazole of His64. These networks consist of four water molecules, W1, W2, W3a, and W3b. W1 is hydrogen bonded to W2, which is in turn hydrogen bonded to W3a and/or W3b. W1 is held in place by interactions with not only W2, but also with the hydroxyl group of Thr200. W3a and W3b interact with Tyr7 and Asn62/Asn67, resp ectively (Figure 2-1) W2 is the only water molecule held entirely in place by in teractions with other water molecules. The configuration of these networks leads to two possibilities for a proton transfer pathway from the zinc-bound solvent to the proton shuttle His64; W1 W2 W3a His64 or W1 W2 W3b His64. However, the distances between the distal waters, W2, W3a, and W3b, and His64 were all too long (> 3.4 ) to constitute viable hydrogen

PAGE 53

38 bonds and it is unclear how the proton will make the jump from one of these waters to His64. Other studies with HC A VI alkylated with an im idazole-containing reagent showed a completed, hydrogen-bonded water chai n and was associated with a very small rate of proton transfer (10 3 s -1 compared to 10 6 s -1 for His64 in wild type HCA II; Jude et al. 2002). Thus, the observation in the crystal structure of a completed, hydrogen-bonded water chain that links the pr oton donor and acceptor is not a requirement for the rapid proton shuttling on the scale determined for HCA II. It should also be noted that the hydrogen bonds in crystal structur es are assigned based on the distance and geometries between hydrogen bond donors and acceptors. Hydrogen atoms are invisible by X-ray crystallography at these resolutions and subsequently these assignments can be extremely misleading as th ere is no way to observe the ionization of waters or amino acid side chains. Currentl y, the only definitive way to observe hydrogen bond are with either ultra-high resolution (< 1 ) X-ray crystallography or macromolecular neutron diffraction studies. The only observed structural difference betw een wild type HCA II at the various pHs was the positional occupancy of His64 (Fi gure 2-1). The occupancies of His64 were determined from F o -F c omit maps, where His64, W1, W 2, W3a, and W3b were removed for the calculation to avoid model bias. The re sidual electron density volumes were used to determine the relative occupancies of Hi s64 in the in and out positions. At pH 5.1, His64 was predominantly in the in position and this could be due to electrostatic effects exerted by the sulfate ion. Consequently, th e out conformer was modeled as a water molecule. At pH 6.1, 7.0, and 7.8 the occupancies were 60, 70, and 80% in the in conformation. There was no increased occupancy at higher pH and at pH 9.0 and 10.0

PAGE 54

39 His64 was still 80% in the in conformation (Figure 2-1). The overall trend from this data shows that with increasing pH, His64 fa vors the in conformation and this result is consistent with the work of others (Nair and Christianson, 1991). The side chain torsion angles for the in and out conformers were ( 1 = 48, 2 = -98) and ( 1 = -43, 2 = 90), respectively. Thus, over a range of pH s it was observed that His64 occupies both positions with equivalent occupancy and this phenomenon is consistent with its function as a proton shuttle. The phys iological function of His64 as a shuttle in catalysis is supported by these observations as hydration and dehydration occur at nearly the same rates. Structural and kinetic studies with 4-methylimidazo le and H64A HCA II suggest that binding of this exogenous proton donor/acce ptor to a position that mimics the out position of His64, is not as efficient in prot on transfer as His64. Perhaps the motion of this residue is somehow related to its efficacy as a proton shuttle. When His64 occupies the in position it can pick up an excess proton from one of the active site solvent molecules and then move rapidly to the out position and deliver it to bulk solvent. Effect of pH on the Mutant HCA II Active Site The hydrogen bonding pattern and solven t network is different for the H64A/N62H and H64A/N67H HCA II mutants co mpared to each othe r and wild type (Figure 2-2). Even though both mutants display similar solvation levels in the active site, the organization of the water molecules has been altered. It is interesting that both mutants have sulfate bound at pH 6.0, while in wild type sulfate wa s only observed at pH 5.1 (Figure 2-2 (a), (c)). However, at ph 7.8 there is not sulfate present in either wild type or mutant HCA II. The sulfate binds in th e same orientation in all the structures determined in this study: it displaces the zinc-bound solvent and simultaneously engages in a hydrogen bond with the hydroxyl group of Th r199, thus maintaining the tetrahedral

PAGE 55

40 coordination of the zinc. Sulfate has been s hown to be inhibitory to HCA II at low pH and it was thought that it inte racted with the zinc (Simonsson and Lindskog, 1982). These data suggest that the inhibition at low pH by sulfate may be due to the binding of a protonated sulfate ion in a similar mode to other protonated inhibitors such as HSO 3 and the sulfonamide drugs (Liljas et al. 1994). There are several structures of mutant HCA II in complex with sulfate but the results presented here show for the first time a sulfate bound to the zinc center of wild type HCA II and this occurs only at very low pH (Xue et al. 1993(a)). In both mutants the introduced His side ch ains extend into the active site cavity towards the zinc. In contrast to His64 in wild type HCA II, neither one of the His in the mutants display any conformational mobility. In H64A/N62H HCA II at pH 7.8 there is a complete hydrogen bonded solvent network that extends from the zinc-bound solvent to His62 (W1 W2 W3b; Figure 2-2 (b)). Due to th e long distance (3.2 ) between W3b and His62 this interaction represen ts a very weak hydrogen bond. The water W3b that His62 is hydrogen bonded to is also connected to the side chain of N67. In H64A/N67H HCA II at pH 7.8, His67 is connected directly to W2 with a bond distance of 3.2 and W3b has been completely displaced by the side chain of His67 (W1 W2; Figure 2-2). His67 is also hydrogen bonde d to the side chain of N62 and this interaction might prohib it rotational freedom of the His during catalysis. The His at position 67 is ~ 6.6 away from the zinc while His62 is over 8 away (Table 2-3). This extra distance is sp anned by an additional wa ter molecule and the solvent network appears more branched in H64A/N62H HCA II compared to H64A/N67H HCA II. The observed network of two water molecules between the proton

PAGE 56

41 donor and acceptor seen in H64A/N67H HCA II at pH 7.8 is similar to one of the possible pathways (W1 W2) as seen in wild type HCA II from pH 6.1 to 10.0, except that in wild type all of the distal waters are over 3.2 away from the proton shuttle. Catalysis The pH dependence of the hydration/dehydration rate, k cat /K M and the proton transfer-dependent rate constant, R H2O /[E], for the release of 18 O labeled water from the active site, was measured. The pH range was from pH 5.0 to 9.0 for wild type and both mutants. These data was compared to 18 O exchange data obtained for a mutant, H64A HCA II, that lacks a proton shuttle in the activ e site. The data for wild type HCA II shows that the enzyme is remarkably stable and hi ghly active over the pH range studied (Figures 2-3 and 2-4). The data in Figure 2-3 show the k cat /K M for CO 2 hydration for each of the mutants (H64A, H64A/N62H, and H64A/N 67H HCA II) superposed onto data for wild type HCA II. The data shows that the mutants have very similar values to wild type and this serves as a control indicating that the mutations do not cause gro ss structural changes that interfere with the first step of catalysis. The data in Figure 2-3 were f it to a single ionization with pK a values as given in Table 2-3 and these are characteristic of the pK a of the zinc-bound water (Lindskog, 1997). The pK a for the proton donor and acceptor in w ild type HCA II are very close in value to each other (6.9 and 7.2) and this mi ght be another requirement for efficient proton transfer between them. In both mutants the differences between the pK a values are larger compared to wild type and could c ontribute to the lower observed rate of proton transfer between His and the zinc-bound water (Table 2-3).

PAGE 57

42 The pK a values for the zinc-bound water given in Table 2-3 are in agreement with values obtained by others that measured th e pH dependent esterase activity of HCA II (Liang et al. 1993). For H64A/N67H HCA II it appears fr om the plot in Figure 2-3 that a single pK a is not sufficient to fit the data a nd this implies that a second ionization, probably that of His67, is influencing the k cat /K M The proton transfer activity data shown in Figure 2-4 indicate that all three mutants are significantly impaired compared to wild type HCA II, with H64A being the slowest. These data showed only slightly better rates for H64A/N 62H HCA II compared to H64A HCA II (Figure 2-4). However, ther e was significant enhancement of catalysis for H64A/N67H HCA II at pH < 6.5 compared with that of H64A HCA II. The solid lines of Figure 2-4 represent least-squares f its to eq 6, which assumes that the observed enhancement of R H2O /[E] above that of H 64A HCA II is due to the proton transfer activity of the inserted Hi s to the zinc-bound hydroxide. Other studies using stopped flow methods at steady state have shown that this mutant had a turnover number, k cat that is 20% of that observed for wild type (Liang et al. 1993). The data presented here was determined by 18 O exchange methods at chemical equilibrium and showed that this mutant had a maximal rate constant for proton transfer that is 25% that of wild type (Table 2-3). It should be noted that there is typically up to 20% error in these measurements and this is due to the scatter of the points and the limited pH range. Conclusion HCA II shows remarkable structural and ki netic stability over a wide range of pH (5.1 to 10.0). The only structural difference is the side chain of His 64 that displays two conformational states that are almost equally occupied at physiologica l pH. In wild type,

PAGE 58

43 H64A/N62H, and H64A/N67H HC A II structures there is ei ther no completed hydrogenbonded water chain between proton donor and accep tor or the distal water is very weakly bound to the proton shuttling residue. A His inserted at position 67 shows appreciable proton shuttling activity (25% of wild type), compared to a His at position 62 (4% of wild type). In these examples, the distance be tween the proton shuttling His and the zincbound solvent and the number of waters that sp ans this distance, might be more important for efficient proton shuttling than observi ng a complete hydrogen-bonded chain of waters to the His residue. The results suggest that th e optimal distance for the His to the zinc is between 6.6 and 7.5 and that the number of intervening water molecules should not exceed two for the support of effi cient and fast proton transfer.

PAGE 59

Table 2-1. Data set and model statistics for wild type HCA II from pH 5.1 to 10.0. HCA II pH 5.1 pH 6.1 pH 7.0 pH 7.8 pH 9.0 pH 10.0 Resolution () 20.0 2.00 (2.07 2.00)* 20.0 2.00 (2.07 2.00) 20.02.00 (2.07 2.00) 20.0 2.00 (2.07 2.00) 20.0 2.00 (2.07 2.00) 20.0 2.00 (2.07 2.00) Total Number of Unique Reflections 15898 (1546) 16365 (1575) 15878 (1554) 16212 (1539) 16751 (1665) 16714 (1598) Completeness (%) 93.1 (90.0) 95.9 (94.0) 93.5 (91.8) 95.1 (91.9) 98.6 (99.2) 97.8 (95.4) Redundancy 1.9 (1.9) 2.7 (2.7) 2.3 (2.3) 3.2 (3.2) 2.6 (2.5) 2.6 (2.6) R symm 0.063 (0.199) 0.120 (0.447) 0.051 (0.140) 0.073 (0.217) 0.092 (0.198) 0.081 (0.143) R cryst / R work 0.195/0.209 0.145/0.205 0.130/0.184 0.128/0.201 44 0.134/0.173 0.134/0.175 Ave B-factor ( 2 ) Main/side/solvent 16/20/28 23/27/30 18/22/28 14/18/27 17/21/28 17/21/28 Number of solvent 112 91 112 150 112 115 Data in parenthesis are for the highest resolution shell. R symm = I / R cryst = |F o | |F c | / |F o | R free is calculated the same as R cryst except it is for data omitted from refineme nt (5% of reflections for all data sets)

PAGE 60

45 Table 2-2. Data set and model statistics for H64A/N62H and H64A/N67H HCA II at pH 6.0 and 7.8. HCA II H64A/N62H pH 6.0 H64A/N62H pH 7.8 H64A/N67H pH 6.0 H64A/N67H pH 7.8 Resolution () 20.0 1.80 (1.86 1.80)* 20.0 1.90 (1.97 1.90) 20.0 1.63 (1.69 1.63) 20.0 1.80 (1.86 1.80) Total Number of Unique Reflections 22162 (2132) 18130 (1813) 29130 (2583) 21547 (2070) Completeness (%) 95.0 (91.60) 91.8 (92.9) 93.5 (83.7) 92.7 (89.0) Redundancy 3.0 (3.0) 2.5 (2.3) 3.1 (2.9) 2.8 (2.6) R symm 0.074 (0.395) 0.072 (0.316) 0.059 (0.329) 0.068 (0.292) R cryst / R work 0.171/0.206 0.168/0.217 0.178/0.210 0.166/0.209 Ave B-factor ( 2 ) Main/side/solvent 16/20/30 16/20/29 17/21/30 16/20/30 Number of solvent 129 137 120 132 Data in parenthesis are for the highest resolution shell. R symm = I / R cryst = |Fo| |Fc| / |Fo| R free is calculated the same as R cryst except it is for data omitted from refinement (5% of reflections for all data sets).

PAGE 61

46 Table 2-3. pH-Independent rate cons tants for proton transfer and pK a for proton donor and acceptors in wild type and mutant HCA II. Enzyme k B ( s) B -1 a (pK a ) His (pK a ) ZnH2O Zn-His distance Wild Type b 0.8 0.1 7.2 0.1 6.9 0.1 7.5 H64A ~0.02 N/a 6.9 0.1 N/a H64A/N62H 0.2 0.1 c 5.3 0.3 7.2 0.1 6.6 H64A/N67H ~ 0.03 c 5.7 0.4 7.3 0.1 8.2 a Values are from a least-squares fit of eq 6 to the data of Figures 2-3 and 2-4. b From Duda et al. 2001 c Values are uncertain due to the scatter of the data.

PAGE 62

47 Figure 2-1. Crystal structures of wild type HCA II active s ite. Panels (a) through (d) are shown in the same orientation, (a) pH 5.1, (b) pH 6.1, (c) pH 7.0, and (d) pH 10.0. Active site residues are shown in yell ow ball-and-stick and the zinc as a black sphere. Water molecules are as la beled and are shown as red spheres. 2F o -F c electron density maps for His64 are shown in blue and are contoured at 1.0 Dashed lines represent inferred hydrogen bonds based on geometry and distance between donor and acceptor atoms. Figure was generated and rendered with Bobscript and Raster3D (Esnouf, 1997; Merritt and Bacon, 1997).

PAGE 63

48 Figure 2-2. Crystal structures of mutant HCA II active site Panels (a) through (d) are shown in the same orientation, (a) H64A/N62H pH 6.0, (b ) H64A/N62H pH 7.8, (c) H64A/N67H pH 6.0, and (d) H64A /N67H pH 7.8. Active site residues are shown in yellow ball-and-stick and the zinc as a black sphere. Water molecules are as labeled and are shown as red spheres. 2F o -F c electron density maps for His62/His67 are shown in blue and are contoured at 1.0 Dashed lines represent inferred hydrogen bon ds based on geometry and distance between donor and acceptor atoms. Fi gure was generated and rendered with Bobscript and Raster3D (Esnouf 1997; Merritt and Bacon, 1997).

PAGE 64

49 Figure 2-3. The pH profiles for k cat /K M catalyzed by wild type and mutant HCA II. ( ) Wild type, ( ) H64A, ( ) H64A/N62H, ( ) H64A/N67H. Data were obtained at 25 in the absence of exogenous buffe rs using a total concentration of all species of CO 2 of 25 mM, with the ionic stre ngth maintained at 0.2 M by addition of sodium sulfate. The solid lin es are fit to a singl e ionization (eq 5) with the pK a given in Table 2-3.

PAGE 65

50 Figure 2-4. The pH profiles for RH 2 O/[E] catalyzed by wild type and mutant HCA II. ( ) Wild type, ( ) H64A, ( ) H64A/N62H, ( ) H64A/N67H. Data were obtained at 25 in the absence of exogenous buffe rs using a total concentration of all species of CO 2 of 25 mM, with the ionic stre ngth maintained at 0.2 M by addition of sodium sulfate.

PAGE 66

CHAPTER 3 STRUCTURAL AND KINETIC EFFECTS OF HYDROPHOBIC MUTATIONS IN THE ACTIVE SITE OF HUMAN CARBONIC ANHYDRASE II In Chapter 2, studies of the structural and kinetic effects of moving the proton shuttle to various locations in the activ e site of human carbonic anhydrase II (HCA II) were presented. Kinetic and structural data over various pHs were obtained and correlated with observed changes in the active s ite and proton transfer rates. In Chapter 3, a different approach to studying proton transf er was undertaken and is discussed. Sitedirect mutagenesis of several key active site residues were performed with the expectation that these mutations would affect proton transfer rates a nd the architecture of the hydrogen bonded solvent network found in the active site of HCA II. Introduction HCA II is the most efficient of all the HCAs with a maximal turnover rate of 10 6 s -1 (Khalifah, 1971). Catalytic ra tes of HCA II and the mechanism has been described in detail in Chapters 1 and 2. Briefly, the s econd part of the catalysis by HCA II involves two proton transfer steps: the first between the zinc-bound solvent and an internal proton acceptor, His64, and the second between His 64 and an exogenous proton acceptor such as the bulk solvent of buffer (Silv erman, 1982; Silverman and Lindskog, 1988, Tu et al. 1989). The first proton transfer event involve s several water molecules (W1, W2, W3a, and W3b; Figure 3-1) and this network is also described in detail in Chapter 1 (Fisher et al. 2005). With the exception of W2, all ot her water molecules are coordinated by hydrogen bonding interactions to several key active site residues such as Tyr7, Asn62, 51

PAGE 67

52 and Asn67. Figure 3-1 shows a view of the w ild type HCA II active site with these residues and water molecules as labeled. The proton shuttle residue, His64, is locat ed on the edge of th e active cavity, about 8 away from the zinc. This distance makes a direct proton transfer impossible and as a result the excess proton has to exit the act ive site via the conserved and ordered water molecules shown in Figure 3-1 (Steiner et al. 1975; Eriksson et al. 1988). In several crystal structures determined at various pHs, His64 has been shown to occupy two distinct conformations, the so-called in a nd out positions. However, the presence of the dual conformation is not always observed and under some conditions (such as pH or in the presence of inhibitors) it appears to be either all in the in or all in the out position (Nair and Christianson, 1991; Krebs and Fierke, 1991; Fisher et al. 2005). It should be noted that at the re solution of these studies, an occupancy change of 0-10% would be impossible to observe. To better understand the role of the conserved water ne twork and the implications of side chain mobility of His64 for efficient proton transfer, three hydrophilic amino acids in the active site of HCA II were replaced by site-directed mutagenesis with hydrophobic residues: Y7F, N62L, and N67L. Effects of these substitutions were evaluated by measuring kinetic proton transfer rates, the conformation of His64, and the structural effects on overall active site ar chitecture as well as the water network. Materials and Methods Enzymes Plasmids with the appropriate mutations in the cDNA of HCA II were produced by site-directed mutagenesis using the Qiagen QuikChange kit. Mutagenic primers were designed and the procedures pe rformed as per manufacturers instructions. Residue Tyr7

PAGE 68

53 was mutated to a Phe, and residues Asn62 and Asn67 were separately mutated to Leu. All mutations were verified as correct by seque ncing the entire codi ng region of the HCA II plasmid. Protein expression was performed in E. coli BL21(DE3) pLysS cells and the resulting enzymes were purified using affinity chro matography (Khalifah, 1977). A detailed description of these procedures can be found in the Materials and Methods section of Chapter 2. Prior to any crystalliz ation or activity assays the purity of the protein was verified by electrophoresis on a 12% polyacrylamide gel stained with Coomassie. The concentration of HCA II was determined by measuring the absorbance at 280 nm and using a molar absorptivity of 5.4 x 10 4 M -1 cm -1 (Coleman, 1967). Crystallography Crystals of mutant HCA II were obtain ed using the hanging drop method at room temperature (McPherson, 1982). The first set of crystals of all three mutants at pH 8.2 were obtained by mixing 5 l of protein (10-15 mg/ml con centrations in 50 mM TrisHCl, pH 7.8) with 5 l of the precipitant solution (50 mM Tris-HCl, pH 8.2, 2.5-2.9 M ammonium sulfate) against 1000 l of the precipitant solutio n. For the N62L and N67L structures determined at pH 6.0, previously gr own crystals were soaked overnight in 50 mM sodium acetate, pH 6.0, 2.6 M ammonium sulfate. For Y7F at pH 10.0, previously grown crystals were soaked overnight in 50 mM CAPS, pH 10.0, 2.6 M ammonium sulfate. Y7F crystals were also grown using a different precipitant solution that consisted of 100 mM Tris-Cl, pH 9.0, 1.3 M sodium citrat e. Useful crystals under all conditions appear within 7 days of crystallization set-up. X-ray diffraction data was collected usi ng a R-AXIS IV++ image plate system with Osmic mirrors and a Rigaku HU-H3R Cu rotating anode operating at 50 kV and 100mA. The detector to crystal distance was set to 100 mm. Each data set was collected at room

PAGE 69

54 temperature from 1-3 crystals mounted in quart z capillaries. The oscillation steps were 1 with a 7 minute exposure time. Diffraction data of the Y7F crystals grown with sodium citrate were collected at the Advanced Photon Source (APS) on beamline SER-CAT 22ID using a wavelength of 0.979 The crysta l to detector distance was set at 120 mm with a 1 second exposure time per image. Im ages were recorded on a Mar300 CCD with 1 oscillations. X-ray data processing was performed using DENZO and scaled and reduced with SCALEPACK (Otwinowski and Minor, 1997). Structure determination and refinement was carried out with Crystallographic and NMR System (CNS) version 1.1 (Brnger et al. 1998). All manual building was performed with Coot (Emsley and Cowtan, 2005). The structure of wild type HCA II (PDB accession code 1TBT) was isomorphous with all the data collected and was used to phase the data sets (Fisher et al. 2005). To avoid phase bias of the model, th e zinc atom, water molecules, and mutated side chains were removed. After one cycle of rigid body refinement, annealing by heating to 3000 K with gradual cooling, geometry restrained position refinement, and temperature factor refinement, the F o -F c and 2F o -F c Fourier maps were maps generated. Visual inspection of these maps clearly showed the electron density for the zinc atom and the mutated side chains and these were subsequently incorporated into their respective models. After several cycles of refinement, solvent molecules were incorporated into the models using the automated water-picking program implemented in CNS until no more waters were found at the 2.0 level. Refinement of the models continued until the Rfactors converged. Tables 3-1 and 3-2 show the data set and final model statistics. Kinetics and Activity Analysis Determination, by membrane-inlet mass spectrometry, of R 1 and R H2O was performed using the 18 O exchange method by Dr. Chingkuang Tu in the lab of Dr. Silverman

PAGE 70

55 (Silverman, 1982). The methodology for obtaining th ese rate constants was described in detail in the Materials and Methods section of Chapter 2 and will not be repeated here. Initial rates of CO 2 hydration were measured by following the change in absorbance of a pH indicator on an Applied Photophysics (SX.18MV) stopped-flow spectrophotometer and these assay were also done by Dr. Chingkuang Tu in the Silverman lab. The pK a values and wavelengths for the pH indicator-buf fer pairs used to create pH profiles were as follows: MES (pK a = 6.1) and chlorophenol red (pK a = 6.3), = 574 nm; MOPS (pK a = 7.2) and p-nitro phenol (pK a = 7.1), = 401 nm; HEPES (pK a = 7.5) and phenol red (pK a = 7.5), = 557 nm; TAPS (pK a = 8.4) and m-cresol purple (pK a = 8.3), = 578 nm; CHES (pK a = 9.3) and thymol blue (pK a = 8.9), = 596 nm. Final buffer concentrations were 50 mM, and total ionic strength was kept at 0.2 M by the addition of sodium sulfate. CO 2 solutions were prepared by bubbling CO 2 into water at 25 C with final concentrations after mixing ranging from 0.7 17 mM. The mean initia l rates at each pH were determined from 5 to 8 reaction traces comprising the initia l 10% of the reaction. The uncatalyzed rates were determined in a similar manner and subtracted from the total observed rates. Determination of the kinetic constants k cat and k cat /K M were carried out by a nonlinear least-squares method (E nzfitter, Elsevier-Biosoft). Results and Discussion Structural Effects of Hydrophobic Mutations All crystals were isomorphous and belonged to the space group P2 1 with the following mean unit cell dimensions: a = 42.7 0.4 b = 41.6 0.5 c = 72.9 0.2 = 104.6 0.5. The HCA II mutant data sets at pH 8.2 were processed to 1.65 1.70 resolution, while the other pH data sets were processed to 1.8 resolution. A summary of the data set and final model statis tics is given in Tables 3-1 and 3-2.

PAGE 71

56 The N62L mutant structure determined at pH 8.2 shows that the water network is conserved compared to wild type (Figure 3-1 and 3-2 (a)) However, as expected, the hydrogen bond between Leu62 and W3b is lost. Th e side chain of His64, which is often observed in a dual conformation in wild type structures, is completely in the in position. Leu62 has moved away from the solv ated area in the act ive site and now occupies a hydrophobic region than also co ntains Leu60. The displacement of Leu62 removes the hydrophobic side chain away from the water network and as a result does not affect it at all. In contrast to N62L, the N67L mutant di splays a disrupted water network compared to wild type HCA II (Figure 3-1 and 3-2 (b)) and His64 appears to be all in the out position. Unlike the orient ation of Leu62 that has been rota ted away from the active site compared to an Asn at that position, Leu67 o ccupies a similar position to Asn67. It might be the presence of this hydrophobic residue that disrupts the water network. The differences in the His64 side chain orient ations and water network should be due to electrostatic changes in the active site and not steric effects as the effective size of the side chains are similar. In the Y7F mutant structure determined at pH 8.2, the water network appears mainly conserved. However, due to the lo ss of the hydroxyl group by changing Tyr7 to a Phe, the essential hydrogen bond to W3a is missi ng and as a result this water molecule is not observed here anymore. Phe7 occupies a si milar position compared to Tyr7 as seen in wild type (Figure 3-1 and 3-3(a)). His64 in this mutant is al so all in the in position and appears to make a very weak hydrogen bond to W2.

PAGE 72

57 An unusual observation is the presence of a sulfate ion bound to the zinc atom. It has been observed that sulfate will only bind to wild type HC A II at very low pH (~ pH 5.1) and it seems that this mutation could be facilitating a higher affinity of the active site for sulfate (Fisher et al. 2005). From other structural studies of wild t ype HCA II it is known th at the presence of sulfate does not affect the wate r network but it is not known if the presence of sulfate affects the orientation of His64 (Fisher et al. 2005). As the Y7F structure at pH 8.2 showed a sulfate bound and His64 is only in the in position, the pH of the other two mutant crystals was lowered to pH 6.0 to get sulfate to bind. These structure revealed that sulfate did bind at the lower pH and that it did not affect the orientation of His64 as His64 was stil l either all in the in (as in the case of N62L) or all in the out (as in the case of N67L) (Figure 3-2 (c) and (d)). The water networks in both N62L and N67L structures at pH 6.0 were completely disrupted (Figure 3-2 (c) and (d)). Leu62 at pH 6.0 had moved co mpared to the struct ure at pH 8.2 and is now pointing into the active site and might be a contributing factor in the water network disruption. The side chain of Leu67 at pH 6.0 also moved compared to its structure at pH 8.2 (Figure 3-2 (c) and (d)). Apart from repellin g solvent from these areas, it is not clear if the movement of these two residu es at lower pH is significant. In an attempt to remove the sulfate in th e Y7F mutant, the pH was increased to pH 10.0. Surprisingly, this structure showed that sulf ate was still present even at this extreme pH. However, His64 and the water network was undisturbed and appeared the same as the structure at pH 8.2 (Figure 3-3 (a) and (b)). These observations rule out the idea that the presence of sulfate affects the orientati on of His64 as it is observed that His64 can

PAGE 73

58 occupy the in conformation in the absence or presence of sulfate (Figure 3-2, compare panels (a) and (b) with (c) and (d)). To furt her study this, it was necessary to see the active site of Y7F in the absence of any sulf ate. To this end, a different crystallization condition using sodium citrate instead of ammonium sulfate was tried and made it possible to determine the structure of Y7F at ph 9.0 in the absence of any sulfate (Figure 3-3 (c)). Except for the lack of sulfate it can be seen that His64 and the water network appears the same as they did in the other two stru ctures at pH 8.2 a nd 10.0 (Figure 3-3 (b) and (c)). The observation of different orientations of His64 can be used to comment on the relationship between this orientation and the efficiency of proton transfer. Inspection of His64 in the more efficient wild type, Y7F, and N62L HCA II this residue occupies the in conformation. This is in contrast with the less efficient N 67L HCA II where His64 appears more in the out conformation. These observations alone do not provide adequate evidence that His64 in the in posi tion is necessary or sufficient for fast proton transfer rates. Kinetic Effects of Hydrophobic Mutations The measurements of k cat /K M for the hydration of CO 2 was obtained by both 18 O exchange methods at chemical equilibrium and by stopped-flow methods at steady state. As expected, these data shown in Table 3-3 indicate no significant changes between wild type and the mutants, however, the values for N67L appear a bit lowe r than the others in Table 3-3. This was found for both the maximal values of k cat /K M and for the apparent pK a near 7.0 of the zinc-bound water calcu lated from the pH profile of k cat /K M (Figure 34). This supports the structural data that there are no substantial changes in the active site structures and implies no changes in the ch emistry of catalyzed interconversion of CO 2

PAGE 74

59 and bicarbonate. It is not surprising that th ese mutations do not affect the first part of catalysis as residues 7, 62, and 67 are more than 7 away from the active site zinc. Other studies involved mutati ng residues closer to the zinc, such as Thr199 and Glu106, and the results show that those residue s are essential for efficient interconversion of CO 2 and bicarbonate (Liang et al. 1993; Krebs et al. 1993). These observations support the mechanism of the fi rst stage of catalysis which involves a direct nucleophilic attack by the zinc-bound hydroxide on CO 2 with Thr199 and Glu106 enhancing the nucleophilicity of the zinc-bound hydroxide and optimally orienting it for reaction with CO 2 (Merz, 1990). In the second stage of catalysis long-range proton transfer occurs between the zincbound solvent and the bulk solution using inte rvening water molecules and His64 as a proton shuttle (Tu et al. 1989; Lindskog, 1997). In c ontrast to the first stage of catalysis, as reflected by the values of k cat /K M (Figure 3-4), the mutants in this study all displayed altered proton transfer rates when compared to wild type HCA II (Figure 3-5). The effect is best seen when inspec ting the pH profiles of R H2O /[E], the rate constant for the release of isotopically labeled water from the enzyme to solution, which in turn depends on the rate of proton tran sfer (Figure 3-5). Proton transfer by N62L HCA II is the most complicated and difficult to interpret as it displays two small bell-shaped curves and seems to have lost its pH dependence. N67L HCA II shows a similar prof ile to wild type HCA II, just at a lower rate. The most surprising result here is the data for Y7F HCA II as it has a rate constant for proton transfer that is appreciably larger than for wild type (Table 3-3 and Figure 3-5).

PAGE 75

60 Addition of an exogenous activator of pr oton transfer activity, 4-methylimidazole (4-MI), does not appear to appreciably increase the rate of N62L or N67L HCA II (Figure 3-6). Wild type HCA II is activated, compared to N62L and N67L HCA II, with maximal activation occurring at around 50 mM 4-MI. Y7F HCA II is activ ated to a larger extent compared to wild type with the addition of 4-MI (Figure 3-6). These data implies that the active site of Y7F HCA II is more accessible to external buffers and that the removal of a hydroxyl group, by mutating Tyr7 to a Phe, thus enhancing the activation by 4-MI over that observed for wild type N62L, and N67L HCA II. Solvent Structure and Implications for Proton Transfer In crystal structures of wild type HCA II, each of the side chains Tyr7, Asn62, and Asn67 appear to make hydrogen bonds with wa ter molecules W3a and W3b (Figure 3-1; Fisher et al. 2005). In fact, the side chains of Asn62 and Asn67 interact with the same water molecule, W3b. This water structure in N62L at pH 8.2 seems unaffected by the mutation as just the hydrogen bond to W3b is lost but the water is still present due to its interaction with Asn67. The reas on why Leu62 does not interfere w ith the water is that it has moved away from the solvated active site into a more hydrophobic region also occupied by Leu60. Overall, the water structur e seems conserved with that observed in wild type HCA II. However, when the pH was lowered to pH 6.0 this residue moved back into the active site and subsequently completely disrupted the solvent structure (Figure 3-2 (c)). The data for N62L, showi ng His64 in the in conformation and an intact solvent structure at pH 8.2, would suggest an unchange d proton transfer rate. Yet, probably due to electrostatic effects, the side chain of Leu62 and the water network is easily disrupted upon a change in pH and this instability of the so lvent structure could account for the slower measured proton transf er rate compared to wild type HCA II.

PAGE 76

61 In contrast to N62L, Leu67 in the N67L HC A II mutant extends into the active site and effectively disrupts the or dered water structure at both pH 6.0 and 8.2 (Figure 3-2 (b) and (d)). The only water molecule that is c onserved is the zinc-bound solvent that could be either a hydroxide or a water molecule. The other two waters have been displaced and do not make contact with any active site resi dues shown here. Also, as mentioned above, His64 is observed to be always in the out positi on, regardless of the presence of sulfate bound to the zinc. Proton transfer by this muta nt displays the characteristic bell-shaped pH dependency, just at a lower rate compared to wild type HCA II (Figure 3-5). Due to these observations it is likely that the typica l pH dependency of proton transfer does not correlate with an ordered solvent st ructure in the HCA II active site. The Y7F HCA II mutant shows the most efficient proton transfer, even when compared to wild type HCA II. At pH 8.2, 9.0, and 10.0 the solvent structure is highly conserved and His64 is always in the in conformation. Th e net effect of the loss of W3a, due to the Tyr7 to Phe mutation, produ ces a single-line array of water molecules that bridges the zincbound solvent to His64. It has been suggested by others that a single arra y of waters is much more efficient than a branched system at facilitating proton transfer. Also, an unbranched hydrogenbonded array of water can promote most e fficient proton transfer by a concerted mechanism rather than a step-wise and appears to proceed through an intermediate with partial hydronium ion character (Cui and Karplus, 2003; Voth et al. 1998). There are various explanations for the very fast moveme nt of protons in solu tion and these include Grotthusss idea of so-called st ructural diffusion. Other struct ural models for the proton in solution, or hydrated proton, incl udes the elementary proton (H + ), hydronium ion

PAGE 77

62 (H 3 O + ), the Zundel, and Eigen catio n models (Eigen, 1964; Marx et al. 1999). The Zundel cation is a H 5 O 2 + complex where the proton is shared between two water molecules. The Eigen cation consists of a H 9 O 4 + complex where the central H 3 O + is strongly hydrogen bonded to three H 2 O molecules in a secondar y hydration configuration (Eigen, 1964; Marx et al. 1999). According to Eigen, weaker hydrogen bonds are formed at the periphery of the H 9 O 4 + complex and that directed breaking and formation of these bonds lead to structural di ffusion of the entire comple x. The diffusion of the hydrate complex is the rate-limiting step in the mob ility of a proton in solution (Eigen, 1964). Recent studies with infrared spectroscopy of acid-base proton transfer reactions suggest the existence of a fairly stable H 3 O + intermediate structure that is coordinated in an Eigen configuration by three water molecule s. Other conclusions were that the proton transfer between acid and bases occur by a sequential, Grotthus-type proton hopping mechanism that is mediated by hydrogen bonded water networks or bridges (Mohammed et al. 2005). The highly conserved core wate r structure as seen in wild type HCA II consists of W1, W2, W3a, and W3b (Figure 37, green box). It shares many structural features of an Eigen cation in that the geometry of the stru cture is almost planar with ~120 between W2 and the surrounding waters. Also, the hydr ogen bond distances be tween W2 and the others is ~2.8 for all three. Based on the structural properties of the solvent core it could be an Eigen cation with W2 repres enting the excess hydrated proton as a H 3 O + The Eigen cation is relatively stable entity a nd can be observed on th e nanoto picosecond scale (Mohammed et al. 2005). This core structure is a somewhat branched water network in that the excess proton can get acce ss to His64 through either one of W2, W3a,

PAGE 78

63 or W3b. The X-ray crystal structure represents an energy minimized, stable structure and perhaps this solvent structure kinetically traps the excess proton. It is likely that in solution during a normal catalytic cycle, thes e solvent structures are constantly being broken and reformed and the crystal structure just shows the most energetically stable species. In the Y7F HCA II mutant, W3a has been removed by the introduction of the hydrophobic Phe residue. This results in a line ar array with only W2 having access to His64 and could also disrupt the stability of the Eigen cation as seen in wild type HCA II. This could serve as a structural explanation for why the proton transfer proceeds so much faster in this mutant compared to wild type HCA II. Even though this mutant is faster than wild type, natural selection did not choos e to have a Phe at position 7 and a possible reason is that it appears to be somewhat unstable, especially at low pH (Figure 3-5). Additionally, in contrast to wild type HCA II, where the distance between W2 and His64 is too long (~ 3.5 ), the distance between W2 and His64 is now ~ 3.2 and constitutes a weak hydrogen bond and this could contribute to the enhanced proton transfer rates. Conclusion To disrupt the hydrogen bonded water network in the active site of HCA II, several key catalytic amino acids, involved with co ordinating these waters, were mutated to hydrophobic residues that are similar in si ze. Asn62 and Asn67 were mutated to Leu (N62L and N67L) and Tyr7 was mutated to a Ph e (Y7F). X-ray crystal structures and rate constants for the hydration of CO 2 and proton transfer were determined for all three mutants at different pHs. The structural and kinetic data for N62L and N67L shows that the water networks were readily disrupted, especially at low pH, and both displayed considerably low proton transfer rates compar ed to wild type and Y7F HCA II. The most surprising result was the enhanced proton transfer rate over th at of wild type observed for

PAGE 79

64 Y7F HCA II. This mutant also displayed a si milar water network as wild type HCA II, except for the loss of one active site water. To examine whether the presence or absence of sulfate affected the orientat ion of His64, structures of the mutants were determined at several different pHs. Lowering the pH for N62L and N67L HCA II caused sulfate to bind and it was determined that it had no eff ect on the orientation of His64. The structure of Y7F HCA II at pH 8.2 and 10.0 showed that sulfate was still bound. Subsequently, Y7F HCA II was crystallized in the absence of any ammonium su lfate and this too showed that sulfate had no effect on His 64. Correlations between the structural and kinetic data suggest that a single, linear array of water bridging His64 and the zinc-bound solvent might be more efficient at proton transfer than a branched structure.

PAGE 80

65 Table 3-1. Data set and final model sta tistics for N62L HCA II and N67L HCA II. HCA II N62L pH 6.0 N62L pH 8.2 N67L pH 6.0 N67L pH 8.2 Resolution () 20.0 1.80 (1.86 1.80)* 20.0 1.70 (1.76 1.70) 20.01.80 (1.86 1.80) 20.0 1.65 (1.71 1.65) Total Number of Unique Reflections 20541 (2031) 25698 (2422) 21266 (2142) 27340 (2590) Completeness (%) 88.2 (86.9) 91.0 (87.4) 91.7 (92.6) 93.1 (88.9) Redundancy 4.1 (4.0) 3.2 (3.3) 3.4 (3.2) 3.1 (3.1) R symm 0.106 (0.339) 0.085 (0.448) 0.075 (0.349) 10.8 (28.6) R cryst / R work 0.189 / 0.226 0.179 / 0.217 0.187 / 0.224 18.6 / 20.1 Ave B-factor ( 2 ) Main/side/solvent 16.8 / 20.1 / 28.7 16.1 / 19.5 / 29.6 18.6 / 22.2 / 29.5 18.2 / 21.4 / 29.6 Number of solvent 108 132 97 116 Data in parenthesis are for the highest resolution shell. R symm = I / R cryst = |F o | |F c | / |F o | R free is calculated the same as R cryst except it is for data omitted from refinement (5% of reflections for all data sets).

PAGE 81

66 Table 3-2. Data set and final model statistics for Y7F HCA II. HCA II Y7F pH 8.2 Y7F pH 10.0 Y7F (citrate) pH 9.0 Resolution () 20.0 1.70 (1.76 1.70)* 20.0 1.80 (1.86 1.80) 20.01.15 (1.19 1.15) Total Number of Unique Reflections 25613 (2444) 20874 (1951) 85073 (8284) Completeness (%) 92.7 (89.8) 89.4 (83.9) 99.1 (96.8) Redundancy 3.9 (3.9) 3.1 (3.2) 4.2 (3.2) R symm 0.060 (0.322) 0.071 (0.238) 0.079 (0.364) R cryst / R work 0.181 / 0.199 0.179 / 0.200 0.193 / 0.184 Ave B-factor ( 2 ) 16.6 / 20.1 / 29.4 15.1 / 18.5 / 28.3 10.5 / 13.5 / 23.4 Main/side/solvent Number of solvent 115 115 270 Data in parenthesis are for the highest resolution shell. R symm = I / R cryst = |F o | |F c | / |F o | R free is calculated the same as R cryst except it is for data omitted from refinement (5% of reflections for all data sets).

PAGE 82

67 Table 3-3. Maximal values of ra te constants for hydration of CO 2 proton transfer, and pK a of the zinc-bound water. k cat /K M (M -1 s -1 ) a pK a R H2O ( s -1 ) Wild type 120 6.9 0.1 0.3 Y7F 120 7.1 0.1 1.0 N62L 140 7.3 0.1 0.1 N67L 90 6.5 0.2 0.1 a The standard errors for k cat /K M are 10% or less.

PAGE 83

68 Figure 3-1. Active site of wild type human car bonic anhydrase II. The zinc atom is shown as a black sphere and active site residues are in yellow ball-and-stick and are as labeled. Water molecules are shown as red spheres and are as numbered. Figure was generated with BobScript and rendered w ith Raster3D (Merritt and Bacon, 1997; Esnouf, 1997).

PAGE 84

69 Figure 3-2. Active site of N62L and N67L at pH 8.2 and pH 6.0. (a) N62L HCA II at pH 8.2; (b) N67L HCA II at pH 8.2; (c) N62L at pH 6.0; (d) N67L HCA II at pH 6.0. Active site residues ar e shown in yellow ball-andstick and are as labeled. The black sphere is the zinc atom and wa ters are shown as red spheres. Sulfate ions are in green ball-and-stick and are labeled. Figure was generated and rendered with BobScript and Raster3D respectively (Esnouf, 1997; Merritt and Bacon, 1997).

PAGE 85

70 Figure 3-3. Active site of Y7F HCA II at various pH. (a) Y7F HCA II at pH 8.2; (b) Y7F HCA II at pH 10.0; (c) Y7F HCA II at pH 9.0 with no sulfate. Active site residues are shown in yellow ball-and-stick and are as labeled. The zinc atom and water molecules are shown as black and red spheres, respectively. Figure was generated with BobScript and re ndered with Raster3D (Esnouf, 1997; Merritt and Bacon, 1997).

PAGE 86

71 Figure 3-4. The pH profiles for k cat /K M catalyzed by wild type and mutant HCA II. ( ) Wild type, ( ) N67L, ( ) N62L, ( ) Y7F HCA II. Data we re obtained at 25 in the absence of exogenous buffers using a total concentration of all species of CO 2 of 25 mM, with the ionic strengt h maintained at 0.2 M by addition of sodium sulfate. The solid lines are fit to a single ionization with the pK a given in Table 3-3.

PAGE 87

72 Figure 3-5. The pH profiles for RH 2 O/[E] catalyzed by wild type and mutant HCA II. ( ) Wild type, ( ) N67L, ( ) N62L, ( ) Y7F HCA II. Data we re obtained at 25 in the absence of exogenous buffers using a total concentration of all species of CO 2 of 25 mM, with the ionic strengt h maintained at 0.2 M by addition of sodium sulfate.

PAGE 88

73 Figure 3-6. Activation of RH 2 O/[E] catalyzed by wild t ype and mutant HCA II by the addition of 4-methylimidazole. ( ) Wild type, ( ) N67L, ( ) N62L, ( ) Y7F HCA II. Data were obtained at 25 by the 18 O exchange method using a total concentration of all species of CO 2 of 25 mM and addi ng increasing amounts of 4-methylimidazole, with the ionic st rength maintained at 0.2 M by addition of sodium sulfate.

PAGE 89

74 Figure 3-7. Active sites of wild type and Y7F HCA II. (a) Wild type, and (b) Y7F HCA II. Residues are shown in yellow ball-andstick with the zinc atom as a black sphere. Water molecules and inferred hydrogen bond are shown as red spheres and dashed lines, respectively. Th e green boxes surround the core solvent structure in the active sites. Figure was generated with BobScript and rendered with Raster3D (Esnouf, 1997; Merritt and Bacon, 1997).

PAGE 90

CHAPTER 4 WORKING TOWARDS A NEUTRON STRUCTURE OF PERDEUTERATED HUMAN CARBONIC ANHYDRASE II Chapters 2 and 3 included work toward s a better understanding of the proton transfer mechanism that is part of the cat alysis by human carbonic anhydrase II (HCA II). The methods employed included kinetic meas urements of rate constants for proton transfer as well as detailed structural st udies of the active si te residues and water structure. However, X-ray crystallography is not the best technique for investigating waters because hydrogen atoms, which make up about half of all the atoms in a protein, are virtually invisible to X-ra ys. Neutron macromolecular crystallography is currently the only direct technique for obs erving hydrogen H (or deuteriu m D) atoms and can give atomic details even at modest resolutions (~ 2.0 ). This chapter will describe how and why neutrons are different from X-rays and will include recent progress towards obtaining a neutron structure of perdeuterated HCA II. Introduction Neutron crystallography can provide unique information about hydration states of proteins, ionization states of key catalytic amino acids, water molecules, and position of H atoms. H atoms are a fundamental part of many enzymatic processes, some of which involve proton transfer between residues in th e protein and substrates products, ligands, as well as mediating the binding, through wate r molecules, of pharmacological agents (Langan et al. 2004). 75

PAGE 91

76 Assigning positions of hydrogen atoms is possible in X-ray crystallography if atomic or sub-atomic resolution data can be collected ( 1.2 ). Even in the case of high resolution X-ray diffraction it is often still ha rd or impossible to confidently assign the positions of hydrogen H atoms associated with water molecules. H atoms that are associated with well-ordered parts of the prot ein (such as the main or side chains) can be assigned, but H atoms in disordered regions or part of water molecules are much harder to assign (Coates et al. 2001; Habash et al. 2000). The extent of diffraction by X-rays depe nds on the number of electrons. As H and D have only 1 electron, they di ffract X-rays very poorly comp ared to more electron-rich atoms such as carbon C, nitrogen N, oxygen O, and sulfur S. The scattering amplitude of atoms by X-rays increase linearly with increas ing numbers of electrons (Table 4-1). This is in contrast to how atoms behave in an neutron beam as each atom has a unique scattering amplitude that is a property of the nucleus and has to be determined experimentally (Table 4-1). H and/or D atoms are more easily located by neutron analysis as the scattering lengths of H (-3.7 x 10 -15 m, or .7 fm) is equal in magnitude but opposite in sign when compared to other atoms found in proteins. D atoms have a scattering length (6.7 fm) that is very close to the range of O (5.8 fm), N (9.4 fm), C (6.6 fm) and S (3.1 fm) making it easier to locate by this method (Coates et al. 2001). The real strength of using neutrons over X-rays are the different scattering properties of H and D atoms. The negative scattering amplitude of H with neutrons can potentially be exploited to observe them by just looking at negative nuclear density. This kind of contrast labeling has been successfully used to determin e levels of H/D exchange and using the data for further

PAGE 92

77 elucidation of the mechanism or activity of enzymes (Habash et al. 1997). A classic example is the study of a carbon monoxide myog lobin derivative that was subject to H/D exchange prior to neutron diffraction. Th e results gave information on the hydration shells around the protein as we ll as region in the heme binding site that did not exchange (Norvell et al. 1975). However, there are two prob lems with having H in proteins crystals for diffraction experiments. The first is the low coherent scattering cross section of H atoms and the second is th e large incoherent sc attering cross sectio n that leads to a very large undesirable background (Table 4-1). The heavy isotope of H, deuterium D, beha ves the same with X-rays but have very different properties with ne utrons. D atoms have a scattering amplitude similar to the other atoms found in proteins and also a small incoherent scattering cross section (Table 4-1). For these reasons, it is often most favorable to replace as many H atoms with D atoms as is practical by either H/D exchange methods, such as soaking crystals in deuterated solutions, or by pr oducing deuterated materials. As the H/D exchange process does not lead to fully deuterated (perdeuterat ed) materials, most often it is better to synthesize or purchase perdeuterated materials. Perdeuteration, as opposed to just soaking crystals in deuterated solutions, and subseque nt neutron diffraction data collection vastly improves location of D atoms because, as mentioned above, D and O have similar scattering lengths and both are positive in si gn. The resulting nuclear density indicates the orientation of D 2 O and thus allows the location of the two D atoms (Coates et al. 2001; Habash et al. 2000; Myles et al. 1998). In the case of H at oms in water molecules the negative density from the H can smear out or cancel the positive s cattering contribution from the O atoms (Habash et al. 2000). Another bonus to using perdeuterated materials

PAGE 93

78 are estimates that, due to the small incohere nt scattering cross section of D atoms, one can achieve at least a 40-fold reduction in the background compared to having any H atoms around. Also, having perdeuterated material s cuts back on the n eed for very large crystals and it is now possible to collect data from crystals that are only 0.15 mm 3 as in the case of aldose reductase (Niimura et al. 2005). Neutron macromolecular crystallography is a powerful technique that complements high resolution X-ray crystallography well as a combination of the two techniques allows analysis of key hydrogen atom positions and solvent structure (Myles et al. 1998). Before collecting neutron diffraction data, crys tals are usually subjec ted to H/D exchange by soaking the crystals in deuterated solutions Some of the advantag es to this procedure include the fact that very high incoherent background scattering from hydrogen is eliminated and that it is a lo t cheaper than purchasing deuterated materials. Soaking also allows the exchange of solvent accessible hyd roxyl and amide groups as well as replacing H 2 O with D 2 O and this information can be used assess solvent accessibility, flexibility or disorder of a protein (Niimura et al. 2005). The major drawback of neutron diffraction is the low flux of neut rons available for sample irradiation. The diffraction intens ity can be calculated from eq 4-1: I = (I 0 x F 2 x V x A) / (v 0 ) 2 (4-1) In eq 4-1, I is the diffraction intensity, I 0 incident neutron intensity, F 2 structure factor, V volume of the crystal, A de tector area covered by sample, and v 0 is the volume of the unit cell (Niimura, 1999). This mean s that the diffraction intensity strongly depends on the size of the crystal, intensity of the incident neutron beam, and the area detector. Recent advances in area detector technology, data collection, and processing

PAGE 94

79 strategies in neutron protein crystallography allow st udies of larger biological molecules and smaller crystals than prev iously thought possible. The biggest challenge is to grow a very large, single crystal (~ 1 mm 3 ) with a small unit cell volume (Niimura, 1999). A data collection strategy that is used to optimize the number of neutrons available, is Laue diffractometry where a range of wavelengths are used. This is in contrast to a monochromatized beam, as used in most X-ray diffraction experiments, where most of the useful neutrons are removed and only a single wavelength is used. Due to the low flux of the beam, this is a wasteful methodol ogy and Laue diffraction allows the user to simultaneously measure diffraction in different directions from different lattice planes. Laue methods allow for a more efficient surv ey of reciprocal space with fewer crystal settings, and along with large area detectors e ffectively increases the flux on the sample (Myles et al. 1998; Schultz et al. 2005). The single biggest advance in neutron protein crystallography, has been the design of better area detectors. Ne utron imaging plate (NIP) tec hnology has been developed in which a neutron converter, such as 6 Li or Gd, is mixed with photostimulated luminescence materials that are layered on a flexible plastic backing. The dynamic range, spatial resolution, and flexibility make the NI P suitable for detecting diffracted neutrons. NIPs are available in different sizes ( 200 mm x 200 mm and 200 mm x 400 mm) and these can be arranged side-by-side to make a combined detector of any desired size (Niimura, 1999). The neutron-sensitive image plate in use at the In stitut Laue-Langevin (European Molecular Biology Laboratories, Grenoble) consists of 4 large Gd-doped phosphor plates that are packed together to give an active area of 400 x 800 mm (Myles, 1998). Image plates have high counting rate s and good dynamic range, but they cannot

PAGE 95

80 relay real-time information and do not have timing characteristics to determine a neutrons wavelength. This is crucial for spallation sources wher e a spectrum of neutron wavelengths are used and time-of-flight is determined in order to resolve which neutrons gave rise to which reflections. For these applications, an advanced cylindrical detector was built by Brookhaven National Laboratory and is in use at the Protein Crysta llography Station at Los Alamos Neutron Science Ce nter (LANSCE). It has a he ight of 200 mm and a curved horizontal dimension that subtends 120 at th e sample position. This detector is filled with 3 He that decay into a charged pairs upon neutron absorption. This induced charge is detected by a two-dimensional multiwire arra y and is used to determine the x and y position of each event (Langan et al. 2004). Figure 4-1 shows a photograph of the detector in use at LANSCE. There are very few examples in the literatur e of neutron protein structures and these include lysozyme, endothiapepsin, xylose isom erase, and aldose reductase. For all these projects, the investigators were unable to obt ain pertinent structural information about catalysis using ultra-high resolution X-ray st ructure alone. Using neutron diffraction, the functionally important H atoms were identif ied and led to elucidation of enzyme mechanism (Langan et al., 2004). However, all of these neutron structures were determined from protein crystals that were subject to H/D exchange. The only perdeuterated neutron structure so far is of myoglobin and was reported in recent years (Shu et al. 2000; Niimura et al. 2005). As discussed in Chapter 1-3, the rate-lim iting step in catalysis of HCA II is the intermolecular transfer of a pr oton from the zinc-bound solvent (H 2 O/OH ) to the proton

PAGE 96

81 shuttling residue, His64. This distance (~ 7.5 ) is spanned by well-defined solvent molecules that are connected to each other and several critical side chains via a hydrogenbonded network (Christianson and Fi erke, 1996; Lindskog, 1997). Despite the availability of high-resolution crystal structures of HCA II to 1.05 there is currently no definitive information available on the absolute positions and orientations of H atoms from either the solvent network or the ioni zation state of active site residues (Duda et al. 2003). As mentioned above, it is very hard to directly observe H atoms even in highresolution X-ray crystal structures and ne utron diffraction studies of perdeuterated crystals can be a powerful complementary t echnique to elucidate proton donors/acceptors and ionization states in macromolecules. Key qu estions that need to be answered include which solvent molecules are H 2 O or OH molecules and which re sidues are proton donors or acceptors. Another controversial topic is th e nature of the zinc-bound solvent in terms of whether it is a OH or H 2 O in the crystal structure. Even in sub-atomic resolution structures of HCA II there have been no conc lusive answers to any of these questions. Despite all the advantages to neutron pr otein crystallography, there are very few neutron studies compared to X-ray. The main reasons are that there are few sources around the world and the available neutr on beams have low flux, compared to synchrotron sources for X-rays. There is also no foreseeable way to increase the flux of neutrons from nuclear reactors due to the inhe rent limitations of the fission reaction that produces them. In contrast to reactor sources, spallation sources produce neutrons by the bombardment of a heavy metal target, such as mercury or tungsten, with pulsed highenergy protons. The main advantage, besides th e higher flux attained, is that the neutrons produced in this way have a time-of-flight comp onent so that fast or high energy neutrons

PAGE 97

82 arrive at the detector before the slower or low energy neutrons. This allows the energy and wavelength of each neutron to be calcula ted and this information can be applied to the structure determination and ma ke the process more efficient. Structure and catalytic activity analysis and comparison of perdeuterated HCA II with hydrogenated HCA II shows that these st ructures are highly isomorphous and active, especially with regards to the active site architecture and the solv ent network, and this indicates that using perdeuterated HCA II is appropriate for neutron macromolecular crystallography. This work lays the foundati on for planned future neutron structure determination and st ructural analysis. Materials and Methods Production and Crystallization of Perdeuterated HCA II Transformed Escherichia coli ( E. coli ) BL21 pLysS(DE3) cells were plated out on Luria broth agar plates supplemented with 1mM ampicillin. The plates were incubated overnight at 37 C. The next day, an appropr iate colony was select ed and placed into 15 mL Spectra 9-d deuterated minimal media supplemented with 1mM ampicillin made in D 2 O. This initial culture was grown for 24 hour s and then the entire 15 mL was used to inoculate 150 mL Spectra 9-d minimal medi a supplemented with 1 mM ampicillin made in D 2 O. After 18 hours, the cells were harv ested by centrifugation and the resulting pellets were resuspended in 5 mL of Spectra 9-d media. The cells were then placed in 800 mL of fresh Spectra 9-d media suppl emented with 1 mM ampicillin in D 2 O. The large scale culture was then grow n for 2-3 hours and protein expression was induced by the addition of 0.6 mM isopropyl-D-1-thiogalactop yranoside in D 2 O and incubated for 6 hours at 37 C while shaking at 220 rpm. Cells were harvested by centrifugation and stored at 0 C. Upon freeze/thawing, the ce ll lysates were processed and the protein

PAGE 98

83 purified as previously described (Khalifah et al. 1977). After purification, the protein was rapidly back-exchanged into deuterat ed buffers and concentrated. Purity of perdeutereated HCA II was determined by SDS-PAGE. The deuteration level of the protein sa mple was determined by time-of-flight electrospray mass spectrometry. The software Isotopic Pattern Calculator v.1.6.5 for Macintosh (http://www.shef.ac.uk/chemistr y/chemputer/isotopes.html) was used to calculate the theoretical isotopic distributi on and the pattern was then matched to the experimental spectrum. Crystals of perdeuterated HCA II were obtained by the hanging drop vapor diffusion method at room temperature (McPherson, 1982). Crystallization drops were prepared by mixing 5 L of concentrated protein so lution (10-15 mg/mL) with 5 L precipitant solution consisti ng of 2.6 M ammonium sulfate, 50 mM Tris-Dl (pD 8.0) made in D 2 O. Useful crystals appeared within 7 days of crystallization set-up. Crystallography Crystals were cryoprotected prior to data collection by quick-dipping them in 30% glycerol in mother liquor. The crystals were then quick frozen to 100 K in a N 2 -gas stream on the beamline. Synchrotron diffrac tion data were collected at the ESRF beamline ID29. Several crystals were used during data collection and the crystal to detector distance was set at 105, 125, and 185 mm. X-ray data processing was performed using DENZO and scaled and reduced w ith SCALEPACK (Otwinowski and Minor, 1997). Several data sets were collected and th ree were ultimately processed and scaled together from 20.0 1.5 for a total of 220 degrees of data. All manual model building was done with Coot and model refinement was carried out using CNS version 1.1 (Emsley and Cowtan, 2004; Brnger et al. 1998). The wild type structure of HCA II

PAGE 99

84 (PDB accession code 1TBT; Fisher et al. 2005) was isomorphous w ith the perdeuterated HCA II data and was subsequently used to phase the experimental data In order to avoid phase bias of the model, the zinc atom and water molecules were removed from the phasing model. After one cycle of rigid body refinement, annealing by heating to 3000 K with gradual cooling, geometry-restrained position, and individual temperature factor refinement, F o -F c and 2F o -F c electron density maps were generated. These maps clearly showed the position of the zinc atom whic h was then included in the model and subsequent refinements. After several further cycles of refinement, water molecules were incorporated into the model using the auto mated water picking program, as implemented in CNS 1.1, until no more waters were found at the 2.0 level (Brnger et al. 1998). Manual model building in Coot and refine ment continued unt il convergence of R free and R work was reached (Emsley and Cowtan, 2004). Ta ble 4-2 contains the data collection and model refinement statistics. Activity Analysis by 18 O Exchange Methods The rate constants for the hydration and proton transfer reactions were determined by Dr. CK Tu in the Silverman lab using 18 O exchange methods. All assays were performed as detailed in Ch apter 2 except it was all done under deuterated conditions. Hydrogenated and perdeuterated HCA II were exchanged into D 2 O and incubated for 3 hours prior to the assays. Results and Discussion Despite a small lag in cell growth, the transformed cells adapted readily to deuterated minimal media conditions and ove rexpression of HCA II with a yield of 30 mg/mL pure protein per liter of cells was achieved. Mass spectrometry of the purified perdeuterated HCA II showed that over 98% deuteration was obtained. By varying the

PAGE 100

85 protein to precipitant ratio, optimal cond itions for growing large (0.2 x 0.2 x 1.0 mm) perdeuterated crystals within a w eek were determined (Figure 4-2). Structural Effects of Perdeuteration The perdeuterated HCA II crystals diffr acted to 1.5 and belonged to the monoclinic space group P2 1 with unit cell parameters: a = 42.1, b = 41.0, c = 72.0 = 104.4. The scaling R merge was 0.090 and isomorphous with hydrogenated wild type HCA II crystals (Fisher et al. 2005). All data collec tion and refinement statistics are given in Table 4-2. The final R cryst and R free to 1.5 resolution were 19.5 and 20.6 %, respectively. The model was of good quality with root mean square deviations (rmsd) for bond lengths and angles of 0.005 and 1.4, respectively. All final model statistics are given in Table 4-2. A superposition of hydrogenated HCA II (PDB accession code 1TBT; Fisher et al. 2005) onto the structure of perdeuterated HCA II gave a rmsd of only 0.2 for all C atoms (Figure 4-3). Visual inspection of the backbone representations for hydrogenated and perdeuterated HCA II clearly s hows that there is no apprec iable difference between the two and this shows that de uteration has minimal effect on overall fold and three dimensional structure of HCA II (Figure 4-3). When only the active site residues (Tyr7, Asn62, Asn67, His64, His94, His96, His119, Thr199, and Thr200) and solvent molecules (Zn-H 2 O/OH W1, W2, W3a, and W3b) are superposed, the rmsd decreases to less than 0.1 See Figure 4-4 (a) and (b) for a visual comparison of the active site arch itecture. The relative positions and distances between the solvent molecules and the active site residues that coordinate them are very similar, within experimental error, and are shown in Table 4-3 and Figure 4-4.

PAGE 101

86 It is well-known and described in the liter ature that perdeuteration of proteins can have subtle effects on the physical, chemical and functional propert ies of proteins. A comparative study with perdeuterated cytoch rome P450cam in terms of structural stability and dynamics showed that enzyme catalysis was unaffected but that the perdeuterated protein had a midpoint transition temperature that was 4 C lower than its hydrogenated counterpart (Brockwell et al. 2001; Meilleur et al. 2004). However, apart from the thermal stability differences, there we re no structural or kinetic differences and these results imply that the perdeuterated en zyme is a good representation of the native hydrogenated protein (Meilleur et al. 2004). Similarly, for HCA II, there appears to be no difference in the overall structure, but more importantly, the active site architectur e is conserved, even down to the solvent structure which is of great importance. Two neutron studies using quasi-Laue di ffraction and H/D exchanged crystals successfully located H/D atoms at medium resolution (around 2 ). Using these methods it is possible to gain more information from medium resolution neutron diffraction than is possible with ultra-high resolu tion X-ray data alone. In one study, the structure of D 2 O soaked concanavalinA solved with neut ron methods showed the location of 62 D 2 O molecules. Furthermore, it was possible to assign both D atoms for all 62 D 2 O molecules. This is a significant improvement to the 12 H 2 O molecules that were assigned using the 0.95 X-ray data (Habash et al. 2000). In another study of endothiapepsin at 1.95 resolution using neutron diffraction methods, it was possible to elucidate the e ssential catalytic hydrogen atoms of the

PAGE 102

87 aspartate residues responsible for protease activity. These hydrogen atoms were unobservable using 1.0 X-ray diffract ion data analysis alone (Coates et al. 2001). As growing a large protein cr ystal is often the most cha llenging aspect of a neutron diffraction experiment, the large size (0.2 x 0.2 x 1.0 mm) and diffraction quality (1.5 with X-rays) of the perdeuterated HCA II crys tals are the first step towards obtaining a neutron structure. The conserva tion of the key active site solvent molecules, as shown in Figure 4-4 (b), is also encouraging, as accurate assignment of the associated H/D atoms is important for an meaningful interpretation. Kinetic Effects of Perdeuteration Activity assays with perdeuterated and hydrogenated HCA II, soaked in D 2 O for 3 hours prior to the assays, using the 18 O exchange method under de uterated conditions did not show any difference in rate constants for the hydration/dehydration reaction (Figure 4-5). This result was to be expected as the nucleophilic attack the by the zinc-OH on the incoming CO 2 does not involved any proton transfer and this step should not be affected by the presence of D atoms. Figure 45 shows the pH profile of the k cat /K M for the hydration of CO 2 catalyzed by hydrogenated and perdeuterated HCA II. As far as proton transfer is concerned, a difference between the two enzyme species was not predicted or expected. The active si te of HCA II is highly hydrated and solvent accessible and, in theory, the aminoand carboxyl groups of residues should be able to rapidly exchange their H atoms for D atoms. Similarly, the solvent molecules should easily exchange for D 2 O. The expectation was that due to the rapid exchange, the hydrogenated HCA II soaked in D 2 O prior to the assays, should not displa y different kinetics compared to the perdeuterated enzyme. However, as show n in Figure 4-6, the pH profile for R H2O /[E]

PAGE 103

88 clearly shows a difference between the two, w ith the perdeuterated HCA II catalyzing the proton transfer reaction at a lower ra te compared to the hydrogenated, D 2 O soaked HCA II. These kinetic isotope effect s suggest that the non-excha ngeable H atoms are affecting the proton transfer reaction. The effect might be an additive effect over the entire structure instead of specific key residues that mediate either catalys is or the coordinating of active site solvent molecules. Overall, there are no appreciable kine tics effects of perdeuteration on the hydration/dehydration reaction cat alyzed by HCA II (Figure 4-5). In addition, the effects on proton transfer are minor with regards to the magnitude, since the overall bell-shape and pH dependency has not been compromised (Figure 4-6). The activity data, along with the structural data, supports the idea that perdeuteration has a very subtle and slight effect, if any, on the structur e and catalysis of HCA II. Conclusion Neutron protein crystallography is a pow erful tool for studying perdeuterated protein crystals for the direct determination of key solvent molecules and H/D atoms. As a first step toward obtaining a neutron stru cture, perdeuterated HCA II was produced and analyzed. Known crystallization conditions were adapted to the deuterated system and the a high-quality X-ray crystal structure to 1.5 resolution was determined. These data showed that the overall structure and fold of perdeuterated HCA II was virtually identical to its hydrogenated counterpart. More impor tantly, the active site residues and key solvent molecules were completely conserved and this model will serve as the starting point for planned neutron studies. Measurem ent of catalysis by perdeuterated HCA II showed that the hydration/de hydration reaction was unaffected and the proton transfer displayed a minor isotope effect. This work shows the proof-of-principle that

PAGE 104

89 perdeuterated HCA II is a good and accurate representation of hydrogenated HCA II and serves as a comparable entity for studies of hydrogenated HCA II. Initial neutron diffraction te sts are underway at the Pr otein Crystallography Station of the Los Alamos Neutron Science Center Low resolution diffraction was observed and suggests that larger crystals are required for more data acquisition. However, with the new high-intensity Spallation Neutron Source at Oak Ridge Nationa l Laboratory nearing completion, the construction of the Macrom olecular Neutron Diffr actometer (MaNDi) should facilitate faster data collec tion times from smaller crystals.

PAGE 105

90 Table 4-1. Scattering amplitudes and cross se ctions of atoms by X-rays and neutrons. Atom type X-rays (fm) a Neutrons (fm) a Neutrons coh (barns) b Neutrons incoh (barns) b Hydrogen 1 -3.8 1.8 82.0 Deuterium 1 6.5 5.4 7.6 Carbon 6 6.6 5.5 5.5 Nitrogen 7 9.4 11 11.4 Oxygen 8 5.8 4.2 4.4 Sulfur 16 3.1 1.2 1.2 Iron 26 9.6 11.4 11.8 a fm = 1 x 10 -15 m b barn = 1 x 10 -24 cm 2

PAGE 106

91 Table 4-2. Data collection and model refinement statistics. Data Collection Statistics Perdeuterated HCA II Molecular Weight (Da) 30664 Temperature (K) 100 Wavelength 1.005 Resolution () 20.0 1.50 (1.55 1.50)* No. unique reflections 37372 (3689) Completeness (%) 97.0 (96.8) Redundancy 3.0 (2.3) R sym 0.09 (0.17) Refinement Statistics No. protein atoms / Solvent atoms 2058 / 222 R free / R work 0.195 / 0.206 Rmsd bond lengths () / angles () 0.005 / 1.396 Average B-factors ( 2 ) Main / Side / Solvent / Zinc atom 10.5 / 12.9 / 23.3 / 4.9 Ramachandran Statistics Most favored (%) 87.6 Additional & Generous ly Allowed (%) 12.4 *Data in parenthesis is for th e highest resolution shell. R sym = I (I) /I R free = F obs k F calc / F obs

PAGE 107

92 Table 4-3. Distances between solvent molecu les and active site residues for hydrogenated and perdeuterated HCA II. Residues and Solvent Molecules H ydrogenated () Perdeuterated () Zn 2+ ZnH 2 O/OH (D 2 O/OD ) 2.0 1.9 ZnH 2 O/OH (D 2 O/OD ) W1 2.6 2.8 W1 W2 2.8 2.8 W2 W3a 2.7 2.8 W2 W3b 2.9 2.8 ZnH 2 O/OH (D 2 O/OD ) T199 2.7 2.6 W1 T200 2.7 2.7 W3a Y7 2.7 2.7 W3b N62 2.7 2.6 W3b N67 3.2 3.0

PAGE 108

93 Figure 4-1. The LANSCE position-sensitive 3 He-filled detector.

PAGE 109

94 Figure 4-2. Optical photograph of pe rdeuterated wild type HCA II.

PAGE 110

95 Figure 4-3. Backbone superposition of hydroge nated and perdeuterated wild type HCA II. The orange an blue coils repres ent the hydrogenated and perdeuterated HCA II, respectively. N and C termini ar e as indicated. Figure was generated with BobScript and rendered with Raster3D (Merritt and Bacon, 1997; Esnouf, 1997).

PAGE 111

96 Figure 4-4. Active site comparison of hydrogena ted and perdeuterated wild type HCA II. (a) Hydrogenated HCA II, (b) Perdeuterated HCA II. Active site residues are shown in yellow ball-and-sti ck with the catalytic zinc atom as a black sphere. Solvent molecules are shown as red spheres and inferred hydrogen bonds are orange dashed lines. Blue 2F o -F c electron density is shown only for active site waters and it contoured at 1.5 Figure was generated with Bobscript and rendered with Raster3D (Merri tt and Bacon, 1997; Esnouf, 1997).

PAGE 112

97 Figure 4-5. The pH profiles for k cat /K M catalyzed by hydrogenated and perdeuterated HCA II. ( ) Perdeuterated soaked in D 2 O, ( ) Hydrogenated soaked in D 2 O. Data were obtained at 25 in the ab sence of exogenous buffers using a total concentration of all species of CO 2 of 25 mM, with the ionic strength maintained at 0.2 M by addition of sodium sulfate. The solid lines are fit to a single ionization as described in Chapter 2.

PAGE 113

98 Figure 4-6. The pH profiles for R H2O /[E] catalyzed by hydrogenated and perdeuterated HCA II. ( ) Perdeuterated soaked in D 2 O, ( ) Hydrogenated soaked in D 2 O. Data were obtained at 25 in the ab sence of exogenous buffers using a total concentration of all species of CO 2 of 25 mM, with the ionic strength maintained at 0.2 M by addition of sodium sulfate.

PAGE 114

CHAPTER 5 NOVEL INHIBITORS AND THEIR BINDING MODES TO HCA II Introduction There is a wealth of information availabl e on various CA inhibitors (CAI) and their kinetic and structural characterizatio ns (For general review, see Mansoor et al. in The Carbonic Anhydrases, 2000; Supuran et al., 2004). Briefly, various anions have historically been useful for st udying the properties of the metal active site of CA and it is clear that most monovalent anions bind and i nhibit CA activity to va rying extents. These CAIs bind to the metal ion and interfere with the ZnOH coordination by either displacing or replacing the hydroxide, thus disrupting catalysis (Bertini et al. 1983). The key group in determining this disp lacement/replacement is the hydroxyl of Thr199 and this residue is sometimes referred to as the gate keeper. Thr199 is a hydrogen bond acceptor and anionic inhibitors (eg. azide and cyanate) that have protonated groups can bind directly to the zinc ion, thus substituting the OH group and efficiently disrupts catalysis (Merz, 1990). Inhib itors that bind in this wa y maintain hydrogen bonding with Thr199 and participate in tetrahedral coordination of the metal center. Another important class of strong, select ive inhibitors is the aromatic and heterocyclic sulfonamides of the R-SO 2 NH 2 or R-SO 2 NH(OH) form. Several crystal structures of complexes of various sulfona mide inhibitors with CA show similar interactions: the ionized NH group binds direc tly to the metal and simultaneously donates a hydrogen bond to the hydroxyl of Thr199 and an oxygen of the sulfonamide interacts 99

PAGE 115

100 with the amide backbone of Thr199 and thus displaces the deep water (Figure 5-1; Nair et al. 1995). Clinically important drugs include acetazolamide (Diamox) and dorzolamide (Trusopt) and these compounds have applica tion in the treatment of congestive heart failure, altitude sickness and glaucoma (Mansoor et al. 2000) and now may also be useful in controlling mosquito popu lations, as described in Chapter 6. CAIs are commonly prescribed for the trea tment of glaucoma symptoms such as increased intraocular pressure. Th e application of topical inhibitors to the eye inhibits CA and decreases the secretion of sodium, bicar bonate, and aqueous humor thus lowering pressure in the eye (Maren, 1987). There is al so considerable intere st in CAIs for the treatment of cancer as aberra nt expression of extracellula r HCA isoform IX (HCA IX) has been associated with certain types of solid tumors. This overexpression has been observed in several cancers, including non-small-cell lung can cer and renal-cell carcinoma. The overabundance of HCA IX in solid tumors could help these cells in adapting to hypoxic conditions and, due to the ac idification of the extracellular matrix, assist in metastasis (Wykoff et al. 2000; Giatromanolaki et al. 2001). Current CAIs that are clinically used are not very water solubl e and it is thought that a drug with better solubility would have improved bioavailability (Kim et al. 2005). Two novel compounds, BB3 (2-amino-1,3,4-thiadiazolyl-5difluoromethanesulfonamide) and TDM (2-dimethylamino-5-sulfonamido(aminomethyl)1,3,4-thiadiazole), have an a dditional spacer group, either CF2or CHNH2-, between the sulfonamide and thiadiazole ring moietie s (Figure 5-2). These drugs have better solubility than the conventi onal glaucoma treatments and st ill bind CA with nanomolar

PAGE 116

101 affinity (Boyle et al. 2005). To investigate BB3 and TD M binding and their interactions with the HCA II active site, the crystal structur es of the co-complexes were determined by X-ray crystallography to 1.60 and 1.15 respectively. Materials and Methods Expression, Purification, and Crystallization HCA II was overexpressed in Eschericia coli BL21(DE3)pLysS and affinity purified using p-aminomethylbe nzenesulfonamide coupled to agarose, as described elsewhere (Khalifah et al. 1977). Fractions were assessed for yield and purity by silver stained SDS-PAGE and fractions containi ng HCA II were pooled accordingly. HCA II was dialyzed against 50 mM Tris-Cl, pH 7.8 and then concentrated to ~ 15 mg/ml with Amicon Ultra centrifugation filtration devices. Protein concentration prior to crystallization was determined by measuring the optical density at 280 nm and assuming a molar absorbance of 5.5 x 10 4 M -1 cm -1 Hanging drop crystals were prepared by mixing 5 l of protein (~ 15 mg/ml in 50 mM Tris-Cl pH 7.8) with 5 l precipitant (50 mM TrisCl pH 7.8, 2.6 M ammonium sulfate) (McPhers on, 1982). The drops were equilibrated at room temperature against 1000 l reservoir solution. Diffraction quality crystals appeared within 2-3 days. The inhibitors were solubili zed in water to a fina l concentration of 20 M. For the drug soaks, sitting drops us ing microbridges, were set up with 10 l drop size. Prior to adding single, pre-grown crysta ls the soluble drugs we re added to a final concentration of 1 M. Crystals were allowed to soak for at least 24 hours prior to data collection. Synchrotron X-ray Data Collection All crystals were quick-dipped in cryopr otectant (30% glycerol in reservoir solution) before flash-freezing at 100 K for X-ray data collection. Diffraction data for the

PAGE 117

102 BB3-complex were collected at Brookhaven National Laboratory beamline X29 on a Quantum 315 CCD detector using a wavelength of 1.100 Crys tal to detector distance was set to 100 mm and a total of 297 images were collected w ith an oscillation angle of 1. Diffraction data for the TDM-complex were collected at Cornell High Energy Synchrotron Source beamline A1 on a Quantu m 210 CCD detector with a wavelength of 0.978 The crystal to detect or distance was 100 mm and a total of 410 images were collected with an oscillation angle of 1. All data set stat istics for both complexes are given in Table 5-1. Structure Determination and Model Refinement Diffraction data were inde xed and reduced with DENZO and SCALEPACK from the HKL program suite (Otwinowski and Mi nor, 1997). The data were phased with molecular replacement methods as implemente d in the software package Crystallography and NMR Systems (CNS) version 1.1 (Rossmann, 1990; Brnger et al. 1998). The structure of wild type HCA II was used as a starting model (PDB accession code 1TBT; Fisher et al. 2005). After initial cycles of rigid b ody refinement, simulated annealing to 3000 K with gradual cooling, geometry-restrain ed positional, and individual temperature factor refinement, F o -F c omit electron density maps were calculated. Visual inspection of these maps clearly revealed the position of the zinc, solvent, and the inhibitors. Coordinates and topology file s for the inhibitors were generated with the PRODRG server (http://davapc1.bioch.dundee.ac.uk/progra ms/prodrg) and were subsequently built into the respective electron de nsity (Schttelkopf and van Aa lten, 2004). After the initial refinement routines in CNS the R work and R free were 0.193 and 0.224 for the BB3 complex and 0.225 and 0.236 for the TDM complex, respectively.

PAGE 118

103 At this point the refinement was moved into SHELXL97 with the conjugate gradient least squares (CGLS) mode using default restra ints for protein geometry (Sheldrick and Schneider, 1997; Engh and Huber, 1991). After each round of CGLS refinement, F o -F c and 2F o -F c electron density maps were generated and the model and maps visually inspected using the molecu lar graphics program Coot (Emsley and Cowtan, 2004). Manual adjustments to side chai ns and inhibitor positions were done at this point. After several rounds of refineme nt, automated water placement was performed with SHELXPRO using F o -F c omit maps with a 4 cut-off (Sheldrick and Schneider, 1997). For the high resolution TDM complex structure all non-hydrogen atoms were refined anisotropically using the SHELXL97 de fault anisotropic disp lacement parameters (Sheldrick and Schneider, 1997). For BB3 only the zinc and sulfur atoms were refined anisotropically. Manual m odel building in Coot involved modeling of alternate conformations (Ile22 and Lys39 for TDM comp lex) and addition/removal of solvent. Refinement of both structures continued until the R factors converged. The final R cryst and R free were 0.167 and 0.212 for BB3, and 0.136 and 0.166 for TDM, respectively. All model geometries were analyzed by PROCHECK and enzyme:inhibitor interactions were determined with LIGPLOT (Laskowski et al. 1993; Wallace et al. 1995). Data refinement and final model stat istics are given in Table 5-1. Results and Discussion The crystal structures of HCA II, in co mplex with BB3 and TDM, reported here were isomorphous with previously determin ed HCA II and had unit cell dimensions as shown in Table 5-1. The final, refined models had good geometry and all backbone torsion angles fell within acceptable regions on a Ramachandran plot. The model of the

PAGE 119

104 BB3 complex (to 1.6 ) had an R cryst and R free of 0.167 and 0.212, while TDM (to 1.15 ) had an R cryst and R free of 0.136 and 0.166, respectively (Table 5-1). Both compounds used in this study cont ain the classically employed sulfonamide moiety that anchors the rest of the inhibitor in the active site. From th e structure it is clear that the ionized NH group displaces the zinc-bound solvent and makes a hydrogen bond with Thr199 in the same way that has b een observed in other complexed structures (Figure 5-1 to 5-4). The N3 from the su lfonamide group of BB3 and TDM is 2.8 and 2.9 away from the hydroxyl of the side chain of Thr199, respectivel y (Figure 5-5). The novel compounds are different from the classical CAIs in that they contain a spacer group between the zinc-anchoring sulfonamide gr oup and thiadiazole ring (Figure 5-2). BB3 contains a difluoromethane group while TDM co ntains an aminomethyl group. It is clear from the high resolution electron density maps that these compounds adopt unique binding modes compared to other clin ically used CAIs (Figure 5-4). In the BB3 complex structure the difluoromethane group allows rotational flexibility between the sulfonamide and thiadi azole ring. This feature allows the aromatic ring of BB3 to fold back and stack with the imidazole of His96, one of the zinc ligands (Figure 5-5 (a)). This folding back is unique as the other CAIs adopt a more extended conformation that leads out of the active site. Also, as can be seen from the features of the electron density, the ring of BB3 is planar. Apart from the stacking interaction, there is also an inferred hydrogen bond between F2 of BB3 and OG1 of Thr200 with a distance of 2.5 There are also two othe r direct hydrogen bonds between N1 and NH2 of the thiadiazole ring and residues Gln92 and Asn62. The onl y indirect hydrogen bond is between NH2 and Asn62 that occurs through a solvent molecule, W369 (Figure 5-5 (a)).

PAGE 120

105 To address the issue of whether the observed stack is a result of the spacer group insertion or a specific feature of the difluoromethane group, an analogous compound (TDM) was synthesized that contained an am inomethyl group as the spacer (Figure 5-2). In the crystal structure of the TDM-HCA II complex, no stacking interaction with any His was observed. Also, the thiadiazole ring sh ows puckering, compared to BB3 that has a planar ring. (Figure 5-3 (b)). The ring also occupied a different position compared to BB3 and acetazolamide (Figure 5-4). To more clearly demonstrate the different binding modes, both BB3 and TDM were superimposed in the active site of HCA II with the view rotated to highlight these differe nces, as indicated in Figure 5-6. There are two indirect hydrogen bonds fr om TDM to Asn67 that are mediated by two solvent molecules, W316 and W599 (Figure 5-5 (b)). In contrast to the hydrogen bond between F2 and Thr200 of BB3, there is no hydrogen bonding in teraction between the aminomethyl spacer group of TDM and Thr2 00. This is due to the ring being in a different position compared to BB3 and this has displaced the aminomethyl group away from any potential interac tions. The dimethyl moiety of TDM occupies a hydrophobic region of the HCA II active site comprise d of Phe131 and Leu198 (Figure 5-5 (b)). Conclusion Carbonic anhydrases are interesting and clinically relevant targets for the development of new inhibitors. Such co mpounds will have implications for both pharmaceutical, industrial, and research applic ations. CAIs are routinely prescribed for the treatment of various diseases, such as glaucoma and altitude sickness, and recently has shown potential for treating certain tumor types. Classical CAIs have poor solubility and bioavailability and new compounds with be tter chemical propert ies could have novel clinical uses. This work showed the bindi ng and interactions of two new water-soluble

PAGE 121

106 compounds, BB3 and TDM, and their interacti ons with HCA II by X-ray crystallographic methods. These compounds are unique in that th ey contain two different kinds of spacer groups between the sulfonamide and thiadiazo le groups. The spacers confer interesting properties on the inhibitors and cause them to bind in different modes than what have been observed before. These alternate binding lo cations in the active site of HCA II could provide useful information for the future development of isoform specific inhibitors.

PAGE 122

107 Table 5-1. Data collection, refineme nt and final model statistics. BB3 TDM Space group P2 1 P2 1 Unit cell parameters (, ) a = 42.3, b = 41.4, c = 72.1, = 104.1 a = 42.5, b = 41.6, c = 72.7, = 103.9 Resolution () 20.0 1.60 (1.66 1.60)* 20.0 1.15 (1.19 1.15) No. unique reflections 32031 (3154) 87379 (7896) Completeness (%) 99.9 (100.0) 97.9 (89.3) Redundancy 4.4 (4.3) 7.0 (4.1) Overall I/ (I) 17.1 (4.2) 26.4 (4.7) R sym 0.066 (0.209) 0.099 (0.323) R cryst / R free 0.167 / 0.212 0.136 / 0.166 Rmsd bond lengths/angles 0.009 / 1.490 0.010 / 1.315 Average B-factors Main/side/solvent/inhibitor 16.7/23.2/31.8/21.5 14.1/17.5/32.9/16.5 No. solvent molecules 246 310 Ramachandran Statistics (%) Most favored/additionally allowed 87.9/12.1 88.8/11.2 *Data in parenthesis is for th e highest resolution shell. R sym = I (I) /I R cryst = F obs k F calc / F obs R free is calculated the same way as R cryst except for 5% of data omitted from refinement.

PAGE 123

108 Figure 5-1. Classical carbonic anhydrase inhibitors bound to HCA II. (a) Acetazolamide (AZM), (b) Brinzolamide (BZO). Active site residues are as labeled and the zinc is shown as a black sphere PDB accession codes for (a) 1AZM (Chakravarty and Kannan, 1994); (b) 1A42 (Stams et al. 1998). Figure was generated and rendered with BobScript and Raster3D (Esnouf, 1997; Merritt and Bacon, 1997).

PAGE 124

109 Figure 5-2. Chemical structure of novel in hibitors. (a) BB3, and (b) TDM. Figure was generated with ChemDraw.

PAGE 125

110 Figure 5-3. Active site structures of HCA II in complex w ith two novel inhibitors. (a) BB3, and (b) TDM. Active site residues are shown in yellow ball-and-stick and are as labeled. The zinc atom is shown as a black sphere. The red mesh electron densities are F o -F c maps contoured at 2 and was calculated without the inhibitor present. Figure was gene rated and rendered with BobScript and Raster3D, respectively (Esnouf, 1997; Merritt and Bacon, 1997).

PAGE 126

111 Figure 5-4. HCA II active site with inhibitors superimpos ed. Active site residues are shown in yellow ball-and-s tick and the zinc atom is a black sphere. BB3 is shown in blue, TDM in red, and acet azolamide is shown in green. PDB accession code for acetazolamide coordinates: 1AZM (Chakravarty and Kannan, 1994). Figure was generated and rendered with BobScript and Raster3D, respectively (Esnouf, 1997; Merritt and Bacon, 1997).

PAGE 127

112 Figure 5-5. Interactions of BB3 and TDM with HCA II. (a) BB3, and (b) TDM. Active site residues are shown in yellow ball-andstick, as labeled, with the zinc atom as a black sphere. Hydrogen bond dist ances are as indicated. Figure was generated and rendered with BobScript and Raster3D, respectively (Esnouf, 1997; Merritt and Bacon, 1997).

PAGE 128

113 Figure 5-6. HCA II with BB3 and TDM superim posed. Zinc and zinc ligands are shown as a black sphere and yellow ball-and-stick, respectively. The two inhibitors are shown in yellow ball-and-stick als o. Figure was generated with BobScript and Raster3D, respectively (Esnouf, 1997; Merritt and Bacon, 1997).

PAGE 129

CHAPTER 6 EXPRESSION, PURIFICATION, KINETIC, AND STRUCTURAL CHARACTERIZATION OF AN ALPHACLASS CA FROM AEDES AEGYPTI (AACA1) Introduction Overall, the class of CAs are present in vert ebrates, but are also found in invertebrates such as fruit flies and mosquitoes (Del Pilar Corena et al. 2002). Recently, another CA was identified from the midgut of Aedes aegypti larvae. Subsequent sequence comparisons showed it to be homolog ous to human isoforms I and II with 39% and 32% sequence identity, respectively (H CA I and HCA II; Figure 6-1). Biochemical analysis also showed the pr otein to be a glycosyl-phosphotidylinositol-linked membrane associated CA, similar to HCA IV with 27% sequence identity (Del Pilar Corena et al. 2002; Seron et al. 2004). Due to the high similarity to HCA I, this protein was named AaCA1 (Genbank accession number AF395662). It ha s been suggested this CA produces bicarbonate ions that are needed in the mi dgut to help buffer the extreme alkaline environment (~ pH 11) found there. This high pH is required for the digestion of plant materials that larvae ingest duri ng this development stage (Martin et al. 1980). Recent experimental evidence that supports the essent ial role of CA in these organisms showed that treatment with CA inhib itors (CAIs) blocks midgut alka linization and cause death of the mosquito larvae (Del Pilar Corena et al. 2004). All CAs use the same zinc-hydroxide mech anism of catalysis for the reversible hydration/dehydration of CO 2 through a nucleophilic attack by the zinc-bound hydroxide. The product, bicarbonate, is easily replaced by a water molecule and freely diffuses out 114

PAGE 130

115 of the active site (Lindsk og, 1997; Christianson and Fierke 1996). The hydroxide at the zinc is then regenerated through a series of proton transfer events that are thought to occur via a well-ordered and conserved wa ter network that co nnects the zinc-bound solvent and proton shuttle Hi s64 (HCA II numbering) (Tu et al. 1989; Fisher et al. 2005). A detailed discussion of the catalytic mechanism can be found in Chapter 2. In catalytically active CAs the zinc ion is tetrahedra lly coordinated by three His residues (His94, 96, and 199) and a solvent ligand. Thr199 assists by making a hydrogen bond with the zinc-bound solvent to optimally orient it for nucleophilic attack of the incoming substrate (Merz, 1990). Based on mutagenesis, kinetic, and st ructural studies of amino acids that line the active site cavit y, a detailed picture has evolved as to the relevance of these residues. In HCA II the side chains of residues Tyr7, Asn62, Asn67, and Thr200 are all believed to be involved in stabilizing several active site waters that constitute the solvation structure that coul d promote efficient proton transfer from the zinc-bound solvent to His64. Mutagenesis on an y of these residues have been shown to have considerable effects on proton transfer (Fisher et al. 2005; Krebs et al. 1991; Bhatt et al. 2005). Mosquitoes act as vectors for many important human pathogens that cause diseases such as dengue fever, malaria, and yellow fever. According to the American Mosquito Control Association, there are more than 2500 different species of mosquitoes worldwide and 150 of these occur in the U.S.A. (Darsie and Ward, 2000; Spielman et al. 2001). Recent reports indicate that between fi fty million and one hundred million cases of dengue fever appear annually. There are al so an additional several hundred thousand occurring cases of the lethal dengue hemo rrhagic fever (Halstead, 1997). Alarmingly, the

PAGE 131

116 incidence of mosquito-borne diseases has b een increasing dramatically in recent years, most likely due to the sp read of certain diseasecarrying species such as Aedes aegypti and Anopheles gambiae (Gubler, 1997). The discovery of an class CA in mosquitoes presents an interesting new drug target for the potential regul ation of mosquito populations. This chapter describes the b acterial expression, pur ification, kinetic characterization, and structural modeling of the soluble CA domain of AaCA1 from Aedes aegypti This work has important applications for the development of drugs specific for mosquito CAs and lays the in itial foundation for future crystallographic structure determinations of such CAs. Materials and Methods Expression and Purification A pET100/D-TOPO (Invitrogen Corporati on) plasmid containing the truncated gene for AaCA1 (residues 17-275, indicated by th e black arrows in Fig. 1) was generated as previously described (Seron et al. 2004). The coding region was removed by digestion of plasmid DNA with restricti on enzymes, NcoI and BamHI (New England Biolabs). The linearized insert was then gel-purified and ligated into a modified pET vector, pET81f1p, which was constructed by Tanhauser et al. as described elsewhere (Tanhauser et al. 1992). Correct insertion and orientation of the gene was verified by sequencing the resulting plasmid. Expression of the truncat ed, soluble region of AaCA1 was performed in E. coli BL21(DE3) pLysS cells th at were grown to an opt ical density of ~ 0.30 as measured at a wavelength of 600 nm. Prot ein expression was then induced by the addition of 1 mM IPTG and 1 mM ZnSO 4 was also added for uptake in the expressed protein. At 4 hours post-induction, cells were harvested by centrifugation and the cell pellets frozen at -20 C overnight. Cell pellets we re lysed by freeze/thawing and

PAGE 132

117 solubilized in 0.2 M sodium sulfate, 50 mM Tris-Cl (pH 9.0). The soluble cell fraction was obtained by centrifuging the lysates at 100,000 x g for 1 hour at 4 C. AaCA1 was further purified from the supernatant by affi nity chromatography using p-amino-methylbenzenesulfonamide (pAMBS; a specifi c binder to the active site of -CAs) coupled to agarose beads as described elsewhere (Khalifah et al. 1977). Purity of the protein was verified by electrophoresis on a 12% polyacrylamide gel stai ned with Coomassie (Figure 6-2). Prior to further analysis, the protein was buffer-exchanged in to 50 mM Tris-Cl (pH 7.8) using Amicon Ultra centrifugation devices. Activity Analysis by 18 O Exchange This method is based on the measurement using membrane-inlet mass spectrometry of the exchange of 18 O between CO 2 and water at chemical equilibrium (eqs 6-1 and 6-2) (Silverman, 1982). HCOO 18 O + EZnH 2 O EZnHCOO 18 O COO + EZn 18 OH (6-1) H 2 O EZn 18 OH + BH + EZnH 2 18 O + B EZnH 2 O + H 2 18 O + B (6-2) An Extrel EXM-200 mass spectrometer with a membrane inlet probe was used to measure the isotopic content of CO 2 Solutions contained a to tal concentration of all species of CO 2 of 25 mM, and the ionic strength wa s maintained by the addition of 0.2 M sodium sulfate. Analysis of 18 O-exchange data are desc ribed in detail elsewhere (Silverman, 1982; An et al. 2002). This approach yields k cat / K M as can obtained by steady state methods. The pH dependence of k cat / K M depends on the ionization state of the zinc-bound water, as shown in eq 6-3. k cat / K M = ( k cat / K M ) max (1 + [H + ]/( K a ) ZnH2O ) -1 (6-3)

PAGE 133

118 A second rate determined by the 18 O exchange method is R H2O the rate of release from the enzyme of water-bearing substrate oxygen (eq 6-2). The pH dependence of R H2O /[E] is often bell-shaped, consistent with the transfer of a proton from a single predominant donor to the zinc-bound hydroxide. In these cases, the pH profile is adequately fit by eq 6-4 in which k B is a pH-independent rate c onstant for proton transfer, and ( K B a ) donor and ( K a ) ZnH2O are the noninteracti ng ionization constants of the proton donor BH of eq 6-2 and the zinc-bound water. + k B obs = k B /{[1 + ( K a ) donor /[H + ]][1 + [H + ]/( K a ) ZnH2O ]} (6-4) All pH data were obtained at 25 C in the absence of buffer using a total concentration of all species of CO 2 of 25 mM, with the ionic st rength maintained at 0.2 M by addition of sodium sulfate. Determination of Inhibition Constants Inhibition constants for tight-binding sulfonamides were determined by titration of the 18 O-exchange activity with methazolamide or ethoxzolamide (Maren and Conroy, 1993; Maren and Sanyal, 1983). Data were analyzed by the method of Henderson and fitting the data with Enzfitter (Biosoft) (Segal, 1975). Model Building A molecular model of AaCA1 was built usi ng the Swiss-Model facility with HCA I as a template (PDB accession code: 1AZM ) (Chakravarty and Kannan, 1994). SwissModel is an automated protein structur e homology modeling server available at http://www.expasy.org (Peitsch, 1996; Guex et al. 1999; Schwede et al. 2003). Model analysis, visualization, and manipulation were performed using the programs Coot and GRASP (Emsley and Cowtan, 2004; Nicholls et al. 1991).

PAGE 134

119 Results and Discussion Expression and purification of AaCA1 from the E. coli BL21(DE3) pLysS expression system typically gave yields of ~ 40 mg purified protein from 1 L of cells (Figure 6-2). Gel lane 2 shows the cell lysa te at the time of induction, and lane 3 was taken at 3 hours post-induction. There is a di stinct band at ~ 30 kDa in lane 3 that corresponds to overexpressed AaCA1. Lane 4 is the soluble fraction of the cell lysate that was loaded onto the affinity resin and AaCA1 was found exclusively here Lane 5 represents the unbound material that did not bind to the affinity resin, and a comparison of lanes 4 and 5 shows that all of the AaCA1 bound to the affinity resin. Lanes 6-8 are different fractions of the protein that were collected at elution of the affinity resin. After buffer exchange and concentration of the enzyme, denaturing polyacrylamide gels were silver-stained, and the protein wa s estimated to be more than 95% pure. The protein had limited solubility at concentrations over 1 mg/ml, but was stable at 1 mg/ml in 50 mM Tris-Cl, pH 7.8 at 4 C. Kinetic analysis reveals that AaCA1 is a highly efficient CA isoform with k cat / K M for hydration of carbon dioxide at 67 3 M -1 s -1 and a kinetic pK a of 6.7 0.1 (Figure 63a). The magnitude of this rate constant is more similar to HCA I than HCA II (Table 61). This was to be expected as AaCA1 ha s an overall higher sequence homology to HCA I than that of HCA II (39% and 32% sequence identity, respectively, Fi gure 6-1). In terms of k B the rate constant for proton transfer, however, AaCA1 displayed a rate constant more typical of HCA II with maximal activity near pH 6.5 (Table 6-1; Figure 6-3b). Determination of inhibition constants (K i ) with sulfonamide drugs methazolamide and ethoxzolamide reveal that these i nhibitors bind tightly to AaCA1, with K i s of 2.7

PAGE 135

120 0.3 and 1.9 0.3 nM, respectively, as determined from the 18 O-exchange method. These results are very similar to those previously obtained for HCA I and HCA II; for each of these enzymes the inhibition constants were 10 nM for methazolamide and 2 nM for ethoxzolamide, determined by a buffer-titration method (Sanyal et al. 1982). The sequence alignment of AaCA1 with HCA I and II shows the high level of conservation between these CAs (Figure 6-1). This high sequence pres ervation led to the construction of a reliable three-dimensiona l molecular model of AaCA1 (Figure 6-4). Superposition of the AaCA1 m odel onto the crystal structures of HCA I and HCA II gave root mean square deviations for C s of 0.8 and 1.2 respectively (Fisher et al. 2005; Chakravarty and Kannan, 1994). As the sequence alignment suggested, the gene ral fold and active site architecture is conserved to that of HCA I and II. An intere sting structural feature of AaCA1 is that there are two additional cystei ne residues (Cys40 and Cys217; highlighted yellow, Figure 6-1) that are not present in the sequences of HCA I or II, and the model predicts that these residues are close enough to form a disulfide bond (Figure 6-4b). Interestingly, HCA XII and murine CA XIV also have cysteine residues at equivalent position compared to AaCA1, and their crystal struct ures have shown that they do form an intermolecular disulfide bond (Whittington et al. 2001; Whittington et al. 2004). The AaCA1 molecular model shows three histidine residues (His111, 113, and 130) that are equivalent to His94, 96, and 119, whic h are the protein ligands to the zinc in HCA I and II (Figure 6-5). Also, residues threonine 213 and 214, and histidine 83 of AaCA1 correspond to Thr199, Thr200, and th e proton shuttling His64 of HCA II, respectively, whereas HCA I has a histidine at position 200. There are, however, key

PAGE 136

121 active site differences between AaCA1 and HC A I (other than the Thr200 position, HCA II numbering) and HCA II. AaCA1 has a threonine at position 81 and a glutamine at position 86, whereas HCA II has asparagines at both positions (HCA II numbering 62 and 67, respectively) while HCA I has a valine at 62 and a histidine at 67. Hence, AaCA1 and HCA II have different polar amino acids at these positions, whereas HCA I has a hydrophobic valine in the active site. In the AaCA1 model, residues Thr81 and Gln86 point towards the active site, and could therefore be reasonably expected to hydrogen-bond and contribute to a solvent molecule architecture in the active site that resembles that of HCA II. In contrast to HCA II and AaCA1, HCA I has a histidine at position 200 (HCA II numbering) instead of a threonine, and a hydrophobic valine in the ac tive site (Figure 6-5). The observation that AaCA1 would have a solvated active site, more like HCA II than I, supports the measurements of the efficient pr oton transfer rate constant, k B for AaCA1, which is comparable to HCA II (Table 6-1) (Khalifah, 1971). Inspection of the AaCA1 model give s no clear explanation of why the k cat / K M for hydration of CO 2 for AaCA1 is more similar to HCA I than HCA II, as the residues that line the proposed substrate hydrophobi c pocket (Val121, Val143, Leu198, Val207, and Trp209, HCA II numbering) are conserved in a ll three CAs (highlighted green, Figure 61). Hence this variation might be influenced by the numerous secondary shell interactions forming this pocket, and as AaCA1 overall sequence similar is more like HCA I than II this might account for the observed differences. The N-terminus of AaCA1 could not be m odeled with certainty because of a six amino acid insertions compared to HCA I and II (Figure 6-1). In the sequence alignment,

PAGE 137

122 tyrosine 7 in HCA II appears conserved among AaCA1, HCA I, and HCA II. However, closer inspection of the AaCA1 model shows Asn27 (highlighted in red, Figure 6-1) at the same structural position (Figure 6-5a, c, and e) as Tyr7 in HCA I and II. There could be structural c onservation of a tyrosine at that position of AaCA1, and the insertion loop could be displaced. Altern ately, Asn27 could potentially be in the position where the model indicates, and this would have implications for active site architecture and catalysis by AaCA1. Despite the high sequence identity among AaCA1, HCA I, and HCA II, there are morphological differences between them that are reflected in their overall surface topology and charge distribution (Figure 65b, d, and f). HCA I and HCA II seem more compact and globular compared to AaCA1 and appear to have a more even charge distribution. AaCA1 has a much higher overall calculated negative charge (-7.0) when compared to HCA I (-1.0) and HCA II (-2.0) at physio logical pH (Figure 6-5b, d, and f). This surface feature might be relate d to the cellular location of this isoform at the cell membrane as a glycosyl-phosphatidylinosit ol-linked protein, while HCA I and HCA II are soluble cytosolic proteins. Conclusion A previously cloned -CA from the mosquito Aedes aegypti AaCA1, has been successfully expressed and purif ied to near homogeneity from a bacterial system (Figure 6-2) and subsequently kinetically characteri zed (Figure 6-3). The kinetic data presented here reveal that AaCA1 is a highly active is oform that has catalytic features resembling both HCA II and HCA I (Table 6-1), despite higher sequence identity with the latter

PAGE 138

123 (Figure 6-1). Specifically, proton transfer capac ity appears similar to HCA II and the rate constant for conversion of CO 2 to bicarbonate k cat /K m appears similar to HCA I. A homology model of AaCA1 has provide d information on the overall fold, active site, and surface charge distribution, compar ed to HCA I and II (Figure 6-4 and 6-5). Based on the sequence alignment and model, it has been shown that critical catalytic residues found in HCA II are conserved in AaCA1. A structural rationale of why AaCA1s proton transfer properties are more like HCA II than HCA I, is the conservation of residues threonine 214 (Thr200 in HCA II) that is a histidine in HCA I, and the conservation of polar amino acids in the active site, whereas HCA I has a valine (Figure 6-5). Also of interest are the amino acid diffe rences in the active site of AaCA1 at positions Thr81 and Gln86, compared to the st ructurally equivalent amino acids for HCA I (Val62 and His67) and HCA II (Asn62 and 67) as these amino acid si de chains could be useful to design selective CAIs as potentia l drug target against mosquitoes and not humans.

PAGE 139

124 Table 6-1. Maximal values of k cat /K m for the hydration of CO 2 and k B for the proton transfer dependent release of H 2 18 O from isoforms of carbonic anhydrase. Isoform k cat /K m ( M -1 s -1 ) k B ( s -1 ) Human CA I 50 a 0.1 a Human CA II 150 a 0.8 b AaCA1 67 +/3 1.1 +/0.2 a From Khalifah, 1971. b From Duda et al. 2001.

PAGE 140

125 Figure 6-1. Sequence alignment of AaCA1 with HCA I and II. The black arrows indicate the Nand C-termini of the AaCA1 expr ession construct. Residues that form the active site and CO 2 hydrophobic pocket are highlighted in blue and green, respectively. In additi on, residues Cys40 and 217, and Asn27 of AaCA1 are highlighted in orange and red, respectively. The represents sequence identity while : represents relative sequence homology between the CAs. Alignment was performed with ClustalW (Thompson et al. 1994).

PAGE 141

126 Figure 6-2. The 12% Coomassie stained polya crylamide gel of AaCA1 expression and purification. Lane 1 lo w molecular weight marker; lane 2 lysate at induction; lane 3 lysate at 3 hours po st-induction; lane 4 soluble lysate loaded onto affinity resin; lane 5 unbound/flow through material from affinity purification; lanes 6 fractio n collected from elution of affinity purification. Arrow indicates AaCA1.

PAGE 142

127 Figure 6-3. The pH profile of rate constants for catalysis by AaCA1. (a) pH profile for k cat /K m (M -1 s -1 ) for hydration of CO 2 catalyzed by AaCA1 determined by 18 O exchange. The solid line is a fit to eq 5 with (k cat /K m ) max = (6.7 0.3) x 10 7 M 1 s -1 and (pK a ) ZnH2O = 6.7 0.1 (see also Table 1). (b) pH profile for the proton transfer dependent rate constant R H2O /[E] (s -1 ) in catalysis by AaCA1 as determined by 18 O-exchange methods. The solid line is a fit to eq 6-6 with k B = (1.1 0.2) x 10 s and (pK B 6 -1 a ) donor = 6.5 0.2 and (pK a ) ZnH2O = 6.5 0.2 (see also Table 1).

PAGE 143

128 Figure 6-4. Molecular model of AaCA1. (a) Grey coil of ov erall structure of AaCA1, with active site residues in yellow ba ll-and-stick, and catalytic zinc atom depicted as a black sphere. N and C termini are as marked; (b) The putative disulfide bond of AaCA1 between residues Cys 40 and 217. Figure was generated and rendered with BobScript and Raster3D (Esnouf, 1997; Merritt and Bacon, 1997).

PAGE 144

129 Figure 6-5. Active sites and surface charge di stribution of AaCA1, HCA I, and HCA II. (a), (c), and (e) Active sites of AaCA1, HCA I, and HCA II, respectively. (b), (d), and (f) Molecular surf ace with charge distributi on, where blue represents positive and red the negative charge of AaCA1, HCA I, and HCA II, respectively. The green arrows point to the active site of each isoform. Figures (a), (c), and (e) were generated and rendered with BobScript and Raster3D, respectively(Esnouf, 1997; Merritt and B acon, 1997). Figures (b), (d), and (f) were made using GRASP (Nicholls et al. 1991).

PAGE 145

CHAPTER 7 CONCLUSIONS AND FUTURE DIRECTIONS Summary and Conclusions The main focus of this dissertation, as di scussed in Chapters 2 to 6, has been to gain insights into active site structure of va rious CAs and relate this knowledge to proton transfer as catalyzed by CA. Di fferent approaches towards thes e ends were taken to assist in the investigation of structural featur es as they affect enzyme function. In Chapter 2, the proton shuttling histid ine residueswere inserted at different positions in the active site of HCA II. The effect on solvent structure and proton transfer kinetics was then determined over a broad pH range for wild type and mutant HCA II (pH 5.1 to 10.0). A striking feature of HCA II as that is shows remarkable structural and kinetic stability over a wide range of pH ( 5.1 to 10.0). The main structural difference is the orientation of His64 in wild type HCA II. The side chain displays two conformations and these two states appear to be equally occupied at physiological pH. In H64A/N62H and H64A/N67H HCA II the introduced His do es not show any mobility and there is no completed hydrogen-bonded water chain be tween proton donor and acceptor. A His inserted at position 67 shows appreciable prot on shuttling activity ( 25% of wild type), compared to a His at position 62 (4% of wild type). In these examples, the distance between the proton shuttling His and the zi nc-bound solvent and the number of waters that spans this distance, might be more im portant for efficient proton shuttling than observing a complete hydrogen-bonded chain of waters to the His residue. The results suggest that the optimal distan ce for the His to the zinc is between 6.6 and 7.5 and that 130

PAGE 146

131 the number of intervening water molecules should not exceed two for the support of efficient and fast proton transfer. In Chapter 3 three key active site re sidues in HCA II (Tyr7, Asn62, and Asn67) were replaced by hydrophobic Phe and Leu residu es. The rationale was to remove any polar groups while at the same time avoidi ng steric problems and then measuring the effect on active site solvation and catalysis by HCA II. Instead of directly altering the proton shuttle His64, the goal was to speci fically affect the hydrogen-bonded water structure. The structural and kinetic data for N62L and N67L mutants show that the water networks were easily disrupted, especially at low pH, and both displayed considerable lower proton transfer rates compared to w ild type and Y7F HCA II. The most surprising result was the enhanced proton tr ansfer rate (up to 7-fold higher) over that of wild type observed for Y7F HCA II. This mutant also displayed a similar water network as wild type HCA II, except for the loss of one activ e site water. A surprising observation was that Y7F had a sulfate bound at pH 8.2 as this is very unexpected and usually only occurs at low pH (Fisher et al. 2005). Due to the presence of sulf ate in the crystal structure and the uncertainty of the effect this has on the orientation of His64 a nd water structure, the protein was crystallized in the absence of any ammonium sulfate. Comparing the structures of all three mutants, determined at various pH and in the presence or absence of sulfate, strongly suggests that sulfate does not affect the conformation of His64 or the water structure. Correlations be tween the structural and kinetic data suggest that a single, linear array of water bridging His64 and the zinc-bound solvent might be more efficient at proton transfer than a branched structure. Also, these mutations caused changes in the

PAGE 147

132 pK a of the proton donor and acceptor and the difference in the pK a could also have an effect on proton transfer efficiency. Chapter 4 describes some initial work done towards obtaining a neutron structure of HCA II. X-ray crystallography is severely limited in its ab ility to reveal the locations of hydrogen or deuterium (H/D) atoms. This is because the extent of diffraction by Xrays depends on the number of electrons and H/D atoms diffract poorly compared to more electron-rich atoms such as carbon C and nitrogen N. Neutrons interact with nuclei of atoms and the diffraction of these at oms, including H/D, does not depend on the number of electrons at all (Habash et al. 2000; Langan et al. 2004). Neutron crystallography can provide information abou t ionization of catalytic residues, water molecules, and the positions of H/D atoms. The benefits and information obtainable by using neutrons versus X-rays can really assi st in the detailed analysis of the CA active site. To this end, fully deuterated enzyme was produced, crystallized, and the structure determined to 1.50 (Budayova-Spano et al. 2006). To show proof-of-principle that the deuterated enzyme is homologous to its hydrog enated counterpart, the X-ray structure and catalytic rate of perdeuterated HCA II was determined. The structure and kinetic measurements revealed that perdeuterated HCA II was, within error, identical to hydrogenated HCA II. The only difference was an expected isotope effect in the proton transfer component of catalysis. Chapter 5 describes a more applied a pproach compared to the fundamental questions being asked in the preceding chapters. Here X-ray crystallography was used to investigate the binding modes of novel, spac er-containing inhibitors to HCA II, BB3 and TDM. CA inhibitors (CAIs) are commonly prescribed for the treatment of glaucoma,

PAGE 148

133 altitude sickness, and epilepsy (Maren, 1987). The classical, clinically available CAIs have very poor solubility and subsequent bioavailability. Novel compounds with better solubility and different chem ical properties could have new applications. Specifically, there is significant interest in using CA Is against certain tumor types as the overexpression of HCA IX is strongly associated with some solid tumors (Wykoff et al. 2000; Giatromanolaki et al. 2001). The work described in th is chapter involves two new water-soluble compounds that employ the cl assical sulfonamide group but have an additional spacer group between the sulfonami de and aromatic ring groups. Structure determination of these com pounds bound to HCA II reveal that the spacer group confers interesting and novel binding modes to HCA II (Fisher et al. 2006). These alternated binding modes and locations in the active site could be useful for the future development of human isoform specific drugs. The recent discovery and cloning of an class CA from the mosquito larvae of Aedes aegypti (AaCA1) presents an interesting and new target for the possible control of mosquito populations (Del Pilar Corena et al. 2002). Chapter 6 illustrates the expression, purification, and subsequent kinetic and structural ch aracterization of AaCA1. The homology model indicates th at this protein is an -CA and inhibition of the enzyme with clinically used sulfonamides, methazola mide and acetazolamide, supports this conclusion. Catalytic rate measurements also revealed that AaCA 1 is a high activity isoform that displays simila r proton transfer kinetics to HCA II. However, the rate constant for CO 2 /bicarbonate interconversion is more closely resembles that of HCA I. This is interesting as AaCA1 shows highe r sequence identity with HCA I compare to HCA II (39% and 32%). The homology model of AaCA1 provides some rationale for the

PAGE 149

134 observed differences in the enzyme activity. Cr itical active site re sidues found in HCA II are conserved in AaCA1, speci fically Thr214 and other polar amino acids that line the active site cavity. In contrast, HCA I has a His at position 214 and other hydrophobic residues in the active site. Other residues of interest are Thr81 and Gln86 in AaCA1 (Val62 and His67 in HCA I, and Asn62 and As n67 in HCA II) that could be exploited for the design of selective CAIs to AaCA1. Future Directions Despite the abundance of information availa ble about CA activity and active site structure, there are still many unanswered que stions. It is unclear which active site residues directly participate in hydrogen bond interactions with water molecules and which are hydrogen bond donors and acceptors involved in proton transfer. These questions would best be answered by a comb ination of ultra-high resolution X-ray and medium resolution neutron crystallography st udies. There are several high resolution crystal structures on hand but these have been determined in complex with either inhibitors or activators (Duda et al. 2003; Jude et al. 2006). Recently, X-ray diffraction data to 1.05 of wild type HCA II, without any inhibitors or activators present, was obtained and the structure determination and analysis is currently underway. These data should give valuable information on the protona tion status of the proton shuttle His64 as well the possible identity of the zincbound solvent. Beam time on the ILL (Grenoble, France) neutron source has also been secure d and it is anticipated that the large, perdeuterated crystals of HCA II will at least diffract. Results from the complementary neutron and X-ray techniques could provide high resolution and specific active site information. Not only with these studies cont ribute an enormous amount of information to our current understanding of the CA active si te and catalysis, the information can also

PAGE 150

135 be used to understand differences between human isoforms. Also, details of the protonation state of the HCA II act ive site will be useful for applications such as rational drug design.

PAGE 151

LIST OF REFERENCES Alber B.E., Ferry J.G. (1996), Characteriza tion of heterologously produced carbonic anhydrase from Methanosarcina thermophila J. Bacteriol. 178 3270-3274. An H., Tu C.K., Duda D.M., Montanez-Clemen te I., Math K., Laipis P.J., McKenna R., Silverman D.N. (2002), Chemical rescue in catalysis by human carbonic anhydrases II and III, Biochem. 41, 3235-3242. Aronsson G., Mrtensson L.G., Carlsson U., Jonsson B.H. (1995), Folding and stability of the N-terminus of human carbonic anhydrase II, Biochem. 34, 2153-2162. Bergenhem N.C.H., Hallberg M., Wisen S. (1998), Molecular char acterization of the human carbonic-anhydrase-related protein (HCA-RP VIII), Biochimica et Biophys. Acta 1384, 294-298. Bertini I., Canti G., Luchinat C., Borghi E. (1983), Investigation of the system copper(II) carbonic anhydrase and HCO 3 /CO 2 J. Inorgan. Biochem. 18, 221-229. Bhatt D., Tu C.K., Fisher S.Z., Hernand ez-Prada J.A., McKenna R., Silverman, D.N. (2005), Proton transfer from His 200 in human carbonic anhydrase II, Proteins: Structure, Function, and Bioinformatics 61, 239-245. Boyle N.A., Chegwidden W.R., Blackbur n G.M. (2005), A new synthesis of difluoromethanesulfonamides: A nove l pharmacophore for carbonic anhydrase inhibition, Org. Biomol. Chem. 3, 222-224. Breton S. (2001), The cellular phys iology of carbonic anhydrases, J. Pancreas 2, 159164. Brockwell D., Yu L., Cooper S., McClela nd S., Cooper A., Attw ood D., Gaskell S.J., Barber J. (2001), Physicochemical consequences of the perdeuteriation of glutathione S-tranferase from S. japonicum Protein Sci. 10, 572-580. Brnger A.T., Adams P.D., Clore G.M., DeLa no W.L., Gros P., Grosse-Kuntsleve R.W., Jiang J.S., Kuszewski J., Nilges M., Pannu N.S., Read R.J., Rice L.M., Simonson T., Warren G.L. (1998), Crystallography & NMR system: A new software suite for macromolecular structure determination, Acta Crystallogr. D54 905-921. 136

PAGE 152

137 Budayova-Spano M., Fisher S.Z., Dauvergne M.T., Agbandje-McKenna M., Silverman D.N., Myles D.A.A., McKenna R. (2006), Production and X-ra y crystallographic analysis of fully deuterated human carbonic anhydrase II, Acta Crystallogr. F62, 69. Burnell J.N. (2000), Carbonic anhydrases in higher plants in: Ch egwidden WR, Carter ND, Edwards YH (Eds), The Carbonic Anhydrases, New Horizons; Birkhauser Verlag, Basel, Switzerland, 501-518. Chakravarty S., Kannan K.K. (1994), Drug-protein interactions: Refined structures of three sulfonamide drug complexes of human carbonic anhydrase I enzyme, J. Mol. Biol. 243, 298-309. Chegwidden W.R. and Carter N.D. (2000), Introduction to the carbonic anhydrases in: Chegwidden WR, Carter ND, Edwards YH (Eds), The Carbonic Anhydrases, New Horizons; Birkhauser Verlag, Basel, Switzerland, 13-28. Christianson D.W., Fierke C.A. (1996), Ca rbonic Anhydrase: Evolution of the zinc binding site by nature and design, Acc. Chem. Res. 29, 331-339. Coates L., Erskine P.T., Wood S.P., Myles D.A.A., Cooper J.B. (2001), A neutron Laue diffraction study of endothiap epsin: Implications for the aspartic protease mechanism, Biochem. 40, 13149-13157. Coleman J.E. (1967), Mechanism of action of carbonic anhydrase. S ubtrate, sulfonamide, and anion binding, J. Biol. Chem. 242, 5212-5219. Cui Q., Karplus M. (2003), Is a proton wir e concerted or stepwi se? A model study of proton transfer in carbonic anhydrase, J. Phys Chem. B 107, 1071-1078. Darsie R.F., Ward R.A. (2000), Summary of new distribution records for mosquito species in the United States a nd Canada for the period 1982-99, J. Am. Mosq. Control Assoc. 16, 1-4. Del Pilar Corena M., Seron T.J., Lehman H.K ., Ochrietor J.D., Kohn A., Tu C.K., Linser P.J. (2002), Carbonic anhydrase in the midgut of larval Aedes aegypti: Cloning, localization, and inhibition, J. Exp. Biol. 205, 591-602. Del Pilar Corena M., Fiedler M.M., Van Eker is L., Linser P.J. (2004), A comparative study of carbonic anhydrase in th e midgut of mosquito larvae, Comp. Biochem. Physiol. 137 207-225. Dodgson S.J. (1991), The carbonic anhydras es: Cellular physiol ogy and molecular genetics: Dodgson S.J., Tashian R.E., Gros G., Carter N.D. (Eds), Plenum, New York, 267-305.

PAGE 153

138 Duda D., McKenna R. (2004), Carbonic anhydrase ( -class) in: Messerschmidt A., Bode W., Cygler M. (Eds.) Handbook of Meta lloproteins, John Wiley & Sons, New York, 249-263. Duda D.M., Tu C.K., Qian M., Laipis P. J., Agbandje-McKenna M., Silverman D.N., McKenna R. (2001), Structural and kinetic analysis of th e chemical rescue of the proton transfer function of carbonic anhydrase II, Biochemistry 40, 1741-1748. Duda D.M., Govindasamy L., Agbandje-McKenna M., Tu C.K., Silverman D.N., McKenna R. (2003), The refined atomic st ructure of carbonic anhydrase II at 1.05A resolution: Implications of chem ical rescue of proton transfer, Acta Crystallogr. D59 93-104. Earnhardt J.N., Silverman D.N. (1998) Carbonic anhydrase in: Sinnott M.(Ed.) Comprehensive biological catalysis Vo lume 1, Academic Press, London, 483. Eigen M. (1964), Proton tran sfer, acid-base catalysis, and enzymatic hydrolysis, Angewandte Chemie 3, 1-19. Emsley P., Cowtan K. (2004), Coot: Model-building tools for molecular graphics, Acta Crystallogr. D60 2126-2132. Engh R.A., Huber R. (1991), Accurate bond an d angle parameters for X-ray protein structure refinement, Acta Cryst. A 47, 392-400. Eriksson A.E., Jones T.A., Liljas A. ( 1988), Refined structure of human carbonic anhydrase II at 2.0 A resolution, Proteins 4, 274-282. Esnouf R.M. (1997), An extensively modified version of MolScript that includes greatly enhanced coloring capabilities, J Mol Graph Model. 15, 132-134 and 112-113. Fisher Z., Hernandez Prada J.A., Tu C.K., Duda D., Yoshioka C., An H., Govindasamy L., Silverman D.N., McKenna R. (2005) Stru ctural and kinetic characterization of active site histidine as a proton shuttle in catalysis by human carbonic anhydrase II, Biochem. 44, 1097-1105. Fisher S.Z., Govindasamy L., Boyle N ., Agbanje-McKenna M., Silverman D.N., Blackburn G.M., McKenna R. (2006), X-ray cr ystallographic studie s reveal that the incorporation of spacer groups in carbonic anhydrase inhibitors cause alternate binding modes, Acta Crystallogr. F62 618-622. Forster R.E. (2000), Remarks on the disc overy of carbonic anhydrase in: Chegwidden WR, Carter ND, Edwards YH (Eds.), The Carbonic Anhdyrases, New Horizons; Birkhauser Verlag Ba sel, Switzerland, 1-11. Fraser C.M., Gocayne J.D., White O., Adam s M.D., Clayton R.A., Fleischmann R.D., Bult C.J., Kerlavage A.R., Sutton G.G., Kelley J.M. et al. (1995), The minimal gene complement of Mycoplasma genitalium Science 270, 397-403.

PAGE 154

139 Fujikawa-Adachi K., Nishimori I., Taguchi T., Onishi S. (1999), Human carbonic anhydrase XIV (CA14): cDNA cloning, mRNA expression, and mapping to chromosome 1, Genomics 61, 74-81. Giatromanolaki A., Koukourakis M.I., Sivridis E., Pastorek J., Wykoff C.C., Gatter K.C., Harris A.L. (2001), Expression of hypoxiainducible carbonic anhydrase-9 relates to angiogenic pathways and independently to poor outcome in non-small cell lung cancer, Cancer Res. 61, 7992-7998. Gubler D.J. (1997), Dengue and dengue hemorrh agic fever. Its histor y and resurgence as a global health problem, in Gubler D.J. and Kuno G. (Eds.) Dengue and Dengue Hemorrhagic Fever. CAB International, Wallingford, United Kingdom, 1-22. Guex N., Diemand A., Peitsch M.C. (1999), Protein modeling for all, TIBS 24, 364-367. Guilloton M.B., Korte J.J., Lamblin A.F., Fuchs J.A., Anderson P.M. (1992), Carbonic anhydrase in Eschericia coli : A product of the cyn operon, J. Biol. Chem. 267, 3731-3734. Habash J., Raftery J., Weisgerber S., Casse tta A., Lehmann M.S., Hghj P., Wilkinson C., Campbell J.W., Helliwell J.R. (19 97), Neutron Laue diffraction study of concanavalin A: The proton of Asp28, J. Chem. Soc Faraday Trans. 93, 43134317. Habash J., Raftery J., Nuttall R., Price H.J ., Wilkinson C., Kalb A.J., Helliwell J.R. (2000), Direct determination of the pos ition of the deuterium atoms of the bound water in concanavalin A by neutron Laue crystallography, Acta Cryst. D65 541550. Hkansson K., Carlsson M., Svensson L.A., Lilj as A. (1992), Structure of native and apo carbonic anhydrase II and structure of some of its anion-ligand complexes, J. Mol. Biol. 227, 1192-1204. Halstead S.B. (1997), Epidemiology of dengue and dengue hemorrhagic fever, in Gubler D.J. and Kuno G. (Eds.) Dengue and Dengue Hemorrhagic Fever. CAB International, Wallingf ord, United Kingdom, 23-44. Henriques O.M. (1928), Die bindungsweise des kohlendioxids im blute, Biochem. Ztschrft 200, 1-24. Henry R.P. (1996), Multiple roles of carbonic anhydrase in cellu lar transport and metabolism, Annu. Rev. Physiol. 58, 523-538. Hewett-Emmett D., Tashian R.E. (1996), Functional diversity, conservation, and convergence in the evolution of the al pha-, beta-, and gamma-carbonic anhydrase gene families, Mol. Phylogen. Evo. 5, 50-77.

PAGE 155

140 Iverson T.M., Alber B.E., Kisker C., Ferry J.G., Rees D.C. (2000), A closer look at the active site of gamma-class carbonic high-reso lution crystallographic studies of the carbonic anhydrase from Methanosarcina thermophila Biochem. 39, 9222-9231. Jewell D.A., Tu C.K., Paranawithana S.R., Ta nhauser S.M., LoGrasso P.V., Laipis P.J., Silverman D.N. (1991), Enhancement of th e catalytic properties of human carbonic anhydrase III by site-directed mutagenesis, Biochem. 30, 1484-1490. Jones T.A., Zou J.Y., Cowan S.W., Kjel dgaard M. (1991), Improved methods for building protein models in electron density maps and the location of errors in these models, Acta Crystallogr. A47 110-119. Jonsson B.M., Hakansson K., Liljas A. (1993), The structure of human carbonic anhydrase II in complex with bromide and azide. FEBS Lett. 322, 186-190. Jude K.M., Wright S.K., Tu C.K., Silverman D.N., Viola R.E., Christianson D.W. (2002), Crystal structure of F65A/Y131C-methyl imidazole carbonic anhydrase V reveals architectural features of an engineered proton shuttle, Biochem. 41, 2485-2491. Jude K.M., Banerjee A.L., Haldar M.K., Ma nokaran S., Roy B., Mallik S., Srivastava D.K., Christianson D.W. (2006), Ultrahigh resolution crystal st ructures of human carbonic anhydrase I and II comp lexed with two-prong reveal the molecular basis of high affinity, J. Am. Chem. Soc. 128, 3011-3018. Kato K. (1990), Sequence of a novel carbon ic anhydrase-related polypeptide and its exclusive presence in the Purkinje cells, FEBS Letters 271, 137-140. Kannan K.K., Notstrand B., Fridborg K., Lvgren S., Ohlsson A., Petef M. (1975), Crystal structure of human erythrocyte carbonic anhydrase B. Three-dimensional structure at a nominal 2.2resolution, PNAS 72 51-55. Kannan K.K., Petef M., Fridbor g K., Cid-Dresdner H., Lovgr en S. (1977), Structure and function of carbonic anhydrases. Imidazole binding to human carbonic anhydrase B and the mechanism of acti on of carbonic anhydrases, FEBS Let 73, 115-119. Khalifah R.G. (1971), The carbon dioxide hydr ation activity of carbonic anhydrase. I. Stop-flow kinetic studies on the na tive human isoenzymes B and C, J. Biol. Chem. 246, 2561-2573. Khalifah R.G., Strader D.J., Bryant S. H., Gibson S.M. (1977), Carbon-13 nuclear magnetic resonance probe of active-site ionizations in hum an carbonic anhydrase B, Biochemistry 16, 2241-2247. Kim G., Lee T.H., Wetzel P., Geers C., R obinson M.A., Myers T.G., Owens J.W., Wehr N.B., Eckhaus M.W., Gros G., WynshawBoris A., Levine R.L. (2004), Carbonic anhydrase III is not required in the mouse for normal growth, development, and life span, Mol Cell Biol. 24, 9942-9947.

PAGE 156

141 Kim S., Rabbani Z., Dewhirst M., Vujaskovic Z., Vollmer R., Schreiber E., Oosterwijk E., Kelley M. (2005), Expression of HIF1alpha, CA IX, VEGF, and MMP-9 in surgically resected non-small cell lung cancer, Lung Cancer 49, 325-335. Kisker C., Schindelin H., Alber B.E., Ferry J. G., Rees D.C. (1996), A left-hand beta-helix revealed by the crystal structure of a carbonic an hydrase from the archaeon Methanosarcina thermophila EMBO J. 15, 2323-2330. Krebs J.F., Fierke C.A. (1991), Conforma tional mobility of His64 in the Thr200 Ser mutant of human carbonic anhydrase II, Biochem. 30, 9153-9160. Krebs J.F., Rana F., Dluhy R.A., Fierke C.A. (1993), Kinetic and sp ectroscopic studies of hydrophilic amino acid substitutions in the hydrophobic pocket of human carbonic anhydrase II, Biochem. 32, 4496-4505. Kresge J.A., Silverman D.N. (1999), Applicatio n of Marcus rate theory to proton tansfer in enzyme-catalyzed reactions, Methods Enzym. 308, 276-297. Langan P., Greene G., Schoenborn B.P. (2004), Protein crystallography with spallation neutrons: the user facility at Lo s Alamos Neutron Science Center, J. Appl. Cryst. 37, 24-31. Laskowski R.A., MacArthur M.W., Moss D.S., Thornton J.M. (1993), PROCHECK: A program to check the stereochemical quality of protein structures, J. Appl. Cryst. 26, 283-291. Lehtonen J., Shen B., Vihinen M., Casini A., Scozzafava A., Supuran C.T., Parkkila A.K., Saarnio J., Kivela A.J., Waheed A., Sly W.S., Parkkila S.J. (2004), Characterization of CA X III, a novel member of the carbonic anhydrase isozyme family, J. Biol Chem. 279, 2719-2727. Liang Z., Xue Y., Behravan G., Jonsson B.H., Lindskog S. (1993), Importance of the conserved active-site resi dues Tyr7, Glu106 and Thr199 fo r the catalytic function of human carbonic anhydrase II, Eur J Biochem. 211, 821-827. Liang J.Y., Lipscomb W.N. (1990), Binding of substrate CO 2 to the active site of human carbonic anhydrase II: A molecular dynamics study, PNAS 87, 3675-3679. Liang Z., Jonsson B.H., Lindskog S. (1993), Prot on transfer in the catalytic mechanism of carbonic anhydrase. Effects of placing histidine re sidues at various positions in the active site of human isozyme II, Biochim. Biophys. Acta 1203 142-146. Liljas A., Kannan K.K., Bergstn P.C., Waar a I., Fridborg K., Strandberg B., Carlbom U., Jrup L., Lvgren S., Petef M. (1972) Crystal structure of human carbonic anhydrase C, Nature New Biol. 235, 131-137. Liljas A., Hkansson K., Jonsson B.H., Linds kog S. (1994), Inhibition and catalysis of carbonic anhydrase. Recent cr ystallographic analyses, Eur. J. Biochem. 219, 1-10.

PAGE 157

142 Lindskog S. (1997), Structure and mechanism of carbonic anhydrase, Pharmacol. Ther. 74, 1-20. Lindskog S., Silverman D.N. (2000), The catalytic mechanism of mammalian carbonic anhydrases in: Chegwidden W.R., Carter N. D., Edwards Y.H. (Eds.), The Carbonic Anhdyrases, New Horizons; Birkhauser Verlag Basel, Switzerland, 175-195. LoGrasso P.V., Tu C.K., Jewell D.A., Wynns G.C., Laipis P.J., Silverman D.N. (1991), Catalytic enhancement of human carboni c anhydrase III by replacement of phenylalanine-198 with leucine, Biochem. 30, 8463-8470. Mann T., Keilin D. (1940), Sulfanilamide as a specific inhibitor of carbonic anhydrase, Nature 146, 164-165. Mansoor U.F., Zhang X.R., Blackburn G. M. (2000), The design of new carbonic anhydrase inhibitors in: Chegwidden W.R., Carter N. D., Edwards Y.H. (Eds.), The Carbonic Anhdyrases, New Horizons; Bir khauser Verlag Basel, Switzerland, 437459. Maren T.H. (1967), Carbonic anhydrase: Chemistry, phyiosology, and inhibition, Physiol. Rev. 47, 595-781. Maren T.H. (1974), HCO 3 formation in aqueous humour. Mechanism and relation to the treatment of glaucoma, Invest. Opthalmol. 13, 479-484. Maren T.H.(1987), Carbonic anhydrase: general perspectives and advances in glaucoma research. Drug Dev. Res. 10, 255-276. Maren T.H., Conroy C.W. (1993), A new class of carbonic anhydrase inhibitor, J. Biol. Chem. 268, 26233-26239. Maren T.H., Sanyal G. (1983), The activity of sulfonamides and anions against the carbonic anhydrases of animals, plants, and bacteria, Ann. Rev. Pharmacol. Toxicol. 23, 439-459. Martin M.M., Martin J.S., Kukor J.J., Merr it R.W. (1980), The digestion of protein and carbohydrate by the stream detritovore, Tipula abdominalis (Diptera, Tipulidae), Oecologia 46 360-364. Marx D., Tuckerman M.E., Hutter J., Parrine llo M. (1999), The nature of the hydrated excess proton in water, Nature 397, 601-604. McPherson A. (1982), Preparation and Analys is of Protein Crystals, Wiley, New York. Meilleur F., Contzen J., Myles D.A.A., Jung C. (2004), Structural stability and dynamics of hydrogenated and perdeuterated cytochrome P450cam (CYP101), Biochem. 43, 8744-8753.

PAGE 158

143 Meldrum N.U., Roughton F.J.W. (1932), Some properties of car bonic anhydrase, the CO 2 -enzyme present in blood, J. Physiol. (London) 75, 15-16. Meldrum N.U., Roughton F.J.W. (1933), Ca rbonic anhydrase: Its preparation and properties, J. Physiol. (London) 80, 113-142. Merritt E.A., Bacon D.J. (1997), Raster3D Ve rsion 2: Photorealistic molecular graphics, Methods of Enzymology 277, 505-524. Merz K.M. (1990), Insights into the f unction of the zinc hydroxide-Thr199-Glu106 hydrogen bonding network in carbonic anhydrases, J Mol Biol. 214, 799-802. Merz K.M. (1991), Carbon dioxide binding to human carbonic anhydrase II J. Am. Chem. Soc. 113, 406-411. Mitsuhashi S., Mizushima T., Yamashita E., Yamamoto M., Kumasaka T., Moriyama H., Ueki T., Miyachi S., Tsukihara T. ( 2000), X-ray structure of beta-carbonic anhydrase from the red alga, Porphyridium purpureum reveals a novel catalytic site for CO(2) hydration, J. Biol. Chem. 275, 5521-5526. Mohammed O.F., Pines D., Dreyer J., Pi nes E., Nibbering E.T.J (2005), Sequential proton transfer through water br idge in acid-base reactions, Science 310, 83-86. Murakami H., Sly W.S. (1987), Purificati on and characterizati on of human salivary carbonic anhydrase, J. Biol. Chem. 262, 1382-1388. Murakami Y., Kanada K., Tsuji M., Kanaya ma H., Kagawa S. (1999), MN/CA9 gene expression as a potential biomar ker in renal cell carcinoma, BJU Int. 83, 743-747. Myles D.A.A., Bon C., Langan P., Cipriani F., Castagna J.C., Lehmann M.S., Wilkinson C. (1998), Neutron Laue diffraction in macrom olecular crystallography, Physica B 241-243, 1122-1130. Nagle J.F., Morowitz H.J. (1978), Molecu lar mechanisms of proton transport in membranes, PNAS 75, 298-302. Nair S.K., Christianson D.W. (1991), Unexp ected pH-dependent conformation of His-64, the proton shuttle of carbonic anhydrase II J. Am. Chem. Soc. 113, 9455-9458. Nair S.K., Ludwig P.A., Christianson D.W. (1994), Two-Site Bi nding of phenol in the active site of human carboni c anhydrase II: Structural im plications for substrate association, J. Am. Chem. Soc. 116, 3659-3660. Nair S.K., Krebs J.F., Christianson D.W., Fierke C.A. (1995), Structural basis of inhibitor affinity to variants of human carbonic anhydrase II, Biochem. 34, 3981-3989.

PAGE 159

144 Nicholls A., Sharp K., Honig B. (1991), Prot ein folding and association: Insights from the interfacial and thermodyna mic properties of hydrocarbons, Proteins 11 281296. Niimura N. (1999), Neutrons expand the field of structural biology, Curr. Opin. Struct. Biol. 9, 602-608. Niimura N., Arai S., Kurihara K., Chatake T ., Tanaka I., Bau R. (2006), Recent results on hydrogen and hydration in biology st udies by neutron macromolecular crystallography, Cell. Mol. Life Sci. 63, 285-300. Norvell J.C., Nunes A.C., Schoenborn B.P. (1975), Neutron diffraction analysis of myoglobin: Structure of the carbon monoxide derivative, Science 190, 568-569. Ortova Gut M., Parkkila S., Vernerov Z.K., Rohde E., Zvada J., Hocker M., Pastorek J., Karttunen T., Gibadulinov A., Zva dov Z., Knobeloch K.P., Wiedenmann B., Svoboda J., Horak I., Pastorekova S. ( 2002), Gastric hyperplasia in mice with targeted disruption of the carbonic anhydrase gene Car9, Gastroenterol. 123 18891903. Osborne W.R.A., Tashian R.E. (1974), Ther mal inactivation studies of normal and variant human erythrocyte carbonic anhydr ases by using a sulphonamide-binding assay, Biochem. J. 141, 219-225 Otwinowski Z., Minor W. (1997) Processing of X-ray diffraction data collected in oscillation mode, Methods Enzymol 276, 307-326. Pastorekov S., Zvadov Z., Kotl M., Babuikov O., Zvada J. (1992), A novel quasi-viral agent, Ma-Tu, is a two-component system, Virology 187, 620-626. Peitsch M.C. (1996), Promod and Swiss-Mode l: Internet-based tools for automated comparative protein modeling, Biochem. Soc. Trans. 24, 372-279. Pines E., Magnes E.Z., Lang M.J., Fleming G.R. (1997), Direct measur ement of intrinsic proton transfer rates in di ffusion-controlled reactions, Chem. Phys. Letters 281, 413-420. Poms R., Roux B. (1996), Theoretical study of H + translocation along a model proton wire, J. Phys. Chem. 100, 2519-2527. Rossmann M.G. (1990), The molecular replacement method, Acta Cryst. A 46, 73-82. Rowlett R.S., Silverman D.N. (1982), Kinetic s of the protonation of buffer and hydration of CO 2 catalyzed by human carbonic anhydrase II, J. Am. Chem. Soc. 104, 67376741. Sanyal G., Swenson E.R., Pessah N.I., Mare n T.H. (1982), The carbon dioxide hydration activity of skeletal mu scle carbonic anhydrase, Mol. Pharmacol. 22, 211-220.

PAGE 160

145 Scheiner S., Yi M. (1996), Proton transfer properties of imidazole, J. Phys. Chem. 100, 9235-9241. Schultz A.J., Thiyagarajan P., Hodges J.P., Rehm C., Myles D.A.A., Langan P., Mesecar A. (2005), Conceptual design of a macrom olecular neutron diffractometer (MaNDi) for the SNS, J. Appl. Cryst. 38, 964-974. Schwede T., Kopp J., Guex N., Peitsch M.C. (2003), Swiss-Model: An automated protein homology-modeling server, Nucleic Acids Research 31, 3381-3385. Segal I.H. (1975), Enzyme kinetics, Wiley-Interscience, New York, 150-159. Seron T.J., Hill J., Linser P.J. (2004), A GPI-linked carbonic anhydrase expressed in the larval mosquito midgut, J. Exp. Biol. 207, 4559-4572. Shah G.U., Ulmasov B., Waheed A., Becker T., Makani S., Svichar N., Chesler M., Sly W.S. (2005), Carbonic anhydrase IV and XIV knock-out mice: Roles of the respective carbonic anhydrases in buffe ring the extracellular space in brain, PNAS 102, 16771-16776. Sheldrick G.M., Schneider T.R. (1997), Macromolecular Crystallography, Part B, Sweet R.M. and Carter C.W. Jr. (Eds.) Methods Enzymol. 277, 319-343. Shu F., Ramakrishnan V., Schoenborn B.P. (2000), Enhanced visibility of hydrogen atoms by neutron crystallography of fully deuterated myoglobin, Proc. Natl. Acad. Sci. USA 97, 3872-3877. Schttelkopf A.W., van Aalten D.M.F. (2004), PRODRG: A tool for high-throughput crystallography of protein-ligand complexes, Acta cryst. D 60, 1355-1363. Silverman D.N. (1982), Carbonic anhydrase: Oxygen-18 exchange catalyzed by an enzyme with rate-c ontributing steps, Methods Enzymol. 87, 732-752 Silverman D.N. (2000), Marcus rate theory applied to enzymatic proton transfer, Biochimica Biophys. Acta 1458, 88-103. Silverman D.N., Tu C.K., Lindskog S., Wynns G.C. (1979), Rate of exchange of water from the active site of human carbonic anhydrase C, J. Am. Chem Soc. 101, 6734 6740. Silverman D.N., Tu C.K., Chen X., Tanhauser S.M., Kresge A.J., Laipis P.J. (1993), Rate-equilibria relationships in intramolecular proton transfer in human carbonic anhydrase III, Biochemistry 32, 10757-10762. Silverman D.N., Lindskog S. (1988), The catalytic mechanism of carbonic anhydrase: Implications of a rate-limiting protolysis of water, Acc. Chem. Res. 21, 30-36.

PAGE 161

146 Simonsson .I., Lindskog S. (1982), The interac tion of sulfate with carbonic anhydrase, Eur. J. Biochem. 123, 29-36. Skaggs L.H., Bergenhem N.C.H., Venta P.J ., Tashian R.E. (1993) The deduced amino acid sequence of human car bonic anhydrase-related protein (CARP) is 98% identical to the mouse homologue, Gene 126, 291-292. Sly W.S., Hu P.Y. (1995), Human carboni c anhydrases and carbonic anhydrase deficiencies, Annu. Rev. Biochem. 64, 375-401. So A.K.C., Espie G.S., Williams E.B., Shiv ely J.M., Heinhorst S., Cannon G.C. (2004), A novel evolutionary lineage of carbonic anhydrase ( class) is a component of the carboxysome shell, J. of Bacteriol. 186 623-630. Spielman A., DAntonio M. (2001), Mosquito: The Story of Mans Deadliest Foe, Hyperion Books, New York. Stams T., Chen Y., Boriack-Sjodin P.A., Hurt J.D., Liao J., May J.A., Dean T., Laipis P., Silverman D.N., Christianson D.W. ( 1998), Structures of murine carbonic anhydrase IV and human carbonic anhydrase II complexed with brinzolamide: molecular basis of isozyme-drug discrimination, Protein Sci. 7, 556-563. Stams T., Nair S.K., Okuyama T., Waheed A., Sly W.S., Christianson D.W. (1996), Crystal structure of the secretory form of membrane-associated human carbonic anhydrase IV at 2.8resolution, PNAS 93, 13589-13594. Steiner H., Jonsson B.H., Lindskog S. ( 1975), The catalytic m echanism of carbonic anhydrase. Hydrogen-isotope effects on th e kinetic parameters of the human C isoenzyme, Eur. J. Biochem. 59, 253-259. Strop P., Smith K.S., Iverson T.M., Ferry J.G ., Rees D.C. (2001), Crystal structure of the "cab"-type beta class carboni c anhydrase from the archaeon Methanobacterium thermoautotrophicum J Biol Chem. 276, 10299-10305. Supuran C.T., Scozzafava A., Conway J. ( 2004), Carbonic Anhydrase: Its Inhibitors and Activators, CRC Press, Boca Raton. Tanhauser S.M., Jewell D.A., Tu C.K., Silv erman D.N., Laipis P.J. (1992), A T7 expression vector optimized for site-directed mutagenesis using oligodeoxyribonucleot ide cassettes, Gene 117, 113-117. Taniuchi K., Nishimori I., Takeuchi T., Fujikawa-Adachi K., Ohtsuki Y., Onishi S. (2002), Developmental expression of carbonic anhydrase-related proteins VIII, X, and XI in the human brain, Neuroscience 112, 93-99. Tashian R.E. (1992), Genetics of mammalian carbonic anhydrases, Adv. Genet. 30, 321356.

PAGE 162

147 Thiel A. (1913), ber die langsame neutralisation des kohlensure, Berichte Deutsche Chem Gesellschaft 46, 241-244. Thompson J.D., Higgins D.G., Gibson T.J. (1994), Using Clustal for multiple sequence alignments, Nucleic Acids Research 22, 4623-4680. Tibell L., Forsman C., Simonsson I., Lindskog S. (1985), The inhibition of human carbonic anhydrase II by some organic compounds, Biochim Biophys Acta. 829, 202-208. Tripp B.C., Ferry J.G. (2000), A structure-f unction study of a proton transport pathway in the gamma-class carbonic anhydrase from Methanosarcina thermophila Biochem. 39, 9232-9240. Tripp B.C., Smith K., Ferry J.G. (2001), Ca rbonic anhydrase: New insights for an ancient enzyme J. Biol. Chem. 276, 48615-48618. Tu C.K., Couton J.M., Van Heeke G., Richards N.G.J., Silverman D.N. (1993), Kinetic analysis of a mutant (His107-Tyr) res ponsible for human carbonic anhydrase II deficiency syndrome, J. Biol. Chem. 268, 4775-4779. Tu C.K., Silverman D.N., Forsman C., J onsson B.H., Lindskog S. (1989), Role of hisitidine 64 in the catalytic mechanis m of human carbonic anhydrase II studied with a site-specific mutant, Biochem. 28, 7913-7918. Tufts B.L., Esbaugh A., Lund S.G. (2003) Comparative physiology and molecular evolution of carbonic anhydrase in the er ythrocytes of early vertebrates, Comp. Biochem. Physiol. Part A 136, 259-269. Treci O., Sahin U., Vollmar E., Siemer S., Gottert E., Seitz G., Parkkila A.K., Shah G.N., Brubb J.H., Pfreundchuh M., Sly W. S. (1998), Human carbonic anhydrase XII: cDNA cloning, expression, and chro mosomal localization of a carbonic anhydrase gene that is overexpressed in some renal cell cancers, PNAS 95, 76087613. Ulmasov B., Waheed A., Shah G.N., Grubb J. H., Sly W.S., Tu C.K., Silverman D.N. (2000), Purification and kinetic analysis of recombinant CA XII, a membrane carbonic anhydrase overexpressed in certain cancers, PNAS 97, 14212-14217. Venkatasubban K.S., Silverman D.N. ( 1980), Carbon dioxide hydration activity of carbonic anhydrase in mixtures of water and deuterium oxide, Biochem. 19, 49844989. Venta P.J. (2000), Inherited deficiencies and activity variants of the mammalian carbonic anhydrase in: Chegwidden WR, Carter ND, Edwards YH (Eds.), The Carbonic Anhdyrases, New Horizons; Birkhauser Verlag Basel, Switzerland, 403-412.

PAGE 163

148 Venta P.J., Welty R.J., Johnson T.M., Sly W. S., Tashian R.E. (1991) Carbonic anhydrase deficiency syndrome in a Belgian family is caused by a point mutation at an invariant hisitidine residue (107 His Ty r): Complete structure of the normal human CA II gene, Am. J. Hum. Genet. 49, 1082-1090. Vidgren J., Liljas A., Walker N.P. (1990), Refined structure of the acetazolamide complex of human carboni c anhydrase II at 1.9 A, Int. J. Biol. Macromol. 12, 342344. Lu D. Voth G.A. (1998), Molecular dynami cs simulations of human carbonic anhydrase II: Insight into experimental re sults and the role of solvation, Proteins 33, 119-134. Wallace A.C., Laskowski R.A., Thornton J.M. (1995), LIGPLOT: A program to generate schematic diagrams of pr otein-ligand interactions, Protein Eng. 8, 127-134. Whittington D.A., Waheed A ., Ulmasov B., Shah G.N., Grubb J.H., Sly W.S., Christianson D.W. (2001), Cr ystal structure of the dimeri c extracellular domain of human carbonic anhydrase XII, a bitopi c membrane protein overexpressed in certain cancer tumor cells, PNAS 98, 9545-9550. Whittington D.A., Grubb J.H., Waheed A., Shah G.N., Sl y W.S., Christianson D.W. (2004), Expression, assay, and structure of the extracellular do main of murine carbonic anhydrase XIV: Implications fo r selective inhibi tion of membraneassociated isozymes, J. Biol. Chem. 279, 7223-7228. Wykoff C.C., Beasley N.J., Watson P.H., Turner K.J., Pastorek J., Sibtain A., Wilson G.D., Turley H., Talks K.L., Maxwell P.H ., Pugh C.W., Ratcliffe P.J., Harris A.L. (2000), Hypoxia-inducible expression of tumor-associated carbonic anhydrases, Cancer Res. 60, 7075-7083. Xue Y., Liljas A., Jonsson B.H., Lindskog S. (1993), Structural analysis of the zinc hydroxide-Thr199-Glu106 hydrogen bond netw ork in human carbonic anhydrase II, Proteins 17, 93-106.(a) Xue Y., Vidgren J., Svensson S., Liljas A., Jonsson B.H., Lindskog S. (1993), Crystallographic analysis of Thr-200--> His human carbonic anhydrase II and its complex with the substrate, HCO 3 Proteins Struct. Funct. Genet. 15, 80-87.(b)

PAGE 164

BIOGRAPHICAL SKETCH Suzanne Zo Fisher was born in Cape To wn, South Africa, in 1976. She graduated from the high school Jan van Riebeeck in 1994 and spent a year afte r that working and traveling around Europe. In 1996 she entere d the University of Stellenbosch and completed her undergraduate B.Sc. in 1999, majoring in biochemistry and animal physiology. In 2000 Zo then went on to obtai n her B.Sc.(Hons.) in the Biochemistry Department under the direction of Prof. Ja net Hapgood. At the end of that year, she moved to the United States to work as a lab technician for Dr. Christopher West, Professor of Anatomy and Cell Biology at the University of Florida. In the Fall of 2002, she started in the Interdisciplinary Program a nd subsequently joined the lab of Dr. Robert McKenna. She worked on the elucidation of the proton transfer mechanism in carbonic anhydrase using a structur al biology approach. 149


Permanent Link: http://ufdc.ufl.edu/UFE0016680/00001

Material Information

Title: Structural Insights into the Active Site of Alpha-Carbonic Anhydrases
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0016680:00001

Permanent Link: http://ufdc.ufl.edu/UFE0016680/00001

Material Information

Title: Structural Insights into the Active Site of Alpha-Carbonic Anhydrases
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0016680:00001


This item has the following downloads:


Table of Contents
    Title Page
        Page i
        Page ii
    Dedication
        Page iii
    Acknowledgement
        Page iv
        Page v
    Table of Contents
        Page vi
        Page vii
        Page viii
    List of Tables
        Page ix
    List of Figures
        Page x
        Page xi
    Abbreviations
        Page xii
        Page xiii
    Abstract
        Page xiv
        Page xv
    Introduction
        Page 1
        Page 2
        Page 3
        Page 4
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
        Page 15
        Page 16
        Page 17
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
        Page 24
        Page 25
        Page 26
        Page 27
    Structural and kinetic analysis of proton shuttling in human carbonic anhydrase II
        Page 28
        Page 29
        Page 30
        Page 31
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
        Page 43
        Page 44
        Page 45
        Page 46
        Page 47
        Page 48
        Page 49
        Page 50
    Structural and kinetic effect of hyrdrophobic mutations in the active site of human carbonic anhydrase II
        Page 51
        Page 52
        Page 53
        Page 54
        Page 55
        Page 56
        Page 57
        Page 58
        Page 59
        Page 60
        Page 61
        Page 62
        Page 63
        Page 64
        Page 65
        Page 66
        Page 67
        Page 68
        Page 69
        Page 70
        Page 71
        Page 72
        Page 73
        Page 74
    Working towards a neutron structure of perdeuterated human carbonic anhydrase II
        Page 75
        Page 76
        Page 77
        Page 78
        Page 79
        Page 80
        Page 81
        Page 82
        Page 83
        Page 84
        Page 85
        Page 86
        Page 87
        Page 88
        Page 89
        Page 90
        Page 91
        Page 92
        Page 93
        Page 94
        Page 95
        Page 96
        Page 97
        Page 98
    Novel inhibitors and their binding modes to HCA II
        Page 99
        Page 100
        Page 101
        Page 102
        Page 103
        Page 104
        Page 105
        Page 106
        Page 107
        Page 108
        Page 109
        Page 110
        Page 111
        Page 112
        Page 113
    Expression, purification, kinetic, and structural characterization of an alpha-class CA from Aedes aegypti
        Page 114
        Page 115
        Page 116
        Page 117
        Page 118
        Page 119
        Page 120
        Page 121
        Page 122
        Page 123
        Page 124
        Page 125
        Page 126
        Page 127
        Page 128
        Page 129
    Conclusions and future directions
        Page 130
        Page 131
        Page 132
        Page 133
        Page 134
        Page 135
    References
        Page 136
        Page 137
        Page 138
        Page 139
        Page 140
        Page 141
        Page 142
        Page 143
        Page 144
        Page 145
        Page 146
        Page 147
        Page 148
    Biographical sketch
        Page 149
Full Text












STRUCTURAL INSIGHTS INTO THE ACTIVE SITE OF ALPHA-CARBONIC
ANHYDRASES


















By

SUZANNE ZOE FISHER


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


2006
































Copyright 2006

By

Suzanne Zoe Fisher
































This document is dedicated to my parents, Alet and Peter Fisher.















ACKNOWLEDGMENTS

I would like to thank my advisor and mentor, Dr. Robert McKenna. Throughout my

stay in his lab, he was an inspiration and a constant source of enthusiasm and

encouragement. He was always available for questions and very patient, to say the least.

Without him, none of this work would have been possible and I thank him deeply and

sincerely for his outstanding guidance over the years. Another key person to thank is Dr.

David Silverman. I have worked closely with him and his lab and appreciate the excellent

advice and insightful discussions that we have had. I would also like to thank my other

committee members, Dr. James Flanegan and Dr. Maurice Swanson. Together, my

committee has shown nothing but support and respect for me and my work and their

supervision has kept me on track and focused on my goals.

I would like to thank some of the other labmembers and friends. Mavis Agbandje-

McKenna, Lakshmanan Govindasamy, Deepa Bhatt, and Huyn-Joo Nam have all given

me a lot of assistance, general advice, and technical expertise. The graduate students John

Domsic, Brittney Whitaker, and Nicolette Case have made a lot of tough times

manageable and provided many a good laugh. I would also like to thank the

undergraduates Edward Miller, Mike DiMattia, and Caroli Genis for great conversations

and being good company. The McKenna lab really is a special and unique place and I

will miss it a great deal.

I would also like to thank Kathy Conture, Susan Gardner, and Wayne McCormack

who are part of the administrative staff for the IDP and have helped me with all things









practical over the years. Also, Pat Jones and Terry Rickey from the Biochemistry and

Molecular Biology Department have been helpful with everyday things, from registering

for class to purchasing chemicals for the lab.

Lastly, I would like to thank my family. Despite living far away in South Africa,

they have always supported and encouraged me to stay in the United States and continue

my studies. A special thank you also goes to my American family, the Hunts. They have

been generous and kind by welcoming me into their home and making me a part of their

family.















TABLE OF CONTENTS



A C K N O W L E D G M E N T S ................................................................................................. iv

LIST OF TABLES .................................. ........... ............................ ix

L IST O F FIG U RE S .............. ......................... ........................... ....................... .. .. .... .x

ABBREVIATIONS ............................. ............ .................................... xii

A B S T R A C T ..................................................................................................................... x iv

CHAPTER

1 IN TR O D U C T IO N ........ .. ......................................... ..........................................1.

The Discovery of Carbonic Anhydrase ....................... ...........................................
A Fine Example of Convergent Evolution ..................................................2...
H um an Carbonic A nhydrases (H CA s) .................................................... ...............5...
Structure of H CA II ...................... ...... ............................. .... ...... ...............9....
Activity and Catalytic M echanism of CAs ................................................... 10
Substrate Binding in the CA A ctive Site ................................................................. 11
M easuring C A A activity. ................................................................... ............... 13
H C A and H um an D disease .......................................... ......................... ............... 14
C A In h ib ito rs .............................................................................................................. 16

2 STRUCTURAL AND KINETIC ANALYSIS OF PROTON SHUTTLING IN
HUMAN CARBONIC ANHYDRASE II.............................................................28

In tro d u c tio n ................................................................................................................ 2 8
M materials and M ethods .. ..................................................................... ................ 32
E n zy m e s .............................................................................................................. 3 2
C ry sta llo g rap h y ................................................................................................... 3 3
Activity Analysis by 180 Exchange................................................35
R results and D discussion .............. .................. ............................................... 36
Crystallography ........ ........ ... .................... ...................... 36
Effect of pH on the Wild Type HCA II Active Site ......................................37
Effect of pH on the Mutant HCA II Active Site.............................................39
C ataly sis .............. ................................................ ................................. . 4 1
C o n c lu sio n ........................................................................................................... 4 2









3 STRUCTURAL AND KINETIC EFFECTS OF HYDROPHOBIC MUTATIONS
IN THE ACTIVE SITE OF HUMAN CARBONIC ANHYDRASE II..................51

In tro d u c tio n ................................................................................................................. 5 1
M materials and M ethods .. ..................................................................... ................ 52
E nzym es .............. ...................................................................... ......52
Crystallography .............................................53
K inetics and A activity A nalysis....................................................... ................ 54
R results and D discussion ........................................................... ...... ........ .. ........ ..... 55
Structural Effects of Hydrophobic M stations ................................ ................ 55
Kinetic Effects of Hydrophobic M stations .................................... ................ 58
Solvent Structure and Implications for Proton Transfer.................................60
C o n c lu sio n ........................................................................................................... 6 3

4 WORKING TOWARDS A NEUTRON STRUCTURE OF PERDEUTERATED
HUMAN CARBONIC ANHYDRASE II.............................................................75

Introduction ................................................................................... ...................... 75
M materials and M ethods .......................................................................... ................ 82
Production and Crystallization of Perdeuterated HCA II ....................................82
C ry stallography ...... .............................. .......................... ..................... 83
Activity Analysis by 180 Exchange Methods ................................................84
Results and Discussion ........................................................ .. .. ................. 84
Structural Effects of Perdeuteration ............................................... ................ 85
K inetic Effects of Perdeuteration ................................................... 87
C conclusion .............. ........................................................................ . ..... 88

5 NOVEL INHIBITORS AND THEIR BINDING MODES TO HCA II.................99

In tro d u c tio n ................................................................................................................. 9 9
M materials and M ethods ................................................. ..... ............... ............... 10 1
Expression, Purification, and Crystallization...... .................. .................. 101
Synchrotron X-ray Data Collection........................................ 101
Structure Determination and Model Refinement ................... ...................102
R results and D discussion .............. ...... ............ ............................................... 103
C conclusion .............. ....................................................................... . .... 105

6 EXPRESSION, PURIFICATION, KINETIC, AND STRUCTURAL
CHARACTERIZATION OF AN ALPHA-CLASS CA FROM AEDES
A E G Y P T I (A A C A 1) ................................................................................................ 114

Introduction ................................................................................. ...................... 114
M materials and M ethods ................... .............................................................. 116
Expression and Purification...... ............. ............ ..................... 116
Activity Analysis by 180 Exchange....................................... 117
Determ nation of Inhibition Constants ...... .... ..................................... 118
M odel B building ......................................................................................... 118









R esu lts an d D iscu ssion ............................................................................................. 1 19
C o n c lu sio n ................................................................................................................ 1 2 2

7 CONCLUSIONS AND FUTURE DIRECTIONS ....................... ...................130

Sum m ary and C onclu sions ....................................................................................... 130
Future Directions ................... .................... ........ ................... 134

LIST OF REFEREN CES ...................... ............................................................... 136

BIOGRAPH ICAL SKETCH .................. .............................................................. 149















LIST OF TABLES


Table page

1-1 Catalytic constants and subcellular locations of active a-CA isozymes ............... 19

2-1 Data set and model statistics for wild type HCA II from pH 5.1 to 10.0..............44

2-2 Data set and model statistics for H64A/N62H and H64A/N67H HCA II at pH
6 .0 a n d 7 .8 ........................................................................................................... . 4 5

2-3 pH-Independent rate constants for proton transfer and pKa for proton donor and
acceptors in w ild type and m utant H CA II.......................................... ................ 46

3-1 Data set and final model statistics for N62L HCA II and N67L HCA II..............65

3-2 Data set and final model statistics for Y7F HCA II. ...........................................66

3-3 Maximal values of rate constants for hydration of C02, proton transfer, and pKa
of the zinc-bound w ater .......................................... ......................... ................ 67

4-1 Scattering amplitudes and cross sections of atoms by X-rays and neutrons............90

4-2 Data collection and model refinement statistics.................................. ................ 91

4-3 Distances between solvent molecules and active site residues for hydrogenated
and perdeuterated H C A II ........................................ ........................ ................ 92

5-1 Data collection, refinement and final model statistics. ............... ....... ............107

6-1 Maximal values of kcat/Km for the hydration of CO2 and kB for the proton
transfer dependent release of H2180 from isoforms of carbonic anhydrase .........124















LIST OF FIGURES


Figure page


1-1 Ribbon diagram of CA from three classes .......................................... ................ 20

1-2 Multiple sequence alignment of fourteen HCA domains....................................21

1-3 C on serv ation of H C A s. ............................................................................................22

1-4 A ctive sites of H C A I, II, and III......................................................... ................ 23

1-5 Molecular and electrostatic surface potentials of HCAs................ ................24

1-6 Ribbon diagram of HCA II with secondary structure elements..............................25

1-7 A ctive site of w ild type H C A II .......................................................... ................ 26

1-8 H CA II in com plex w ith inhibitors ..................................................... ................ 27

2-1 Crystal structures of wild type HCA II active site .............................................47

2-2 Crystal structures of mutant HCA II active site ............................ ..................... 48

2-3 The pH profiles for kcat/KM catalyzed by wild type and mutant HCA II. ..............49

2-4 The pH profiles for RH20/[E] catalyzed by wild type and mutant HCA II.............50

3-1 Active site of wild type human carbonic anhydrase II........................................68

3-2 Active site of N62L and N67L at pH 8.2 and pH 6.0 ........................................69

3-3 A active site of Y 7F H CA II at various pH ...................................... ..................... 70

3-4 The pH profiles for kcat/KM catalyzed by wild type and mutant HCA II ..............71

3-5 The pH profiles for RH20/[E] catalyzed by wild type and mutant HCA II.............72

3-6 Activation of RH20/[E] catalyzed by wild type and mutant HCA II by the
addition of 4-m ethylim idazole ......................................................... 73

3-7 Active sites of wild type and Y7F HCA II.......................................... ................ 74









4-1 The LANSCE position-sensitive 3He-filled detector. ........................................93

4-2 Optical photograph of perdeuterated wild type HCA II......................................94

4-3 Backbone superposition of hydrogenated and perdeuterated wild type HCA II...... 95

4-4 Active site comparison of hydrogenated and perdeuterated wild type HCA II.......96

4-5 The pH profiles for kcat/KM catalyzed by hydrogenated and perdeuterated HCA
II. ............................................................................................................ ....... .. 9 7

4-6 The pH profiles for RH2o/[E] catalyzed by hydrogenated and perdeuterated HCA
II. ............................................................................................................ ....... .. 9 8

5-1 Classical carbonic anhydrase inhibitors bound to HCA II............................... 108

5-2 Chemical structure of novel inhibitors ............ .. ......................................... 109

5-3 Active site structures of HCA II in complex with two novel inhibitors ..............110

5-4 HCA II active site with inhibitors superimposed .................................................111

5-5 Interactions of BB3 and TDM with HCA II ..................................... ................ 112

5-6 HCA II with BB3 and TDM superimposed ...................................... ................ 113

6-1 Sequence alignment of AaCA1 with HCA I and II..................... ...................125

6-2 The 12% Coomassie stained polyacrylamide gel of AaCA1 expression and
p u rifi c atio n ............................................................................................................. 12 6

6-3 The pH profile of rate constants for catalysis by AaCAl .................................127

6-4 M olecular m odel of A aC A ................................................................................... 128

6-5 Active sites and surface charge distribution of AaCA1, HCA I, and HCA II .......129















ABBREVIATIONS

A Angstrom

A. aegypti Aedes aegypti

CA carbonic anhydrase

CAPS N-cyclohexyl-3-aminopropanesulfonic acid

CHES N-cyclohexyl-2-aminoethanesulfonic acid

CO2 carbon dioxide

cm centimeter

E. coli Eschericia cob

g gravitational force

gm gram

H proton/ hydrogen ion

HCA human carbonic anhydrase

HCI hydrochloric acid

HCO3- bicarbonate ion

HEPES N-(2-hydroxyethyl)-piperazine-N'-2-ethanesulfonic acid

IPTG isopropyl-P-D-thiogalactopyranoside

kcat turnover number

kcat/KM specifity constant

kD kilo Daltons

Ki inhibition constant









KM Michaelis-Menten constant

kV kilovolt

LB luria broth

M molar

MES 2-(4-morpholino)-ethane sulfonic acid

MOPS 3-(N-morpholino)-propanesulfonic acid

[Ig microgram

pl microliter

IM micromolar

mA milliampere

mg milligram

ml milliliter

mm millimeter

mM millimolar

nm nanometer

nM nanomolar

pAMBS para-aminomethylbenzenesulfonamide

pH negative log of the hydrogen ion concentration

rmsd root mean square deviation

TAPS N-Tris(hydroxymethyl)methyl-3-aminopropanesulfonic

acid

Tris tris(hydroxymethyl)aminomethane

Zn zinc









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 INSIGHTS INTO THE ACTIVE SITE OF ALPHA-CARBONIC
ANHYDRASES

By

Suzanne Zoe Fisher

December 2006

Chair: Robert McKenna
Major Department: Medical Sciences--Biochemistry and Molecular Biology

Carbonic anhydrases (CA) are ubiquitously expressed metalloenzymes that are

found in all organisms, ranging from bacteria to humans. Human CA II (HCA II) is the

most well-studied and utilizes a zinc-hydroxide mechanism to catalyze the reversible

hydration of carbon dioxide to produce bicarbonate and a proton. Catalysis involves an

intramolecular proton transfer event that delivers an excess proton from the zinc-bound

water to an internal proton acceptor, His64. His64 then shuttles this proton to the bulk

solvent, thus regenerating the active site for the next round of catalysis.

An extensive analysis of the structural and kinetic stability of wild type and several

mutants of HCA II was conducted over a broad pH range. The results show that the

enzyme, and the water network in the active site, is extremely stable. It is also the first

observation of sulfate ion binding in the active site of wild type HCA II. Attempts to

disrupt not only the proton shuttle His64, but other residues involved in stabilizing the

water network were also successful as reflected in changes of the measured proton

transfer rates. Overall, the results give insights into the structural requirements for

efficient proton transfer as catalyzed by CA. To directly observe the active site waters









and protonation state of His64, perdeuterated wild type HCA II was produced,

crystallized and the X-ray structure determined. This work lays the foundation for future

proposed neutron diffraction experiments.

Classical, clinically used CA inhibitors (CAI) are not very water-soluble and this

feature has implications for bioavailability of these drugs. The X-ray structures of two

novel, water-soluble CAIs bound to HCA II were determined. They reveal that

incorporation of spacer groups and fluorines can change the binding modes of CAIs. This

work has implications for the clinical use and bioavailability of systemically applied

CAIs and for targeting different isozymes of HCA.

A CA from mosquito larvae (AaCA1) was also expressed, purified, and structurally

and kinetically characterized. AaCA1 is a high activity CA that shows inhibition with all

the classical sulfonamide-based CAIs. This enzyme represents an interesting new drug

target for the control of mosquito populations and further understanding of CA function

in other organisms.














CHAPTER 1
INTRODUCTION

The Discovery of Carbonic Anhydrase

Prior to the discovery of carbonic anhydrase (CA) there were two main theories as

to the physiological mechanism of carbon dioxide transport in blood. The first

(bicarbonate theory) stated that carbon dioxide was transported as bicarbonate and upon

reaching the lung, was converted by blood proteins to carbon dioxide and expelled. The

second (direct combination theory) declared that carbon dioxide is carried directly by

blood proteins, primarily hemoglobin, and can reversibly dissociate from them, upon

reaching the lung (Forster, 2000). From 1917 to 1921 the bicarbonate theory gained a lot

of support through the efforts of many physiologists. However, early experiments by

Thiel in 1913 studied the uncatalyzed rate of carbon dioxide hydration and found it slow

at near 0.02 s-1. This low rate was obviously insufficient as red blood cells (RBCs) only

have about 1 second in the lung to exchange carbon dioxide. In 1926 Henriques

calculated the hydration rate under physiological conditions using velocity constants

obtained by others and concluded that there must be a catalyst in blood that speeds up the

hydration reaction. He also predicted that there must be another mechanism that does not

purely rely on the bicarbonate theory (Henriques, 1928). Later experiments involved

determination of the rate of carbon dioxide production from hemoglobin solutions. These

showed that there was a dramatic increase in the rate compared to just buffer alone. These

observations gave support to the direct combination theory, which led to the suggestion

that carbamate binds to hemoglobin giving rise to a hypothetical complex called









carbhemoglobin (Henriques, 1928). As it later turned out, the hemoglobin purification

technique was not perfect and actually contained a contaminant, carbonic anhydrase, that

was mediating the observed catalysis. It was not until 1932 that Meldrum and Roughton

isolated a non-hemoglobin protein from ox blood and showed that it was catalyzing the

carbon dioxide hydration reaction. They named this protein carbonic anhydrase

(Meldrum and Roughton, 1932). Then in 1933 the same authors published a paper that

described the details of CA preparation and its catalytic properties. Their rather crude

initial experiments tested an impressive array of CA properties including temperature and

pH stability. Their results indicated that pure preparations of CA that were heated to 65

C for 30 minutes still retained 40% activity and the protein was stable and active from

pH 4.0 to 12.0. In the same paper they also showed that CA was potently inhibited by

cyanide and azide (Meldrum and Roughton, 1933). Since the early discoveries and

observations in the 1930's the CA field has grown enormously and a lot is known about

the various isoforms, physiological functions, catalytic activity, and crystal structures.

A Fine Example of Convergent Evolution

From humans to plants, the ubiquitous zinc metalloenzyme CA catalyzes the

reversible hydration of carbon dioxide (C02) to form bicarbonate (HCO3-) and a proton

(H ). There are five evolutionary distinct classes of CAs: ca, P3, y, 6, and P. The a-class

was discovered first and is found primarily in mammals but has also been identified in

such diverse organisms as the mosquito and plant green algae, Chlamydomonas

reinhardtii. The P-class is found in plants but there are also examples found in

Escherichia coli (E. coli) and Synechococcus (Hewett-Emmett and Tashian, 1996). The

y-class is found mainly in archaebacteria and was initially discovered in the archeon









1 'ithmiiiii, L ilia thermophila (Alber and Ferry, 1996). Arabidopsis thaliana and

Synechococcus seem to be the exceptions by having sequences that have similarity to all

three classes (Hewett-Emmett and Tashian, 1996). The recently discovered 6- and g-

classes have not been extensively studied and are found in diatoms and cyanobacteria,

respectively (Tripp et al., 2001; So et al., 2004). Most living organisms have genes that

encode for CAs, except for Mycoplasma genitalium that appears to lack any CA-

encoding gene (Fraser et al., 1995).

There are numerous crystal structures of ca-, P3-, and y-class CAs (Figure 1-1) that

are available in the Protein Data Bank (www.pdb.org). Currently, there are well over 200

CA structures; most of these are ca-class CAs and are, more specifically, represented by 5

human CA (HCA) isozyme structures, HCA I, II, as well as extracellular domains of

HCA IV, XII, and XIV (Liljas et al., 1972; Kannan et al., 1975; Stams et al., 1996;

Whittington et al., 2001; Whittington et al., 2004). Crystal structures of the a-class

(Figure 1-1(a)) include various mutants of different isoforms as well as complexes of the

protein with inhibitors and/or activators. The first CA crystal structure of human isozyme

II was determined in 1972 (Liljas et al., 1972). More recently, the first P3- and y-CA

crystal structures have been reported (Mitsuhashi et al., 2000; Kisker et al., 1996).

Visual inspection of the crystal structures of the three classes (Uc-, P-, and y-classes)

reveals a dramatic picture of the variation in topology of these zinc-containing enzymes

(Figure 1-1 (a)-(c)) and it is no surprise that the different class members are found in

phylogenetically diverse organisms (Strop et al., 2001). The a-CAs act mainly as

monomers except in the case of HCA XII that was shown to be a dimer (Whittington et

al., 2001). The structures of 3- and y-class CA reveal that they oligomerize to form









pseudo-dimers and trimers, respectively (Mitsuhashi et al., 2000; Iverson et al., 2000).

Despite the obvious overall structural differences between these classes, closer inspection

of the active sites reveal a remarkably similar architecture of the catalytic zinc center.

The carbonic anhydrases are an excellent example of convergent evolution where

distinctly varied life forms have found structurally alternative ways to construct an

enzyme that has the same catalytic mechanism (Figure 1-1; Lindskog, 1997).

The many physiological functions of CA have been most extensively characterized

in plants and mammals. In mammals CA is involved in many physiological processes

such as acid/base homeostasis, renal acidification, bone resorption, cellular respiration,

gluconeogenesis, formation of gastric acid, cerebrospinal fluid and aqueous humour

production, tumor metastasis, and the interconversion of C02/HC03- in red blood cells

(RBCs) during respiration (Henry, 1996; Breton, 2001; Chegwidden and Carter, 2000).

Due to all these functions that are vital to all life processes CAs are found in

virtually all tissue types. In plants CA is found in both the cytosol and the chloroplast

where it is primarily involved with providing carbon for the fixation of inorganic carbon

into sugar. Carbonic anhydrase provides either HCO3- or CO2 as a source of inorganic

carbon for either PEP carboxylase or RuBisCO (ribulose-1,5-bisphosphate

carboxylase/oxygenase), respectively (Burnell, 2000). The functions of CA in bacteria

are less well understood except for two well-documented cases. The first is the role of

CA in E. coli where, as a part of the cyn operon, it prevents depletion of CO2 during

cyanate breakdown (Guilloton et al., 1992). The second is in cyanobacteria where CA is

part of the carboxysome shell. Carboxysomes are polyhedral microcompartents that

consist of several shell proteins that "package" RuBisCO for carbon fixation. RuBisCO is









a notoriously slow enzyme and it is thought that CA, present in the carboxysomal shell,

can convert HCO3- into CO2 and deliver this substrate directly to RuBisCO in high local

concentrations (So et al., 2004).

Human Carbonic Anhydrases (HCAs)

In the human ca-class there are fourteen identified expressed CA isoforms (HCA I -

XIV) and there are examples of cytosolic (HCA I, II, III, and VII),

transmembrane/membrane anchored (HCA IV, IX, XII and XIV), secretary (HCA VI and

XI) and mitohondrial isoforms (HCA VA and VB) (Chegwidden and Carter, 2000; Duda

and McKenna, 2004). The percentage sequence identity varies from 26 to 61% among all

fourteen isoforms and they differ in catalytic efficiency and subcellular location (Table 1-

1, Figure 1-2). Table 1-1 summarizes some of the properties for representative active a-

CAs. Some of the isozymes, such as HCA II and IV, are high activity forms that work

near the diffusion limit. Most of the ones listed in Table 1-1 fall in the medium activity

level range while HCA III represents the isozyme with the lowest catalytic rate constants.

Figure 1-2 is a multiple sequence alignment of the CA domains of all fourteen

HCAs, and the conserved and similar residues are marked as described in the figure

legend. Figure 1-3 indicates the location, on the backbone of a CA domain, of all the

conserved residues found in all fourteen HCAs. It is striking that these residues are not

located directly in the active site but seem to cluster around it. This distribution implies

that the variation seen in the active sites is necessary for the wide range of catalytic rates

for different isozymes (Table 1-1). The conserved residues overall appear to have a role

in maintaining the distinctive CA fold. Very few of these residues are found on the









surface and the resulting surface heterogeneity lends support for the different cellular

locations and functions of the various HCAs (Table 1-1; Figure 1-3)

HCA I is found predominantly in RBCs while HCA II, also abundant in RBCs, is

also found in cells of all tissue types (Tashian, 1992). HCA I and HCA II's presence in

the RBCs are very important for converting HCO3- into CO2 during respiration. Figure 1-

4 shows the active sites of cytoplasmic HCA I, HCA II, and HCA III. The active sites of

these three isozymes are highly conserved and the differences indicated in Figure 1-4

could account for the observed differences in their respective activities (Table 1-1). HCA

III is a major part of the soluble protein in adipose and muscle tissue, but its function in

these tissues remains elusive. Despite is abundance, recent studies with mouse knock-out

models of HCA III showed no phenotype (Sly and Hu, 1995; Kim et al., 2004).

HCA IV is the only glycosylphosphatidylinositol(GPI)-linked isoform and, in

contrast to CA IV in other species, not glycosylated. The CA domain of HCA IV is

located on the extracellular face of cells in many tissue types that include kidney, lung,

and the eye (Sly and Hu, 1995; Chegwidden and Carter, 2000).

Of the mitochondrial CAs there are two HCA V variants, CA VA and CA VB.

HCA VA is expressed only in the liver while HCA VB expression is everywhere except

the liver. HCA V is mainly involved with providing HCO3- to metabolic enzymes in the

gluconeogenesis and ureagenesis pathways (Dodgson, 1991).

HCA VI is heavily glycosylated and the only secreted isoform in humans. It is

found predominantly in saliva where it is involved with pH control of the mouth

(Murakami and Sly, 1987). HCA VII is another cytoplasmic, soluble isoform that is

expressed mainly in the brain, salivary gland, and lung. It is the most highly conserved









(compared to the consensus CA domain sequence) of all the active CA isoforms, but its

function remains unknown (Chegwidden and Carter, 2000).

HCA VIII, X, and XI have no catalytic activity but contain a CA domain. These

inactive isoforms are subsequently known as the CA-Related Proteins, or CA-RPs

(Khalifah, 1971; Jewell et al., 1991; Kato, 1990; Skaggs et al., 1993). The lack of

catalytic activity in the HCA-RPs is due to deleterious mutations of the zinc ligand

histidines (Figure 1-2). CA-RP VIII is highly conserved across species and shares about

98% sequence identity between humans and mice (Bergenhem et al., 1998). HCA-RP is

widely expressed throughout the brain and its expression is developmentally controlled

(Taniuchi et al., 2002).

HCA IX was first identified in HeLa cells and is a glycosylated extracellular

enzyme with a membrane-spanning region (Pastorekova et al., 1992). It is normally

expressed only in the gastrointestinal epithelial lining but is usually absent in normal

tissues. HCA IX displays constitutive expression in tumors such as clear cell renal

carcinoma and has potential as a biomarker for certain tumors (Murakami et al., 1999;

Ortova Gut, 2002). HCA XII is another transmembrane glycoprotein with its CA domain

located on the extracellular side of cells. It is expressed mainly in colon, kidney, and

prostrate, but it was discovered due to overexpression of its mRNA in renal and lung

cancer cells (Ttireci et al., 1998; Ulmasov et al., 2000). The crystal structure revealed that

this membrane protein exists as a dimer (Whittington et al., 2001).

HCA XIII is the most recent addition to the HCA family and is predicted to be a

cytosolic isoform, similar to HCA I and II. Not much is known about its function but it

has been identified in thymus, spleen, and colon (Lehtonen et al., 2004).









HCA XIV is the last of the human isoforms and shows 45% sequence identity to

HCA XII (Fujikawa-Adachi et al., 1999). It is also a single-membrane spanning

glycoprotein with an extracellular catalytic domain but, unlike HCA XII, the crystal

structure of murine CA XIV shows it to be monomeric (Whittington et al., 2004). In

humans, it is widely expressed but is found mainly in heart and kidney (Chegwidden and

Carter, 2000).

Figure 1-5 shows the molecular surface and electrostatic charge distributions of

several representative HCAs. HCA II and III are the most and least efficient of the

catalytically active cytosolic HCAs, respectively. HCA II appears to be more negatively

charged around the active site region compared to HCA III. HCA VI (Figure 1-5 (c)) is

the only secreted isoform and is more hydrophobic around the active site compared to the

other soluble forms shown in Figure 1-5 (a) and (b).

HCA-RP VIII is one of the acatalytic isoforms and its subcellular location is

unknown. Compared to the other isoforms shown in Figure 1-5, HCA-RP VIII has a

much higher overall negative charge distribution.

HCA IX is a membrane protein and the surface shown is for the catalytic CA

domain (Figure 1-5 (e)). It has an asymmetric distribution of negative charge and

hydrophobicity at the region adjacent to the active site. HCA XII is dimeric membrane

protein (Figure 1-5 (f)) with the two active sites located on the same side of the dimer.

These surface features of the HCAs reflect the variation in amino acid sequence among

them. As is shown in Figure 1-3, the most conservation of sequence is not in the active

site or the surface, but in the regions that form the scaffold of the metal binding center.









Structure of HCA II

The first crystal structure of HCA II was determined over thirty years ago by Liljas

and colleagues (Liljas et al., 1972). The overall fold of HCA II can be described as a

single-domain, mixed a/p, globular protein that is almost spherical with approximate

dimensions of 5 x 4 x 4 nm3 (Figure 1-6; Lindskog, 1997). These enzymes are overall

small and compact proteins with the possible exception of the N-terminal region (residues

1 to 24) that is more loosely connected to the rest of the molecule. N-terminal residues 1-

4 are usually disordered in the crystal structures. A cluster of aromatic residues at the N-

terminus of the enzyme consisting of Trp5, Tyr7, Trpl6, and Phe20 has been suggested

to assist in the anchoring of this region to the rest of the enzyme (Lindskog, 1997). It has

been shown that the removal of the N-terminal region does not result in a major loss of

protein stability or enzyme activity. Another cluster of aromatic residues located under

the P-sheet consists of Phe66, Phe70, Phe93, Phe95, Trp97, Phel76, Phel79, and Phe226

(Lindskog, 1997). The active site folds first and independently of the N-terminal region

(Aronsson et al., 1995). The central feature of the HCA II structure can be described as a

10-stranded (PA-PJ) twisted P-sheet, which is decorated on the surface by seven

a-helices (Figure 1-6; aA- aG). The strands of the P-sheet are mainly antiparallel, with

the exception of two pairs of parallel strands (Figure 1-6; OF and 3G, P1 and PJ). There is

a conserved loop region extending towards the active site that contains the proton

shuttling residue, His64. The active site consists of a conical cleft that is ~ 15 A deep

with the catalytic Zn2+ placed at the bottom of the cleft (Figure 1-6).

The Zn2+ is tetrahedrally coordinated by four direct ligands: the imidazole groups

of three conserved His residues (His94, His96, and His 119) and a H20/ OH- molecule









(Figure 1-7). The direct metal ligand histidines, in turn, are held in position by hydrogen

bonding interactions with other residues, the indirect or second shell metal ligands. The

side chains of Gln92 and Glul 17 interact with His94 and Hisi 19, respectively, while the

carbonyl oxygen of Asn244 coordinates His96. Thrl99 makes a hydrogen bond with the

solvent metal ligand and is optimally oriented for this by interacting with Glu106

(Christianson and Fierke, 1996).

Activity and Catalytic Mechanism of CAs

HCA II kinetics and catalytic mechanism have been studied extensively; however,

it is thought that all a-CAs perform the same general activity, which is known as the

zinc-hydroxide mechanism. The interconversion between CO2 and HCO3- is a two-step

reaction that can be described as a ping-pong mechanism (eq 1-1 and 1-2; Silverman and

Lindskog, 1988; Lindskog, 1997; Christianson and Fierke, 1996).

H20O
EZnOH- + CO2 EZn(OH-)CO2 EZnHCO3- EZnH20 + HCO3- (1-1)

EZnH20 HTEZnOH- + B EZnOH- + BH+ (1-2)

The first step (eq 1-1) in the hydration direction is the binding of CO2 in a

hydrophobic region adjacent to the zinc atom. Substrate binding is then followed by a

nucleophilic attack on the carbon by the zinc-bound hydroxide leading to the formation

of HCO3-. Water can freely diffuse into the active site and displace HCO3- leaving a water

molecule bound at the zinc atom. The second part of the reaction (eq 1-2) occurs at 106 s-1

and is the rate-limiting step of the overall reaction and involves the removal of an excess

proton from the zinc-bound water to regenerate the hydroxide needed for catalysis

(Khalifah, 1971; Steiner et al., 1975; Silverman and Lindskog, 1988). The transfer of a

proton out of the active site involves an intramolecular and intermolecular proton transfer









event. The intramolecular proton transport occurs between the Zn-bound solvent and the

side chain of His64 through an intervening chain of hydrogen-bonded water molecules,

Wl, W2, W3a, and W3b (Figure 1-6; Lindskog and Silverman, 2000; Fisher et al., 2005).

Mutation of His64 to an alanine results in a 10-50 fold reduction in the proton transfer

rate (Tu et al., 1989). From His64 the intermolecular transfer event delivers the proton to

bulk solvent/buffer. In eq 1-2, B signifies either an acceptor on the protein (His64) or an

exogenous acceptor that becomes protonated, BH+ (Silverman and Lindskog, 1988).

For CO2 hydration in HCA II, both the kcat and kcat/Km have pH profiles that appear

as simple titration curves with a pKa ~7 and with maximal activity at high pH (Silverman

and Lindskog, 1988). HCA II is the most efficient isozyme (kcat = 1.4 xl06 S-1 at 25 C) of

this class while HCA III is the slowest (kcat = 8 x 103 s 1 at 25 C) (Khalifah, 1971; Jewell

et al., 1991). HCA III has similar catalytic features to HCA II in that catalysis also occurs

in the same two, separate steps (eq 1-1 and 1-2), but the enzyme is resistant to the classic

HCA II inhibitors, the sulfonamides (e.g., acetazolamide and HCA III yield a Ki of 40

tM vs. 0.06 tM for HCA II; LoGrasso et al., 1991). In fact, there seems to be a

consistent, inverse relationship between the turnover number (kcat) and level of inhibition

by sulfonamide of CAs across species (Tufts et al., 2003).

Substrate Binding in the CA Active Site

The precise location of CO2 binding in the CA active site remains elusive. The

binding site has been narrowed down to a hydrophobic region behind the Zn, but the

exact interactions that mediate substrate binding are still unknown. The binding of CO2 is

weak and the binding of HCO3- is even weaker with approximate Kd of 100 and 500 mM,

respectively (Krebs et al., 1993; Lindskog and Silverman, 2000). It is thought that CO2

binding in the hydrophobic pocket displaces the deep water that is normally hydrogen









bonded to the amide group of Thrl199 (Hikansson et al., 1992). CO2 binding causes it to

be polarized by the interaction with the backbone amide of Thrl99 in addition to the

electrostatic effects exerted by the zinc. This polarization causes the CO2 to become

susceptible to nucleophilic attack by the Zn-bound solvent, as described above.

The hydrophobic substrate binding pocket in HCA II is defined by four residues:

Vall21, Val143, Leu198, and Trp209 (Lipscomb, 1990; Merz, 1991). The crystal

structure of a HCA II mutant (Thr200--His) in complex with HCO3- (PDB accession

code: IBIC, Xue et al., 1993). The structure shows one of the HCO3- oxygen atoms

acting as a metal ligand, replacing the Zn-bound solvent while the hydrogen is involved

in an H-bond with the side chain of Thrl99. The second oxygen atom acts as a fifth

ligand to the Zn2+ and the third seems to be in an H-bond with the backbone amide of

Hisl99 (Earnhardt and Silverman, 1998). The structure of this mutant enzyme:product

complex implicates Thrl99 as a very important residue for orienting water molecules in

the active site as well as the direct interaction with HCO3-. A similar complex with the

native enzyme has not been obtained, as the native form probably does not exhibit the

same extent of binding to HCO3-.

The use of competitive inhibitors to elucidate the substrate-binding site has yielded

interesting but conflicting results. Imidazole is a competitive inhibitor of HCA I with a Ki

of 20 mM and phenol is a competitive inhibitor of HCA II with a Ki of 10 mM (Khalifah,

1971; Tibell et al., 1985). Crystal structures of HCA I in complex with imidazole and

HCA II in complex with phenol show different binding modes of these inhibitors in the

active site, making it hard to elucidate the possible binding mode of CO2 (Kannan et al.,

1977; Nair et al., 1994; Earnhardt and Silverman, 1998).









Measuring CA Activity

One of the techniques used to measure CA activity is stopped-flow

spectrophotometry under steady state conditions using pH-indicator pairs. This technique

is used to determine kcat and KM by measuring the initial rates of CO2 hydration. The

turn-over number kcat reflects the part of catalysis that involves rate-limiting proton

transfer, while kcat / KM reflects the steps involved in CO2 / HCO3- interconversion

(Khalifah, 1971; Steiner etal., 1975).

Another technique used to measure CA activity is 18O-exchange and this is done

at chemical equilibrium. This method is based on the exchange of 180 between 12C and

13C-containing species of CO2 and water that occurs because of the hydration-

dehydration reaction of CA. Two rates can be determined by this method, Ri and RH20o.

The first rate, Ri, is a measure of the exchange between CO2 and HCO3- at chemical

equilibrium. The second rate determined by these methods, RH20, indicates the rate of

release from the enzyme of water carrying substrate oxygen. The rate of water release

from the active site depends on the rate of proton transfer, thus RH2o is used to measure

proton transfer activity (Silverman et al., 1979; Silverman 1982).

Analysis of the catalyzed reaction by CA at steady state and chemical equilibrium

has led to a model of its mechanism of action that implicates two ionizing groups in the

active site that have pKa values near 7 (Steiner et al., 1975; Silverman and Lindskog,

1988; Lindskog 1997). One of these groups corresponds to the zinc-bound water, which

ionizes to a hydroxyl ion, and is responsible for the interconversion of CO2 / HCO3-,

while the other group, His64, is involved in proton transfer (Tu and Silverman, 1989).









HCA and Human Disease

Due to their wide-spread distribution and various physiological functions, the

HCAs support many systemic and cellular HCO3- / CO2 transport processes as well as

biosynthetic pathways. There are not many examples of CA-associated diseases and this

reflects its crucial role in so many fundamental life processes. In general, HCA

deficiencies are rare and a possible reason is that HCAs are ubiquitously expressed and,

in some cases, a different HCA can possibly compensate for the loss of a particular

isoform (Sly and Hu, 1995).

A number of genetic variants of HCA I have been described and they generally

differ only in one amino acid due to nucleotide substitutions. HCA I variants are less heat

stable compared to the wild type version but the catalytic differences are very small or

even insignificant. Individuals that either have a variant or lack HCA I completely, show

no phenotype (Osborne and Tashian, 1974; Venta, 2000). Despite there being a lot more

HCA I than HCA II in RBCs, maybe a loss of HCA I shows no phenotype because HCA

II, a very active isoform compared to HCA I (Table 1-1), can functionally compensate for

this loss.

HCA II is the most well-studied isoform of all the HCAs and part of the reason is

that there are some severe diseases associated with mutant versions of this isoform. One

of these is an inherited diseases called HCA II deficiency syndrome. It manifests as a lack

of erythrocyte HCA II and is associated with osteopetrosis, renal tubular acidosis and, in

extreme cases, cerebral calcification that leads to mental retardation (Sly and Hu, 1995).

A number of mutations that include nonsense, frameshift, and splicing mutations can lead

to a lack of HCA II. Most of these changes cause HCA II deficiency syndrome, but a few

other variants have been identified that do not lead to a change in activity or amount of









enzyme (Venta, 2000). The first mutation identified that is associated with the disease

results in Hisl07--Tyr substitution and leads to an unstable enzyme that shows a three-

fold lower catalytic rate (Venta et al., 1991; Tu et al., 1993). HCA II deficiency

syndrome is quite rare and is mostly found in homozygous individuals from families

where some level of inbreeding has occurred (Sly and Hu, 1995).

No variants or deficiencies have been described for other HCAs. There is a

polymorphism in HCA III where Ile3 1- Val and variation in activity is predicted based

on the location of this substitution. Genetic knock-out studies in mice of murine CA III

did not show any phenotype under all the standard muscle-stress tests and longevity of

these animals was also not affected (Kim et al., 2004).

Recent studies with CA IV and CA IX single and double knock-out mice

implicated these isoforms in buffering and pH regulation of the extracellular space in the

hippocampus. The CA IV knock-out mice offspring were produced in lower than

expected numbers and females seemed to die more frequently than males during gestation

and immediately after birth. CA IX knock-out mice had normal fertility and viability. The

double knock-out mice were smaller than wild type and most of the females died before

ten months of age. Electrophysiological measurements on brain slices from these animals

showed that either one of these membrane-bound CAs can buffer the hippocampus after

synaptic firing. A loss of both in the double knock-out mice completely ablated this

buffering effect (Shah et al. 2005).

CA IX is a highly active isoform with an extracellular CA domain that has

functionally been implicated with acid/base balance and intercellular communication.

Aberrant expression of CA IX is associated with various tumors and has become of









significant clinical interest. In 2002 Ortova Gut et al. constructed a CA IX knock-out

mouse and investigated the effects on gastrointestinal epithelia cells. Although these mice

had normal stomach pH and acid secretion, they developed gastric hyperplasias and

several cysts. These studies highlight the important role of CA IX in cell proliferation and

differentiation (Ortova Gut et al., 2002).

CA Inhibitors

Not long after the discovery of CA by Meldrum and Roughton (Meldrum and

Roughton, 1932) these authors also investigated inhibition of CA by small molecule

inhibitors such as azide and cyanate (Figure 1-8 (a) and (b); Meldrum and Roughton,

1933). In the 1940's, Mann and Keilin found that sulfonamide-based compounds are

specific and strong inhibitors of CA (Mann and Keilin, 1940). Figure 1-8 shows several

crystal structures of HCA II with small molecule inhibitors as well as the clinically used

sulfonamide-based drugs. Since the early findings, many other strong and selective

inhibitors have been investigated and these are the aromatic and heterocyclic

sulfonamides of the R-S02NH2 or R-S02NH(OH) form (Maren, 1967; Maren 1974). All

the sulfonamide-based inhibitors interact with CA by the same mechanism: they bind to

the metal ion and interfere with the ZnOH- coordination by either displacing or replacing

the hydroxide, thus disrupting the interconversion of CO2 and HCO3- (Figure 1-8; Bertini

and Luchinat, 1983).

Several crystal structures of complexes of various sulfonamide-based inhibitors

with CA show similar interactions: the -NH group of the sulfonamide moiety binds

directly to the metal and simultaneously donates a hydrogen bond to hydroxyl of Thrl99.

An oxygen of the sulfonamide also interacts with the amide backbone of Thrl99 and thus

displaces the deep water (Figure 1-8; Lindskog, 1997). The key group in determining this









displacement/replacement is the hydroxyl of Thr199 and this residue is sometimes

referred to as the "gate keeper".

Examples of clinically important drugs include acetazolamide (Diamox) and

brinzolamide (Azopt) that have applications in the treatment of congestive heart failure,

altitude sickness and epilepsy (Figure 1-8 (c) and (d); Mansoor et al., 2000). The most

common sulfonamide inhibitor in clinical use is acetazolamide, which is a strong

inhibitor of HCA II with a Ki value near 0.01 [LM (Maren and Conroy 1993). Analysis of

acetazolamide bound to HCA II revealed the binding interactions of this compound. The

thiadiazole ring is in van der Waals contact with Vall21, Leul98, and Thr200 and the

carbonyl oxygen of the amido group shares a hydrogen bond interaction with the side

chain amide of Gln92. The methyl group was shown to interact with the side chain of Phe

131 (Vidgren et al., 1990).

CA inhibitors are commonly prescribed to treat a major symptom of glaucoma, i.e.

increased intraocular pressure. The inhibition of CA in the eye by topical application of

the drug suppresses the secretion of Na HCO3-, and subsequently production of aqueous

humor thus lowering intraocular pressure. (Maren, 1987).

In the following chapters the detailed active site structures of wild type and mutant

CAs, and how it relates to proton transfer processes, will be presented. Chapter 2 will

deal with pH stability of wild type and site-specific mutants as well as the position of a

proton shuttling residue in the active site. These structural features will be correlated with

kinetic measurements. In Chapter 3 a different approach will be discussed where, instead

of moving the proton shuttle to different positions, mutations were made in the active site

of HCA II in order to disrupt solvent networks that mediate proton transfer. That data will









also be correlated with kinetic measurements and implications for proton transfer will be

discussed. Chapter 4 will be a discussion of initial experiments performed for the

eventual determination of a neutron diffraction structure of perdeuterated wild type HCA

II. A detailed structural comparison of overall and active site features between

hydrogenated and perdeuterated HCA II will be presented. This work shows proof-of-

principle for using HCA II crystals for neutron diffraction experiments. In Chapter 5, the

unique binding modes of two novel CA inhibitors, as revealed by X-ray crystallography,

will be discussed. These are new compounds and these inhibitors target different

residues compared to other canonical CA inhibitors. Chapter 6 will deal with the

characterization of a CA from the mosquito, Aedes aegypti. This work includes kinetic

characterization, inhibition studies, and a homology model of the enzyme. Work

presented in Chapters 5 and 6 have implications for the search and design of novel CA

inhibitors with possible applications for controlling mosquito populations and treating

certain cancers that have associated CA overexpression. Finally, Chapter 7 will contain a

summary and concluding remarks on the work presented elsewhere in this thesis, as well

as possible future directions for the topics discussed.













Table 1-1. Catalytic constants and subcellular locations of active a-CA isozymes.*
Isoform kcat (s') kcat/KM (M's') Subcellular Activity Level
Location
HCA I 2.0 x 105 5.0 x 107 Cytoplasm Medium
HCA II 1.4 x06 1.5 x 108 Cytoplasm High
HCA III 1.0 x 104 3.0 x 105 Cytoplasm Low
HCA IV 1.1 x 106 5.0 x 10 Membrane-bound High
Murine CA V 3.0 x 105 3.0 x 107 Mitochondrial Medium
Rat CA VI 7.0 x 104 1.6 x 107 Secreted Medium
Murine CA VII 9.4 x 105 7.6 x 107 Cytoplasm High
HCA IX 3.8 x 105 5.5 x 107 Transmembrane Medium-high
HCA XII 4.0 x 10' 7.4 x 107 Transmembrane Medium-high
Adapted from Chegwidden and Carter (2000). HCA XIII and XIV are not included as definitive rate constants have not been
determined at the time of this writing.














































Figure 1-1. Ribbon diagram of CA from three classes. (a) a-class CA; (b) P-class CA; (c)
y-class CA. Panels on the right of (b) and (c) represent the biological assembly
as a dimer and trimer, respectively. Coloring is from blue for the N-terminus
to red for the C-terminus, gray spheres are Zn 2+. Figures were generated and
rendered with Bobscript and Raster3D (Esnouf 1997; Merritt and Bacon,
1997). PDB accession codes for (a), (b), and (c) are IMOO, 1DDZ and
1QRG, respectively (Duda et al., 2003; Mitsuhashi et al., 2000; Iverson et al.,
2000).













1 -0 20 30 40 50 40 70 s0

CAI ASDiGYMB C .------EOQWSKLY?.AN -----GKQ5 Q7D-KT SET&-DTSKPISS 75---YN1?A-".KE I aGH5HFECQ-R5CnKGPFS
"CA&I -5Ma.GYGKHG ------ PEHIMIDE0PIAK-----.-GRQSP"7DDTIHTAiKYD.PS LSVS---YDQA$SLRIILNNG iNVEFDDSQBK,'.LKC-GPLD
HCAII2 -AKElGYASHNG ------ PEWH*ELYPIAK ---- -GOSP ELHTKD- I.DP$SIQPNSVS---YDPGSAXTILHDGKTCRWvFDDTFDRSMLRGCPLS
HCAIV AESHWYEVOAE----- SSNYPCLVP"r6 %GNCCKDROSP:NITTK .VDiKLGRFFFSG--YtKKQTWTO' 0NGHS'.MlLLE--- -KAS-SGGGLP
MCAV ---------------------C---- T---GRQS?:NQWKD3;YDPQ2APLRV-S- -YDAASCaRY I.GYFrQVEFDD5CEDSG:GSGPLG
HCAVI HGVrETYSEC .D .-----.EAfPLWEYKCG -----RQS ZDLQMI*EY; P SI.RAINLTGY-GLWHGEFPV-.I.% GHTQ IS' S TMS TSD-rQYL.
HCAVII SHHG3YGQDlG 3----- PSWHKLYP:AQ----- GDRQOS?:PIISSQAVYSPSLQPLELS-- -YEACMSLSITHNGHS'.'VQVDFDSDRVVTIGSPLE
HeAII I EGV6 rYEEG ---------- VG'lVFPDACN -----GEYOS PL;NSEARYDPS LDVRLSPN-YWCROCEVTMBHTI0VILKS---KSVLSS3PL?
HCA;X .NaRDEGDDQSH: YGGDPWBRVSACA-----GR3QS W'IRQI.AAFCPARLELLGFQLPa.- ELRB2JGESVQLTL PG---L- :M61PGR
HCAX JEG-nRAYKEVVQGS/WV.SE-GLaSA.iNL-CSVGKROSWIETSHNIFDPF-.TPUN-T-GG^KSGTiliyOTGEWSLRLDK-tHL.WiSGS T
HCAXI -EDSYKE.LOGNFVYGPPF-LVAAWSSL-CAVK-ROS-ELDRVLKYDPVFL S-T-GELTLTNGHSLNLPS---VHN GP
HCAXI -ASKi-.YrGSD.-----ENS3SKKYSCG----SGLL.QS?-DL.SD-LQYDASITPL.EFQG .-SA.KQFLLT IGaSPELa.SLS-----...D GLQ
KCAXI I SRi. GYREPWnCG------PH WpEMa PAD -----GDQQS?.E:KTKE'YDSS2LSIK---YEPS$5(.ISNMSGKS.YDFDDTENKSV.RGGPLT
HCAXIV GGQO'7YEGPHG .---- ODHWPASYPECG ----- IAOSPIDIQTDSVITFDPDLPALOPHGYD3PGTEPLDLHNGHTVQLSLPS----- TLYLGLP
***4* 4 4 #
90 100 110 1 130 140 150 160 170

HCAI D--SYRLF FHIFR ---STENHSSEHTVD>OWYSMAELVH,?Sl-AKYSSLAEAAS KAGLIV:SV --EAN PKLrv--DALQ: A KGI
HCATI G---TYR(.IQFHFHG--SLDC-6SEH4TVKKKYAAELHLVAIN--TKYGDFGWAVOQPDGLAVEGIFLKN G-SA- KW--DVLDSIE-KGKSAD
CAiIII G--PYRLRQF' E..G--55DDHGSEHYID-V!KYAAElHLVHh.i--PKY.NTTGEAlKQ PGIA :FL1G6-RSKGEFaQLL-D-OALBK 5KGK2A
HCAIV A P-- YCKOQL LMS DLPYKGSEH SLDGEJIFAMEZI VHEKEKGTS VMEALDPEDE IAIAFL1-TQVNEGFOPLV- EALS PRPEMS TT
HCAV N--HYRLKQFHFHWG--ATDEWGSEHAVDGHTYPAELHlV',)ITS-TKYENTKKASVE[SGLAVI6FLK- -AHHQAL KLV--DVLPEVRHKDWTA
-aVI A------KQ.FWGASSEiSGSEaTCMRY*a-E:fFrMiVT--SKiVNStE ftEEGLAVi.ArALVK''SDYBYsEIS--HirDiRY -QsI

HCAVIII QGHEFELYEV'FH-WG--REPNORGSEHIVNFKAFPMELH Il'AS T LES SIDEAVGKPHG IAI:ALFVOIG --KEHVGLaVT- EILQDIOQYKGKS KT
HCAIX E--Y-AI.SQLMr-G--AAGRPGSEHrVGOHRFPAE: H S --ITAFARVDEA GR? MCAX Y--SHRLEEIRI.FG--SEDSQGSEB^LNGQASGEQLCHKy-ELY VTEAAK5?!I6LV G5F;` SO-SSNPiAFGIU)RMLC-TI7TB DYL
HCAXI Y--SHRLSELRILFG --ARDGAGSEHO NHE.FSAEVQLIHraQ-ELYGNLSAASRGPSiGLAILSLFVlNAG -SSSNPFLSRLLRT ITRrISYKDAYF
HCAXII S--RYSATOLMH!'rGN-PNDPMGSEHrTVSGHFAAESHEIVHYS-DLYPDASTASNKSEGLA"'JVLIEM3--$SFPSYDKIFS--HlQHVMYK5EAF
HCAXIII ---SYRLROL FWG-SADDaGSEHIVrDGiHSYaAELHFir/sID-EKYPSHnIEM:0.IFZA-'M-IVQIK-EPNSaLVKIT--TI.DSIKEKSKQTR
aCAXIV R--KyvAAQL."_.'.Q-KGSSSES .QZNSEAT AEJ.H VHyDS-DSYDSLSEAAERPOSI.AVI.GIL -E TiKAY ILS--HL E'5t KDQIS
# 99 # #* ""o* S *# ##, #### ### 5 4 4
IS0 193 200 210 220 230 240 250 260

HCAI FrNrDPSTLLPS--- SLD tYPGSLTapp*,YESVTWICKES$SVSSEQLAOMFSLLSNSVEGDIAW ----- .HHNRPTOPLKGRTVBASF-
HCAM I FTNFD itGLT.E---SLDYWiYP6rSLITT PP2LECVnlVLEEPSrSSEQVLt LNFrGEGSE?EEL------f:DNGRPAQ PL QASFK
HCATII rEN.DPsC.FPA---CRDY'NTB:GEsrrPPCEC 19-^usEC V55Ea K!.25r.fASAE O-----LsMRnRQiKGWRV
HCAIV MAESSLLDLLP)EEKLRHYFRYLGSLT"TPTCICKWTVTiREIOQL QOILAFSOKLYYDKEQTISHK------ IVRPLOQLGCRTVIKS--
HCAV MGPFDPSCL.MPA---CRDYafTYPGSLTPPI.AE2TrIV-rQK. ESPSOLSMFRTLLFSGRGC-EEED .-----. NM YL fR PLLRCE RSSFR
"CAVI RGLD60EBELG- Da- LRY3SYLeGSlT.PCTESHW- aBDTD'-LSKTQVESK=NSL QmEQNn ..Q.-------D-yBIRRTOUiRWEA.NFM
iCiAVI; FSCniEKCLLBA---SR^YWGS1 PLr5P?!SESVISIVLREP;C;SEaiGKSLE;SEDDERtI"-----MV?!IFRPoQLKGRVV'KASFR
HCAVI:II IPCFnPNTLLPDPL-LRDYVYEGSLT PPCSEGVIILFRYPLTIISQLOIEEFRRL7THKOAELVESCDGI31LGSNRPOPLSCRVIRAAFQ
HCAIX PGLBD :SALLPS-D-F5RYT-QYE G$TSPGQG rI r'TSFN .LS aQLBLBDTLWG-PG6DSRLQ.------.NFRATOQUGRVIEASFP
HCAX LQGLI:EELYPE;---55.-ITYDG3S..;FPPCYE5TASWIIMiSKPVY;TRMQNHSLRLLSiOQPSQmFL5---. M5DNFRFVQPLNaiRC;RNI:N
HCAXI IQDLS2IlLFKES---F GSTYQGSLStPPC SErT'ILIDRA2IITS L SLRLLSONPPSatFQS-----.LSGNGRLOP!i.RAIARGNRDS
HCAXI: VPGFt:EELLPER--TAEYYYRRGSSLT'P pOTPTVI-FRNW'OISCEOLLALESTALYCTHMDDPSPRE---M-INtROVOKFDERVYTSFS
HCAXIII FTNFDLLSLLPP---SWDYaWTYGSLTISPPLr LIESMTIVLKQP:NII SSLAKORSLLCTAEG6SAAU -----LVSSRPP QPLGRWVRASSF
HCAXIV VPPE IRELLPK-- !GYF--QYTPYN' GS"L.TPPCYQS VLTV'FYRRSQK i-- KQ GTLFS-F.-EEPSKL ....ALLO PI.N '!"TASFI
# 4* $ **"## # # $ *





Figure 1-2. Multiple sequence alignment of fourteen HCA domains. Alignment was

performed using ClustalW and residue numbering is according to HCA II

(Thompson et al., 1994). Conserved and similar substituted residues are

indicated by and #, respectively. Regions in bold that are red and blue

signify P-strand and a-helical regions of HCA II.





































Figure 1-3. Conservation of HCAs. Gray ribbon backbone representation of HCA II with
completely conserved residues of fourteen HCAs indicated as red spheres.
Active site residues are shown as yellow ball-and-stick (HCA II numbering),
the zinc atom is a black sphere and the N- and C-terminus as labeled.














a) TIW
H200


H64





V-62
v 6


1


b)
T200
*Y7 I


H119 t H94


A


H67


T199

--


H64 H119 Z



H96

N62 N67


H94
H94


c)
T20



K64






N62


T199



119 H


Figure 1-4. Active sites of HCA I, II, and III. (a) HCA I, (b) HCA II, and (c) HCA III.
Active site residues are in yellow ball-and-stick and are as labeled, zinc atom
= black sphere. Figure was generated with Bobscript and Raster3D (Esnouf,
1997; Merritt and Bacon, 1997).















































Figure 1-5. Molecular and electrostatic surface potentials of HCAs. (a) HCA II; (b) HCA
III; (c) HCA VI; (d) HCA-RP VIII; (e) HCA IX; (f) HCA XII dimer. Al
models are shown in the same orientation with the active site in the center.
Negative and positive charge is represented by red and blue, respectively.
Figures was generated with GRASP (Nicholls et al., 1991).





































Figure 1-6. Ribbon diagram of HCA II with secondary structure elements. Red regions
are P-strands (PA PJ), blue regions are a-helices (aA aG). Zinc atom is
shown as a black sphere and the N- and C-termini are labeled. Figure
generated with Bobscript and Raster3D (Esnouf, 1997; Merritt and Bacon,
1997).











T19

T200 ZnOH-


Sr '. W1 H94



Proton ShuttleX3a H119l

"out" N "il" ", H96
H64 / '



/62 N67



Figure 1-7. Active site of wild type HCA II. Catalytic residues are shown in yellow ball-
and-stick, solvent molecules = red spheres, zinc atom = black sphere.
Residues are as labeled and inferred hydrogen bonds are indicated by the
dashed orange lines. Figure was generated with Bobscript and rendered with
Raster3D (Esnouf, 1997; Merritt and Bacon, 1997).












T199

Y7 T200 sulfate



H64 H96


6,-


\ N67


T199
Y7 T200
T azde

64 e H94
H64
H96


-'N62


N7
\ 'N6


Y T200




H64 N62


/ N62


T199

acetazol amide


T200

11_


T199

'b-rinzolamide


N67 /N62 67


Figure 1-8. HCA II in complex with inhibitors. (a) Sulfate, (b) Azide, (c) acetazolamide,
and (d) brinzolamide. HCA II active site residues are shown as yellow ball-
and-stick and the black sphere is the zinc atom. PDB accession codes used (a)-
(d): 1T9N, 1RAY, 1YDA, and 1A42 (Fisher et al., 2005; Jonsson et al., 1993;
Nair et al., 1995; Stams et al., 1998). Figure was generated with Bobscript and
Raster3D (Esnouf, 1997; Merritt and Bacon, 1997).














CHAPTER 2
STRUCTURAL AND KINETIC ANALYSIS OF PROTON SHUTTLING IN HUMAN
CARBONIC ANHYDRASE II

Introduction

Human carbonic anhydrase II (HCA II) is one of the most efficient enzymes with

a turnover number near 106 s-1 and kcat/KM of 1 x 108 M ^s1. This fast rate indicates that

the catalysis is limited by the rate of diffusion of substrate into the active site of HCA II

(Khalifah, 1971). The overall catalytic mechanism was presented in some detail in

Chapter 1 and will only be discussed here briefly. The first step in the interconversion of

CO2 and HCO3- is the hydration of CO2 by the zinc-bound OH- which is then followed by

the displacement of HCO3- by water. The second step is the deprotonation of the zinc-

bound water to regenerate the zinc-bound OH- and involves both intramolecular and

intermolecular proton transfer steps (Silverman, 1982; Silverman and Lindskog, 1988).

Imidazole and derivatives act as nucleophilic and general base catalysts. The

structure of imidazole, with two almost identical N atoms, allow it to pick up a proton off

one of its N atoms forming a cation and then delivering it to the second N atom. As the

functional group of histidine, it is commonly associated with proton transport in proteins

(Scheiner and Yi, 1996). His64 in HCA II acts as a proton shuttle between the zinc-bound

solvent and buffer in solution and mutation of His64 to an alanine reduces enzymatic

activity 10-50 fold. This mutation does not affect the hydration/dehydration part of

catalysis, but the observed decrease in proton transfer can be rescued by supplying free

imidazole in the reaction buffer. It has been postulated that the proton transfer between









the zinc-bound water and His64 occurs through intervening water molecules

(Venkatasubban and Silverman, 1980; Tu et al., 1989).

In other protein systems, such as cytochrome c oxidase and the bacterial

photosynthetic reaction center, hydrogen-bonded water chains have been observed in

several crystal structures and the location and geometry of these chains suggest that they

participate in proton transfer reactions (Pomes and Roux, 1996). These chains are thought

to be effective in long-range proton translocation and such hydrogen-bonded water chains

have been called proton wires (Nagle and Morowitz, 1978). Protons do not move by

diffusion as a hydrated proton (hydronium ion, H30 ), instead the high observed mobility

of protons is thought to occur from successive protonation-deprotonation events. In these

cases, protons "hop" from oxygen to oxygen along a pre-existing hydrogen-bonded water

chain in a Grotthus-type mechanism (Pomes and Roux, 1996). Similar proton wires are

observed in the active site of HCA II and these water molecules bridge the distance

between the zinc-bound solvent and the proton shuttle, His64. The water molecules that

span the 8 A distance between zinc and His64 appears to be hydrogen bonded to each

other but not to the imidazole group of His64 (Eriksson et al., 1988). The proton transfer

path in HCA II was first described by Steiner et al. based on assays that showed

significant solvent hydrogen isotope effects. These experiments indicated that maximum

velocity, kcat, was limited in rate by the intermolecular proton transfer (Steiner et al.

1975).

The Bronsted relation, as applied to proton transfer, is a linear free-energy

relationship that correlates the rate constant (kB) for proton transfer with the difference in









acid/base strength (ApKa) of the proton acceptor and donor, is shown in eq 2-1

(Silverman et al., 1993; Silverman, 2000)

log(kB) = P[pKa (acceptor) pKa (donor)] + constant (2-1)

The slope (0) of a Bronsted plot of log(kB) versus ApKa is used to investigate a

reaction mechanism and can be used as an estimate of the extent of proton transfer

between reactants and products in the transition state. Observed variation in the slope

over a range of pKa values has been interpreted through Marcus theory and yields 1) an

intrinsic energy barrier for proton transfer, and 2) two work terms, wr and wp, that

quantitate the energy required for aligning the reactants in the reaction complex in both

the forward and reverse directions, respectively (Silverman et al., 1993; Silverman, 2000;

Kresge and Silverman, 1999).

Application of Marcus rate theory to Bronsted plots of the proton transfer

processes in carbonic anhydrase has provided a way to experimentally determine the

intrinsic energy barrier as well as separating different thermodynamic contributions from

the observed activation energy (Silverman, 2000). The rate constant for proton transfer is

maximal when the difference between the pKa values of the proton acceptor and donor as

near zero (Rowlett and Silverman, 1982; Silverman et al., 1993). Further analysis of free

energy plots has shown there is a significant energetically unfavorable pre-equilibrium

that exists before a very rapid proton transfer event occurs. It was determined that

catalysis proceeds with an intrinsic free energy of activation that is very small (- 1.5

kcal/mol) but with a large work function wr (~ 10 kcal/mol) (Kresge and Silverman,

1999; Silverman, 2000). The magnitude of this work function accounts for the relative

low proton transfer rate in HCA II, at most 106 s-1, compared to maximal rates of 1011 s1









measured for proton transfer between naphtol-related photo acids to acetate in solution

(Pines et al., 1997). It is not well understood what processes add to the work functions,

but in HCA II this may involve the rearrangement and formation of a hydrogen bonded

water chain as well as orientation of His64. These factors imply that setting up the active

site environment for proton transfer takes far more energy than the actual proton transfer

itself (Silverman et al., 1993).

The shuttling of protons between bulk solvent and the zinc-bound solvent

molecule may require some conformational mobility of His64 as the proton

donor/acceptor. Considerable support for this comes from various crystal structures in

which the proton shuttle residue His64 in HCA II and Glu84 in the archaeal carbonic

anhydrase from 1 lethaini,, ina thermophila show two conformational rotamers in the

active-site cavity (Nair and Christianson, 1991; Tripp and Ferry, 2000; Iverson et al.,

2000). A chemically modified cysteine residue acting as a proton shuttle in a mutant of

CA V also shows evidence of multiple orientations (Jude et al., 2002).

Structural studies by Nair and Christianson showed that His64 occupies different

positions depending on the pH. At pH 5.7 the side chain of His64 rotates around x1 by

640 to occupy what is termed the "out" position (~ 12 A away from the active site)

compared to the structure at pH 8.5 that shows the side chain of His64 pointing towards

the active site (~ 8 A away from the active site) in what is known as the "in"

conformation (Nair and Christianson, 1991). A similar observation was made for the

Thr200--Ser mutant at pH 8.0 where His64 was observed to occupy an even further

"out" position due to a rotation about X1 of 1050 (Krebs and Fierke, 1991). In both cases,









the conformational mobility of His64 as a function of pH or the mutation at position 200

did not seem to adversely affect proton transfer kinetics (Nair and Christianson, 1991).

The current understanding of the requirements for rapid proton transfer, as

measured for HCA II, involves several aspects that could be investigated by comparisons

of kinetics and structure. The kinetic effect of the location of a histidine proton shuttle

has been studied by introducing histidine residues at different positions in the active-site

cavity of HCA II. These results show that a histidine residue at sites other than position

64 is able to participate in proton transfer; specifically, His67 appears capable of more

efficient proton transfer than His62 (Liang et al., 1993).

To further characterize and understand the relationship between efficient proton

transfer and the existence of proton wires in HCA II, detailed comparisons of kinetics and

active site water structure were conducted. These included structure determinations of

wild-type HCA II and HCA II mutants (H64A/N62H and H64A/N67H) from pH 5.1 to

10.0. These data were correlated with catalytic activities that were measured by the

exchange of 180 between CO2 and water from pH 5.0 to 9.0 as catalyzed by these

enzymes to determine what common structural features are important for proton transfer.

Materials and Methods

Enzymes

Plasmids with the appropriate mutations in the cDNA of HCA II were provided by

Professor Sven Lindskog, Umed University, Sweden. Expression of wild type and mutant

HCA II was performed in E. coli BL21(DE3) pLysS cells that were grown to an optical

density of- 0.60 as measured at a wavelength of 600 nm. Protein expression was induced

by the addition of 1 mM isopropyl-P-D-1-thiogalactopyranoside (IPTG) and 1 mM zinc

sulfate was also added for uptake in the expressed protein. At 4 hours post-induction,









cells were harvested by centrifugation and the cell pellets frozen at -20 C overnight. Cell

pellets were lysed by freeze/thawing and solubilized in 0.2 M sodium sulfate, 50 mM

Tris-Cl (pH 9.0). The soluble cell fraction was obtained by centrifuging the lysates at

100,000 x g for 1 hour at 4 C. Enzymes were further purified from the supernatant by

affinity chromatography using p-amino-methyl-benzenesulfonamide (pAMBS; a specific

binder to the active site of co-CAs) coupled to agarose beads as described elsewhere

(Khalifah, 1977). Purity of the protein was verified by electrophoresis on a 12%

polyacrylamide gel stained with Coomassie. The concentration of HCA II was

determined by measuring the absorbance at 280 nm and using a molar absorptivity of 5.4

x 104 M-cm1 (Coleman, 1967).

Crystallography

Crystals of wild type and mutant HCA II were obtained using the hanging drop

method (McPherson, 1982). The crystallization drops were prepared by mixing 5 tl of

protein (10.5 mg/ml concentration in 50 mM Tris-HCl, pH 7.8) with 5 tl of the

precipitant solution (50 mM Tris-HCl, pH 7.8, 2.5-2.9 M ammonium sulfate) at 4 C

against 600 tl of the precipitant solution. Useful crystals were observed within five days

after crystallization setup. The pH of the wild type crystals were obtained by equilibrating

crystals in appropriate buffers (50 mM sodium acetate, pH 5.1; 50 mM MES, pH 6.1; 50

mM Tris-HCl, pH 7.0, 7.8 and 9.3; 50 mM CAPS, pH 10.0) and 3.0 M ammonium

sulfate. The pH of the double mutant crystals were obtained by using the same approach

as above using the buffers (50 mM Tris-HCl, pH 6.0 and 7.8) and 3.0 M ammonium

sulfate. Crystals were allowed to equilibrate for 4-12 hours at 40C before data collection









to ensure that complete solvent exchange in the crystal lattice occurred (the pH stated was

that as measured at the start of the experiment).

X-ray-diffraction data sets were obtained using an R-AXIS IV++ image plate system

with Osmic mirrors and a Rigaku HU-H3R Cu rotating anode operating at 50 kV and 100

mA. The detector to crystal distance was set to 100 mm for H64A/N67H, and 120 mm for

H64A/N62H and wild type HCA II X-ray data collection. Each data set was collected at

room temperature from a single crystal mounted in a quartz capillary. The oscillation

steps were 1 with a 3 minute exposure per image. X-ray data processing was performed

using DENZO and scaled and reduced with SCALEPACK software (Otwinowski and

Minor, 1997). Data set statistics for the wild type and mutant structures at different pHs

are given in Tables 2-1 and 2-2, respectively.

All models were built using the program 0, version 7 (Jones et al., 1991). Refinement

was carried out with the software package CNS, version 1.1 (Brunger et al., 1998). The

wild type HCA II structure (Protein Data Bank accession number 2CBA; Hikansson,

1992), which was isomorphous with all the data sets collected, was used to phase the data

sets. To avoid phase bias of the model, the zinc ion, mutated side chains and water

molecules were removed. After one cycle of rigid body refinement, annealing by heating

to 3000 K with gradual cooling, geometry-restrained position refinement, and

temperature factor refinement, the 2Fo F, Fourier maps were generated. These density

maps clearly showed the position of the zinc and the mutated residues, which were

subsequently built into the respective models. After several cycles of refinement, solvent

molecules were incorporated into the models using the automatic water-picking program

in CNS until no more water molecules were found at a 2.0O level. Refinement of the









models continued until convergence of the Rwork and Rfree was reached. Final model

statistics for wild type and mutant structures are given in Tables 2-1 and 2-2,

respectively.

Activity Analysis by 80 Exchange

These assays were performed by Dr. Chingkuang Tu in the Silverman lab and the

details of determinating the rate constants for catalysis by HCA II are discussed in detail

below. The 180 exchange method is based on the measurement, using membrane-inlet

mass spectrometry, of the exchange of 180 between CO2 and H20 at chemical

equilibrium (Silverman, 1982) (eqs 2-2 and 2-3).

HCOO O- + EZnH20 EZnHCOO' O- COO + EZn OH- (2-2)

H20O

EZn OH- + BH+ EZnH2180 + B EZnH20 + H2180 + B (2-3)

An Extrel EXM-200 mass spectrometer with a membrane inlet probe was used to

measure the isotopic content of CO2. Solutions contained a total concentration of all

species of CO2 of 25 mM and the ionic strength was maintained by the addition of 0.2 M

sodium sulfate. This approach yields two rates for the 180 exchange catalyzed by

carbonic anhydrase. The first is Ri, the rate of exchange of CO2 and HCO3- at chemical

equilibrium, as shown in eq 2-4.

Ri/[E] = katex[S]/(Keff + [S]) (2-4)

Here kcatex is a rate constant for maximal interconversion of substrate and product,

Keffs is an apparent binding constant for substrate to enzyme, and [S] is the concentration

of substrate, either CO2 or HCO3-. The kcatex/Ke f ratio is, in theory and in practice, equal

to kcat/KM obtained by steady state methods. The binding of CO2 and HCO3- to the active

site of HCA II is weak (Krebs et al., 1993), and in this work too it is assumed that [C02]









<< Keffs. The pH dependence of kcat/KM depends on the ionization state of the zinc-bound

water, as shown in eq 2-5.

kcat/KM = (kcat/KM)max(1 + [H ]/(Ka)ZnH20)-1 (2-5)

A second rate determined by the 180 exchange method is RH20, the rate of release

from the enzyme of water bearing substrate oxygen (eq 2-3). This is the component of the

180 exchange that is enhanced by exogenous proton donors (Silverman, 1982). The pH

dependence of RH20/[E] is often bell-shaped, consistent with the transfer of a proton from

a single predominant donor to the zinc-bound hydroxide. In these cases, the pH profile is

adequately fit by eq 2-6 in which kB is a pH-independent rate constant for proton transfer,

and (Ka)donor and (Ka)ZnH2o are the noninteracting ionization constants of the proton donor

BH+ of eq 2-3 and the zinc-bound water.

kBobs = kB/{[1 + (Ka)donor/[H ]][1 + [H ]/(Ka)ZnH20] } (2-6)

Results and Discussion

Crystallography

All crystals of wild type, H64A/N62H, and H64A/N67H HCA II were isomorphous in

the P21 space group with mean unit cell dimensions: a = 42.7 2.0 A, b = 41.6 1.0 A, c

= 72.9 2.0 A, and f = 104.6 2.00. Wild type crystals from pH 5.1 to 10.0 all diffracted

to 2.0 A, while crystals of H64A/N62H and H64A/N67H diffracted to between 1.63 and

1.90 A. Tables 2-1 and 2-2 contain summaries of the diffraction data set and model

statistics from wild type and mutant HCA II crystals, respectively.

A least squares superposition of wild type HCA II structures (at pH 5.1, 6.1, 7.0,

9.0, and 10.0) onto the structure at pH 7.8 indicated no significant differences between

them and had an average root mean square deviation (rmsd) of less than 0.1 A for all

atoms. Superposition of H64A/N62H HCA II at pH 6.0 and 7.8, as well as H64A/N67H









at pH 6.0 and 7.8, onto the pH 7.8 wild type HCA II structure showed an average rmsd

near 0.15 A for all atoms. When the mutants were compared to each other, the rmsd was

only 0.07 A.

The rmsd of bond lengths and angles for all structures were between 0.038 and

0.005 A and between 1.3 and 2.1, respectively. Ramachandran statistics for all amino

acids were ~ 90% in most favored region and 10% in the additionally allowed region

with no residues in the generously and disallowed regions.

Effect of pH on the Wild Type HCA II Active Site

The tetrahedral coordination of the zinc atom by His94, His96, His199, and a non-

protein atom (either H20/OH- or sulfate) was intact over the broad pH range from 5.1 -

10.0 used in this study. At pH 5.1 a sulfate ion bound directly to the zinc atom was

observed (Figure 2-1). The sulfate binding displaced the zinc-bound H20/OH- but had no

effect on the coordination of the zinc atom. In all the wild type structures determined

from pH 5.1 to 10.0, there was a highly conserved, well-ordered water network extending

from the zinc-bound solvent (or sulfate) to the imidazole of His64. These networks

consist of four water molecules, WI, W2, W3a, and W3b. WI is hydrogen bonded to

W2, which is in turn hydrogen bonded to W3a and/or W3b. WI is held in place by

interactions with not only W2, but also with the hydroxyl group of Thr200. W3a and

W3b interact with Tyr7 and Asn62/Asn67, respectively (Figure 2-1). W2 is the only

water molecule held entirely in place by interactions with other water molecules. The

configuration of these networks leads to two possibilities for a proton transfer pathway

from the zinc-bound solvent to the proton shuttle His64; WI -1 W2 -- W3a -- His64 or

Wl -* W2 -* W3b -* His64. However, the distances between the distal waters, W2,

W3a, and W3b, and His64 were all too long (> 3.4 A) to constitute viable hydrogen









bonds and it is unclear how the proton will make the jump from one of these waters to

His64. Other studies with HCA VI alkylated with an imidazole-containing reagent

showed a completed, hydrogen-bonded water chain and was associated with a very small

rate of proton transfer (103 s-1 compared to 106 s-1 for His64 in wild type HCA II; Jude et

al., 2002). Thus, the observation in the crystal structure of a completed, hydrogen-bonded

water chain that links the proton donor and acceptor is not a requirement for the rapid

proton shuttling on the scale determined for HCA II.

It should also be noted that the hydrogen bonds in crystal structures are assigned based

on the distance and geometries between hydrogen bond donors and acceptors. Hydrogen

atoms are invisible by X-ray crystallography at these resolutions and subsequently these

assignments can be extremely misleading as there is no way to observe the ionization of

waters or amino acid side chains. Currently, the only definitive way to observe hydrogen

bond are with either ultra-high resolution (< 1 A) X-ray crystallography or

macromolecular neutron diffraction studies.

The only observed structural difference between wild type HCA II at the various pHs

was the positional occupancy of His64 (Figure 2-1). The occupancies of His64 were

determined from Fo-Fo omit maps, where His64, WI, W2, W3a, and W3b were removed

for the calculation to avoid model bias. The residual electron density volumes were used

to determine the relative occupancies of His64 in the "in" and "out" positions. At pH 5.1,

His64 was predominantly in the "in" position and this could be due to electrostatic effects

exerted by the sulfate ion. Consequently, the "out" conformer was modeled as a water

molecule. At pH 6.1, 7.0, and 7.8 the occupancies were 60, 70, and 80% in the "in"

conformation. There was no increased occupancy at higher pH and at pH 9.0 and 10.0









His64 was still 80% in the "in" conformation (Figure 2-1). The overall trend from this

data shows that with increasing pH, His64 favors the "in" conformation and this result is

consistent with the work of others (Nair and Christianson, 1991). The side chain torsion

angles for the "in" and "out" conformers were (X1 = 480, X2 = -980) and (X1 = -430, X2 = -

900), respectively. Thus, over a range of pHs it was observed that His64 occupies both

positions with equivalent occupancy and this phenomenon is consistent with its function

as a proton shuttle. The physiological function of His64 as a shuttle in catalysis is

supported by these observations as hydration and dehydration occur at nearly the same

rates. Structural and kinetic studies with 4-methylimidazole and H64A HCA II suggest

that binding of this exogenous proton donor/acceptor to a position that mimics the "out"

position of His64, is not as efficient in proton transfer as His64. Perhaps the motion of

this residue is somehow related to its efficacy as a proton shuttle. When His64 occupies

the "in" position it can pick up an excess proton from one of the active site solvent

molecules and then move rapidly to the "out" position and deliver it to bulk solvent.

Effect of pH on the Mutant HCA II Active Site

The hydrogen bonding pattern and solvent network is different for the

H64A/N62H and H64A/N67H HCA II mutants compared to each other and wild type

(Figure 2-2). Even though both mutants display similar solvation levels in the active site,

the organization of the water molecules has been altered. It is interesting that both

mutants have sulfate bound at pH 6.0, while in wild type sulfate was only observed at pH

5.1 (Figure 2-2 (a), (c)). However, at ph 7.8 there is not sulfate present in either wild type

or mutant HCA II. The sulfate binds in the same orientation in all the structures

determined in this study: it displaces the zinc-bound solvent and simultaneously engages

in a hydrogen bond with the hydroxyl group of Thr199, thus maintaining the tetrahedral









coordination of the zinc. Sulfate has been shown to be inhibitory to HCA II at low pH

and it was thought that it interacted with the zinc (Simonsson and Lindskog, 1982). These

data suggest that the inhibition at low pH by sulfate may be due to the binding of a

protonated sulfate ion in a similar mode to other protonated inhibitors such as HSO3- and

the sulfonamide drugs (Liljas et al., 1994).

There are several structures of mutant HCA II in complex with sulfate but the

results presented here show for the first time a sulfate bound to the zinc center of wild

type HCA II and this occurs only at very low pH (Xue et al., 1993(a)).

In both mutants the introduced His side chains extend into the active site cavity

towards the zinc. In contrast to His64 in wild type HCA II, neither one of the His in the

mutants display any conformational mobility. In H64A/N62H HCA II at pH 7.8 there is a

complete hydrogen bonded solvent network that extends from the zinc-bound solvent to

His62 (WI -- W2 -- W3b; Figure 2-2 (b)). Due to the long distance (3.2 A) between

W3b and His62 this interaction represents a very weak hydrogen bond. The water W3b

that His62 is hydrogen bonded to is also connected to the side chain of N67.

In H64A/N67H HCA II at pH 7.8, His67 is connected directly to W2 with a bond

distance of 3.2 A and W3b has been completely displaced by the side chain of His67 (WI

-- W2; Figure 2-2). His67 is also hydrogen bonded to the side chain of N62 and this

interaction might prohibit rotational freedom of the His during catalysis.

The His at position 67 is ~ 6.6 A away from the zinc while His62 is over 8 A

away (Table 2-3). This extra distance is spanned by an additional water molecule and the

solvent network appears more branched in H64A/N62H HCA II compared to

H64A/N67H HCA II. The observed network of two water molecules between the proton









donor and acceptor seen in H64A/N67H HCA II at pH 7.8 is similar to one of the

possible pathways (WI -- W2) as seen in wild type HCA II from pH 6.1 to 10.0, except

that in wild type all of the distal waters are over 3.2 A away from the proton shuttle.

Catalysis

The pH dependence of the hydration/dehydration rate, kcat/KM, and the proton

transfer-dependent rate constant, RH2o/[E], for the release of 180 labeled water from the

active site, was measured. The pH range was from pH 5.0 to 9.0 for wild type and both

mutants. These data was compared to 180 exchange data obtained for a mutant, H64A

HCA II, that lacks a proton shuttle in the active site. The data for wild type HCA II shows

that the enzyme is remarkably stable and highly active over the pH range studied (Figures

2-3 and 2-4).

The data in Figure 2-3 show the kcat/KM for CO2 hydration for each of the mutants

(H64A, H64A/N62H, and H64A/N67H HCA II) superposed onto data for wild type HCA

II. The data shows that the mutants have very similar values to wild type and this serves

as a control indicating that the mutations do not cause gross structural changes that

interfere with the first step of catalysis.

The data in Figure 2-3 were fit to a single ionization with pKa values as given in

Table 2-3 and these are characteristic of the pKa of the zinc-bound water (Lindskog,

1997). The pKa for the proton donor and acceptor in wild type HCA II are very close in

value to each other (6.9 and 7.2) and this might be another requirement for efficient

proton transfer between them. In both mutants the differences between the pKa values are

larger compared to wild type and could contribute to the lower observed rate of proton

transfer between His and the zinc-bound water (Table 2-3).









The pKa values for the zinc-bound water given in Table 2-3 are in agreement with

values obtained by others that measured the pH dependent esterase activity of HCA II

(Liang et al., 1993). For H64A/N67H HCA II it appears from the plot in Figure 2-3 that a

single pKa is not sufficient to fit the data and this implies that a second ionization,

probably that of His67, is influencing the kcat/KM.

The proton transfer activity data shown in Figure 2-4 indicate that all three

mutants are significantly impaired compared to wild type HCA II, with H64A being the

slowest. These data showed only slightly better rates for H64A/N62H HCA II compared

to H64A HCA II (Figure 2-4). However, there was significant enhancement of catalysis

for H64A/N67H HCA II at pH < 6.5 compared with that of H64A HCA II. The solid

lines of Figure 2-4 represent least-squares fits to eq 6, which assumes that the observed

enhancement of RH2o/[E] above that of H64A HCA II is due to the proton transfer

activity of the inserted His to the zinc-bound hydroxide.

Other studies using stopped flow methods at steady state have shown that this

mutant had a turnover number, kcat, that is 20% of that observed for wild type (Liang et

al., 1993). The data presented here was determined by 180 exchange methods at chemical

equilibrium and showed that this mutant had a maximal rate constant for proton transfer

that is 25% that of wild type (Table 2-3). It should be noted that there is typically up to

20% error in these measurements and this is due to the scatter of the points and the

limited pH range.

Conclusion

HCA II shows remarkable structural and kinetic stability over a wide range of pH

(5.1 to 10.0). The only structural difference is the side chain of His64 that displays two

conformational states that are almost equally occupied at physiological pH. In wild type,









H64A/N62H, and H64A/N67H HCA II structures there is either no completed hydrogen-

bonded water chain between proton donor and acceptor or the distal water is very weakly

bound to the proton shuttling residue. A His inserted at position 67 shows appreciable

proton shuttling activity (25% of wild type), compared to a His at position 62 (4% of wild

type). In these examples, the distance between the proton shuttling His and the zinc-

bound solvent and the number of waters that spans this distance, might be more important

for efficient proton shuttling than observing a complete hydrogen-bonded chain of waters

to the His residue. The results suggest that the optimal distance for the His to the zinc is

between 6.6 and 7.5 A and that the number of intervening water molecules should not

exceed two for the support of efficient and fast proton transfer.















Table 2-1. Data set and model statistics for wild type HCA II from pH 5.1 to 10.0.
HCA II pH 5.1 pH 6.1 pH 7.0 pH 7.8 pH 9.0 pH 10.0
Resolution (A) 20.0 -2.00 20.0 2.00 20.0- 2.00 20.0 -2.00 20.0 -2.00 20.0 -2.00
(2.07- (2.07- (2.07- (2.07- (2.07 2.00) (2.07-
2.00)* 2.00) 2.00) 2.00) 2.00)
Total Number of 15898 16365 15878 16212 16751 16714
Unique (1546) (1575) (1554) (1539) (1665) (1598)
Reflections
Completeness 93.1 (90.0) 95.9 (94.0) 93.5 (91.8) 95.1 (91.9) 98.6 (99.2) 97.8 (95.4)
(%)
Redundancy 1.9 (1.9) 2.7 (2.7) 2.3 (2.3) 3.2 (3.2) 2.6 (2.5) 2.6 (2.6)

Rsymm 0.063 0.120 0.051 0.073 0.092 0.081
(0.199) (0.447) (0.140) (0.217) (0.198) (0.143)
Rcryt /Rwor 0.195/0.209 0.145/0.205 0.130/0.184 0.128/0.201 0.134/0.173 0.134/0.175
Ave B-factor 16/20/28 23/27/30 18/22/28 14/18/27 17/21/28 17/21/28
(A2)
Main/side/solvent
Number of 112 91 112 150 112 115
solvent
Data in parenthesis are for the highest resolution shell.
,Rsymm = I /
Rcryst |Fo| |F|c / I Fo|
Rfree is calculated the same as Rcryst except it is for data omitted from refinement (5% of reflections for all data sets)













Table 2-2. Data set and model statistics for H64A/N62H and H64A/N67H HCA II at pH
6.0 and 7.8.


HCA II


Resolution (A)

Total Number of
Unique Reflections
Completeness (%)
Redundancy

Rsymmt
cryst / Rwork
Ave B-factor (A2)
Main/side/solvent
Number of solvent


H64A/N62H
pH 6.0
20.0 -1.80
(1.86- 1.80)*
22162 (2132)


H64A/N62H
pH 7.8
20.0 -1.90
(1.97- 1.90)
18130 (1813)


H64A/N67H
pH 6.0
20.0- 1.63
(1.69- 1.63)
29130 (2583)


H64A/N67H
pH 7.8
20.0- 1.80
(1.86- 1.80)
21547 (2070)


95.0 (91.60) 91.8 (92.9) 93.5 (83.7) 92.7 (89.0)


3.0 (3.0)


2.5 (2.3)


3.1 (2.9)


0.074 (0.395) 0.072 (0.316) 0.059 (0.329)


2.8 (2.6)


0.068 (0.292)


0.171/0.206 0.168/0.217 0.178/0.210 0.166/0.209
16/20/30 16/20/29 17/21/30 16/20/30


* Data in parenthesis are for the highest resolution shell.
?Rsymm I I / Y
SRcryst |Fol Fc| / I |Fol
Rfree is calculated the same as Rcryst except it is for data omitted from refinement (5% of
reflections for all data sets).







46




Table 2-3. pH-Independent rate constants for proton transfer and pKa for proton donor
and acceptors in wild type and mutant HCA II.
Enzyme kB (pS-1) a (pKa)His (pKa)znH2o Zn-His distance
Wild Typeb 0.8 0.1 7.2 0.1 6.9 0.1 7.5
H64A -0.02 N/a 6.9 0.1 N/a
H64A/N62H 0.2 0.1 5.3 0.3 7.2 0.1 6.6


H64A/N67H 0.03 5.7 0.4 7.3 0.1 8.2
a Values are from a least-squares fit of eq 6 to the data of Figures 2-3 and 2-4.
b From Duda et al., 2001
' Values are uncertain due to the scatter of the data.













a) T199 b) T199 -

T200 Sulfate T200 \ Zn---H,O OH-


W2
"out" W3a ...-. 4t 1t 193bW
out3b H96 "out" *i

H is64N6 His64 N67


c) T199 d) T199
T200 Zn--H:OOH- T200 Zn"-H O OH-

W j 'HS 9 Wl V1 H94


1 62 2n
"out- n > b H96 "out" .. W3b H96

His64 \ 7 His64 \7 N







Figure 2-1. Crystal structures of wild type HCA II active site. Panels (a) through (d) are
shown in the same orientation, (a) pH 5.1, (b) pH 6.1, (c) pH 7.0, and (d) pH
10.0. Active site residues are shown in yellow ball-and-stick and the zinc as a
black sphere. Water molecules are as labeled and are shown as red spheres.
2Fo-Fc electron density maps for His64 are shown in blue and are contoured at
1.0 G. Dashed lines represent inferred hydrogen bonds based on geometry and
distance between donor and acceptor atoms. Figure was generated and
rendered with Bobscript and Raster3D (Esnouf, 1997; Merritt and Bacon,
1997).







48





a) T199 b) T199

T200 Sulfate T200 Zn -HO/OH-

T200 ` W1 H4

4 H119 W-2 .
A64 W2 *W3b H96 A64 W3b H96

6-. 2 A .N6\
N6%7
H62 N6 H62
T199 T199
) -t d) -Y
T200200 ulfate d) T200 Zn2-H2O/OH-



W3a V-3a *W2
A64 W2 A64 H
/ H67

N62 -




Figure 2-2. Crystal structures of mutant HCA II active site. Panels (a) through (d) are
shown in the same orientation, (a) H64A/N62H pH 6.0, (b) H64A/N62H pH
7.8, (c) H64A/N67H pH 6.0, and (d) H64A/N67H pH 7.8. Active site residues
are shown in yellow ball-and-stick and the zinc as a black sphere. Water
molecules are as labeled and are shown as red spheres. 2Fo-Fc electron density
maps for His62/His67 are shown in blue and are contoured at 1.0 a. Dashed
lines represent inferred hydrogen bonds based on geometry and distance
between donor and acceptor atoms. Figure was generated and rendered with
Bobscript and Raster3D (Esnouf, 1997; Merritt and Bacon, 1997).






49




9




8-















5 6 7 8 9
pH



Figure 2-3. The pH profiles for kcat/KM catalyzed by wild type and mutant HCA II. (e)
Wild type, (.m) H64A, (o) H64A/N62H, (o) H64A/N67H. Data were obtained
at 250 in the absence of exogenous buffers using a total concentration of all
species of CO2 of 25 mM, with the ionic strength maintained at 0.2 M by
addition of sodium sulfate. The solid lines are fit to a single ionization (eq 5)
with the pKa given in Table 2-3.







50





6



6,---------------5---



















3
5 6 7 8 9

pH




Figure 2-4. The pH profiles for RH20/[E] catalyzed by wild type and mutant HCA II. (*)
Wild type, (m) H64A, (o) H64A/N62H, (n) H64A/N67H. Data were obtained
at 250 in the absence of exogenous buffers using a total concentration of all
species of CO2 of 25 mM, with the ionic strength maintained at 0.2 M by
addition of sodium sulfate.














CHAPTER 3
STRUCTURAL AND KINETIC EFFECTS OF HYDROPHOBIC MUTATIONS IN
THE ACTIVE SITE OF HUMAN CARBONIC ANHYDRASE II

In Chapter 2, studies of the structural and kinetic effects of moving the proton

shuttle to various locations in the active site of human carbonic anhydrase II (HCA II)

were presented. Kinetic and structural data over various pHs were obtained and

correlated with observed changes in the active site and proton transfer rates. In Chapter 3,

a different approach to studying proton transfer was undertaken and is discussed. Site-

direct mutagenesis of several key active site residues were performed with the

expectation that these mutations would affect proton transfer rates and the architecture of

the hydrogen bonded solvent network found in the active site of HCA II.

Introduction

HCA II is the most efficient of all the HCAs with a maximal turnover rate of 106 S-1

(Khalifah, 1971). Catalytic rates of HCA II and the mechanism has been described in

detail in Chapters 1 and 2. Briefly, the second part of the catalysis by HCA II involves

two proton transfer steps: the first between the zinc-bound solvent and an internal proton

acceptor, His64, and the second between His64 and an exogenous proton acceptor such

as the bulk solvent of buffer (Silverman, 1982; Silverman and Lindskog, 1988, Tu et al.,

1989). The first proton transfer event involves several water molecules (WI, W2, W3a,

and W3b; Figure 3-1) and this network is also described in detail in Chapter 1 (Fisher et

al., 2005). With the exception of W2, all other water molecules are coordinated by

hydrogen bonding interactions to several key active site residues such as Tyr7, Asn62,









and Asn67. Figure 3-1 shows a view of the wild type HCA II active site with these

residues and water molecules as labeled.

The proton shuttle residue, His64, is located on the edge of the active cavity, about

8 A away from the zinc. This distance makes a direct proton transfer impossible and as a

result the excess proton has to exit the active site via the conserved and ordered water

molecules shown in Figure 3-1 (Steiner et al., 1975; Eriksson et al., 1988). In several

crystal structures determined at various pHs, His64 has been shown to occupy two

distinct conformations, the so-called "in" and "out" positions. However, the presence of

the dual conformation is not always observed and under some conditions (such as pH or

in the presence of inhibitors) it appears to be either all in the "in" or all in the "out"

position (Nair and Christianson, 1991; Krebs and Fierke, 1991; Fisher et al., 2005). It

should be noted that at the resolution of these studies, an occupancy change of 0-10%

would be impossible to observe.

To better understand the role of the conserved water network and the implications

of side chain mobility of His64 for efficient proton transfer, three hydrophilic amino

acids in the active site of HCA II were replaced by site-directed mutagenesis with

hydrophobic residues: Y7F, N62L, and N67L. Effects of these substitutions were

evaluated by measuring kinetic proton transfer rates, the conformation of His64, and the

structural effects on overall active site architecture as well as the water network.

Materials and Methods

Enzymes

Plasmids with the appropriate mutations in the cDNA of HCA II were produced by

site-directed mutagenesis using the Qiagen QuikChange kit. Mutagenic primers were

designed and the procedures performed as per manufacturers instructions. Residue Tyr7









was mutated to a Phe, and residues Asn62 and Asn67 were separately mutated to Leu. All

mutations were verified as correct by sequencing the entire coding region of the HCA II

plasmid. Protein expression was performed in E. coli BL21(DE3) pLysS cells and the

resulting enzymes were purified using affinity chromatography (Khalifah, 1977). A

detailed description of these procedures can be found in the Materials and Methods

section of Chapter 2. Prior to any crystallization or activity assays the purity of the

protein was verified by electrophoresis on a 12% polyacrylamide gel stained with

Coomassie. The concentration of HCA II was determined by measuring the absorbance at

280 nm and using a molar absorptivity of 5.4 x 104 M-cm-1 (Coleman, 1967).

Crystallography

Crystals of mutant HCA II were obtained using the hanging drop method at room

temperature (McPherson, 1982). The first set of crystals of all three mutants at pH 8.2

were obtained by mixing 5 ul of protein (10-15 mg/ml concentrations in 50 mM Tris-

HC1, pH 7.8) with 5 [l of the precipitant solution (50 mM Tris-HCl, pH 8.2, 2.5-2.9 M

ammonium sulfate) against 1000 ld of the precipitant solution. For the N62L and N67L

structures determined at pH 6.0, previously grown crystals were soaked overnight in 50

mM sodium acetate, pH 6.0, 2.6 M ammonium sulfate. For Y7F at pH 10.0, previously

grown crystals were soaked overnight in 50 mM CAPS, pH 10.0, 2.6 M ammonium

sulfate. Y7F crystals were also grown using a different precipitant solution that consisted

of 100 mM Tris-Cl, pH 9.0, 1.3 M sodium citrate. Useful crystals under all conditions

appear within 7 days of crystallization set-up.

X-ray diffraction data was collected using a R-AXIS IV++ image plate system with

Osmic mirrors and a Rigaku HU-H3R Cu rotating anode operating at 50 kV and 100mA.

The detector to crystal distance was set to 100 mm. Each data set was collected at room









temperature from 1-3 crystals mounted in quartz capillaries. The oscillation steps were 1

with a 7 minute exposure time. Diffraction data of the Y7F crystals grown with sodium

citrate were collected at the Advanced Photon Source (APS) on beamline SER-CAT 22-

ID using a wavelength of 0.979 A. The crystal to detector distance was set at 120 mm

with a 1 second exposure time per image. Images were recorded on a Mar300 CCD with

1 oscillations. X-ray data processing was performed using DENZO and scaled and

reduced with SCALEPACK (Otwinowski and Minor, 1997). Structure determination and

refinement was carried out with Crystallographic and NMR System (CNS) version 1.1

(Brunger et al., 1998). All manual building was performed with Coot (Emsley and

Cowtan, 2005). The structure of wild type HCA II (PDB accession code 1TBT) was

isomorphous with all the data collected and was used to phase the data sets (Fisher et al.,

2005). To avoid phase bias of the model, the zinc atom, water molecules, and mutated

side chains were removed. After one cycle of rigid body refinement, annealing by heating

to 3000 K with gradual cooling, geometry restrained position refinement, and

temperature factor refinement, the Fo-Fo and 2Fo-Fc Fourier maps were maps generated.

Visual inspection of these maps clearly showed the electron density for the zinc atom and

the mutated side chains and these were subsequently incorporated into their respective

models. After several cycles of refinement, solvent molecules were incorporated into the

models using the automated water-picking program implemented in CNS until no more

waters were found at the 2.0 a level. Refinement of the models continued until the R-

factors converged. Tables 3-1 and 3-2 show the data set and final model statistics.

Kinetics and Activity Analysis

Determination, by membrane-inlet mass spectrometry, of Ri and RH2o was performed

using the 180 exchange method by Dr. Chingkuang Tu in the lab of Dr. Silverman









(Silverman, 1982). The methodology for obtaining these rate constants was described in

detail in the Materials and Methods section of Chapter 2 and will not be repeated here.

Initial rates of CO2 hydration were measured by following the change in absorbance of a

pH indicator on an Applied Photophysics (SX. 18MV) stopped-flow spectrophotometer

and these assay were also done by Dr. Chingkuang Tu in the Silverman lab. The pKa

values and wavelengths for the pH indicator-buffer pairs used to create pH profiles were

as follows: MES (pKa = 6.1) and chlorophenol red (pKa = 6.3), X = 574 nm; MOPS (pKa

= 7.2) and p-nitro phenol (pKa = 7.1), X = 401 nm; HEPES (pKa = 7.5) and phenol red

(pKa = 7.5), X = 557 nm; TAPS (pKa = 8.4) and m-cresol purple (pKa = 8.3), X = 578 nm;

CHES (pKa = 9.3) and thymol blue (pKa = 8.9), X = 596 nm. Final buffer concentrations

were 50 mM, and total ionic strength was kept at 0.2 M by the addition of sodium sulfate.

CO2 solutions were prepared by bubbling CO2 into water at 25 C with final

concentrations after mixing ranging from 0.7 17 mM. The mean initial rates at each pH

were determined from 5 to 8 reaction traces comprising the initial 10% of the reaction.

The uncatalyzed rates were determined in a similar manner and subtracted from the total

observed rates. Determination of the kinetic constants kcat and kcat/KM were carried out by

a nonlinear least-squares method (Enzfitter, Elsevier-Biosoft).

Results and Discussion

Structural Effects of Hydrophobic Mutations

All crystals were isomorphous and belonged to the space group P21 with the

following mean unit cell dimensions: a = 42.7 0.4 A, b = 41.6 0.5 A, c = 72.9 0.2

A, p = 104.6 0.50. The HCA II mutant data sets at pH 8.2 were processed to 1.65 1.70

A resolution, while the other pH data sets were processed to 1.8 A resolution. A summary

of the data set and final model statistics is given in Tables 3-1 and 3-2.









The N62L mutant structure determined at pH 8.2 shows that the water network is

conserved compared to wild type (Figure 3-1 and 3-2 (a)). However, as expected, the

hydrogen bond between Leu62 and W3b is lost. The side chain of His64, which is often

observed in a dual conformation in wild type structures, is completely in the "in"

position. Leu62 has moved away from the solvated area in the active site and now

occupies a hydrophobic region than also contains Leu60. The displacement of Leu62

removes the hydrophobic side chain away from the water network and as a result does

not affect it at all.

In contrast to N62L, the N67L mutant displays a disrupted water network compared

to wild type HCA II (Figure 3-1 and 3-2 (b)) and His64 appears to be all in the "out"

position. Unlike the orientation of Leu62 that has been rotated away from the active site

compared to an Asn at that position, Leu67 occupies a similar position to Asn67. It might

be the presence of this hydrophobic residue that disrupts the water network. The

differences in the His64 side chain orientations and water network should be due to

electrostatic changes in the active site and not steric effects as the effective size of the

side chains are similar.

In the Y7F mutant structure determined at pH 8.2, the water network appears

mainly conserved. However, due to the loss of the hydroxyl group by changing Tyr7 to a

Phe, the essential hydrogen bond to W3a is missing and as a result this water molecule is

not observed here anymore. Phe7 occupies a similar position compared to Tyr7 as seen in

wild type (Figure 3-1 and 3-3(a)). His64 in this mutant is also all in the "in" position and

appears to make a very weak hydrogen bond to W2.









An unusual observation is the presence of a sulfate ion bound to the zinc atom. It

has been observed that sulfate will only bind to wild type HCA II at very low pH (~ pH

5.1) and it seems that this mutation could be facilitating a higher affinity of the active site

for sulfate (Fisher et al., 2005).

From other structural studies of wild type HCA II it is known that the presence of

sulfate does not affect the water network but it is not known if the presence of sulfate

affects the orientation of His64 (Fisher et al., 2005). As the Y7F structure at pH 8.2

showed a sulfate bound and His64 is only in the "in" position, the pH of the other two

mutant crystals was lowered to pH 6.0 to get sulfate to bind.

These structure revealed that sulfate did bind at the lower pH and that it did not

affect the orientation of His64 as His64 was still either all in the "in" (as in the case of

N62L) or all in the "out" (as in the case of N67L) (Figure 3-2 (c) and (d)). The water

networks in both N62L and N67L structures at pH 6.0 were completely disrupted (Figure

3-2 (c) and (d)). Leu62 at pH 6.0 had moved compared to the structure at pH 8.2 and is

now pointing into the active site and might be a contributing factor in the water network

disruption. The side chain of Leu67 at pH 6.0 also moved compared to its structure at pH

8.2 (Figure 3-2 (c) and (d)). Apart from repelling solvent from these areas, it is not clear

if the movement of these two residues at lower pH is significant.

In an attempt to remove the sulfate in the Y7F mutant, the pH was increased to pH

10.0. Surprisingly, this structure showed that sulfate was still present even at this extreme

pH. However, His64 and the water network was undisturbed and appeared the same as

the structure at pH 8.2 (Figure 3-3 (a) and (b)). These observations rule out the idea that

the presence of sulfate affects the orientation of His64 as it is observed that His64 can









occupy the "in" conformation in the absence or presence of sulfate (Figure 3-2, compare

panels (a) and (b) with (c) and (d)). To further study this, it was necessary to see the

active site of Y7F in the absence of any sulfate. To this end, a different crystallization

condition using sodium citrate instead of ammonium sulfate was tried and made it

possible to determine the structure of Y7F at ph 9.0 in the absence of any sulfate (Figure

3-3 (c)). Except for the lack of sulfate it can be seen that His64 and the water network

appears the same as they did in the other two structures at pH 8.2 and 10.0 (Figure 3-3 (b)

and (c)). The observation of different orientations of His64 can be used to comment on

the relationship between this orientation and the efficiency of proton transfer. Inspection

of His64 in the more efficient wild type, Y7F, and N62L HCA II this residue occupies the

"in" conformation. This is in contrast with the less efficient N67L HCA II where His64

appears more in the "out" conformation. These observations alone do not provide

adequate evidence that His64 in the "in" position is necessary or sufficient for fast proton

transfer rates.

Kinetic Effects of Hydrophobic Mutations

The measurements of kcat/KM for the hydration of CO2 was obtained by both 180

exchange methods at chemical equilibrium and by stopped-flow methods at steady state.

As expected, these data shown in Table 3-3 indicate no significant changes between wild

type and the mutants, however, the values for N67L appear a bit lower than the others in

Table 3-3. This was found for both the maximal values of kcat/KM and for the apparent

pKa near 7.0 of the zinc-bound water calculated from the pH profile of kcat/KM (Figure 3-

4). This supports the structural data that there are no substantial changes in the active site

structures and implies no changes in the chemistry of catalyzed interconversion of CO2









and bicarbonate. It is not surprising that these mutations do not affect the first part of

catalysis as residues 7, 62, and 67 are more than 7 A away from the active site zinc.

Other studies involved mutating residues closer to the zinc, such as Thrl99 and

Glu 106, and the results show that those residues are essential for efficient interconversion

of CO2 and bicarbonate (Liang et al., 1993; Krebs et al., 1993). These observations

support the mechanism of the first stage of catalysis which involves a direct nucleophilic

attack by the zinc-bound hydroxide on C02, with Thrl99 and Glul06 enhancing the

nucleophilicity of the zinc-bound hydroxide and optimally orienting it for reaction with

CO2 (Merz, 1990).

In the second stage of catalysis long-range proton transfer occurs between the zinc-

bound solvent and the bulk solution using intervening water molecules and His64 as a

proton shuttle (Tu et al., 1989; Lindskog, 1997). In contrast to the first stage of catalysis,

as reflected by the values of kcat/KM (Figure 3-4), the mutants in this study all displayed

altered proton transfer rates when compared to wild type HCA II (Figure 3-5). The effect

is best seen when inspecting the pH profiles of RH2o/[E], the rate constant for the release

of isotopically labeled water from the enzyme to solution, which in turn depends on the

rate of proton transfer (Figure 3-5).

Proton transfer by N62L HCA II is the most complicated and difficult to interpret

as it displays two small bell-shaped curves and seems to have lost its pH dependence.

N67L HCA II shows a similar profile to wild type HCA II, just at a lower rate. The most

surprising result here is the data for Y7F HCA II as it has a rate constant for proton

transfer that is appreciably larger than for wild type (Table 3-3 and Figure 3-5).









Addition of an exogenous activator of proton transfer activity, 4-methylimidazole

(4-MI), does not appear to appreciably increase the rate of N62L or N67L HCA II (Figure

3-6). Wild type HCA II is activated, compared to N62L and N67L HCA II, with maximal

activation occurring at around 50 mM 4-MI. Y7F HCA II is activated to a larger extent

compared to wild type with the addition of 4-MI (Figure 3-6). These data implies that the

active site of Y7F HCA II is more accessible to external buffers and that the removal of a

hydroxyl group, by mutating Tyr7 to a Phe, thus enhancing the activation by 4-MI over

that observed for wild type, N62L, and N67L HCA II.

Solvent Structure and Implications for Proton Transfer

In crystal structures of wild type HCA II, each of the side chains Tyr7, Asn62, and

Asn67 appear to make hydrogen bonds with water molecules W3a and W3b (Figure 3-1;

Fisher et al., 2005). In fact, the side chains of Asn62 and Asn67 interact with the same

water molecule, W3b. This water structure in N62L at pH 8.2 seems unaffected by the

mutation as just the hydrogen bond to W3b is lost but the water is still present due to its

interaction with Asn67. The reason why Leu62 does not interfere with the water is that it

has moved away from the solvated active site into a more hydrophobic region also

occupied by Leu60. Overall, the water structure seems conserved with that observed in

wild type HCA II. However, when the pH was lowered to pH 6.0 this residue moved

back into the active site and subsequently completely disrupted the solvent structure

(Figure 3-2 (c)). The data for N62L, showing His64 in the "in" conformation and an

intact solvent structure at pH 8.2, would suggest an unchanged proton transfer rate. Yet,

probably due to electrostatic effects, the side chain of Leu62 and the water network is

easily disrupted upon a change in pH and this instability of the solvent structure could

account for the slower measured proton transfer rate compared to wild type HCA II.









In contrast to N62L, Leu67 in the N67L HCA II mutant extends into the active site

and effectively disrupts the ordered water structure at both pH 6.0 and 8.2 (Figure 3-2 (b)

and (d)). The only water molecule that is conserved is the zinc-bound solvent that could

be either a hydroxide or a water molecule. The other two waters have been displaced and

do not make contact with any active site residues shown here. Also, as mentioned above,

His64 is observed to be always in the "out" position, regardless of the presence of sulfate

bound to the zinc. Proton transfer by this mutant displays the characteristic bell-shaped

pH dependency, just at a lower rate compared to wild type HCA II (Figure 3-5). Due to

these observations it is likely that the typical pH dependency of proton transfer does not

correlate with an ordered solvent structure in the HCA II active site.

The Y7F HCA II mutant shows the most efficient proton transfer, even when

compared to wild type HCA II. At pH 8.2, 9.0, and 10.0 the solvent structure is highly

conserved and His64 is always in the "in" conformation. The net effect of the loss of

W3a, due to the Tyr7 to Phe mutation, produces a single-line array of water molecules

that bridges the zinc-bound solvent to His64.

It has been suggested by others that a single array of waters is much more efficient

than a branched system at facilitating proton transfer. Also, an unbranched hydrogen-

bonded array of water can promote most efficient proton transfer by a concerted

mechanism rather than a step-wise and appears to proceed through an intermediate with

partial hydronium ion character (Cui and Karplus, 2003; Voth et al., 1998). There are

various explanations for the very fast movement of protons in solution and these include

Grotthuss's idea of so-called structural diffusion. Other structural models for the proton

in solution, or hydrated proton, includes the elementary proton (H ), hydronium ion









(H30 ), the Zundel, and Eigen cation models (Eigen, 1964; Marx et al., 1999). The

Zundel cation is a H502 complex where the proton is shared between two water

molecules. The Eigen cation consists of a H904 complex where the central H30+ is

strongly hydrogen bonded to three H20 molecules in a secondary hydration configuration

(Eigen, 1964; Marx et al., 1999). According to Eigen, weaker hydrogen bonds are formed

at the periphery of the H904 complex and that directed breaking and formation of these

bonds lead to structural diffusion of the entire complex. The diffusion of the hydrate

complex is the rate-limiting step in the mobility of a proton in solution (Eigen, 1964).

Recent studies with infrared spectroscopy of acid-base proton transfer reactions

suggest the existence of a fairly stable H30 intermediate structure that is coordinated in

an Eigen configuration by three water molecules. Other conclusions were that the proton

transfer between acid and bases occur by a sequential, Grotthus-type proton hopping

mechanism that is mediated by hydrogen bonded water networks or bridges (Mohammed

et al., 2005).

The highly conserved core water structure as seen in wild type HCA II consists of

Wl, W2, W3a, and W3b (Figure 3-7, green box). It shares many structural features of an

Eigen cation in that the geometry of the structure is almost planar with -1200 between

W2 and the surrounding waters. Also, the hydrogen bond distances between W2 and the

others is -2.8 A for all three. Based on the structural properties of the solvent core it

could be an Eigen cation with W2 representing the excess hydrated proton as a H30 The

Eigen cation is relatively stable entity and can be observed on the nano- to picosecond

scale (Mohammed et al., 2005). This core structure is a somewhat branched water

network in that the excess proton can get access to His64 through either one of W2, W3a,









or W3b. The X-ray crystal structure represents an energy minimized, stable structure and

perhaps this solvent structure kinetically traps the excess proton. It is likely that in

solution during a normal catalytic cycle, these solvent structures are constantly being

broken and reformed and the crystal structure just shows the most energetically stable

species. In the Y7F HCA II mutant, W3a has been removed by the introduction of the

hydrophobic Phe residue. This results in a linear array with only W2 having access to

His64 and could also disrupt the stability of the Eigen cation as seen in wild type HCA II.

This could serve as a structural explanation for why the proton transfer proceeds so much

faster in this mutant compared to wild type HCA II. Even though this mutant is faster

than wild type, natural selection did not choose to have a Phe at position 7 and a possible

reason is that it appears to be somewhat unstable, especially at low pH (Figure 3-5).

Additionally, in contrast to wild type HCA II, where the distance between W2 and His64

is too long (- 3.5 A), the distance between W2 and His64 is now ~ 3.2 A and constitutes

a weak hydrogen bond and this could contribute to the enhanced proton transfer rates.

Conclusion

To disrupt the hydrogen bonded water network in the active site of HCA II, several

key catalytic amino acids, involved with coordinating these waters, were mutated to

hydrophobic residues that are similar in size. Asn62 and Asn67 were mutated to Leu

(N62L and N67L) and Tyr7 was mutated to a Phe (Y7F). X-ray crystal structures and rate

constants for the hydration of CO2 and proton transfer were determined for all three

mutants at different pHs. The structural and kinetic data for N62L and N67L shows that

the water networks were readily disrupted, especially at low pH, and both displayed

considerably low proton transfer rates compared to wild type and Y7F HCA II. The most

surprising result was the enhanced proton transfer rate over that of wild type observed for









Y7F HCA II. This mutant also displayed a similar water network as wild type HCA II,

except for the loss of one active site water. To examine whether the presence or absence

of sulfate affected the orientation of His64, structures of the mutants were determined at

several different pHs. Lowering the pH for N62L and N67L HCA II caused sulfate to

bind and it was determined that it had no effect on the orientation of His64. The structure

of Y7F HCA II at pH 8.2 and 10.0 showed that sulfate was still bound. Subsequently,

Y7F HCA II was crystallized in the absence of any ammonium sulfate and this too

showed that sulfate had no effect on His64. Correlations between the structural and

kinetic data suggest that a single, linear array of water bridging His64 and the zinc-bound

solvent might be more efficient at proton transfer than a branched structure.












Table 3-1. Data set and final model statistics for N62L HCA II and N67L HCA II.
HCA II N62L N62L N67L N67L
pH 6.0 pH 8.2 pH 6.0 pH 8.2
Resolution (A) 20.0 -1.80 20.0 -1.70 20.0- 1.80 20.0 -1.65
(1.86- 1.80)* (1.76- 1.70) (1.86- 1.80) (1.71- 1.65)
Total Number of 20541 (2031) 25698 (2422) 21266 (2142) 27340 (2590)
Unique Reflections
Completeness (%) 88.2 (86.9) 91.0 (87.4) 91.7 (92.6) 93.1 (88.9)

Redundancy 4.1(4.0) 3.2 (3.3) 3.4(3.2) 3.1(3.1)

Rsymm 0.106 (0.339) 0.085 (0.448) 0.075 (0.349) 10.8 (28.6)
Rcryst / Rwork 0.189 / 0.226 0.179 / 0.217 0.187 / 0.224 18.6 / 20.1
Ave B-factor (A2) 16.8 / 20.1 /28.7 16.1 / 19.5 / 29.6 18.6 / 22.2 / 29.5 18.2 / 21.4 /29.6
Main/side/solvent
Number of solvent 108 132 97 116
Data in parenthesis are for the highest resolution shell.
,Rsymm I I- /
Rcryst |Fo| |Fe| / I Fo|
Rfree is calculated the same as Rcryst except it is for data omitted from refinement (5% of
reflections for all data sets).













Table 3-2. Data set and final model statistics for Y7F HCA II.


HCA TT


Resolution (A)

Total Number of Unique
Reflections
Completeness (%)

Redundancy

Rsymm
Rcryst: / Rwork
Ave B-factor (A2)
Main/side/solvent
Number of solvent


Y7F
pH 8.2
20.0- 1.70
(1.76- 1.70)*
25613 (2444)

92.7 (89.8)

3.9 (3.9)

0.060 (0.322)
0.181 / 0.199
16.6 / 20.1 / 29.4

115


Y7F
pH 10.0
20.0 -1.80
(1.86- 1.80)
20874 (1951)

89.4 (83.9)

3.1 (3.2)

0.071 (0.238)
0.179 / 0.200
15.1 / 18.5 /28.3

115


Y7F (citrate)
pH 9.0
20.0- 1.15
(1.19- 1.15)
85073 (8284)

99.1 (96.8)

4.2 (3.2)

0.079 (0.364)
0.193 / 0.184
10.5 / 13.5 / 23.4

270


* Data in parenthesis are for the highest resolution shell.
tRsymm I /
SRryst X |Fo| |Fe| / X Fo|
Rfree is calculated the same as Rcryst except it is for data omitted from refinement (5% of
reflections for all data sets).






67




Table 3-3. Maximal values of rate constants for hydration of C02, proton transfer, and
pKa of the zinc-bound water.
kcat/KM (M-1^i)a pKa RH20 (ts-1)
Wild type 120 6.9 0.1 0.3
Y7F 120 7.1 0.1 1.0
N62L 140 7.3 0.1 0.1
N67L 90 6.5 0.2 0.1
aThe standard errors for kcat/KM are 10% or less.















T2 00ZnOHAI1,O


Y7F


W3a s------
"out" "in"



His64


N62


Figure 3-1. Active site of wild type human carbonic anhydrase II. The zinc atom is shown
as a black sphere and active site residues are in yellow ball-and-stick and are
as labeled. Water molecules are shown as red spheres and are as numbered.
Figure was generated with BobScript and rendered with Raster3D (Merritt and
Bacon, 1997; Esnouf, 1997).


His94


N67


T199


T200













T199


T-2 0


ZnOH-iH.O


-U


H64

N62 /
T

d) T200





H64 /

N62 /


199
4 sulfate




H1196

lH96
L67


Figure 3-2. Active site of N62L and N67L at pH 8.2 and pH 6.0. (a) N62L HCA II at pH
8.2; (b) N67L HCA II at pH 8.2; (c) N62L at pH 6.0; (d) N67L HCA II at pH
6.0. Active site residues are shown in yellow ball-and-stick and are as labeled.
The black sphere is the zinc atom and waters are shown as red spheres. Sulfate
ions are in green ball-and-stick and are labeled. Figure was generated and
rendered with Bob Script and Raster3D, respectively (Esnouf, 1997; Merritt
and Bacon, 1997).


T199
T200


H64 5

L62 /


T199


T20


Y 7











a) T199
T200 N': sulfate


H64 H96
H64b..

N62/ N67
b) T199
T200 sulfate

"""4 W2*n, Tj_

H64r W3b- H96

N62/





N62/ N67






Figure 3-3. Active site of Y7F HCA II at various pH. (a) Y7F HCA II at pH 8.2; (b) Y7F
HCA II at pH 10.0; (c) Y7F HCA II at pH 9.0 with no sulfate. Active site
residues are shown in yellow ball-and-stick and are as labeled. The zinc atom
and water molecules are shown as black and red spheres, respectively. Figure
was generated with Bob Script and rendered with Raster3D (Esnouf, 1997;
Merritt and Bacon, 1997).












1.E+09



1.E+08- -








1.E+06 "



L.E+05
5 6 7 8 9
pH




Figure 3-4. The pH profiles for kcat/KM catalyzed by wild type and mutant HCA II. (*)
Wild type, (m) N67L, (o) N62L, (A) Y7F HCA II. Data were obtained at 250
in the absence of exogenous buffers using a total concentration of all species
of CO2 of 25 mM, with the ionic strength maintained at 0.2 M by addition of
sodium sulfate. The solid lines are fit to a single ionization with the pKa given
in Table 3-3.












1,E+06







o 1E+05







1LE+04


5 6


Figure 3-5. The pH profiles for RH20/[E] catalyzed by wild type and mutant HCA II. ()
Wild type, (m) N67L, (o) N62L, (A) Y7F HCA II. Data were obtained at 250
in the absence of exogenous buffers using a total concentration of all species
of CO2 of 25 mM, with the ionic strength maintained at 0.2 M by addition of
sodium sulfate.











4.E+05



3.E+05








1.E+05



O.E+00
0 50 100 150 200
[4-methylimidazole] mM



Figure 3-6. Activation of RH20/[E] catalyzed by wild type and mutant HCA II by the
addition of 4-methylimidazole. (*) Wild type, (m) N67L, (o) N62L, (A) Y7F
HCA II. Data were obtained at 250 by the 180 exchange method using a total
concentration of all species of CO2 of 25 mM and adding increasing amounts
of 4-methylimidazole, with the ionic strength maintained at 0.2 M by addition
of sodium sulfate.












a) TI99 b) T199

T200 .Zn-OH- T200 n-OHIHa













Figure 3-7. Active sites of wild type and Y7F HCA II. (a) Wild type, and (b) Y7F HCA
II. Residues are shown in yellow ball-and-stick with the zinc atom as a black
sphere. Water molecules and inferred hydrogen bond are shown as red spheres
and dashed lines, respectively. The green boxes surround the core solvent
structure in the active sites. Figure was generated with BobScript and rendered
with Raster3D (Esnouf, 1997; Merritt and Bacon, 1997).
with Raster3D (Esnouf, 1997; Merritt and Bacon, 1997).














CHAPTER 4
WORKING TOWARDS A NEUTRON STRUCTURE OF PERDEUTERATED
HUMAN CARBONIC ANHYDRASE II

Chapters 2 and 3 included work towards a better understanding of the proton

transfer mechanism that is part of the catalysis by human carbonic anhydrase II (HCA II).

The methods employed included kinetic measurements of rate constants for proton

transfer as well as detailed structural studies of the active site residues and water

structure. However, X-ray crystallography is not the best technique for investigating

waters because hydrogen atoms, which make up about half of all the atoms in a protein,

are virtually invisible to X-rays. Neutron macromolecular crystallography is currently the

only direct technique for observing hydrogen H (or deuterium D) atoms and can give

atomic details even at modest resolutions (~ 2.0 A). This chapter will describe how and

why neutrons are different from X-rays and will include recent progress towards

obtaining a neutron structure of perdeuterated HCA II.

Introduction

Neutron crystallography can provide unique information about hydration states of

proteins, ionization states of key catalytic amino acids, water molecules, and position of

H atoms. H atoms are a fundamental part of many enzymatic processes, some of which

involve proton transfer between residues in the protein and substrates, products, ligands,

as well as mediating the binding, through water molecules, of pharmacological agents

(Langan et al., 2004).









Assigning positions of hydrogen atoms is possible in X-ray crystallography if

atomic or sub-atomic resolution data can be collected (< 1.2 A). Even in the case of high

resolution X-ray diffraction it is often still hard or impossible to confidently assign the

positions of hydrogen H atoms associated with water molecules. H atoms that are

associated with well-ordered parts of the protein (such as the main or side chains) can be

assigned, but H atoms in disordered regions or part of water molecules are much harder

to assign (Coates et al., 2001; Habash et al., 2000).

The extent of diffraction by X-rays depends on the number of electrons. As H and

D have only 1 electron, they diffract X-rays very poorly compared to more electron-rich

atoms such as carbon C, nitrogen N, oxygen 0, and sulfur S. The scattering amplitude of

atoms by X-rays increase linearly with increasing numbers of electrons (Table 4-1). This

is in contrast to how atoms behave in an neutron beam as each atom has a unique

scattering amplitude that is a property of the nucleus and has to be determined

experimentally (Table 4-1).

H and/or D atoms are more easily located by neutron analysis as the scattering

lengths of H (-3.7 x 10-15 m, or -3.7 fm) is equal in magnitude but opposite in sign when

compared to other atoms found in proteins. D atoms have a scattering length (6.7 fm) that

is very close to the range of O (5.8 fm), N (9.4 fmn), C (6.6 fm) and S (3.1 fmn) making it

easier to locate by this method (Coates et al., 2001). The real strength of using neutrons

over X-rays are the different scattering properties of H and D atoms. The negative

scattering amplitude of H with neutrons can potentially be exploited to observe them by

just looking at negative nuclear density. This kind of contrast labeling has been

successfully used to determine levels of H/D exchange and using the data for further









elucidation of the mechanism or activity of enzymes (Habash et al., 1997). A classic

example is the study of a carbon monoxide myoglobin derivative that was subject to H/D

exchange prior to neutron diffraction. The results gave information on the hydration

shells around the protein as well as region in the heme binding site that did not exchange

(Norvell et al., 1975). However, there are two problems with having H in proteins

crystals for diffraction experiments. The first is the low coherent scattering cross section

of H atoms and the second is the large incoherent scattering cross section that leads to a

very large undesirable background (Table 4-1).

The heavy isotope of H, deuterium D, behaves the same with X-rays but have very

different properties with neutrons. D atoms have a scattering amplitude similar to the

other atoms found in proteins and also a small incoherent scattering cross section (Table

4-1). For these reasons, it is often most favorable to replace as many H atoms with D

atoms as is practical by either H/D exchange methods, such as soaking crystals in

deuterated solutions, or by producing deuterated materials. As the H/D exchange process

does not lead to fully deuterated (perdeuterated) materials, most often it is better to

synthesize or purchase perdeuterated materials. Perdeuteration, as opposed to just soaking

crystals in deuterated solutions, and subsequent neutron diffraction data collection vastly

improves location of D atoms because, as mentioned above, D and 0 have similar

scattering lengths and both are positive in sign. The resulting nuclear density indicates the

orientation of D20 and thus allows the location of the two D atoms (Coates et al., 2001;

Habash et al., 2000; Myles et al., 1998). In the case of H atoms in water molecules the

negative density from the H can smear out or cancel the positive scattering contribution

from the 0 atoms (Habash et al., 2000). Another bonus to using perdeuterated materials









are estimates that, due to the small incoherent scattering cross section of D atoms, one

can achieve at least a 40-fold reduction in the background compared to having any H

atoms around. Also, having perdeuterated materials cuts back on the need for very large

crystals and it is now possible to collect data from crystals that are only 0.15 mm3, as in

the case of aldose reductase (Niimura et al., 2005).

Neutron macromolecular crystallography is a powerful technique that complements

high resolution X-ray crystallography well as a combination of the two techniques allows

analysis of key hydrogen atom positions and solvent structure (Myles et al., 1998).

Before collecting neutron diffraction data, crystals are usually subjected to H/D exchange

by soaking the crystals in deuterated solutions. Some of the advantages to this procedure

include the fact that very high incoherent background scattering from hydrogen is

eliminated and that it is a lot cheaper than purchasing deuterated materials. Soaking also

allows the exchange of solvent accessible hydroxyl and amide groups as well as replacing

H20 with D20 and this information can be used assess solvent accessibility, flexibility or

disorder of a protein (Niimura et al., 2005).

The major drawback of neutron diffraction is the low flux of neutrons available for

sample irradiation. The diffraction intensity can be calculated from eq 4-1:

I = (Io x F2 xVxA) / (vo)2 (4-1)

In eq 4-1, I is the diffraction intensity, Io incident neutron intensity, F2 structure

factor, V volume of the crystal, A detector area covered by sample, and vo is the volume

of the unit cell (Niimura, 1999). This means that the diffraction intensity strongly

depends on the size of the crystal, intensity of the incident neutron beam, and the area

detector. Recent advances in area detector technology, data collection, and processing









strategies in neutron protein crystallography allow studies of larger biological molecules

and smaller crystals than previously thought possible. The biggest challenge is to grow a

very large, single crystal (~ 1 mm3) with a small unit cell volume (Niimura, 1999). A data

collection strategy that is used to optimize the number of neutrons available, is Laue

diffractometry where a range of wavelengths are used. This is in contrast to a

monochromatized beam, as used in most X-ray diffraction experiments, where most of

the useful neutrons are removed and only a single wavelength is used. Due to the low

flux of the beam, this is a wasteful methodology and Laue diffraction allows the user to

simultaneously measure diffraction in different directions from different lattice planes.

Laue methods allow for a more efficient survey of reciprocal space with fewer crystal

settings, and along with large area detectors effectively increases the flux on the sample

(Myles et al., 1998; Schultz et al., 2005).

The single biggest advance in neutron protein crystallography, has been the design

of better area detectors. Neutron imaging plate (NIP) technology has been developed in

which a neutron converter, such as 6Li or Gd, is mixed with photostimulated

luminescence materials that are layered on a flexible plastic backing. The dynamic range,

spatial resolution, and flexibility make the NIP suitable for detecting diffracted neutrons.

NIPs are available in different sizes (200 mm x 200 mm and 200 mm x 400 mm) and

these can be arranged side-by-side to make a combined detector of any desired size

(Niimura, 1999). The neutron-sensitive image plate in use at the Institut Laue-Langevin

(European Molecular Biology Laboratories, Grenoble) consists of 4 large Gd-doped

phosphor plates that are packed together to give an active area of 400 x 800 mm (Myles,

1998). Image plates have high counting rates and good dynamic range, but they cannot









relay real-time information and do not have timing characteristics to determine a

neutron's wavelength.

This is crucial for spallation sources where a spectrum of neutron wavelengths are

used and time-of-flight is determined in order to resolve which neutrons gave rise to

which reflections. For these applications, an advanced cylindrical detector was built by

Brookhaven National Laboratory and is in use at the Protein Crystallography Station at

Los Alamos Neutron Science Center (LANSCE). It has a height of 200 mm and a curved

horizontal dimension that subtends 1200 at the sample position. This detector is filled

with 3He that decay into a charged pairs upon neutron absorption. This induced charge is

detected by a two-dimensional multiwire array and is used to determine the x and y

position of each event (Langan et al., 2004). Figure 4-1 shows a photograph of the

detector in use at LANSCE.

There are very few examples in the literature of neutron protein structures and these

include lysozyme, endothiapepsin, xylose isomerase, and aldose reductase. For all these

projects, the investigators were unable to obtain pertinent structural information about

catalysis using ultra-high resolution X-ray structure alone. Using neutron diffraction, the

functionally important H atoms were identified and led to elucidation of enzyme

mechanism (Langan et al., 2004). However, all of these neutron structures were

determined from protein crystals that were subject to H/D exchange. The only

perdeuterated neutron structure so far is of myoglobin and was reported in recent years

(Shu et al., 2000; Niimura et al., 2005).

As discussed in Chapter 1-3, the rate-limiting step in catalysis of HCA II is the

intermolecular transfer of a proton from the zinc-bound solvent (H20/OH-) to the proton









shuttling residue, His64. This distance (~ 7.5 A) is spanned by well-defined solvent

molecules that are connected to each other and several critical side chains via a hydrogen-

bonded network (Christianson and Fierke, 1996; Lindskog, 1997). Despite the

availability of high-resolution crystal structures of HCA II to 1.05 A, there is currently no

definitive information available on the absolute positions and orientations of H atoms

from either the solvent network or the ionization state of active site residues (Duda et al.,

2003). As mentioned above, it is very hard to directly observe H atoms even in high-

resolution X-ray crystal structures and neutron diffraction studies of perdeuterated

crystals can be a powerful complementary technique to elucidate proton donors/acceptors

and ionization states in macromolecules. Key questions that need to be answered include

which solvent molecules are H20 or OH- molecules and which residues are proton donors

or acceptors. Another controversial topic is the nature of the zinc-bound solvent in terms

of whether it is a OH- or H20 in the crystal structure. Even in sub-atomic resolution

structures of HCA II there have been no conclusive answers to any of these questions.

Despite all the advantages to neutron protein crystallography, there are very few

neutron studies compared to X-ray. The main reasons are that there are few sources

around the world and the available neutron beams have low flux, compared to

synchrotron sources for X-rays. There is also no foreseeable way to increase the flux of

neutrons from nuclear reactors due to the inherent limitations of the fission reaction that

produces them. In contrast to reactor sources, spallation sources produce neutrons by the

bombardment of a heavy metal target, such as mercury or tungsten, with pulsed high-

energy protons. The main advantage, besides the higher flux attained, is that the neutrons

produced in this way have a time-of-flight component so that fast or high energy neutrons









arrive at the detector before the slower or low energy neutrons. This allows the energy

and wavelength of each neutron to be calculated and this information can be applied to

the structure determination and make the process more efficient.

Structure and catalytic activity analysis and comparison of perdeuterated HCA II

with hydrogenated HCA II shows that these structures are highly isomorphous and active,

especially with regards to the active site architecture and the solvent network, and this

indicates that using perdeuterated HCA II is appropriate for neutron macromolecular

crystallography. This work lays the foundation for planned future neutron structure

determination and structural analysis.

Materials and Methods

Production and Crystallization of Perdeuterated HCA II

Transformed Escherichia coli (E. coli) BL21 pLysS(DE3) cells were plated out on

Luria broth agar plates supplemented with ImM ampicillin. The plates were incubated

overnight at 37 C. The next day, an appropriate colony was selected and placed into 15

mL Spectra 9-d deuterated minimal media supplemented with ImM ampicillin made in

D20. This initial culture was grown for 24 hours and then the entire 15 mL was used to

inoculate 150 mL Spectra 9-d minimal media supplemented with 1 mM ampicillin made

in D20. After 18 hours, the cells were harvested by centrifugation and the resulting

pellets were resuspended in 5 mL of Spectra 9-d media. The cells were then placed in 800

mL of fresh Spectra 9-d media supplemented with 1 mM ampicillin in D20. The large

scale culture was then grown for 2-3 hours and protein expression was induced by the

addition of 0.6 mM isopropyl-P-D-1-thiogalactopyranoside in D20 and incubated for 6

hours at 37 C while shaking at 220 rpm. Cells were harvested by centrifugation and

stored at -20 C. Upon freeze/thawing, the cell lysates were processed and the protein









purified as previously described (Khalifah et al., 1977). After purification, the protein

was rapidly back-exchanged into deuterated buffers and concentrated. Purity of

perdeutereated HCA II was determined by SDS-PAGE.

The deuteration level of the protein sample was determined by time-of-flight

electrospray mass spectrometry. The software Isotopic Pattern Calculator v.1.6.5 for

Macintosh (http://www.shef.ac.uk/chemistry/chemputer/isotopes.html) was used to

calculate the theoretical isotopic distribution and the pattern was then matched to the

experimental spectrum.

Crystals of perdeuterated HCA II were obtained by the hanging drop vapor

diffusion method at room temperature (McPherson, 1982). Crystallization drops were

prepared by mixing 5 [tL of concentrated protein solution (10-15 mg/mL) with 5 [tL

precipitant solution consisting of 2.6 M ammonium sulfate, 50 mM Tris-Dl (pD 8.0)

made in D20. Useful crystals appeared within 7 days of crystallization set-up.

Crystallography

Crystals were cryoprotected prior to data collection by quick-dipping them in 30%

glycerol in mother liquor. The crystals were then quick frozen to 100 K in a N2-gas

stream on the beamline. Synchrotron diffraction data were collected at the ESRF

beamline ID29. Several crystals were used during data collection and the crystal to

detector distance was set at 105, 125, and 185 mm. X-ray data processing was performed

using DENZO and scaled and reduced with SCALEPACK (Otwinowski and Minor,

1997). Several data sets were collected and three were ultimately processed and scaled

together from 20.0 1.5 A for a total of 220 degrees of data. All manual model building

was done with Coot and model refinement was carried out using CNS version 1.1

(Emsley and Cowtan, 2004; Brunger et al., 1998). The wild type structure of HCA II









(PDB accession code 1TBT; Fisher et al., 2005) was isomorphous with the perdeuterated

HCA II data and was subsequently used to phase the experimental data. In order to avoid

phase bias of the model, the zinc atom and water molecules were removed from the

phasing model. After one cycle of rigid body refinement, annealing by heating to 3000 K

with gradual cooling, geometry-restrained position, and individual temperature factor

refinement, Fo-Fc and 2Fo-Fc electron density maps were generated. These maps clearly

showed the position of the zinc atom which was then included in the model and

subsequent refinements. After several further cycles of refinement, water molecules were

incorporated into the model using the automated water picking program, as implemented

in CNS 1.1, until no more waters were found at the 2.0 a level (Bringer et al., 1998).

Manual model building in Coot and refinement continued until convergence of Rfree and

Rwork was reached (Emsley and Cowtan, 2004). Table 4-2 contains the data collection

and model refinement statistics.

Activity Analysis by 80 Exchange Methods

The rate constants for the hydration and proton transfer reactions were determined

by Dr. CK Tu in the Silverman lab using 180 exchange methods. All assays were

performed as detailed in Chapter 2 except it was all done under deuterated conditions.

Hydrogenated and perdeuterated HCA II were exchanged into D20 and incubated for 3

hours prior to the assays.

Results and Discussion

Despite a small lag in cell growth, the transformed cells adapted readily to

deuterated minimal media conditions and overexpression of HCA II with a yield of 30

mg/mL pure protein per liter of cells was achieved. Mass spectrometry of the purified

perdeuterated HCA II showed that over 98% deuteration was obtained. By varying the









protein to precipitant ratio, optimal conditions for growing large (0.2 x 0.2 x 1.0 mm)

perdeuterated crystals within a week were determined (Figure 4-2).

Structural Effects of Perdeuteration

The perdeuterated HCA II crystals diffracted to 1.5 A and belonged to the

monoclinic space group P21, with unit cell parameters: a = 42.1, b = 41.0, c = 72.0 A, =

104.40. The scaling Rmerge was 0.090 and isomorphous with hydrogenated wild type HCA

II crystals (Fisher et al., 2005). All data collection and refinement statistics are given in

Table 4-2. The final Rcryst and Rfree to 1.5 A resolution were 19.5 and 20.6 %,

respectively. The model was of good quality with root mean square deviations (rmsd) for

bond lengths and angles of 0.005 A and 1.40, respectively.

All final model statistics are given in Table 4-2. A superposition of hydrogenated

HCA II (PDB accession code 1TBT; Fisher et al., 2005) onto the structure of

perdeuterated HCA II gave a rmsd of only 0.2 A for all Ca atoms (Figure 4-3).

Visual inspection of the backbone representations for hydrogenated and

perdeuterated HCA II clearly shows that there is no appreciable difference between the

two and this shows that deuteration has minimal effect on overall fold and three

dimensional structure of HCA II (Figure 4-3).

When only the active site residues (Tyr7, Asn62, Asn67, His64, His94, His96,

Hisi 19, Thr199, and Thr200) and solvent molecules (Zn-H20/OH-, WI, W2, W3a, and

W3b) are superposed, the rmsd decreases to less than 0.1 A. See Figure 4-4 (a) and (b)

for a visual comparison of the active site architecture. The relative positions and distances

between the solvent molecules and the active site residues that coordinate them are very

similar, within experimental error, and are shown in Table 4-3 and Figure 4-4.