KINETICS AND STRUCTURE OF PROTON TRANSFER PATHWAYS IN CARBONIC ANHYDRASE

Transcription

1 KINETICS AND STRUCTURE OF PROTON TRANSFER PATHWAYS IN CARBONIC ANHYDRASE By ROSE LYNN MIKULSKI 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

2 2010 Rose Lynn Mikulski 2

3 To Christopher, Geralyn, Jasmine, Michael, and Marigold 3

4 ACKNOWLEDGMENTS I thank my mentor Dr. David Silverman for his endless support of my pursuits over these year. Dr. Silverman exposed me to the many facets of academia and allowed me to explore my interests ensuring that I have the confidence in myself as a scientist to pursue the career of my choice. I thank my co-chair Dr. Robert McKenna for his mentoring and enduring enthusiasm. I would also like to thank my other committee members, Dr. William Kem and Dr. Susan Frost. My committee supported me with honest critiques and challenging inquisitiveness to direct a successful path to my PhD. I also sincerely thank Dr. Chingkuang Tu for his patience and sharing his tremendous knowledge of enzyme kinetics, and general biochemistry me with. I would like to thank Dr. Mavis Agbandje-McKenna, Deepa Bhatt, Patrick Quint, Zoe Fisher, and John Domsic for all their assistance, and advice. The graduate students Nicolette Case, Katherine Sipple, Balu Avaur, Dayne West, Balasubramanian Venkatakrishnan, Mayank Aggarwal, Ha-Long Mguyen, Joannalyn Delacruz, Erin Mack-Humphrey, and Karlie Bonstaff have been sounding boards, sources of advice and expertise and I thank them for being my friends. I would also like to thank Wayne McCormack for improving the program and being available to talk during the successes and struggles. I thank the ladies who run the pharmacology department for taking care of the bureaucratic ins and outs for me. Lastly, I would like to thank my family to whom this is dedicated. My younger siblings have been a source of escape from the stress, words of encouragement, and the best medicine in the world laughter. I thank my mother and father for fostering my curious nature, helping me to examine scientifically the world around me from a young age, and for inspiring me to succeed and do what I love. 4

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS... 4 LIST OF TABLES... 7 LIST OF FIGURES... 8 LIST OF ABBREVIATIONS ABSTRACT CHAPTER 1 INTRODUCTION The Carbonic Anhdrases Catalytic Mechanism Structure Solvated Active Site Proton Shuttle Residue Extended Active Site Speeding Up Proton Transfer Physiological and Medical Significance KINETICS OF SPEEDING UP PROTON TRANSFER IN HUMAN CARBONIC ANHYDRASE II Extended Active Site Hydrophilic Residues Faster Rates of Proton Transfer Methods Expression and Purification of Enzymes Kinetics O Exchange Esterase Activity Results Discussion STRUCTURE OF PROTON TRANSFER PATHWAYS IN HUMAN CARBONIC ANHYDRASE II Structural Role of Water Network in Proton Transfer Proton Shuttle Residue Methods X-Ray Crystallography Thermodynamic Stability

6 Results Structure of Y7I HCA II Role of Tyr7 in Thermal Stability of HCA II Structures of N62Q, N67Q and Y7F+N67Q HCA II Discussion PHYSIOLOGIAL ROLE OF HUMAN RED BLOOD CELLS IN THE GENERATION OF NITRIC OXIDE Generation of NO Catalyzed by CA Deoxy-Hemoglobin Mechanism of Nitrite Reductase Methods Materials Inlet probe and Kinetic Measurements Results NO Generation from CA Deoxy-Hemoglobin Catalyzed Generation of NO Human Red Blood Cell Suspensions Generation of NO from Nitrite Effect of EAZ on NO Accumulation Effect of DIDS on NO Accumulation Effect of PCMBS on NO Accumulation Discussion Accumulation of Extracellular NO in Red Cell Suspensions Inhibition of Band 3 Anion Exchanger of Red Cells Inhibition of Aquaporin-1 Channel of Red Cells CONCLUSIONS AND FUTURE DIRECTIONS LIST OF REFERENCES BIOGRAPHICAL SKETCH

7 LIST OF TABLES Table page 2-1 Kinetic rate constants for HCA II and mutants Values of apparent pk a obtained by various kinetic measurements of catalysis in HCA II and mutants Dataset, refinement, and final model statistics for the crystallographic study of N62Q, N67Q HCA II, N67Q+Y7F HCA II, and Y7I HCA II Thermodynamics of unfolding of wild type and Y7 variants of HCA II. a Calorimetric parameters determined by DSC Comparison of proposed hydrogen bond networks for wild type, Y7F, Y7F+N67Q, N67Q, N62Q, and Y7I HCA II Comparison of the features proposed to regulate proton transfer rates in the protein environment in HCA II and selected variants

8 LIST OF FIGURES Figure page 1-1 The active site of HCA II The ph profiles for k cat ex /K eff CO2 (M -1 s -1 ) for the hydration of CO 2 catalyzed by HCA II and hydrophobic Tyr7 substitutions The ph profiles for k cat ex /K eff CO2 (M -1 s -1 ) for the hydration of CO 2 catalyzed by HCA II and polar substitutions at Tyr The ph profiles for k cat ex /K eff CO2 (M -1 s -1 ) for the hydration of CO 2 catalyzed by HCA II and conservative substitutions at Tyr7, Asn62 and Asn The ph profiles for the proton-transfer dependent rate of release of 18 O- labeled water by HCA II and hydrophobic substitutions The ph profiles for the proton-transfer dependent rate of release of 18 O- labeled water by HCA II polar substitutions at Tyr The ph profiles for the proton-transfer dependent rate of release of 18 O- labeled water by HCA II and conservative substitutions at Tyr7, Asn62 and Asn Free energy plot of proton transfer in HCA II and variants Double mutant catalytic cycle Crystal structures of the active sites of Y7I and Y7F HCA II Overall comparison of wild-type HCA II and Y7I HCA II Representation of crystal contacts in Y7I HCA II Differential scanning calorimetry profiles for wt HCA II and mutants Crystal structures of the active sites of N62Q and N67Q and Y7F+N67Q HCA II NO accumulation from nitrite at various concentrations of HCA II NO accumulation from various globular proteins and metal ion contamination The time course of NO accumulation from various forms of hemoglobin Extracellular NO accumulation obtained from a suspension of degassed human red cells

9 4-5 The ph dependence of extracellular NO accumulation The dependence on hematocrit of extracellular NO accumulation The effect of EZA on extracellular NO by degassed human erythrocytes The effect of DIDS on extracellular NO by degassed human erythrocytes The effect of PCMBS on extracellular NO by degassed human erythrocytes

10 LIST OF ABBREVIATIONS Å A AE ACZ ACES BH + BSA C CAM CA CAPS CHES Cl cm CO 2 ΔCp Angstrom alanine anion exchange acetazolamideatm N-(2-Acetamido)-2-aminoethanesulfonic acid protonated base bovine serum albumin Celsius carbonic anhydrase from Methanosarcina thermophila carbonic anhydrase N-cyclohexyl-3-aminopropanesulfonic acid N-cyclohexyl-2-aminoethanesulfonic acid chloride centimeter carbon dioxide change in eat capacity degree D deoxy DIDS DNA DNDs DTPA aspartic acid deoxygenated 4,4'-diisothiocyano-stilbene-2,2'-disulfonic acid deoxyribonucleic acid 4,4'-dinitrostilbene-2,2'-disulfonic acid, disodium salt diethylenetriaminepentaacetic acid 10

11 DSC E differential scanning calorimetry enzyme E. coli Escherichia coli EDTA enos EPR ev EZA F Fe 2+ GI ΔG H H + Hb(Fe II ) HCT HEPES His64 HCA HCO 3 ΔH m I IC50 IPTG K ethylenediaminetetraacetic acid endothelial nitric oxide synthase electron paramagnetic resonance electron volt ethoxyzolamide phenlyalanine Iron gastrointestinal Gibbs free energy histidine proton / hydrogen ion hemoglobin ferrous iron hematocrit 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid proton shuttle residue histidine 64 in HCA II human carbonic anhydrase bicarbonate ion melting enthalpy isoleucine concentration required for 50% inhibition isopropyl-β-d-thiogalactopyranoside rate constant 11

12 k B kcal k cat K M k cat /K M kda k exch CO2 cat /K eff LB M MES Met MOPS μm mm ma mol m/z N nm NO N 2 O 3 rate constant for proton transfer kilocalorie turnover number Michaelis constant specificity constant kilodalton specificity constant determined for hydration of CO 2 by CA luria broth molar 2-(4-morpholino)-ethane sulfonic acid reduced to ferric iron 3-(N-morpholino)-propanesulfonic acid micromolar millimolar milliamp mole mass to charge ratio asparagines nanomolar nitric oxide dinitrogen trioxide 18 O stable isotope of oxygen with atomic mass of 18 OD OH - PCMBS optical density hydroxide ion Para-chloromercuribenzene sulfonic acid, sodium salt 12

13 PDB ph pka Q R rmsd RT S SITS Protein Data Bank negative log of proton concentration acid dissociation constant glutamine arginine root mean square deviation room temperature serine 4-acetamido-4 -isothiocyanstilbene-2,2 -disulfonic acid disodium salt TAPS Tm Tris torr W wt Zn N-tris(hydroxymethyl)methyl-3-aminopropanesulfonic acid temperature melting tris(hydroxymethyl)aminomethane Torr (unit of pressure) tryptophan wild type zinc 13

14 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 KINETICS AND STRUCTURE OF PROTON TRANSFER PATHWAYS IN CARBONIC ANHYDRASE By Rose Lynn Mikulski December 2010 Chair: David Silverman Major: Medical Sciences Physiology and Pharmacology Human carbonic anhydrase II (HCA II) is a zinc metalloenzyme that catalyzes the reversible hydration of carbon dioxide to bicarbonate and a proton. Catalysis involves an intramolecular proton transfer that delivers an excess proton from the zinc-bound water to an internal proton acceptor, His64. His64 shuttles this proton to the bulk solvent, thus regenerating the active site for the next catalysis. The ability to experimentally increase the rate of proton transfer within the HCA II active site can provide insight into the biophysical properties for this process. The three factors proposed to influence the rate limiting step of proton transfer (k B ) are the active site water network, conformation of His64, and the pk a of both the zinc bound solvent and His64. An extensive analysis of the kinetic and structure of wild type and several mutants of HCA II were conducted over a broad ph range. The results show that the enzyme active site is very stable. Several mutants altered the proton shuttle His64 orientation, the water network, and the pk a of the defined proton donor and acceptor which resulted in altered proton transfer rates. Faster rates of proton transfer were observed in all the mutants. The 2-4-fold increases in proton transfer followed the theoretical values 14

15 predicated by Marcus theory based on changes in the pk a. The 7-15 fold increase in proton transfer over wild type showed residue His64 primarily in the inward conformation decreasing the distance to the zinc. The less branched water network with more conventional hydrogen bonds lengths connecting zinc solvent through only two water molecules to His64 appeared in mutants with enhanced proton transfer rates compared to wild type HCA II. Classically the physiological role of CA is in acid-base balance throughout the body. A recent proposal that CA could catalyze the conversion of nitrite into nitric oxide (NO) a potent vasodilator prompted the design of a system to directly measure NO concentrations in red cell suspensions. The examination of CA alone as well as CA within whole RBC as well as hemoglobin alone did not show sufficient generation of NO at physiological levels of nitrite for vasodilation. 15

16 CHAPTER 1 INTRODUCTION The Carbonic Anhdrases The carbonic anhydrases (CA)s are commonly characterized as zinc metalloenzymes that rapidly catalyze the reversible hydration of carbon dioxide to form bicarbonate and a proton [1]. The well studied classes of the enzyme are: α, β, and γ [2]. The α-class was first discovered in and still predominantly found in mammals [3]. The β-class includes both plant and several bacterial CAs such as that of Escherichia coli. Finally the γ-class consists of archaeal CAs. These three classes of CA have been extensively studied kinetically and are believed to share a similar mechanism. The classes of CA vary significantly in their structures which make them an example of convergent evolution. The α CA s contain one zinc per molecule, a characteristic of the conserved active site coordination by three histidine residues and a solvent molecule [2]. The α class are predominatly monomeric, near 30 kda molecular weight except human CA IX which has recently been shown to be a functional dimmer of about kda [4, 5]. The β class shows the widest structural deviation with varying oligermerization leading to molecular weights up to 200 kda. The zinc coordination of the β-class consists of one histidine and two cysteines residues, while the fourth coordination site is occupied by a solvent molecule or an aspartate in type I and II enzymes respectively. The γ-class are trimers with metal coordination analogous in structure to the α CAs and in the case of the CAM enzymes contain iron in place of the zinc under anaerobic growth[6]. The human genome is shown to code for at least 15 different isozymes of CA [7]. These isozymes vary in their expression, distribution, localization, and kinetic activity. 16

17 The cytosolic isoforms are I, II, III, and VII, and include the mitochondrial enzymes VA and VB. The extracellular isoforms include: the membrane bound enzymes by a transmembrane domain (IX, XII and XIV) or GPI-anchor (IV and XV) as well as the only secreted isoform VI. The isoforms VIII, X, and XI are considered CA-related proteins because they are missing one or more of the conserved Zn coordinating histidine residues rendering them catalytically dead. Catalytic Mechanism The catalytic conversion between carbon dioxide and bicarbonate by CA is a well studied mechanism that is believed to be shared by all CA s [1, 7, 8]. Human isozyme II of CA is one of the most extensively studied and is very fast with catalytic turnover of ~ s -1. A wide body of evidence supports a ping-pong mechanism with two steps [1]. + H 2 O EZnOH - + CO 2 EZnHCO EZnH 2 O + HCO 3 (1) His64-EZnH 2 O + B H + His64-EZnOH - + B His64-EZnOH - + BH + (2) The first stage is the conversion of CO 2 into HCO - 3 which is completely reversible. The hydration direction begins when the catalytically active zinc-bound hydroxide encounters a CO 2 as shown in eq 1. The literature often describes this process as a direct nucleophilic attack of the zinc-bound hydroxide on CO 2. The CO 2 may otherwise interact with the zinc through a general base mechanism. A better understanding of the mechanism can be obtained through solvent H/D isotope effects and the structure of a bicarbonate-ca complex [9, 10]. The absence of an H/D isotope effect on kcat/km for the initial stage of the catalysis (eq 1) was confirmed by both 18 O exchange [11] and 13 C 17

18 NMR [12]. The lack of a measurable isotope effect supports a direct nucleophilic attack mechanism of the zinc-bound hydroxide [9, 13]. There does not appear to be a rate influencing proton transfer event in this stage of the catalysis by CA [12]. The second stage of the catalytic mechanism involves the regeneration of the zinc bound hydroxide. This step is also the rate limiting intramolecular proton transfer in HCA II. The presence of an intramolecular proton shuttle in CA was predicted to be a side chain such as histidine with a pka near 7.0 similar to that of the zinc-bound water [14]. The ability to rapidly transfer the protons in the catalysis of CA is now attributed in part to the side chain of His64 based on several important experiments [15]. Through site directed mutagenesis of the human CAII protein, the effect of substituting the His64 with an alanine reduced the rate of proton transfer (eq 2) up to 50-fold [16]. The rate of CO 2 and HCO - 3 (eq 1) interconversion remained constant in the H64A mutant [16]. Further supporting evidence were chemical rescue studies that show recovery of activity in the H64A mutant upon the addition of imidazole [16]. The rate of the slower HCA III isozyme can be increased by 10-fold by replacing the endogenous Lys64 with a histidine [14]. The mutation to a histidine in HCA III also shows a kinetic profile including solvent H/D isotope effects and ph-rate profiles more like HCA II [14]. These experiments signify the extent to which CA s active site can be tuned to carry out the relatively simple hydrolysis of water. Structure The first crystal structure of HCA II was determined by Liljas and colleagues in 1972 and now there are more than 200 CA crystal structures in the protein data bank [17]. Overall HCA II is a single-domain, globular protein that is almost spherical with 18

19 approximate dimensions of 80 nm 3 [18, 19]. The HCA II structure can be described as a 10-stranded twisted β-sheet, which is decorated on the surface by seven α helices. The strands of the β-sheet are mainly anti-parallel, with the exception of two pairs of parallel strands. There is a conserved loop region extending towards the active site that contains the proton shuttling residue, His64. The active site is a conical cavity centrally located about 15 Å deep inside the spherical protein (Figure 1-1) [17]. The catalytic zinc ion (Zn 2+ ) is located at the bottom of the cleft, with tetrahedral coordination by three histidine residues (94, 96, and 119) and a highly polarized water molecule or hydroxyl group (Figure 1-1). The cavity contains a region of hydrophobic residues (Val121, Val143, Leu198, Thr199, Val203 and Trp209) where the carbon dioxide is held in orientation to interact with the zinc [20]. Leading out of the cavity is a patch of more hydrophilic residues (Tyr7, Asn62, His64, Ans67, Thr199 and Thr200). These hydrophilic residues are believed to help stabilize an active site water network though hydrogen bonding. Solvated Active Site The ordered network between the zinc-bound solvent and His64 is depicted in the crystal structure at 1.05 Å resolution (Figure 1-1) [19]. Several amino acids (Tyr7, Asn62, Asn67, Thr199, and Thr200) appear close enough to participate in coordinating a solvent network (W1, W2, W3a, and W3b). Thr199 is purported to hydrogen bond to the zinc-bound solvent that, in turn, is hydrogen bonded to W1. Thr200 and the next solvent in the chain, W2, further stabilize W1 through hydrogen bonds [9, 21]. The solvent network then apparently branches as W2 hydrogen bonds to both W3a and W3b. The hydroxyl group of Tyr7 further orientates the W3a, while Asn62 and Asn67 19

20 stabilize W3b. This solvent network is conserved in numerous crystallographic structures; which localizes W2, W3a, and W3b in close proximity to the side chain of His64 (Figure 1-1)[19, 22, 23]. Examination of the structure of HCA II at near atomic resolution reveals aspects not apparent in lower resolution studies [19, 22]. For instance, the solvent molecule W2 (the only ordered solvent molecule in the active site stabilized exclusively by other solvent molecules) is trigonally coordinated with equal distance (2.75 Å) by W1, W3a, and W3b. Only W2 is in the plane of the imidazole ring of His64 and within hydrogenbonding distance of the D1 nitrogen in the His64 imidazol ring when in the inward conformation (Figures 1-1). A more recent study also examined a water molecule in the hydrophobic region of the active site that forms a short strong hydrogen bond (2.45 Å) to the zinc bound solvent [22]. This deep water may contribute to the catalysis through providing an energy barrier to the binding of CO2 in the enzyme substrate complex, which can enhance the conversion rate of CO2 into bicarbonate, as well as to lower the pka of the zinc-bound water and to promote proton transfer. Proton Shuttle Residue Early structures of HCA II implicated His64 in intramolecular proton transfer despite the ~7 Å distance separating it from the zinc [23]. A highly refined crystallographic structure showing a water bridge between the zinc and His64 supported the above suggestion of intramolecular proton transfer in HCA II [19]. Evidence of the ability of His64 to help shuttle the protons from the active site out into the solution has been strengthen by crystallographic analysis confirming the in and out or flip-flop of the imidazole side chain of the histidine residue [19, 24]. The conformations of His64 are characterized by side chain torsion angle rotamers of 1 around the Cα-Cβ bond 20

21 between the side chain and protein back bone. The inward orientation pointing towards the zinc in the active site is defined by dihedral angles χ 1 of 44 and of χ The outward orientation points toward the bulk solvent surrounding the surface of the protein and is defined by dihedral angles of χ 1 of -39 and χ 2 of 98. Other CA s including the γ- class CAM, show two rotamers of the proton transfer residue Glu84 residue [25]. Several orientations of a chemically modified cysteine residue in a mutant of murine CA V are observed and implicated in proton transfer [26]. The occupancy of the H64 side chain in HCA II appears distributed between the inward and outward conformation to transfer protons intramolecularly. The orientation of the His64 has shown some ph dependence such that the inward conformation is increasingly occupied in increasting ph from 6.0 to 9.0 in crystal structures [19]. The Grotthuss mechanism explains that high proton mobility can be achieved in the active site by protons hopping along the water network s hydrogen bonds with minimized movement of the oxygen atoms [27]. The migration of protons through the active site of CA is highly interesting to theorists as it can be applied to other biologically important proton transfers in proteins and their catalysis. Several computational labs have applied theory to the simulation of proton transfers with CA as the model [27-31]. Extended Active Site The activity of CA may start at the catalytic zinc but the entire environment allows the efficient catalysis of the enzyme. Recent mutational studies in the active site of HCA II show how other residues within the H64 environment modulate the flexibility of the imidazole side chain altering the ability to transfer protons. Insight into the highly evolved active site was gained through mutation of Tyr7, Asn62 and Asn67, which surround the His64 on either side in the hydrophilic portion of the cavity [23, 32]. 21

22 Substitution of hydrophobic residues at these positions was shown to cause changes in the rate of proton transfer, the orientation of His64, the pka of His64, disruption of the water network, and further steric changes to other residue side chains [18]. A typical case is the mutagenesis of Asn67 to leucine in HCA II, which altered the solvent structure, caused nearly complete out orientation of the His 64 side chain, and decreased the proton transfer rate constant to 0.2 μs -1 [23]. A recent examination of several hydrophobic mutations at position 62 showed maintained water networks and concluded that a role of Asn62 in HCA II is to permit two conformations of the side chain of His64 [32]. Speeding Up Proton Transfer The difficulty in understanding the speed at which proton transfer occurs within HCA II is predicated on accurate identification of the proton-conducting pathway. The proposed intramolecular proton channel is ZnH 2 O through W1 to W2 and bridged from W3a and W3b to the His64 and ultimately out to the bulk solvent (Figure 1-1). Identifying the proton donor as the zinc bound solvent and His64 as the intramolecular proton acceptor allows us to correlate changes in the electrostatics of the active site by the ΔpK a (ΔpK a ZnH 2 O - ΔpK a His64). Future mutational analysis should help to confirm this pathway and important residues that help maintain the electrostatics of the active site. The mutation of Tyr7 into a non-polar hydrophobic phenylalanine exhibited a rate constant for proton transfer seven-fold greater than the wild type HCA II [23]. The Y7F substitution did not have an effect on the catalysis of the hydration-dehydration step thus being consistent with wild type serves as a control and indicated that no perturbations are caused at the Zn-OH/H 2 O ~7.0 Å away [23]. The kinetics also showed that the His64 pk a was lowered from 7.0 in wild type to about 6.0. The more acidic side 22

23 chain may correlate with a faster proton transfer in the dehydration direction. Structurally, the Y7F HCA II differed in that His64 was mainly inward and the W3a was missing. The absence of W3a possibly allows a more streamlined water network. A simplified water wire leading to faster proton transfer agrees with in silico studies where branched arrays form a less efficient Eigen -like solvent structures (H 9 O + 4, that is H 3 O + -3(H 2 O) [28]. The hydroxyl of Tyr7 appears to be within hydrogen bonding distance of water molecule W3a. However, the recent neutron crystal structure recently determined at ph 9 shows the tyrosine hydroxyl unprotonated and no such hydrogen bond to solvent [33]. Physiological and Medical Significance. Depending on where the particular CA resides it can be implicated in a number of important physiological functions including: 1) the rapid conversion of HCO - 3 into CO 2, as in red cells and in photosynthesis; 2) generation of HCO - 3 for secretory fluids, as in ocular and cerebrospinal fluids; 3) ph regulation by production of H + from water and CO 2, as in renal acidification of urine and gastric acid secretion; and 4) facilitation of diffusion of CO 2 across membranes as in the lens of the eye [34]. The number of CA isoforms results in functional redundancy such that organisms may survive even when a mutation renders one isoform dysfunctional. This has been the case in humans when deleterious mutations that occur in HCA I do not change physiological fitness because HCA II compensates. HCA II is well known for its presence in red cells where it efficiently stores and converts CO 2 waste in respiration as well as its maintenance of secretory fluids in the eye. HCA II deficiency syndrome results in a variety of symptoms including renal tubular acidosis, cerebral calcification, and osteopetrosis when inherited as an autosomal 23

24 recessive trait [35]. Recently the over-expression of HCA IX has been associated with several malignant tumor types; CA acidification implicated in tumor growth and metastasis; and proposed as a possible cancer marker [36]. A drive to develop a new class of inhibitors and markers for the enzyme may benefit from the identification of the active site features important for proton transfer or isozyme specificity. 24

25 A B Figure 1-1. The structure of HCA II. A) cartoon representation of the over all globular structure of the alpha class Hunam CAII. The active site residues are shown as sticks Orange for the hydrophobic CO 2 binding pocket and blue for the hydrophilic resudues; the zinc ion and the oxygen molecule of waters are shown as gray and red spheres, respectively. B) Close up ball-and-stick diagram of coordinating active site residues as labeled; the zinc ion and the oxygen molecule of waters are shown as gray and red spheres, respectively. The water network of the active-site is labeled W1, W2, etc. Presumed hydrogen bonds are represented as dashed red lines. The dual conformation of His64 side chain is shown in both inward and outward conformations[19]. 25

26 26

27 CHAPTER 2 KINETICS OF SPEEDING UP PROTON TRANSFER IN HUMAN CARBONIC ANHYDRASE II The previous chapter introduces the mechanism by which human carbonic anhydrase II (HCA II) interconverts carbon dioxide (CO 2 ) and bicarbonate (HCO 3 - ). This chapter will discuss the second stage of the reaction, the regeneration of zincbound hydroxide by proton transfer. Specifically, the roles of active site hydrophilic residues Tyr7 and Asn62 and Asn67 in tuning the pk a of the proton acceptor, His64 and the effect on proton transfer efficiency. As was discussed previously, the reaction catalyzed by HCA II is a two-step cycle. The first step, in the hydration direction, is the hydration of CO2 to form bicarbonate at the zinc. The bicarbonate becomes displaced when a water molecule diffuses into the active site. The zinc-bound water is catalytically inactive, so a proton must be transferred off of this water molecule to regenerate a zinc-bound hydroxide, a step that is rate limiting for the overall catalytic cycle. This is accomplished through an ordered network of solvent molecules that form a hydrogen-bonded wire to connect from the zinc bound solvent to the proton shuttle residue His64 (Figure 1-1). The water network is maintained in part by interactions with hydrophilic residues Tyr7 and Asn62 and Asn67 of the extended active site. These residues located on either side of the proton shuttle reside His64 (Figure 1-1), can tune the proton transfer efficiency of His64 in equation 2 directly as well as though hydrogen boding with the water network. Extended Active Site Hydrophilic Residues Although the effect of the replacement of residues lining the active site cavity on catalysis is complex, the kinetic data may indicate influences conducive to increased proton transfer rates. Additionally Tyr7, Asn62, and Asn67 have shown little or no effect 27

28 on the interconversion of CO 2 and bicarbonate that occurs 7-10 Å away from the zinc [23]. It is notable that Try7 is invariant in α class CA II from a wide range of species from chicken, rodents, bovine and the human isozymes [2]. Asn62 and Asn67 are not as conserved in isozyme of HCA but are consistently hydrophilic except in the case of HCA I and HCA IV (containing Val62 and Met67 respectively). Asn62 and Asn67 vary particularly in conjunction with histidine substitutions at residue 64, such as HCA III with Lys64 and Arg67, and HCA V with Tyr64, Thr62, and Gln67[23]. Interestingly in human isozymes position 67 is more often a glutamine (HCA VI, VII, VIII, IX, XIV) than asparagine (HCA II, XIII). Hydrophobic substitutions done in these hydrophilic residues resulted in interesting kinetics that will guide the future substitutions at these residues [23, 32]. Interest in this region of the protein is based on the mutant Y7F HCA II which showed that the proton transfer component of catalytic dehydration was enhanced as much as 9-fold compared to wild type [23]. This increased rate of proton transfer occurred with changes in the pk a of donor and acceptor as well as structural differences. Similar hydrophobic mutations at position Asn62 also showed the changes in the pk a of the donor and acceptor groups but with fewer structural changes [32]. Faster Rates of Proton Transfer Here we discuss a number of substitutions at position 7, 62, and 67 in HCA II that affect rate of the catalysis. Catalysis by each of the variants was studied by 18 O exchange between CO 2 and water using membrane inlet mass spectrometry. We have found that substitution at Tyr7 had no effect on the first stage of catalysis (eq 1), but substitutions at position 62 and 67 showed some decrease in efficiency compared with wild type. Mutations at Tyr7, Asn62, and Asn67 are shown to affect the pka of His64. 28

29 These electrostatic changes alter the rate of proton transfer according to Marcus theory.these studies emphasize the role of Tyr7 and Asn67 on long range, intramolecular proton transfer. Methods Expression and Purification of Enzymes Several variants of HCA II were created with the Stratagene Quick Change II sitedirected mutagenesis kit (La Jolla, CA) on the expression vector coding the full-length wild type HCA II [37]. Tyr7 was replaced with Ala, Ile, Trp, Asp, Asn, Ser, and Arg. Conservative mutations were also examined by replacing Asn62 and Asn67 with Gln and the double substitution Tyr7 to Phe with Asn67 to Gln. The DNA sequences of each mutant were confirmed for the entire region coding of CA in the vector. The verified plasmids were then transformed for expression into Escherichia coli BL21(DE3)pLysS cells from Stratagene. The transformed cells were grown in LB media containing 1.0 mm ZnSO 4 and induced with IPTG to a final concentration of 1.0 mm at an OD 600 of 0.6 AU [32]. Each variant was purified by affinity chromatography using p-(aminomethyl) benzenesulfonamide coupled to agarose beads (Sigma) [38]. Protein concentrations of variants were determined by titration with the tight-binding inhibitor ethoxzolamide and detecting activity by 18 O exchange between CO 2 and water. Kinetics 18 O Exchange Catalysis was measured by the 18 O exchange method based on the measurement by membrane inlet mass spectrometry of the depletion of 18 O from species of CO 2 [39]. The apparatus uses a membrane probe of silastic tubing, which is permeable to dissolved gases and is submerged in the reaction solution. The probe is connected to 29

30 glass tubing that runs through a dry-ice and acetone water trap and continues to the mass spectrometer (Extrel EXM-200) [40]. In the first of two independent stages of catalysis, the dehydration of labeled bicarbonate has a probability of transiently labeling the active site with 18 O (eq 3). In a second stage, the protonation of the zinc-bound, 18 O- labeled hydroxide results in the release of H 2 18 O to the solvent (eq 4). HCOO 18 O + EZnH 2 O EZnHCOO 18 O COO + EZn 18 OH (3) - H 2 O H + His64-EZn 18 OH His64-EZnH 2 18 O His64-EZnH 2 O + H 2 18 O (4) Two rates for the 18 O exchange catalyzed by CA are obtained by this method. The first is R 1, the rate of exchange of CO 2 and HCO 3 at chemical equilibrium, as shown in eq 5. Here k cat exch is a rate constant for maximal interconversion of substrate and product, K eff S is an apparent binding constant for substrate to enzyme, and S indicates substrate, either CO 2 or bicarbonate. The ratio k cat exch /K eff CO2 is, in theory and in practice, equal to k cat /K m for hydration obtained by steady-state methods [12]. R 1 /[E] = k cat exch [CO 2 ]/( K eff CO2 + [CO 2 ] ) (5) The second rate determined by this method is R H2O, which is the rate of release of 18 O labeled water from the active site (eq 4). R H2O is a measure that is dependent upon the donation of protons to the 18 O-labeled zinc-bound hydroxide. In eq 6, k B is the rate constant for proton transfer to the zinc-bound hydroxide, and (K a ) His64 and (K a ) ZnH2O are the ionization constants of the proton donor and zinc-bound water molecule. R H2O /[E] = k B / ([1 + (K a ) His64 /[H + ]][1 + [H + ]/(K a ) ZnH2O ]) + R H2O (6) Here R H2O is a ph independent contribution introduced to provide a fit to data obs showing a plateau in k B at ph > 8. Fits of eqs 5 and 6 to the data were carried out using Enzfitter (Biosoft). 30

31 The catalyzed and uncatalyzed exchanges of 18 O between CO 2 and water at chemical equilibrium were measured in the absence of buffer at a total substrate concentration of 25 mm of all species of CO 2 using membrane-inlet mass spectrometry. The total ionic strength of solution was kept at a minimum of 0.2 M by the addition of Na 2 SO 4. The kinetics on Y7F HCA II was reported earlier [23]. Esterase Activity The catalysis by HCA II and mutants of the hydrolysis of 4-nitrophenylacetate was measured by the method of Verpoorte et al. [41]. The hydrolysis was followed at 348 nm, the isosbestic point of nitrophenol and the conjugate nitrophenylate ion using the molar absorbtivity M -1 cm -1. A Beckman Coulter DU 800 spectrophotometer was used to measure both the uncatalysed and carbonic anhydrase-catalysed initial velocities. A range of buffers from ph 5.0 to 9.0 were used at 100 mm (MES, ACES, MOPS, HEPES, TAPS, CHES, CAPS). The kinetic constants and ionization constants were determined from the ph profiles of k cat /K M by nonlinear least-squares methods to a single ionization with a maximum at high ph (Enzfitter, Elsevier-Biosoft, Cambridge, U.K.).The rate constants k enz reported here represent k cat /K m for the catalyzed hydrolysis. The value of K M is too large to measure k cat. Results The site-specific mutants of HCA II in Table 2-1 were investigated for their catalytic properties using membrane inlet mass spectrometry measuring the rate of exchange of 18 O between CO 2 and water. The efficiency of the catalyzed hydration of CO 2 was measured as the rate constant k cat exch /K eff CO2 (eq 5) over a ph range of 5.0 to 9.0. The ph profiles for k cat exch /K eff CO2 by the Tyr7 mutants under study were adequately fit to a single ionization and appeared similar to that of the wild type (Figure 2-1, 2-2). The 31

32 Asn62 and Asn67 mutations shown in Figure 2-3 some deviations compared to the wild type. The resulting maximal values of k exch cat /K CO2 eff (Table 2-1) represent catalytic activity in the hydration direction and were essentially identical to that of wild type except for N62Q and N67Q that showed a slight decrease in efficiency. This lower k exch cat /K CO2 eff was similar to those observed during the substitution of hydrophobic residues at position 62 [32]. The variation was greater for the activities k cat /K m for the catalyzed hydrolysis of p-nitrophenylacetate (Table 2-1) but followed the trend of the maximal values of k exch cat /K CO2 eff. This probably reflects the larger size of the substrate for the ester hydrolysis which is influenced more by amino acid changes near the mouth of the active site cavity. The resulting values of pk a ZnH2O representing the ionization of the zinc-bound water for each mutant are listed in Table 2-2. The data demonstrate that the values of pk aznh2o of the zinc-bound water are near 7 for wild type and mutants. These pk a ZnH2O values were determined both by measurement of esterase activity and 18 O exchange (Table 2-2). The ph profiles of the rate constant R H2O /[E] (Figures 2-4, 2-5, and 2-6) provide three constants relevant to the catalysis, according to eq 6. The first is an estimate of pk a ZnH2O, values which are given in Table 2-2 and generally are consistent with the values described in the above paragraph. Exceptions are Y7A, Y7S, Y7R and Y7F+N67Q (Table 2-2; Figure 2-2). These differences may be due to different properties of the active site for the processes of eq 3 and eq 4, or possibly difficulty interpreting the irregular curves for R H2O during catalysis by Y7A and Y7S (Figure 2-2). The R H2O /[E] profile of Y7D did not have sufficient bell shape to assign the pk a values. In later construction of a free energy plot, we have used data only from R H2O. 32

33 The second constant is a value of pk a His64 (Table 2-2). These values are uniformly lower for replacements at residue 7, 62, and 67 than the value of pk a His64 at 7.2 in wild type HCA II (Table 2-2). Such a shift in pk a His64 is associated with the inward coordination of the side chain of His 64 [23, 32] and its more hydrophobic environment than in wild type HCA II in which this side chain appears about equally in inward and outward orientations[19, 24]. The third constant is the maximal, ph-independent value of k B describing intramolecular proton transfer in the dehydration direction obtained by a fit to eq 6 to the ph profiles for R H2O /[E] (Table 2-2; Figure 2-4, 2-5, and 2-6). The ph profiles for R H2O, from which k B values were obtained, were more difficult to fit for Y7D and Y7N than the bell curve of wild type HCA II and other variants (Figure 2-5). All of the variants at position 7, 62, and 67 showed rates of proton transfer equally to (Y7A, Y7D) or greater than the wild type enzyme (Y7R, N62Q, Y7S, N67Q, Y7W, Y7I, Y7N, Y7F, and Y7F+N67Q in order of increasing k B ). The data for Y7I were consistent with a value of k B of 2.3 μs -1. The greatest changes in kb among these variants were Y7F with a value of 7 μs -1 and Y7F+N67Q with a value of 12.5 μs -1 (Table 2-1). Discussion A notable result of this study of variants at residue 7 of HCA II is that the rate constants k exch cat /K CO2 eff for the first stage of catalysis were unchanged. The substitutions of Tyr7 did not affect the ability of the protein to carry out the first stage of the catalysis. This appears to be a result of the amino acid substitutions being at least 7Å away from the zinc where the interconversion between CO 2 and bicarbonate occurs. The substitutions N62Q and N67Q caused substantial change in the rate constants k exch cat /K CO2 eff which is not explained by values of the pk a of the zinc-bound water. However, none of the variants in Table 2-1, had a k B for proton transfer lower than that 33

34 of the wild type. This supports a role for Tyr7, Asn62, and Asn67 in fine tuning the rate of intramolecular proton transfer. The changes in k B were examined in relationship to the donor and acceptor pk a values on a free energy plot in Figure 2-3. The open circles are rate constants for proton transfer during catalysis by H64A HCA II determined by enhancement of catalysis when proton donors are exogenous derivatives of imidazole and pyridine [16]. These data are fit by Marcus theory applied to proton transfer [42, 43], represented by the solid line of Figure 2-3. We assume this line represents the dependence of k B on ΔpK a within the active site of HCA II. The significance of this fit is that these values of k B are for proton transfer from donors not attached to the enzyme through chemical bonds, free of many restraints. Interestingly, the values of k B for wild type HCA II and nearly all of the variants of Table 2-1, except Y7F, and Y7F+N67Q HCA II, fall on or near this Marcus line (Figure 2-7). We can conclude that the small variation in k B for the majority of these mutants were possibly due to small changes in the electrostatics of the active site thourgh possible changes in His64 or the water network that altered its pk a. The large increase in the k B for the Y7F, and Y7F+N67Q HCA II mutants was not explained simply by Marcus theory but will be discussed with respect to structural changes in Conclusions (Chapter 5). A double mutant catalytic cycle showed the change in proton transfer rates between the single and double mutants from wild type HCA II. This compares the changes in free energy barriers ΔG according to ΔG =- RTln[k B (mutant 2 )/k B (mutant 1 ). The large negative ΔG values in the cycle of these mutants indicate that the increases in activity compared to wild type result in significant decreases in the free energy barriers of proton transfer. The change in the activation 34

35 barrier for the double mutant ΔG 1+2 in Figure 2-8 for Y7F+N67Q HCA II (-1360 cal/mol) was greater than sum of the individual mutation Y7F and N67Q HCA II (-820 cal/mol). Thus we classify this double mutant interaction as synergistic (greater than the sum of the individual changes) and results in an increase in the proton shuttle efficiency of His64. Kinetic changes we observed maybe better understood in the light of structural changes in the active site water network as well as the conformation of His64 in these mutations and variation of other key residues (See Conclusions). The results help confirm a major driving force of proton transfer is the electrostatics of the active site that defines the pk a of both the zinc-bound solvent and His64. These changes in pk a can directly affect the change in proton transfer consistent with Marcus theory. The ability to increase the rate of proton transfer by nearly 15 fold in Y7F+N67Q HCA II has never been seen before. 35

36 Table 2-1. Kinetic rate constants for HCA II and mutants. The maximal values of the rate constants for hydration of CO 2, hydrolysis of 4-nitrophenylacetate, and proton transfer catalyzed by HCA II and variants. Rate constants were derived from the kinetic curves for each substitution by a fit of the data. a Measured from the exchange of 18 O between CO 2 and water using eq 5 in the hydration direction. b Measured from the fit of the rate constants for ester hydrolysis to a single ionization. c Measured from the exchange of 18 O between CO 2 and water using eq 6 in the dehydration direction. d Data are from Fisher et al. [23]. e These are maximal values of R H2O /[E] since incomplete ph profiles (Figure 2-2) did not allow an adequate determination of k B by a fit of eq 6. Enzyme k exch CO2 cat /K eff CO 2 hydration (μm -1 s -1 ) a k cat /K m esterase (M -1 s -1 ) b k B proton transfer (μs -1 ) c Wild Type 120± ± ± 0.1 Y7I 130± ± ± 0.2 Y7A 140± ± ± 0.3 Y7W 140± ± ± 0.1 Y7F d 120± ± ± 0.2 Y7D 130± ± e ± 0.1 Y7N 120± ± e ± 0.1 Y7S 120± ± ± 0.1 Y7R 120± ± ± 0.1 N62Q 50± ± ± 0.4 N67Q 50± ± ± 0.3 Y7F+N67Q 100± ± ± 2 36

37 Table 2-2. Values of apparent pk a obtained by various kinetic measurements of catalysis in HCA II and mutants. a The pk a was derived from the fits of eq 6. b The pk a was determined from a fit of eq 5. The observed pk a values from a second ionization by some mutants; were not included in this table. c These data from Fisher et al.[23]. d The Y7D HCA II data for R H2O /[E] did not have sufficient bell-shape to be adequately fit by eq 6 (Figure 2-2). Enzyme pk a His64 a pk a ZnH2O a pk a ZnH2O b pk a ZnH2O (esterase) (eq 6) (eq 6) (k exch cat /K CO2 eff ) Wild Type 7.2± ± ± ± 0.1 Y7I 6.2± ± ± ± 0.1 Y7A 6.4± ± ± ± 0.1 Y7W 6.9± ± ± ± 0.1 Y7F e 6.0 c ± c ± c ± ± 0.1 Y7D -- d -- d 6.4± ± 0.2 Y7N 6.2± ± ± ± 0.1 Y7S 6.2± ± ± ± 0.1 Y7R 6.4± ± ± ± 0.1 N62Q 6.6± ± ± ± 0.3 N67Q 6.6± ± ± ± 0.1 Y7F+N67Q 6.3± ± ± ±

38 Figure 2-1. The ph profiles for k cat ex /K eff CO2 (M -1 s -1 ) for the hydration of CO 2 catalyzed by HCA II and hydrophobic Tyr7 substitutions. The log of hydration of CO 2 relative to the amount of enzyme is described from ph 5.0 to 9.0 for: Y7A HCA II ( ), Y7I HCA II ( ), Y7W HCA II ( ), Y7F HCA II (Δ), and wild type ( ). Data were obtained by 18 O exchange between CO 2 and water using solutions containing 25 mm of all species of CO 2 and sufficient Na 2 SO 4 to maintain 0.2 M ionic strength. 38

39 Figure 2-2. The ph profiles for k cat ex /K eff CO2 (M -1 s -1 ) for the hydration of CO 2 catalyzed by HCA II and polar substitutions at Tyr7. The log of hydration of CO 2 relative to the amount of enzyme is described from ph 5.0 to 9.0 for: ( )Y7D HCA II (Δ), Y7N HCA II ( ), Y7R HCA II ( ), Y7S HCA II ( ), and wild type HCA II ( ). Data were obtained by 18 O exchange between CO 2 and water using solutions containing 25 mm of all species of CO 2 and sufficient Na 2 SO 4 to maintain 0.2 M ionic strength. 39

40 Figure 2-3. The ph profiles for k cat ex /K eff CO2 (M -1 s -1 ) for the hydration of CO 2 catalyzed by HCA II and conservative substitutions at Tyr7, Asn62 and Asn67. The log of hydration of CO 2 relative to the amount of enzyme is described from ph 5.0 to 9.0 for: N62Q HCA II ( );N67Q HCA II ( );Y7F HCA II (Δ); Y7F+N67Q HCA II ( ); and wild-type HCA II ( ). Data were obtained by 18 O exchange between CO 2 and water using solutions containing 25 mm of all species of CO 2 and sufficient Na 2 SO 4 to maintain 0.2 M ionic strength. 40

41 Figure 2-4. The ph profiles for the proton-transfer dependent rate of release of 18 O- labeled water by HCA II and hydrophobic substitutions. Kinetic profiles were determined by membrane inlet mass spectrometry for: Y7A HCA II ( ), Y7I HCA II ( ), Y7W HCA II ( ), Y7F HCA II (Δ), and wild type ( ). Data were obtained by 18 O exchange between CO2 and water using solutions containing 25 mm of all species of CO 2 at sufficient Na 2 SO 4 to maintain 0.2 M ionic strength. No buffers were added. 41

42 Figure 2-5. The ph profiles for the proton-transfer dependent rate of release of 18 O- labeled water by HCA II polar substitutions at Tyr7. Kinetic profiles were determined by membrane inlet mass spectrometry for: ( )Y7D HCA II (Δ), Y7N HCA II ( ), Y7R HCA II ( ), Y7S HCA II ( ), and wild type HCA II ( ). Data were obtained by 18 O exchange between CO2 and water using solutions containing 25 mm of all species of CO 2 at sufficient Na 2 SO 4 to maintain 0.2 M ionic strength. No buffers were added. 42

43 Figure 2-6. The ph profiles for the proton-transfer dependent rate of release of 18 O- labeled water by HCA II and conservative substitutions at Tyr7, Asn62 and Asn67. The kinetic profiles were determined by membrane inlet mass spectrometry for: N62Q HCA II ( );N67Q HCA II ( );Y7F HCA II (Δ); Y7F+N67Q HCA II ( ); and wild-type HCA II ( ).Data were obtained by 18 O exchange between CO2 and water using solutions containing 25 mm of all species of CO 2 at sufficient Na 2 SO 4 to maintain 0.2 M ionic strength. No buffers were added. 43

44 Figure 2-7. Free energy plot of proton transfer in HCA II and variants. The logarithm of the rate constant for proton transfer k B (s -1 ) versus ΔpKa (from pk a ZnH2O pk a His64) determined from the R H2O /[E] ph profiles for the wild type and the mutants of HCA II identified on the plot ( 25 C data, 10 C data); and H64A HCA II ( ) from An et al. [44] with proton transfer provided predominantly by derivatives of imidazole and pyridine acting as exogenous proton donors with the solid line a best fit of Marcus Rate Theory [42]. 44

45 Figure 2-8. Double mutant catalytic cycle. Comparisons of the rate constants for intramolecular proton transfer for HCA II and mutants. The values of k B (ms -1 ) appear beneath each designated mutant, and the values adjacent to the arrows are the free energy changes (ΔG in kcal/mol) for the barriers to proton transfer for the corresponding mutations calculated from k B (ΔG =-RTln[k B (mutant2)/k B (mutant1)]. Data were obtained by 18 O exchange between CO 2 and water using solutions at 10 C containing 25 mm of all species of CO 2 at sufficient Na 2 SO 4 to maintain 0.2 M ionic strength. No buffers were added. 45

46 CHAPTER 3 STRUCTURE OF PROTON TRANSFER PATHWAYS IN HUMAN CARBONIC ANHYDRASE II The previous chapter identified changes in proton transfer rates for mutants of HCA II in the region of His64.This chapter will describe structural changes in the active site due to these same mutations in HCA II. The X-ray crystal structures allow us to examine the active site water network, conformation of His64 and changes in active site residues. The discussion will examine the kinetic changes observed in several of the mutants in Chapter 2 with the crystal structures of Y7I, N62Q, N67Q, and Y7F+N67Q HCA II reported here. Structural Role of Water Network in Proton Transfer The active site cavity of HCA II has distinct hydrophilic and hydrophobic surfaces with a zinc atom at the bottom of a conical cavity. An ordered chain of hydrogen bonded, water molecules (W1, W2, W3a and W3b) extends from the zinc-bound water to the proton shuttle residue His64 (Figure 1-1). The water network is maintained in part by interactions with hydrophilic residues within the active site cavity (Tyr7, Asn62, Asn67, Thr200). The significance of this ordered water network in rapid transfer of protons has been studied in various mutants of HCA II [9, 19, 23, 32] as well as molecular dynamic simulations[28]. Computational studies indicate a more rapid proton transfer through a single, non-branched chain of water molecules compared with branched chains [28, 45, 46]. Proton Shuttle Residue The final step in the transport of protons out of the active site is the proton shuttle residue His64, which is believed to bridge the active site with the bulk solvent. Crystal structures of HCA II show that the side chain of His64 exists in two conformations 46

47 referred to as the inward and outward conformations [19, 24]. They also observed a ph dependence of the His64 residue orientation where structures at ph 8.0 to 9.0 had mostly inward His64 and at ph 5.7 had mostly outward His64 in wild type HCA II. The rotation of His64 from inward to outward has been limited by steric interactions with Trp5 and Asn62 such that a rotation about χ 1 appears limited to 100 [47]. The implications of the inward and outward orientations of the proton shuttle for facile proton transfer in CA are still under investigation. Overall HCA II is a small and compact protein 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 1-3 are usually disordered in the crystal structures. A cluster of aromatic residues at the N-terminus of HCA II consisting of Trp5, Tyr7, Trp16, and Phe20 has been suggested to assist in the anchoring of this region to the rest of the enzyme [18]. It has also been shown that the removal of up to 24 residues from the N- terminal region does not result in a major loss of protein stability or enzyme activity. The active site folds first and independently of the N-terminal region. Methods X-Ray Crystallography Crystals of the mutants Y7I, N62Q, N67Q and Y7F+N67Q HCA II were obtained using the hanging drop method [48]. The crystallization drops were prepared by mixing 5 μl of protein [concentration of ~15 mg/ml in 10 mm Tris-HCl (ph 8.0)] with 5 μl of the precipitant solution. The precipitants were mutant specific: 2.8 M ammonium sulfate with 50 mm Tris-Cl (ph 8.5) for Y7I, 1.25 M sodium citrate with 100 mm Tris-Cl (ph 8.0) for N62Q and 2.6 M ammonium sulfate, 5.0 mm HgCl, and 100 mm Tris-Cl (ph 8.0) for N67Q]. The drops were equilibrated at room temperature against a well of 1000 μl 47

48 precipitant solution. Crystals were observed within a week of the crystallization setup at 293K. The x-ray diffraction of Y7I HCA II was obtained at room temperature, 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 80 mm. The oscillation steps were 1 with a 7 min exposure per image. The N67Q, N62Q, and Y7F+N67Q HCA II crystals underwent a quick cryo cooling with a 30% glycerol cryo-protection before mounting for diffraction. The x-ray diffraction of these crystals was obtained at 100K, with a liquid nitrogen stream, using an R-AXIS V ++ optic system from Viarimax HR a Rigaku HU-H3R Cu rotating anode operating at 50 kv and 100 ma. The detector to crystal distance was set to 80 mm. The oscillation steps were 1 with a 6 min exposure per image for 360. Data set statistics for all three crystals are given in Table 3-1. Diffraction data were indexed, integrated, and scaled using the HKL2000 program package [49]. The model building was done manually with the program Coot [50], and refinement was carried out with Phenix suite. The data sets were phased to the 1.54 Å resolution structure of wild-type HCA II (PDB ID: 1tbt) with the waters removed and both His64 and the appropriate Asn 62 or 67 residues mutated to Ala to prevent any introduction of bias. The initial maps were used to confirm and fit the appropriate side chain conformations of His64 and other mutated residues. Waters were then placed and confirmed in Coot during several rounds of refinement until the R cryst reached convergence. Final refinement statistics for all four variants are given in Table

49 These structures are compared to the 2ILI (1.05 Å resolution wt HCA II structure) and the 2NXT (1.15 Å resolution Y7F HCA II structure). Thermodynamic Stability Differential scanning calorimetry (DSC) experiments were performed using a VP- DSC calorimeter (Microcal, Inc., North Hampton, MA) with a cell volume of ~0.5 ml. Tyr7 variants of HCA II were buffered in 50 mm Tris-HCl, ph 8.0 at protein concentrations of 50 μm. DSC scans were collected from 30 to 90 C with a scan rate of 60 C/hour. The calorimetric enthalpies (ΔH ºm) of unfolding were calculated by integrating the area under the peaks in the thermograms after adjusting the pre- and post-transition baselines. Heat capacity (ΔC p ) of protein unfolding obtained by plotting calorimetric enthalphy (ΔH ºm) vs melting temperature (T m ). Results The crystal structure of the four variants N62Q, N67Q, Y7F+N67Q and Y7I were solved to resolution Å (Table 3-1). Structure of Y7I HCA II The Y7I HCA II crystal was well ordered, and diffracted X-rays to 1.5 Å resolution. The structure of Y7I HCA II was similar to wild type in that the residues coordinating the zinc, and those in the active site cavity (with the exception of residue seven) are unchanged (Figure 3-1). His64 showed a predominantly inward orientation similar to Y7F HCA II reported earlier [23]. However, the structure of Y7I HCA II showed changes in the N-terminus that have not been observed in other Tyr7 mutants. The altered location of His10 in this structure replaces Trp5 in the wild types pie stacking interaction in wild type with His64 which may help stabilize the inward conformation of His64 in this mutant. Most notably there was a novel conformation of residues 4 to 11 of 49

50 the N-terminal chain (Figure 3-2). Several new interactions are observed here for residues 4,6,8-11 that are different from those in the wild type which may help stabilize this new N-terminal conformation. The amino-terminal (N-terminal) residues 1 to 3 were not observed in the crystal structure presumably because of disorder. This N-terminal conformation in Y7I HCA II caused the side chain of Ile7 to point away from the metal in the active site; the side chain of Ile7 does not occupy a position analogous to the side chain of Tyr7 in wild type, which lies between Thr200 and His64 (Figure 3-2). Despite the differences in structure of the N-terminal residues of Y7I and wild type, the geometry of the residues coordinating the zinc and the active site solvent structures were not altered (Figures 3-1). The observed changes in the N-terminal residues may be due to protein stability which was examined by DSC described below. The altered conformation in Y7I was not expected because the Y7F structure had an N- terminal fold similar to wild type. The N-terminal fold in wild type and Y7F position side chain of residue seven toward the active site near W3a. The residue at position seven is held by week van der waals forces of the surrounding amino acids side chains and the hydrogen bond to W3a but no other direst interactions are observed. The structure also showed that Ile7 of Y7I interacted with other residues in one of the crystal contact regions. Its amide nitrogen forms a hydrogen bond with the carboxyl side chain of Glu 238 and its side chain is in hydrophobic contact with Leu 240 (Figure 3-3). These interacting amino-acids are located in a single crystal-contact patch. It is possible that the altered orientation of the N-terminus of Y7I HCA II is a consequence of crystal contacts. It is interesting to note however, that this mutant crystallizes in the same space group as the wild type enzyme. 50

51 Role of Tyr7 in Thermal Stability of HCA II The thermal stability of HCA II mutants was determined by differential scanning calorimetry (DSC). A single peak representing the main unfolding transition was observed for Y7A, Y7W, Y7D, Y7N, Y7S,and Y7R HCA II as well as wild type, Y7I and Y7F HCA II as shown in Figure 3-4. The main unfolding transition of the wild-type enzyme T m occurred at 59.5 ± 0.5 C and that of Y7I occurred at 51.8 ± 0.5 C. The values of T m for each of the remaining variants of were between 49.3 and 52.5 C except Y7F which had a distinctly broad transition with a T m of 55.8 ± 0.5 C (Table 3-2). Structures of N62Q, N67Q and Y7F+N67Q HCA II The crystal structure of the three variants N62Q, N67Q, Y7F+N67Q of HCA II were solved to Å resolution and completeness greater than 90% (Figure 3-5). The final statistics of the structure including the R crysts of 17% and R frees 20% are listed in Table 3-1. Overall, no major structural perturbations were observed with C α rmsd less than 0.27 Å for all three variants when compared to the wild-type HCA II (2ili)[47]. The structures of N62Q, N67Q, Y7F+N67Q HCA II were similar to wild type in that the residues coordinating the zinc, and those in the active site cavity other than the mutated residue are unchanged (Figure 3-5). The proton shuttle residue, His64 has been shown to occupy two conformations in wild-type HCA II. The crystal structures of the mutants determined here showed only one orientation of the side chain. The outward confirmation was dominant in both the N62Q and N67Q while the inward conformation was observed in the Y7F+N67Q. In N67Q and N67Q the distance from the zinc to the proton shuttle residue was greater by about 0.5 Å more than wild type in its out conformation (Table 3-3). In the case of the Y7F+N67Q we do not see any density in the outward conformation and a hydrogen bond of 2.8 Å occurs with W2 (Table 3-3). 51

52 The hydrogen bonded solvent network in the active site cavity was altered to various extents in these mutants (Table 3-2). The N62Q structure showed additional waters in the region of W2 which have lower occupancy (0.3) and may be other conformations of W2 or a result of H64 being in the out conformation (Figure 3-5A). N67Q had a water network most similar to wild type and despite having a predominantly outward His64 orientation did not show an extra water molecule. The Y7F+N67Q mutant had a less branched water as W3a was more than 3.5 Å away from W2. This makes the network in Y7F+N67Q more similar to Y7F HCA II with His64 being within hydrogen bonding distance to W2. Discussion The goal of this work was to elucidate the role in catalysis of residues 7, 62, and 67 on the hydrophilic side of the active site cavity in HCA II, in particular its influence on the structure of ordered solvent, and on the properties of the proton shuttle His64. Despite the large changes in the N-terminal region of Y7I HCA II, the structure of the active site is remarkably similar to that of the wild type enzyme. This includes the network of ordered water molecules (Figures 3-1,3-2), unlike Y7F HCA II in which water molecule W3a is not observed. In the crystal lattice, Ile7 rests in one of the crystal contact regions. Its amide nitrogen forms a hydrogen bond with the carboxyl side chain of Glu 238 and its side chain is in hydrophobic contact with Leu 240 (Figure 3-4). These interacting amino-acids are present in a single crystal-contact patch. It is possible that the altered orientation of the N-terminus of Y7I HCA II is a consequence of crystal contacts. It is interesting to note however, that this mutant crystallizes in the same space group as the wild type enzyme. After repeated attempts to crystallize other Tyr7 mutants listed in Table 2-1, only Y7I, Y7F, and Y7F+N67Q HCA II have been 52

53 successful. It is possible that mutations at Try 7 produce a disordered (or alternate) N- terminal conformation giving rise to structural heterogeneity in the samples, which is likely to deter crystal growth. Truncation of as many as 24 residues of the N-terminal end of HCA II does not prevent the remaining protein from folding correctly, and the structure of the N-terminus has been shown to form very late in folding [51]. A comparison of the characteristics of the three structures with Tyr7 mutations reveals that Y7I, Y7F, and Y7F+N67Q HCA II all had His64 in the inward orientation. However, they differ in their values of k B, almost 3 and 7, and 15 fold greater than wild type respectively (Table 2-1). The Y7I structure was different from both the other Tyr7 mutants in that the water network was nearly the same as in the wild type possibly due to the altered N-terminus. Y7I HCA II also showed increased distance between the zinc and His64 (7.9 Ǻ) as well as between W2 and His64 (3.7 Ǻ) relative to Y7F, Y7F+N67Q, and wild type HCA II. These structural changes may increase the energetic barrier to proton transfer and explain why the increase in ΔpK a over wild type followed the theoretical Marcus in Y7I HCA II (Table 3-4). Interestingly, the values of k B for wild type HCA II and nearly all of the variants of Table 2-1, except Y7F, and Y7F+N67Q HCA II fall on or near this Marcus line (Figure 2-7). These significantly increased k B values were not simply explained the changes observed in ΔpK a by the application of Marucs theory as described in Chapter 2. The structural features in contrast to Y7I (which followed Marcus theory) were decreased distance between His64 and the zinc bound solvent, and altered water networks. The loss of W3a in 7F or the 4.3 Ǻ distance preventing hydrogen bonding between W2 and W3a in Y7F+N67Q decrease the branching at W2 and showed a possible single strand 53

54 of hydrogen bonding waters between the zinc and His64. The Y7F+N67Q structure showed a significantly shorter distance (2.8 Ǻ) from W2 to His64 than in Y7F and wild type HCA II (3.2 Ǻ) structures. It is very difficult to isolate the effect of the conformation of His64 on the rate of proton transfer. The N67Q and N62Q HCA II both showed outward conformations compared to the Try7 mutants. Despite the outward orientation of His64 these enzymes still show proton transfer rates faster than wild type. The increase in k B for N67Q follows the theoretical Marcus curve for HCA II based mainly on the change in ΔpK a. The N62Q mutant on the other hand falls slightly below the Marcus theory line which may be due to the additional water filling the additional space due to the outward conformation of His64. It is interesting that we do not see additional water in the N67Q structure to fill this space and this may indicate why the proton transfer rate followed Marcus theory more closely than the N62Q mutant. Previous substitutions at Asn62 also showed decreases in k B regardless of the conformation of His64[32]. We suggest that the reason the value of k B for Y7F and Y7F+N67Q HCA II were considerably above the Marcus line of Figure 2-7 while the other mutants lie closer to the line was due to structural changes in the proton transfer pathway that decrease the energetic barrier to proton transfer. The significant difference first observed in Y7F was a less branched, hydrogen-bonded water structure in the active site cavity (Figure 3-1). In solution these water structures have a lifetime in the picosecond range; however, the ordered water in the crystal structures provides a clue to the more stable water structures in catalysis. Computational studies of the proton transfer step in catalysis by carbonic anhydrase show that water wires consisting of fewer molecules transfer 54

55 protons more efficiently [52]. Wild-type HCA II has an identical water structure to Y7I and N67Q in the active-site cavity with a branched cluster of four ordered water molecules between His64 and the zinc bound water (Figures 3-1, 3-4). The additional water in the active site of N62Q lengthens the water network between zinc and the His64 and may reduce the proton transfer rate to 1.0 μs -1 (Table 3.4). In contrast, Y7F has a smaller, unbranched cluster of three water molecules that is more stable and provides a proton transfer pathway of lower energy barrier than wild type [52]. This was similar to the Y7F+N67Q double mutant which had a distance between W2 and W3a greater than that of a hydrogen bond (Table 3-2). This provides an explanation of why Y7F and Y7F+N67Q do not lie on the line of Figure 2-7 like wild type, as well as other mutants. The combination of more conventional hydrogen bond distances connecting the zinc solvent to His64 though only two molecules of water compared to the branched network in wild type HCA II may provide significantly more favorable pathway for proton transfer. It is interesting to speculate why Tyr7 is conserved when substitution can enhance the rate of maximal catalysis. Calorimetry showed that the replacements of Try7 in HCA II decreased the thermal stability of the protein by 7 10 C, except for Y7F which was decreased about 4 C. This decrease indicates that Tyr 7 stabilizes the enzyme, although it is unclear how it does this since the side chain of Tyr 7 has no apparent interactions with other residues in the crystal structure. At any rate, this decreased stabilization may be a factor to explain the occurrence of Tyr7 in many of the carbonic anhydrases in the α class. Tyr7 also promotes folding of the N-terminal chain, which appears to have another role. A conserved N-terminus is consistent with this region of 55

56 HCA II binding to the Cl /bicarbonate anion exchange proteins (AE) [53]. Recent deletion and mutation studies have shown the numerous histidine residues (His3,His4, His10, His15 and His 17) found in the N-terminus of HCA II represent an acidic motif important for binding the C-terminus of AE-1 and AE-2 [53]. 56

57 Table 3-1. Dataset, refinement, and final model statistics for the crystallographic study of N62Q, N67Q HCA II, N67Q+Y7F HCA II, and Y7I HCA II. a R sym = Σ I - <I> / Σ <I>. b R cryst = (Σ Fo - Fc / Σ F obs ) x 100. c R free is calculated in same manner as R cryst, except that it uses 5% of the reflection data omitted from refinement. Values in parenthesis represent highest resolution bin. N62Q N67Q Y7F+N67Q Y7I space group P2 1 P2 1 P2 1 P2 1 unit cell dimensions: a b c β (deg)] resolution (Å) highest shell ( ) ( ) ( ) ( ) number of unique reflections (3410) (3408) (3143) (2142) completeness (%) 92.5(80.5) 90.7 (80.6) 96.5 (85.8) 91.7 (92.6) redundancy 3.6 (2.5) 4.2 (3.0) 3.8 (3.6) a R symm (0.286) (0.448) (0.339) (0.349) number of protein atoms number of solvent atoms R b cryst c R free average B factor (Å 2 ): main chain side chain Zn solvent

58 Table 3-2. Thermodynamics of unfolding of Y7 variants and wild type HCA II. a Calorimetric parameters determined by DSC. T 1/2 is the temperature at the width of half-peak height Enzyme T m (ºC) a ΔH ºm (kcal mol -1 ) a Range (ºC) a T 1/2 (ºC) a wt 59.5± ± ±0.1 Y7I 51.8± ± ±0.1 Y7A 52.5± ± ±0.1 Y7W 52.0± ± ±0.1 Y7F 55.8± ± ±0.1 Y7D 52.0± ± ±0.1 Y7N 51.8± ± ±0.1 Y7R 49.3± ± ±0.1 Y7S 50.4± ± ±0.1 58

59 Table 3-3. Comparison of proposed hydrogen bond networks for wild type, Y7F, Y7F+N67Q, N67Q, N62Q, and Y7I HCA II. The distances (Å) between oxygen atoms of water molecules and hydrophilic residues in the active site of wild type (wt) HCA II (PDB ID: 2ILI), Y7F (PDB ID: 2NXT), Y7F+N67Q, N67Q, N62Q, and Y7I HCA II crystal structures. O z is the oxygen atom of the Zn-OH - /H 2 O. The 67 and 62 distances are those represented in Figuess 1-1, 3-1, and 3-5 by the red dashed line to the respective residue. The values for both the in and out conformation of His64 are listed for wild type as in/out. Distances for W4 in N62Q HCA II are not listed. wt Y7F Y7F+N67Q N67Q N62Q Y7I Zn - O z O z - O γ1 of T O z -W W1- O γ1 of T W1-W W2-W3a 2.8 n/a W3a- OH of Y7 2.8 n/a n/a n/a W2-W3b W3b W3b W2- N δ1 of H64 3.2/ O z - N δ1 of H64 7.2/ Zn N δ1 of H64 7.4/

60 Table 3-4. Comparison of the features proposed to regulate proton transfer rates in the protein environment in HCA II and selected variants. wt Y7F Y7F+N67Q N67Q N62Q Y7I k B (μs -1 ) ΔpK a Waters in network Zinc-O z to N δ His64 (Å) H64 conformation W1, W2, W3a, W3b W1, W2, W3b W1, W2, W3a, W3b W1, W2, W3a, W3b W1, W2, W3a, W3b, W4 W1, W2, W3a, W3b in/out in in out out in 60

61 A B Figure 3-1. Crystal structures of the active sites of Y7I and Y7F HCA II. A) The structure for Y7I HCA II in yellow overlay the wild type in gray. B) The structure for Y7F HCA II (PDB ID 2NXT) [21]. A ball-and-stick diagram of coordinating active site residues as labeled; the zinc ion and the oxygen molecule of waters are shown as gray and red spheres, respectively. The water network of the active-site is labeled W1, W2, etc. Presumed hydrogen bonds are represented as dashed red lines. 61

62 A B Figure 3-2. Overall comparison of crystal structures of wild-type HCA II and Y7I HCA II A) The superimposed crystal structures of wild-type HCA II and Y7I HCA II. B) The close up of the N-terminal region of superimposed crystal structures of wild-type HCA II and Y7I HCA II. The superimposed enzyme was represented as a surface except the N-terminus region was represented as a ribbon. The N-terminus of (yellow) wild type; and (green) Y7I HCA II is represented as ribbon, while the respective amino acids at position 7 as sticks. The hydrophobic and hydrophilic regions of the active-site are rendered orange and blue respectively. The active site zinc is depicted as a grey sphere. 62

63 A B Figure 3-3. Representation of crystal contacts in Y7I HCA II. A) A cartoon and stick view through a small section of the crystal lattice. B) Zoom-in of the inset in part A depicting the residues involved in crystal contacts. Dashed red line (2.9Å), hydrogen bond; dashed green lines ( 3.6Å), all other contacts. 63