Abstract Bla2 from Bacillus Anthracis

ABSTRACT

SPECTROCOPIC AND MECHANISTIC STUDIES ON METALLO-β-LACTAMASE BLA2 FROM BACILLUS ANTHRACIS

by Megan Hawk

In an effort to probe the structure, mechanism, and biochemical properties of metallo-- lactamase (EC 3.5.2.6) Bla2 from Bacillus anthracis, the enzyme was over-expressed, purified, and characterized. Metal analyses demonstrated that recombinant Bla2 tightly binds 1 equivalent of Zn(II). Steady-state kinetic studies showed that mononuclear Zn(II)-containing Bla2 (1Zn-Bla2) had the highest activity, while the dinuclear Zn(II)- containing Bla2 (ZnZn-Bla2) was unstable. However, dinuclear Co(II)-containing Bla2 (CoCo-Bla2) is more active than the mononuclear Co(II)-containing analog. UV-Vis, 1H NMR, EPR, and EXAFS spectroscopic studies were used to structurally characterize Bla2, and the resulting data show that Co(II) binding to Bla2 is cooperative, while Zn(II) binding is sequential. These spectroscopic studies were integral in determining which analog of Bla2 was used in our pre-steady state kinetic studies. 1Zn-Bla2 utilizes a two- step kinetic mechanism when nitrocefin is the substrate, while the enzyme uses a one-step kinetic mechanism when cefaclor or imipenem is used as the substrate.

SPECTROCOPIC AND MECHANISTIC STUDIES ON METALLO-β-LACTAMASE

BLA2 FROM BACILLUS ANTHRACIS

A Thesis

Submitted to the

Faculty of Miami University

in partial fulfillment of

the requirements for the degree of

Masters of Science

Department of Chemistry and Biochemistry

Megan June Hawk

Miami University

Oxford, Ohio

2008

Advisor ______Dr. Michael Crowder Reader ______Dr. Ann Hagerman Reader ______Dr. Michael Kennedy Reader ______Dr. Richard Taylor Table of Contents

Chapter 1 Introduction

1.1 Introduction 1 1.2 Antibiotic Development 1 1.3 Prokaryotic Bacteria 3 1.4 β-Lactam-containing antibiotics 3 1.5 Antibiotic resistance 5 1.6 Classification of β-lactamases 11 1.7 Metallo-β-lactamases 12 1.7.1 Classification of metallo-β-lactamases 12 1.7.2 Structure of metallo-β-lactamases 12 1.7.3 Reaction mechanism of MβL 14 1.8 Bacillus anthracis and Bla2 16 1.9 Hypothesis and description of thesis 19 1.10 References 21

ii Chapter 2 Spectroscopic and mechanistic studies on metallo--lactamase Bla2 from Bacillus anthracis

2.1 Introduction 26 2.2 Experimental Procedures 29 2.2.1 Materials 29 2.2.2 Over-expression, purification, and biochemical 29 characterization of Bla2 2.2.3 Metal analyses 30 2.2.4 Steady-state kinetic studies 30 2.2.5 Preparation of apo-Bla2 31 2.2.6 UV-Vis spectrophotometry 31 2.2.7 1H NMR spectroscopy 31 2.2.8 EPR spectroscopy 32 2.2.9 EXAFS spectroscopy 32 2.2.10 Stopped-flow UV-Vis studies 33 2.3 Results 35 2.3.1 Over-expression, purification, and biochemical 35 characterization of Bla2 2.3.2 Steady-state kinetic studies on Bla2 35 2.3.3 UV-Vis spectroscopy 36 2.3.4 1H NMR spectroscopy 40 2.3.5 EPR spectroscopy 43 2.3.6 EXAFS spectroscopy 43 2.3.7 Stopped-flow UV-Vis kinetic studies 45 2.4 Discussion 49 2.5 References 61

iii Chapter 3 Conclusions

3.1 Conclusion 65 3.2 References 70

iv List of Tables

1-1: The historic development of classes of antibiotics 4 1-2: Characteristics of different metallo-β-lactamase subgroups 13 2-1: Best fits to Co(II) and Zn(II) Bla2 EXAFS. α 34 2-2: Steady-state kinetic parametersa for nitrocefin, imipenem, cefaclor, and 38 meropenem hydrolysis by Bla2 containing 1 equivalent of Zn(II) 2-3: Steady-state kinetic parameters for Bla2 containing 1 or 2 equivalents 39 of Zn(II) or Co(II) 2-4: Kinetic constants used in KINSIM simulations 50

v List of Figures

1-1: Structures of common β-lactam antibiotics 6 1-2: Structure of D-alanyl-D-alanine 7 1-3: Cross-linking of the peptidoglycan cell 8 1-4: Hydrolysis of nitrocefin 10 1-5: Crystal structures from each of the representative metallo-β-lactamase 15 subgroups 1-6: Proposed mechanisms for MβLs 17 1-7: The crystal structure of BcII 18 2-1: SDS-PAGE gel of purification of recombinant Bla2 37 2-2: UV-Vis difference spectrum of apo-Bla2 titrated with increasing 41 amounts of Co(II) 1 2-3: H NMR spectra of 2Co(II)-Bla2 in 10% D2O and 90% D2O 42 2-4: EPR spectra from Co(II)-containing Bla2 44 2-5: Fourier transformed EXAFS spectra of Co(II)-substituted Bla2 46 2-6: Fourier transformed EXAFS spectra of Zn(II)-substituted Bla2 47 2-7: Progress curves of the reaction of nitrocefin and Bla2 containing 51 1 eq. Zn(II) at 4 oC 2-8: Progress curves of the reaction of imipenem and Bla2 containing 52 1 eq. Zn(II) at 4 oC 2-9: Progress curves of the reaction of cefaclor and Bla2 containing 53 1 eq. Zn(II) at 4 oC 2-10: The proposed active site of Bla2 after the addition of 1 or 2 59 equivalents of Zn(II) or Co(II) to apo-Bla2. 3-1: A penicillin derivative inhibitor for metallo--lactamases with a 69 phosphinate group at the -lactam carbonyl position.

vi List of Schemes

2-1: Proposed mechanism for nitrocefin 48 2-2: Proposed mechanism for imipenem and cefaclor 49

vii Acknowledgements

I would like to thank Dr. Michael Crowder for allowing me to be a part of his group. My experience at Miami University has helped me identify my weaknesses and strengths. I learned that it is important to understand why an experiment is performed and to look at past journal articles to guide you in explaining your current work. I would like to thank my group members who helped me learn proper lab and instrumentation techniques. I would also like to thank Christine Hajdin and Katie Bender who were great assets during the characterization of Bla2. My experience at Miami University gave me an opportunity to work for the Center for Chemical Education (CCE). My mentors Mickey Sarquis, Lynn Hogue, Dr. Susan Hershberger, and Ed Smith have helped me grow and become comfortable with public speaking. The work at the Center has helped me understand the importance in working as a team to achieve greatness. I hope the Center’s contributions to education will inspire the youth in Ohio and increase the number of students who focus in science. Finally, I thank my parents who have allowed me to choose my own path in life. I appreciate their love and support throughout the years.

viii Chapter 1

Introduction

1.1 Introduction

The 20th century marked an age of discovery through luck and human ingenuity. While suffering from a sinus infection in 1922, Alexander Fleming, a bacteriologist, cultured secretions from his nose. When Fleming examined his culture plate, he allowed a tear to fall on the petri dish. The next day, Fleming observed a cleared space where the tear had landed. Fleming concluded that the tear was toxic to bacteria and produced a type of antibiotic. The tear contained an enzyme called lysozyme, which breaks down bacterial cell walls and kills certain types of bacteria. The “body’s own antibiotic,” lysozyme, was found to be of little clinical importance since this enzyme could not kill potent types of bacteria. In 1928, Fleming, returned from a vacation to find a unique type of fungus growing on his culture plate. The fungus had a ring around it where bacteria did not grow. Since the fungus on the contaminated plate was from the Penicillium family, Fleming named the substance produced from the mold penicillin and found that penicillin was toxic to many strains of bacteria. Fleming repeatedly tried to isolate penicillin; however, he was not successful and concluded that penicillin could not be used as a clinical therapeutic. A few years later, Howard Florey and Ernst Chain developed a procedure to isolate and concentrate penicillin, and penicillin was subsequently shown to have medicinal purposes, particularly in fighting bacterial infections in wounded World War II soldiers. Based on this work, Fleming, Florey, and Chain were awarded the 1945 Nobel Prize in Physiology or Medicine. With the use of penicillin, the age of modern antibiotics commenced, and the penicillin family of antibiotics, which includes cephalosporins and carbapenems, is the largest class of effective and inexpensive antimicrobial agents (1) . 1.2 Antibiotic development After the discovery and clinical use of penicillin, many other antibiotics were marketed by pharmaceutical companies. For example, Eli Lilly & Co. developed and marketed antibiotics such as erythromycin, vancomycin, and cephalosporins. By the late

1 1960’s, there were numerous antibiotics that could be used in the clinic, and the U.S. Surgeon general, William H. Stewart, asserted that we should “close the book on infectious disease (2).” Over the past 40 years, only two new classes of antibiotics have emerged: one in 2000 called the oxazolidinones and the other in 2003 called the lipopeptides (Table 1-1) (3). Unfortunately from the 1980’s until the present, infectious diseases have become the 3rd leading cause of death in the world. In addition, the emergence of bacteria that are resistant to most or all known clinical antibiotics has exacerbated the problem (4). Large pharmaceutical companies like Eli Lilly & Co. lost interest in developing new antibiotics in the 1980’s and 1990’s due to low profit margins. The large pharmaceutical companies have not responded well to the reduction of useful antibiotics. The development of a new antibiotics takes an average of ten years and costs $800 million before the antibiotic enters the market (4). Once the drug is introduced into the clinic, the lifetime of the drug is very short. For example, linezolid, a non-natural product, was used in limited situations to combat vancomycin-resistant bacterial infections (5). In four months, there was clinical resistance to linezolid. Even vancomycin, the “antibiotic of last resort” originally found in soil samples from a Borneo jungle and marketed to treat penicillin-resistant bacterial infections, was ineffective at treating all bacterial infections (6). The short clinical lifetime does not allow the pharmaceutical companies to recover the high costs of developing the antibiotic. Yet another drawback to antibiotic development is attributed to physicians saving novel antibiotics as the last resort for treating various bacterial infections. When physicians save new antibiotics, it results in a loss in profit based on the antibiotics not being widely distributed to the masses. The lack of profit in antibiotic development prompted big pharmaceutical companies to focus their resources on developing drugs for chronic illnesses, which lead to higher earnings. All in all, these factors leave physicians with few options in treating bacterial infections. The current threat of antibiotic resistance and the lack of novel antibiotics on the market has prompted Cubist, Targanta Therapeutics, and Optimer Pharmaceuticals to develop derivatives of old antibiotics or to use new technologies to screen for new

2 antibiotics (3). Optimer Pharmaceuticals is currently developing feropenem, which was designed to treat Streptococcus pneumoniae infections (7). The renewed interest in developing new antibiotics by small pharmaceutical companies and continued research by academic institutions are important in the fight against infectious diseases. 1.3 Prokaryotic bacteria Antibiotics are used to inhibit bacterial cell wall synthesis, folic acid metabolism, DNA gyrase, DNA-directed RNA polymerase, and protein synthesis at the ribosome (Table 1-1). Most bacteria are 0.2 μm to 600 μm in length with chromosomal DNA filling the inside of the nucleoid (8). Gram-positive bacteria differ from Gram-negative bacteria in that the former have a cell wall made up of an inner membrane and the peptidoglycan, while the latter have an additional outer membrane. The peptidoglycan layer provides structural stability and prevents cells from bursting due to high internal osmotic pressures. The extra membrane in Gram-negative bacteria prevents many antibiotics from reaching their intracellular targets. Two increasingly serious clinical, Gram-negative pathogens are Pseudomonas aeruginosa and Acinetobacter baumannii (9). 1.4 β-Lactam containing antibiotics Over 50% of all of the clinically-available antibiotics are β-lactam containing compounds, such as penicillins, cephalosporins, and carbapenems (Figure 1-1) (10). All β-lactam containing antibiotics have a four-membered β-lactam ring; however, the ring fused next to the β-lactam ring distinguishes the different classes of β-lactam containing compounds. Penicillins have a five-membered thiazoldine ring fused to the common four-membered β-lactam ring, and cephalosporins have a six-membered dihydrothiazine ring with a sulfur atom at position five and a double bond between C2 and C3. Carbapenems have a five-membered ring like penicillin, but a carbon atom replaces the sulfur atom at position four (10, 11). In 1965 James Park discovered that penicillin inhibits the synthesis of the peptidoglycan layer. The cell wall is a network of linear polysaccharide chains crossed- linked by short peptides. The linear polysaccharide chain consists of the alternating sugars, N-acetylglucosamine (NAG) and N-acetylmuramic acid (NAM), which are connected with β-(1,4) linkages (11).

Mechanism of Year of Development Class of Antibiotic action 1935 Sulphonamides Folic acid metabolism 1941 Penicillins Cell wall synthesis 1944 Aminoglycosides 1945 Cephalosporins Cell wall synthesis 1949 Chloramphenicol Protein synthesis (50S Inhibitors) 1950 Tetracyclines Protein synthesis (30S Inhibitors) 1952 Macrolides/lincosamides/streptogramins 1956 Glycopeptides 1957 Rifamycins DNA-directed RNA polymerase 1959 Nitroimidazoles 1962 Quinolones DNA gyrase 1968 Trimethoprim Folic acid metabolism 2000 Oxazolidinones inhibits the initiation of protein synthesis 2003 Lipopeptides Inserts into the cell wall cause membrane depolarization and the release of intracellular ions resulting in cell death

Table 1-1: The historic development of classes of antibiotics

4 NAM has a lactyl side chain to which a pentapeptide chain is attached. The pentapeptide terminates in two D-alanine residues (Figure 1-2). The disaccharides of the peptidoglycan layer are cross-linked by an enzyme called transglycosylase. The pentapeptides are cross-linked by an enzyme called transpeptidase, which links the third amino acid (Lys) to the non-terminal D-ala on an adjacent pentapeptide. β-Lactam containing antibiotics are structural analogs of the D-alanyl-D-alanine terminus of the peptidoglycan pentapeptide (Figures 1-2 and 1-3) (12). Therefore, transpeptidase binds to β-lactam containing antibiotics and hydrolyzes the β-lactam bond by utilizing an active site serine for the nucleophilic attack. Since the β-lactam containing compound contains a ring, the hydrolyzed compound covalently attaches to the transpeptidase active site and irreversibly inactivates the enzyme. In other words, β-lactam containing compounds are mechanism-based (suicide) inhibitors of transpeptidases. Inactivated transpeptidase cannot crosslink the peptidoglycan layer, resulting in a weakened cell wall, which lyses when the osmotic pressure in the cell becomes too high. 1.5 Antibiotic resistance Penicillin was first introduced into the clinic to fight staph infections in wounded World War II soldiers in 1942, and by 1945, one-half of all isolated strains of S. aureus did not respond to penicillin (13). Bacteria become resistant to antibiotics by spontaneous DNA mutations, as in P. aeruginosa, which has a mutated efflux pump that exports antibiotics out of the cell, or by obtaining gene(s) (via conjugation or transformation) from organisms that are already antibiotic-resistant. The alteration in the genetic make-up of bacteria has resulted in three distinct ways that bacteria resist the antimicrobial activities of antibiotics: (1) the new gene encodes for a protein that chemically-alters the antibiotic, (2) the new gene encodes for an efflux pump that pumps the antibiotic out of the cells before it reaches its intracellular target, or (4) the new gene encodes for a protein that alters the target site (13, 11). Many Gram-negative bacteria also reduce the import of common β-lactams and other antibiotics by lowering the expression levels of porins, which transport low molecular weight species into the periplasmic space. Bacteria often become resistant to aminoglycoside antibiotics, which bind to the 30S ribosome subunit and inhibit the synthesis of vital bacterial proteins, by expressing aminoglycoside-modifying enzymes (AMEs) (13).

R S R' R S R 6 3 7 6 3 R' N 3 R'' N N R' O O O CO2 CO2 CO2 penicillin carbapenem cephalosporin

Figure 1-1: Structures of common β-lactam antibiotics

ROCHN

H3C NH

O CH 3 CO2H H D-alanyl-D-alanine

Figure 1-2: Structure of D-alanyl-D-alanine

7 NAM NAG NAM NAG

NAM NAG NAM NAG

Transpeptidase

NAM NAG NAM NAG

NAM N-acetylmuramic acid

NAG N-acetylglucosamine

Gly

Figure 1-3: Cross-linking of the peptidoglycan cell. The cross-linking of the peptidoglycan cell wall catalyzed by transpeptidase (11).

8 The most common antibiotic-resistance mechanism, observed in both Gram-negative and Gram-positive bacteria, is the over-expression of an enzyme called β-lactamase. β- Lactamases hydrolyze β-lactam containing antibiotics by cleaving the amide bond in the four-membered β-lactam ring (Figure 1-4) (11). The hydrolyzed β-lactam no longer exhibits antibacterial activity. The spread of antibiotic resistance by β-lactamases is dependent on the type of bacterium. Gram-negative bacteria export expressed β- lactamases into the periplasmic space, while Gram-positive bacteria are expressed as membrane-associated proteins (13, 11, 14). Bacterial resistance to antibiotics is due to natural selection, and the use of antibiotics in the clinic leads to antibiotic resistant phenotypes. However, the rapid emergence of antibiotic resistant bacterial strains has been caused by the improper use of antibiotics (13,11). For example, soaps and household cleaners contain antibiotics. Our environment is full of bacteria, and many of these bacteria are harmless. The use of these antibacterial products, however, selectively kills the harmless bacteria, leaving nutrients for the antibiotic resistant bacteria to thrive and drives the evolution of harmless bacteria to become antibiotic-resistant (13). The single biggest contributor to the emergence of clinical antibiotic resistance is the misuse of antibiotics by patients. Patients demand antibiotics from their physicians for all illnesses, including viral infections that do not respond to antibiotics, and the physicians often comply with their patients, rationalizing that secondary bacterial infections follow viral infections (11). Patients take their antibiotics until they feel better and then stop taking their medication. This practice once again selectively kills the “good” bacteria in their bodies, leaving the antibiotic-resistant bacteria to thrive (11). Lastly, the antibiotics are used at sub-therapeutic levels to enhance the growth of fruits, vegetables, and livestock (11). This practice has lead to the detection of antibiotics in rivers and in drinking water. All of these misuses of antibiotics have lead to the rapid emergence of antibiotic resistant phenotypes and the emergence of the “superbug.” The first instance of antibiotic resistance was reported in 1947, when Staphylococcus aureus was reported to exhibit penicillin resistance (9). As different β- lactam containing antibiotics have been introduced to combat infections caused by S. aureus, the organism developed resistance to cephalosporins and to methicillin.

S NH S NO2

O N O

COOH

NO 2

-1 -1 ε485 = 17,422 M cm

S NH S NO 2

O NH O O COOH

NO2

Figure 1-4: Hydrolysis of nitrocefin. The hydrolysis of the substrate nitrocefin, a chromogenic cephalosporin.

10 Resistance to the latter drug led to significant concerns about widespread outbreaks of superbugs like methicillin-resistant S. aureus (MRSA) (13). An example of the clinical emergence of a “superbug” occurred recently at the Walter Reed Army Medical Center. Clinically-relevant, Gram-negative Acinetobacter and Pseudomonas aeruginosa strains are currently a threat to hospitalized patients, since these strains of bacteria show multidrug resistance (15). Acinetobacter causes nosocomial infections, and this bacterium was isolated from a high number of wounded soldiers and civilians stationed in Iraq, Kuwait, or Afghanistan. Robert A. Bonomo at Case Western Reserve screened 75 isolates for antibiotic resistance genes. Bonomo’s group reported that 89% of the patients who contracted Acinetobacter infections showed multi-drug resistance to ciprofloxacin, ceftazidime, cefepime, amikacin, tobramycin, imipenem, or meropenem (15). Genes from these isolates encoded for OXA-58, OXA-23, and ABA1, which are β-lactamases that often confer carbapenem resistance. All in all, “superbugs” like Acinetobacter are resistant to a wide spectrum of antibiotics (15). 1.6 Classification of β-lactamases β-Lactamases are a diverse group of enzymes known to hydrolyze and inactivate a wide spectrum of β-lactam antibiotics. These enzymes are broadly categorized into two groups. The largest group is made up of enzymes that utilize a serine residue in the active site during the hydrolysis reaction and are called serine-β-lactamases. The other group is called metallo-β-lactamases, which require 1 or 2 Zn(II) ions for catalytic activity. Classification schemes for β-lactamases were broadened when new β- lactamases were discovered (16, 17). The first classification scheme for β-lactamases was devised by Richmond and Sykes in 1973; they classified β-lactamases found in Gram- negative bacteria (16). β-Lactamases were further classified based on the addition of isoelectric focusing results by Sykes and Matthews in 1976 (16). In 1980 Ambler placed serine-β-lactamases and MβLs into distinct classes based on their molecular structures (18). Ambler initially placed serine-β-lactamases into class A, and metallo-β-lactamases were placed into class B. Later, Ambler further divided serine-β-lactamases into class C and D, based on the enzymes’ preferences for a given substrate (19). Bush further classified β-lactamases in 1989 by dividing the β-lactamases into four groups based on preferred substrates and affinities for inhibitors. Bush placed serine-β-lactamases into

11 Groups 1, 2, and 4, while metallo-β-lactamases were placed into Group 3(20, 21). Group 3 metallo-β-lactamases, were further subgrouped by Bush in 1998 based on differences in molecular properties (22). 1.7 Metallo-β-lactamases 1.7.1 Classification of metallo-β-lactamases Metallo-β-lactamases (MβLs) were first identified in the 1960’s and isolated from a clinically-irrelevant B. cereus strain (22). Clinical inhibitors, such as clavulanic acid, sulbactam, and tazobactam, inhibit the activity of some serine-β-lactamases; however, these inhibitors do not inactivate metallo-β-lactamases. Bush classified the metallo-β- lactamases into three subgroups called Ba, Bb, and Bc (Table 1-2). Subgroup Ba comprises a group of enzymes that contain 1 or 2 Zn(II) ions in the active site and prefer penicillins as substrates. Enzymes from this subgroup include BcII from B. cereus, CcrA from B. fragilis, IMP-1 from P. aeruginosa, and Bla2 from B. anthracis. ImiS from A. sobria and CphA from A. hydrophila have 1 Zn(II) ion in the active site, belong to subgroup Bb, and are carbapenemases. An interesting characteristic of subgroup Bb enzymes is a second Zn(II) ion inhibits the activity of the enzyme. Finally, subgroup Bc includes L1 from S. maltophilia and Fez-1 from Legionella gormanii. These enzymes prefer penicillins and contain 2 Zn(II) ions (17, 22). Different subgroups of MβLs also contain different metal binding sites. The Ba enzymes have a Zn1 site consisting of His116, His118, and His196 residues and a Zn2 site consisting of Asp120, Cys221, and His263 (Table 1-2). The Bb enzymes have the same metal binding ligands as subgroup Ba enzymes, except His116 is replaced by Asn116. The Bc enzymes have the same metal binding ligands as the Ba enzymes, expect Cys221 is replaced by His112 (23). 1.7.2 Structure of metallo-β-lactamases A crystal structure of at least one member from each of the distinct MβL subgroups has been published. All MβLs have an αββα motif, which has been called the

Enzyme Source Subgroup Zn(II) Preferred Zn1 site Zn2 site content substrate BcII B. cereus Ba 1.0 or 2.0 Penicillins His116, Asp 120, His 118, Cys 221, His 196 His 263 ImiS A. sobria Bb 1.0 Carbapenems Asn 116, Asp 120, His 118, Cys 221, His 196 His 263 L1 S. maltophilia Bc 2.0 Penicillins His 116, Asp 120, His 118, His 121, His 196 His 263

Table 1-2: Characteristics of different metallo-β-lactamase subgroups.

13 β-lactamase fold. The Zn(II) ions in the MβLs bind at the interface of the αβ domains (17). The crystal structures have revealed some differences in metal binding between the enzymes from different subgroups (Figure 1-5). The Ba enzymes normally bind two zinc ions; however, the initial crystal structure of BcII showed that only one Zn(II) ion was bound to the Zn1 site (24). Subsequent structures of BcII have shown that this enzyme in fact binds two equivalents of Zn(II), and it was speculated that the initial crystal structure had low amounts of Zn(II) because the crystals were grown at low pH (26). The crystal structures of all other Ba enzymes have shown 2 equivalents of Zn(II) bound to the enzymes (19). The only crystal structure of a Bb enzyme is of CphA, and this structure showed that the enzyme binds a single Zn(II) ion in the Zn2 site (27). All of the crystal structures of Bc enzymes have shown dinuclear Zn(II) centers in these enzymes. The crystal structures have lead to the hypothesis that the Ba and Bc enzymes required Zn(II) in the Zn1 site and that the Bb enzymes require Zn(II) in the Zn2 site for catalytic activity. However, Vila and coworkers have suggested that GOB, a Bc subgroup member, binds

Zn(II) in the Zn2 site and that BcII requires Zn(II) in the Zn2 site for activity (28). Vila, however, did not address how this latter hypothesis contradicts data from his lab or data from other labs. 1.7.3 Reaction mechanism of MβLs Extensive kinetic studies on MβLs have been reported on a number of MβLs. Stopped-flow UV-Vis studies with nitrocefin as the substrate and L1 (subgroup Bc) and CcrA (subgroup Ba) MβLs revealed the formation of a ring-opened, anionic intermediate (29-31). The rate-determining step in this reaction was the protonation of the anionic intermediate. Later, Spencer et al. used clinically-relevant substrates with MβL L1 to propose that nitrocefin is the only substrate that produced the observed intermediate (32). Spencer’s work with other substrates showed that the rate-determining step was the cleavage of the amide bond in the β-lactam ring. Various mechanistic studies have been conducted on BcII (subgroup Ba); it was initially proposed that the physiological form of BcII was the mononuclear Zn(II) form (18, 33, 34). However, recent publications from Vila indicated that the mononuclear Zn(II) enzyme is the most active, while Page conducted pH dependent studies claiming that the dinuclear Zn(II) form is the only active form of BcII (26, 34-36).

Figure 1-5: Crystal structures from each of the representative metallo-β-lactamase subgroups. A crystal structures from each of the metallo-β-lactamase subgroups: Pseudomonas aeruginosa IMP-1 (PDB accession 1DDK) (subgroup Ba, left), A. hydrophila CphA (PDB accession 1X8G) (subgroup Bb, center), and S. maltophilia L1 (PDB accession 1SML) (subgroup Bc, right) (37-39). The Zn(II) ions are in gray.

15 Vila and coworkers also conducted stopped-flow UV-Vis studies on BcII similar to those earlier reported on L1 and CcrA, and their data showed that BcII does not stabilize the nitrocefin-derived, anionic intermediate (Figure 1-6) (40). Mechanistic studies on mono- Zn(II) ImiS (subgroup Bb) were used to probe the reaction mechanism of this enzyme with carbapenems (41). ImiS also does not stabilize the formation of a reaction intermediate, and the rate-limiting step is β-lactam hydrolysis. In conclusion, the rate- limiting step of all tested MβLs is amide bond cleavage, except when nitrocefin is used as the substrate. 1.8 Bacillus anthracis and Bla2 Bacillus anthracis (a.k.a. anthrax) is a Gram-positive bacterium that causes skin and lung infections in humans and animals. Numerous outbreaks of anthrax poisoning were reported in the 1800s and 1900s in Europe and the United States, due to the increased use of animal furs and hides. The spread of the disease was caused by humans handling fur or hides contaminated with Bacillus anthracis spores. Anthrax infections were contracted by inhalation, ingestion, or contact with open sores on the skin. The spread of anthrax infections and the death rates began to diminish during the 1900s because of improved hygiene and the advent of penicillin to treat infections. However, the first reported release of military-grade, inhalation anthrax occurred in 1979 in the former Soviet Union from a facility that produced biological warfare agents (42). In 2001, a bioterrorist produced military-grade anthrax spores and sent the spores in the mail to several senators in the U.S. Congress. Postal workers and workers in the Capitol became sick, and there were 10 confirmed cases of inhalation anthrax from this domestic bioterrorism tragedy (42). Antimicrobial susceptibility tests on over 50 Bacillus anthracis isolates and 15 different samples from the 2001 bioterrorism attack showed that one human isolate tested positive for a β-lactamase (43). Later, the Sterne strain of B. anthracis was screened for the presence of β-lactamases, and two chromosomally-encoded enzymes were identified: Bla1, which is a class A β-lactamase, and Bla2, is a class B metallo-β-lactamase (43).

k k1 2 k3 k4 (a.) E + S ES EI EP E + P k-1 k-2 k-3 k-4

k k1 2 k3 (b.) E + S ES1 ES2 E+P k-1 k-2

k1 k2 k k (c.) E + S 3 4 ES ES* EI2 E+P k-1 k-2 k-3

k1 k2 (d.) E + S ES E+P k-1

k k1 2 k3 k4 (e.) E + S ES ES* EI E+P k k k -1 -2 -3

Figure 1-6: Proposed mechanisms for MβLs. (a)di-Zn(II) L1 and CcrA with nitrocefin and (b) BcII with nitrocefin (c.) L1 with meropenem and cefaclor (d.) BcII with cefaclor (e.) ImiS with imipenem

Figure 1-7: The crystal structure of BcII. Three indicated amino acid residues near the active site of BcII are different in Bla2. The amino acids found in BcII (PDB accession 1BC2) are IIe39, Thr182, Glu 151, while Bla2 has Val 39, Ala 182 and Glu 151. The Zn(II) ions are green.

18 Amino acid sequence comparisons demonstrated that Bla2 shares 92% amino acid sequence homology and 89% amino acid sequence identity with BcII (44). Nonetheless, preliminary steady-state kinetic data by Palzkill suggest that Bla2 and BcII are significantly different (44). An examination of the crystal structure of BcII and the amino acid sequence comparison of BcII and Bla2 shows that there are a number of amino acid differences at or near the active site of the enzyme (Figure 1-7). It is not clear if the steady-state kinetic differences are due to Bla2 containing different amounts of Zn(II) than BcII or to operator-dependent differences. 1.9 Hypothesis and description of thesis Based on the amino acid sequence similarities, we hypothesize that Bla2 and BcII share similar structures, metal binding properties, and mechanisms. One problem with comparing results on BcII and Bla2 is that previous results on BcII are conflicting, even those from the same research labs. The conflicting, inconsistent results may be due to BcII containing different amounts of metal ions in the studies. The only reliable way to determine whether BcII and Bla2 share similar structures, metal binding properties, and reaction mechanisms is for us to characterize both enzymes in our lab. As a first step in this goal, the research described in Chapter 2 involves spectroscopic and kinetic studies on Bla2. Future studies would require similar studies on BcII in the Crowder lab. Chapter 2 describes the over-expression, purification, and characterization of Bla2 from B. anthracis. Metal analyses were used to show that the recombinant enzyme tightly binds only 1 equivalent of Zn(II). Steady-state kinetic studies showed that the mononuclear Zn(II) enzyme is the most active Zn(II)-containing analog, most likely because the addition of the second equivalent of Zn(II) causes protein degradation. In contrast, the dinuclear Co(II)-containing analog is more active than the mononuclear Co(II)-analog. UV-Vis, 1H NMR, EPR, and EXAFS spectroscopic studies were used to structurally characterize Bla2. These studies represent the first step in addressing our hypothesis. The second step in our study was proposing a reaction mechanism for Bla2. Stopped-flow UV-Vis studies under single turnover and pseudo first-order conditions were used to propose a mechanism for Bla2. The substrates used in this study were nitrocefin, imipenem, and cefaclor. The results indicated that Bla2 utilizes a rapid-

19 equilibrium, two-step binding mechanism with nitrocefin as the substrate, while Bla2 utilized a one-step binding mechanism when imipenem and cefaclor were the substrates. All in all, our studies indicated that there are distinct differences in the kinetic and biochemical properties of Bla2 when compared to the reported properties of BcII.

20 1.10 References 1. Drews, J. (2000) Drug Discovery: A Historical Perspective, Science 287, 1960- 1964. 2. Spellberg, B. (2008) Dr. William H. Stewart: Mistaken or Maligned?, Clin.. Infect. Dis. 47, 294-294. 3. Jarvis, L. M. (2008) An Uphill Battle, in Chem. Eng. News, 85, 15-20. 4. Conly, J. M., and Johnston, B. L. (2005) Where are all the new antibiotics? The new antibiotic paradox, Can. J. Infect. Dis. Med Microbiol. 16, 159-160. 5. Horne, J., Jamshed, N., and Ament, P. (2002) Linezold: its role in the treatment of Gram-positive, drug-resistant bacterial infections., Am. Fam. Physician 65, 663- 670. 6. Hubbard, B. K., and Walsh, C. T. (2003) Vancomycin assembly: Nature's way, Angew. Chem. Intern. Ed. 42, 730-765. 7. Jarvis, L. M. (2008) Imminent Threat, in Chem. Eng. News, 85, 22-24. 8. Angert, E. R., Clements, K. D., and Pace, N. R. (1993) The largest bacterium, Nature 362, 239-241. 9. Bonomo, R. A., and Szabo, D. (2006) Mechanisms of multidrug resistance in Acinetobacter species and Pseudomonas aeruginosa, Clin. Infect. Dis. 43, S49- S56. 10. Page, M. I., and Laws, A. P. (1998) The mechanism of catalysis and the inhibition of β-lactamases, Chem. Comm., 1609-1617. 11. Walsh, C. (2000) Molecular mechanisms that confer antibacterial drug resistance, Nature 406, 775-781. 12. Proctor, P., Gensmantel, N., and Page, M. (1982) "The Chemical-Reactivity of Penicillins and other β -lactam antibiotics.", J. Am. Chem. Soc., 1185-1192. 13. Neu, H. C. (1992) The Crisis in Antibiotic-Resistance, Science 257, 1064-1073. 14. Walsh, T. R., Toleman, M. A., Poirel, L., and Nordmann, P. (2005) Metallo- β - lactamases: the quiet before the storm?, Clin. Microbiol. Rev. 18, 306-+. 15. Hujer, K. M., Hujer, A. M., Hulten, E. A., Bajaksouzian, S., Adams, J. M., Donskey, C. J., Ecker, D. J., Massire, C., Eshoo, M. W., Sampath, R., Thomson, J. M., Rather, P. N., Craft, D. W., Fishbain, J. T., Ewell, A. J., Jacobs, M. R.,

21 Paterson, D. L., and Bonomo, R. A. (2006) Analysis of antibiotic resistance genes in multidrug-resistant Acinetobacter isolates from military and civilian patients treated at the Walter Reed Army Medical Center, Antimicro. Agents Chemo. 50, 4114-4123. 16. Bush, K. (1989) Characterization of β -Lactamases, Antimicro. Agents Chemo. 33, 259-263. 17. Crowder, M. W., Spencer, J., and Vila, A. J. (2006) Metallo- β -lactamases: Novel weaponry for antibiotic resistance in bacteria, Acc. Chem. Res. 39, 721-728. 18. Ambler, R. P. (1980) The Structure of β -Lactamases, Phil. Trans. Royal Soc. London B 289, 321-331. 19. Hall, B. G., and Barlow, M. (2005) Revised Ambler classification of β - lactamases, J. Antimicro. Chemo. 55, 1050-1051. 20. Bush, K. (1989) Classification of β -Lactamases - Group-1, Group-2a, Group-2b, and Group-2b', Antimicro. Agents Chemo. 33, 264-270. 21. Bush, K. (1989) Classification of β -Lactamases - Group-2c, Group-2d, Group-2e, Group-3, and Group-4, Antimicro. Agents Chemo. 33, 271-276. 22. Bush, K. (1998) Metallo- β -lactamases: A class apart, Clin. Infect. Dis. 27, S48- S53. 23. Galleni, M., Lamotte-Brasseur, J., Rossolini, G. M., Spencer, J., Dideberg, O., and Frere, J. M. (2001) Standard numbering scheme for class B β -lactamases, Antimicro. Agents Chemo. 45, 660-663. 24. Fabiane, S. M., Sohi, M. K., Wan, T., Payne, D. J., Bateson, J. H., Mitchell, T., and Sutton, B. J. (1998) Crystal structure of the zinc-dependent β -lactamase from Bacillus cereus at 1.9 angstrom resolution: Binuclear active site with features of a mononuclear enzyme, Biochemistry 37, 12404-12411. 26. Badarau, A., and Page, M. I. (2006) Enzyme deactivation due to metal-ion dissociation during turnover of the cobalt- β -lactamase catalyzed hydrolysis of β -lactams, Biochemistry 45, 11012-11020. 27. Garau, G., Lemaire, D., Vernet, T., Dideberg, O., and Di Guilmi, A. M. (2005) Crystal structure of phosphorylcholine esterase domain of the virulence factor choline-binding protein E from Streptococcus pneumoniae - New structural

22 features among the metallo- β -lactamase superfamily, J. Biol. Chem. 280, 28591- 28600. 28. Moran-Barrio, J., Gonzalez, J. M., Lisa, M. N., Costello, A. L., Dal Peraro, M., Carloni, P., Bennett, B., Tierney, D. L., Limansky, A. S., Viale, A. M., and Vila, A. J. (2007) The metallo- β -lactamase GOB is a mono-Zn(II) enzyme with a novel active site, J. Biol. Chem. 282, 18286-18293. 29. McManus-Munoz, S., and Crowder, M. W. (1999) Kinetic mechanism of metallo- β -lactamase L1 from Stenotrophomonas maltophilia, Biochemistry 38, 1547- 1553. 30. Wang, Z. G., and Benkovic, S. J. (1998) Purification, characterization, and kinetic studies of a soluble Bacteroides fragilis metallo- β -lactamase that provides multiple antibiotic resistance, J. Biol. Chem. 273, 22402-22408. 31. Wang, Z. G., Fast, W., and Benkovic, S. J. (1999) On the mechanism of the metallo- β -lactamase from Bacteroides fragilis, Biochemistry 38, 10013-10023. 32. Spencer, J., Clarke, A. R., and Walsh, T. R. (2001) Novel mechanism of hydrolysis of therapeutic β -lactams by Stenotrophomonas maltophilia L1 metallo- β -lactamase, J. Biol. Chem. 276, 33638-33644. 33. Rasia, R. M., and Vila, A. J. (2002) Exploring the role and the binding affinity of a second zinc equivalent in B. cereus metallo- β -lactamase, Biochemistry 41, 1853-1860. 34. Badarau, A., Damblon, C., and Page, M. I. (2007) The activity of the dinuclear cobalt- β -lactamase from Bacillus cereus in catalyzing the hydrolysis of β - lactams, Biochem. J. 401, 197-203. 35. Badarau, A., and Page, M. I. (2006) The variation of catalytic efficiency of Bacillus cereus metallo- β -lactamase with different active site metal ions, Biochemistry 45, 10654-10666. 36. Gonzalez, J. M., Martin, F. J. M., Costello, A. L., Tierney, D. L., and Vila, A. J.

(2007) The Zn2 position in metallo- β -lactamases is critical for activity: A study on chimeric metal sites on a conserved protein scaffold, J. Mol. Biol. 373, 1141- 1156.

23 37. Concha, N. O., Janson, C. A., Rowling, P., Pearson, S., Cheever, C. A., Clarke, B. P., Lewis, C., Galleni, M., Frere, J. M., Payne, D. J., Bateson, J. H., and Abdel- Meguid, S. S. (2000) Crystal structure of the IMP-1 metallo- β -lactamase from Pseudomonas aeruginosa and its complex with a mercaptocarboxylate inhibitor: Binding determinants of a potent, broad-spectrum inhibitor, Biochemistry 39, 4288-4298. 38. Garau, G., Bebrone, C., Anne, C., Galleni, M., Frere, J. M., and Dideberg, O. (2005) A metallo- β -lactamase enzyme in action: Crystal structures of the monozinc carbapenemase CphA and its complex with biapenem, J. Mol. Biol. 345, 785-795. 39. Ullah, J. H., Walsh, T. R., Taylor, I. A., Emery, D. C., Verma, C. S., Gamblin, S. J., and Spencer, J. (1998) The crystal structure of the L1 metallo- β -lactamase from Stenotrophomonas maltophilia at 1.7 angstrom resolution, J. Mol. Biol. 284, 125-136. 40. Rasia, R. M., and Vila, A. J. (2003) Mechanistic study of the hydrolysis of nitrocefin mediated by B.cereus metallo- β -lactamase, ARKIVOC, 507-516. 41. Sharma, N. P., Hajdin, C., Chandrasekar, S., Bennett, B., Yang, K. W., and Crowder, M. W. (2006) Mechanistic studies on the mononuclear Zn-II-containing metallo- β -lactamase ImiS from Aeromonas sobria, Biochemistry 45, 10729- 10738. 42. Jernigan, J. A., Stephens, D. S., Ashford, D. A., Omenaca, C., Topiel, M. S., Galbraith, M., Tapper, M., Fisk, T. L., Zaki, S., Popovic, T., Meyer, R. F., Quinn, C. P., Harper, S. A., Fridkin, S. K., Sejvar, J. J., Shepard, C. W., McConnell, M., Guarner, J., Shieh, W. J., Malecki, J. M., Gerberding, J. L., Hughes, J. M., and Perkins, B. A. (2001) Bioterrorism-related inhalational anthrax: The first 10 cases reported in the United States, Emerging Infect. Dis. 7, 933-944. 43. Chen, Y. H., Succi, J., Tenover, F. C., and Koehler, T. M. (2003) β -lactamase genes of the penicillin-susceptible Bacillus anthracis sterne strain, J. Bacteriol. 185, 823-830.

24 44. Materon, I. C., Queenan, A. M., Koehler, T. M., Bush, K., and Palzkill, T. (2003) Biochemical characterization of β -lactamases bla1 and bla2 from Bacillus anthracis, Antimicro. Agents Chemo. 47, 2040-2042.

25 Chapter 2

Spectroscopic and mechanistic studies on metallo-β- lactamase Bla2 from Bacillus anthracis

The initial over-expression of Bla2 in Dr. Crowder’s lab was conducted by Dr. Zhenxin Hu. Dr. Brian Bennett and Dr. Zhenxin Hu ran the EPR samples at Milwaukee, Wisconsin. Dr. Brian Bennett performed computer simulations on the EPR spectra and wrote the initial draft of the EPR section in chapter 2. Dr. David L. Tierney and Matt Breece ran the EXAFS samples at the National Synchrotron Light Source (NSLS). Matt Breece performed computer simulations on the EXAFS samples and wrote the initial draft of the EXAFS section in chapter 2. Megan Hawk guided both Christine Hajdin and Katherine Bender in the protein over-expression and kinetic studies. The initial over- expression used by Zhenxin Hu was optimized by Megan Hawk and taught to Christine and Katherine. All protein samples used in the spectroscopic studies were prepared by Megan, Christine and Katherine. Megan Hawk performed the UV-Vis and 1H NMR studies along with multiple steady-state and per-steady state kinetic studies. Christine Hajdin ran computer simulations to propose the kinetic mechanism for Bla2. Finally, the sections written by Dr. Bennett and Dr. Tierney were modified by Dr. Michael Crowder and Megan Hawk. 2.1 Introduction Anthrax is a disease that affects cattle and other herbivores and is caused by the spore-forming bacterium Bacillus anthracis. In humans, anthrax infections, which are most often caused by contact with infected animals, lead to blackish-colored pustules on the skin. However, anthrax infections can also be caused by ingestion or inhalation of B. anthracis spores found in contaminated animals. Since anthrax infections can be caused by airborne transmissions and since anthrax has been developed as a possible biological warfare agent, B. anthracis has been labeled as a category A bioterrorism agent by the Centers of Disease Control and Prevention (CDC). Two significant outbreaks of anthrax

infections have been reported due to non-environmental sources. An anthrax outbreak in Sverdlovsk, USSR in 1979 was due to an accidental release of military-grade anthrax that caused inhalation anthrax infections (1). This outbreak was originally reported to be due to the ingestion of contaminated meat; however, President Borin Yeltsin later admitted the outbreak was caused by the accidental release of the pathogen from a bio-military facility. In October, 2001, the Processing and Distribution Center (DCPDC) of the U.S. Postal Service (USPS) and several congressional offices in Washington, D.C. were contaminated with a highly-processed form of anthrax that was sent in several envelopes (1). This anthrax was from the Ames strain, which was tested for weapons by the U.S. military. This bioterrorism incident exposed more than 2,000 workers and lead to 10 cases of inhalation anthrax and 5 deaths (1, 2). The Centers for Disease Control and Prevention (CDC) recommend that physicians prescribe a 60 day antimicrobial regiment of ciprofloxacin, doxycycline, or amoxicillin for anthrax infections. Cephalosporins, trimethoprim, or sulfamethoxazole were not prescribed in 2001 due to the potential development of bacterial resistance to these drugs (1). Over the past seven years, the FBI has been investigating scientists who were working at Fort Detrick, Md on an anthrax vaccine. Dr. Bruce E. Ivins, the FBI’s main suspect, committed suicide on August 5, 2008. Dr. Ivins had been working at Fort Detrick, Md. for 35 years, and for the past decade, he had been performing experiments on an effective anthrax vaccine (2). Dr. Ivins was facing criminal charges for the 2001 anthrax attack. Dr. Ivins’s death has not resulted in the official closing of the 2001 anthrax case by the FBI (2). Over 50% of the antibiotics prescribed by physicians are β-lactam-containing compounds (simply called β-lactams), such as penicillins, cephalosporins, and carbapenems (3). However, there is an increase in the number of patients who develop infections that are resistant to the β-lactams. The most common pathway for resistance is the bacterial production of β-lactamases, which hydrolyze the invariant four-membered ring in β-lactams and render the antibiotics ineffective. To date, there have been over 500 β-lactamases found in clinical settings, and these enzymes have been classified in four groups: A, B, C, and D. While exhibiting different kinetic and inhibition properties, group A, C, and D β-lactamases are similar in the fact that they all utilize an active site serine in the reaction to hydrolyze the β-lactam bond (4-6). The group A, C, and D β-

lactamases are currently the most clinically-significant, and there are some clinical inhibitors that are active towards many of the enzymes in these groups. The group B β- lactamases require 1-2 Zn(II) ions for catalytic activity, and these enzymes are called metallo-β-lactamases (mβl’s) (4, 7). There have been about 40 mβls reported in the literature, and these enzymes have been further classified into 3 subgroups. The B1 enzymes require 1-2 Zn(II) for full activity, prefer penicillins as substrates, bind 1 Zn(II) in the Zn1 site, which is made up of His116, His118, His196, and a bridging hydroxide.

The B1 enzymes bind a second Zn(II) in the Zn2 site, which is made up of His263, Asp120, Cys221, a bridging hydroxide, and a terminally-bound water, and are represented by BcII from Bacillus cereus, CcrA from Bacteroides fragilis, and IMP-1 from Pseudomonas aeruginosa (8). The B2 enzymes require 1 Zn(II) for full activity, prefer carbapenems as substrates, bind one Zn(II) in the Zn2 site (see above), and are represented by CphA from Aeromonas hydrophila and ImiS from Aeromonas sobria (7). The B3 enzymes require 2 Zn(II)’s for full activity, prefer penicillins as substrates, bind

one Zn(II) in the Zn1 site (see above), bind one Zn(II) in a Zn2 site made up of Asp120, His121, His263, a bridging hydroxide, and a terminally-bound water, and are represented by L1 from Stenotrophomonas maltophilia and FEZ-1 from Legionella gormanii (8). Recently, a new member of the B3 subclass has been reported, and unlike the other B3

members, GOB from Elizabethkingia meningoseptica only utilizes the Zn2 site (9). The mβls have been extensively studied using structural, computational, and mechanistic studies. Recently, the Sterne strain of Bacillus anthracis has been reported to produce a class A (Bla1) and a class B (Bla2) β-lactamase (10). Although the Sterne strain is not one of the strains (Ames and Sverdlovsk strains) that caused human infections, the transfer of the Bla2 gene from the Sterne strain to a more pathogenic Bacillus strain is expected to be facile (10). Bla2 shares 89% amino acid sequence identity and 92% amino acid sequence homology with BcII from B. cereus¸ and it would seem that all structural, mechanistic, and computational studies already reported for BcII would be applicable to Bla2 (10, 11). In spite of having very similar amino acid sequences, steady-state kinetic constants reported for Bla2 and BcII are different, and several active site amino acids (Ile39, Thr182 and Gly151) in BcII are not conserved in Bla2. In addition, there is a

great deal of conflicting data on BcII, and it is not clear which set of data applies to Bla2 (12). Therefore, we performed studies to characterize recombinant Bla2 from B. anthracis. The metal content of the enzyme was ascertained by using ICP-AES, and the steady-state kinetic constants of various analogs of the enzyme were determined. To probe the structure of Bla2, the Co(II)-substituted analog was prepared and characterized using UV-Vis, 1H NMR, EPR, and EXAFS spectroscopies. The spectroscopic studies were used in determining that 1Zn(II)-Bla2 had a higher stability than ZnZn-Bla2. The 1Zn(II)-Bla2 was used due to its stability for pre-steady state kinetic studies with various substrates. These results indicate that Co(II) binds differentially to Bla2 than Zn(II), suggesting that caution should be used when extrapolating spectroscopic data on Co(II)- substituted proteins to the corresponding Zn(II)-containing analogs. 2.2 Experimental Procedures 2.2.1 Materials The pET24bOmpA-Bla2 over-expression plasmid was generously supplied by Professor Timothy Palzkill of the Baylor School of Medicine, and E. coli strain BL21(DE3) was purchased from Novagen, Madison, WI. A Minitan II concentrator system was purchased from Fisher Scientific, Pittsburgh, PA, with 1000 Nominal Molecular Weight Limits (NMWL) plates from Millipore, Bedford, MA. All chromatographic steps were conducted with a fast protein liquid chromatography (FPLC) system purchased from Amersham Pharmacia Biotech. Nitrocefin and cefaclor were purchased from Becton Dickinson (Franklin Lakes, NJ) and Sigma (St. Louis, MO), respectively. Meropenem and imipenem were donated by Zeneca Pharmaceuticals, Wilmington, DE and Merck and Co., Rahway, NJ, respectively. 2.2.2 Over-expression, purification, and biochemical characterization of Bla2 The pET24b-OmpA-Bla2 plasmid was transformed into E. coli BL21(DE3) cells via electroporation. Colonies from the transformation were used to inoculate 10 mL of LB medium containing 25 μg/mL kanamycin, and the culture was shaken overnight at 37 oC (11). The overnight culture was used to inoculate 4X1L of LB medium containing 25 μg/ml kanamycin at 37 ◦C. The cells were allowed to shake at 37 ◦C until reaching an optical density at 600 nm of 0.75-0.85. Protein production was induced by making the

cultures 1 mM in isopropyl-β-D-thiogalactopyranoside (IPTG), and the cultures were allowed to shake overnight at 25 ◦C. The resulting cells were centrifuged (10 min at 7500 xg), and the supernatant was collected. The protein solution was kept on ice and concentrated with a Minitan II system equipped with 1000 NMWL plates until the volume reached ca. 50 mL (13). The concentrated protein mixture was centrifuged (30 min at 14500 xg), and the supernatant was dialyzed versus 2L of 50 mM 4-(2- hydroxymethyl)-1-piperazineethanesulfonic acid (Hepes), pH 6.5, overnight at 4 ◦C. The supernatant was centrifuged (30 min at 14,500 xg) to remove insoluble matter and loaded onto a SP-Sepharose column (1.5 X 12 cm with a 25-mL bed volume), which was pre- equilibrated with 50 mM Hepes, pH 6.5. Bound proteins were eluted with a 0-1 M NaCl gradient in 50 mM Hepes, pH 6.5, at 2 ml/min. Fractions (6 mL) containing Bla2, as determined by SDS-PAGE gels, were pooled and concentrated with an Amicon ultrafiltration cell equipped with a YM-10 membrane. SDS-PAGE gels were used to ascertain protein purity. The molecular mass of a 10 μM sample of recombinant Bla2 in 50 mM Hepes, pH 6.5, was determined by MALDI-TOF mass spectrometry. Amino acid analyses were conducted by the Protein Separation and Analysis Laboratory at Purdue University, West

Lafayette, IN and used to determine the molar extinction coefficient (ε280) of Bla2. 2.2.3 Metal analyses The metal content of Bla2 samples was determined by using a Varian Liberty 150 inductively coupled plasma spectrometer with atomic emission spectroscopy detection (ICP-AES). Bla2 samples were diluted to 10 μM or 20 μM in 50 mM Hepes, pH 6.5. Calibration curves of five standards (Zn(II), Cu(II), Ni(II), Mn(II), Co(II), and Fe) were prepared, and calibration curves with a correlation coefficients of 0.9999 or better were used to quantitative the metal ions in the Bla2 samples, as previously reported (14). 2.2.4 Steady-state kinetic studies Steady-state kinetic studies were conducted at 25 ◦C in 50 mM Hepes, pH 6.5, on a Hewlett-Packard model 5480A diode array UV-Vis spectrophotometer. The molar -1 -1 absorptivities of the antibiotics used were Δε485nm = 17,400 M cm for nitrocefin, -1 -1 -1 -1 Δε280nm = -6,410 M cm for cefaclor, Δε305nm = 7,600 M cm for meropenem, and -1 -1 Δε300nm = -9000 M cm for imipenem (14). The substrate concentrations ranged from 10

μM to 200 μM, and the enzyme concentration was ca. 10 nM. Steady-state kinetic constants were determined as previously described (14). 2.2.5 Preparation of apo-Bla2 Bla2 (1-2 mL of 1 mM enzyme) was dialyzed versus 4 X 2L (12 hours each step) of 15 mM Hepes, pH 6.5, containing 10 mM EDTA at 4 ◦C. The EDTA was removed by dialysis versus 2 X 2L of Chelex-treated, 15 mM Hepes, pH 6.5, containing 150 mM NaCl and 2 X 2L of Chelex-treated, 15 mM Hepes, pH 6.5, containing 100 mM NaCl for 6 h at 4 ◦C. After the last dialysis step, the protein solution was centrifuged (30 min at 14,500 xg) to remove precipitated protein and concentrated by using ultrafiltration to 1 to 2 mL. The resulting solution was passed through a Sephadex G-25 column (1.5 x 68 cm column; bed volume 120 mL), equilibrated with Chelex-treated 15 mM Hepes, pH 6.5, containing 100 mM NaCl. Fractions (6 mL) were collected, and the samples containing Bla2, as determined by SDS-PAGE, were pooled and concentrated with ultrafiltration. The metal content of apo-Bla2 was determined by ICP-AES (15). 2.2.6 UV-Vis spectrophotometry Apo-Bla2 (ca. 1 mM) in Chelex-treated, 15 mM Hepes, pH 6.5, containing 100

mM NaCl was titrated with CoCl2 (Strem Chemicals, 99.999% pure). The samples of Bla2 containing 1 and 2 equivalents of Co(II) were incubated on ice for 5 min. before being centrifuged (10 minutes at 14,500 x g) to remove precipitated protein. Difference spectra of the Co(II)-Bla2 samples were obtained by subtracting the spectrum of apo- Bla2 from those of the Co(II)-added samples (16, 17). 2.2.7 1H NMR spectroscopy NMR spectra were collected on a Bruker Avance 500 Spectrometer at 500.13MHz, 298K, and a magnetic field of 11.7T. The concentration of Bla2 samples was 1 mM, and the buffer was Chelex-treated, 15 mM Hepes, pH. 6.5, containing 100

mM NaCl and 10% or 90% D2O. All samples were incubated for 30 min on ice after D2O addition and centrifuged, before being placed in a 5 mm Wilmad NMR tube. The spectra were collected using a presaturation pulse sequence (zgpr) for water suppression. A recycle delay (AQ) of 41 ms, a sweep width of 400 ppm or 800ppm, a receiver gain of 129, and line broadening of 80 Hz were used to collect and process the spectra, as previously described (16, 17).

2.2.8 EPR spectroscopy EPR spectra were recorded using a Bruker EleXsys E600 EPR spectrometer equipped with an Oxford Instruments ITC4 temperature controller and an ESR-900 helium flow cryostat. A Bruker ER-4116DM cavity was used, with a resonant frequency of 9.63 GHz (in perpendicular mode) and 10 G (1 mT) field modulation at 100 kHz was employed. Other recording conditions are given in the legend to Figure 4. Computer simulations of EPR spectra were carried out using the matrix diagonalization program XSophe (Bruker Biospin GmbH; (18)) assuming a spin Hamiltonian H = βg.B.S + 3 1 S.D.S, where S = /2 and D > 0 corresponds to an MS = |± /2〉 ground state Kramers’ doublet, as previously described (16, 17). 2.2.9 EXAFS spectroscopy Samples of Bla2 (~ 1 mM, including 20% (v/v) glycerol added as a glassing agent) were loaded in Lucite cuvettes with 6 μm polypropylene windows and frozen rapidly in liquid nitrogen. X-ray absorption spectra were measured at the National Synchrotron Light Source (NSLS), beamline X3B, with a Si(111) double crystal monochromator; harmonic rejection was accomplished using a Ni focusing mirror. Fluorescence excitation spectra for all samples were measured with a 13-element solid- state Ge detector array. Samples were held at ~ 15 K in a Displex cryostat during XAS measurements. X-ray energies were calibrated by reference to the absorption spectrum of the appropriate metal foil, measured concurrently with the protein spectra. The data in Figure 2-5 (Co) represent the average of 30 scans from four independently prepared samples each; the data in Figure 2-6 (Zn) represent the average of 24 scans on three independently prepared samples each. Data collection and reduction were performed according to published

procedures,(19) with E0 set to 9680 eV for Zn and 7725 eV for Co. The Fourier filtered first shell EXAFS were fitted to equation (1) using the nonlinear least-squares engine of IFEFFIT that is distributed with SixPack (20). )( SkAN k)( csas k 2 2 R kkR )](2sin[)/2exp()2exp( χ = ∑ 2 − σ as − as λ +φasas (1) kRas

In Equation (1), Ns is the number of scatterers within a given radius (Ras, ± σas),

As(k) is the backscattering amplitude of the absorber-scatterer (as) pair, Sc is a scale

factor, φas k)( is the phase shift experienced by the photoelectron, λ is the photoelectron mean free-path, and the sum is taken over all shells of scatterering atoms included in the

fit. Theoretical amplitude and phase functions, As(k)exp(-2Ras/λ) and φas k)( , were

calculated using FEFF v. 8.00 (21). Optimal values for the Zn-N Sc (0.78) and ΔE0 (-21

eV), the Zn-S Sc (0.85) and ΔE0 (-21 eV), the Co-N Sc (0.74) and ΔE0 (-11 eV), the Co-S

Sc (0.85) and ΔE0 (-11 eV), were determined previously (17, 19) and held fixed throughout this analysis. Fits to the current data were obtained for all reasonable 2 coordination numbers, varying Ras and σas for each shell. Multiple scattering contributions from histidine ligands were approximated according to published procedures, fixing the number of imidazole ligands per zinc ion at half-integral values 2 while varying Ras and σas for each of the combined pathways (Table 2-1). Metal- metal (Zn-Zn) scattering was modeled by fitting calculated amplitude and phase

functions to the experimental EXAFS of Zn2(sapln)2. 2.2.10 Stopped-flow UV-Vis studies Diode array UV-Vis spectra between 200 – 700 nm were collected at 4 oC on an Applied Photophysics SX.18-MVR stopped flow spectrophotometer. The buffer used in these experiments was 50 mM Hepes, pH 7.5, and the substrates were nitrocefin, cefaclor, and imipenem. Substrate concentrations ranged from 10 – 50 μM, and 1Zn-Bla2 was held constant at 20 μM. All experiments were conducted in triplicate, and the data from the experiments were averaged. Stopped-flow absorbance data were converted to concentration data as previously described (22), and data were corrected for the instrument dead time of 2 ms. The resulting progress curves were fitted to single

exponential equations, and kobs values were plotted against substrate concentration. When using nitrocefin as substrate, the plots were hyperbolic and fitted to kobs = K1[S]k2 /

(K1[S] + k-2), as previously described (23). When using cefaclor and imipenem as

substrates, the plots were linear and fitted to kobs = k1[S] + k-1. The progress curves were simulated using KINSIM, as previously described. Dynafit was used to verify the simulations, and ProK was used to obtain errors for the rate constants, as previously described (22, 24, 25).

Table 2-1: Best fits to Co(II) and Zn(II) Bla2 EXAFS. a

b c Model M-N/O M-S M-His M-M Rf

1Co-Bla2 2.80 (10), 3.07 (6.8) 2.06 (4.0) 64 4 N/O (2 His) 4.09 (6.2), 4.46 (16)

2.81 (10), 3.11 (5.0) CoCo-Bla2 2.04 (6.5) 89 4.05 (3.6), 4.44 (7.2) 4 N/O (3 His) 1Zn-Bla2 2.92 (5.0), 3.18 (3.4) 2.02 (4.6) 37 4 N/O (3 His) 4.10 (17), 4.42 (22)

ZnZn-Bla2 2.89 (9.1), 3.09 (3.0) 4 N/O (2.5 His) 2.04 (6.5) 2.28 (1.4) 3.44 (5.4) 17 4.05 (21), 4.43 (22) + 0.5 S + Zn-Zn

a Distances (Å) and disorder parameters (in parentheses, σ2 (10-3 Å2)) shown derive from integer or half-integer coordination number fits to filtered EXAFS data [k = 1-14 Å-1; R = 0.3-4.3 Å] b Multiple scattering paths represent combined paths, as described previously (see Materials and Methods)

N 2 2 ∑ {[ χi ] + []χi )Im()Re( } c calc calc Goodness of fit (Rf for fits to filtered data) defined as 1000* i=1 , where N χ 2 + χ )Im()Re( 2 ∑ {}[][]iobs iobs i=1 N is the number of data points

2.3 Results 2.3.1 Over-expression, purification, and biochemical characterization of Bla2 Previously, Palzkill and coworkers designed a pET24-based over-expression plasmid that produces an OmpA-Bla2 fusion protein (11). The addition of the OmpA protein at the N-terminus of Bla2 causes the fusion protein to be exported initially into the periplasm of E. coli. The subsequent insertion of OmpA into the outer membrane and cleavage of the fusion protein cause Bla2 to be exported ultimately into the growth medium. By using the procedure described in the Materials and Methods, we were able to obtain 17-23 mg of >95% pure Bla2 from 1 liter of LB medium (Figure 2-1). Amino acid analyses were used to determine the extinction coefficient (ε280nm) of Bla2, which is 25,400 M-1cm-1. MALDI-TOF mass spectrometry demonstrated that recombinant Bla2 exhibits a m/z of 24,249, which is consistent with the predicted molecular mass of 25,360 amu determined from the amino acid sequence. After purification and concentration of recombinant Bla2, metal analyses were used to demonstrate that the enzyme binds 1.0 + 0.2 equivalents of Zn(II) and that there are insignificant amounts (<0.1 equivalents) of Co(II), Cu(II), Ni(II), Mn(II), or Fe. To determine whether Bla2 lost metal ion during the purification/concentration protocol, the as-isolated enzyme was incubated with 4 equivalents of Zn(II), and loosely-bound and unbound Zn(II) was removed from the sample by 3 X 1L dialysis steps versus Chelexed- treated, 50 mM Hepes, pH 6.5, for 3 days at 4 ◦C. The resulting enzyme was shown to tightly bind 1.4 + 0.1 equivalents of Zn(II). The incubation of Bla2 with higher equivalents of Zn(II) resulted in irreversible protein precipitation. 2.3.2 Steady-state kinetic studies on Bla2 To evaluate the catalytic activity of recombinant Bla2, steady-state kinetic studies were conducted in 50 mM Hepes, pH 6.5 (Table 2-2), using nitrocefin, imipenem, cefaclor, and meropenem as substrates. Previous kinetic studies by Palzkill were conducted in 50 mM Hepes, pH 7.5, containing 50 μM Zn(II) and 20 μg bovine serum albumin (BSA) (11). Our initial steady-state kinetic studies were conducted using the same conditions; however, we found that the enzyme was extremely unstable, and it was impossible to obtain reproducible data. The inclusion of BSA did not stabilize the enzyme in our hands. Since the predicted pI of Bla2 is 7.6, we speculated that the protein

was precipitating because it is zwitterionic at pH 7.5. Therefore, we lowered the pH of the assay buffer to 6.5, and we were able to obtain reproducible kinetic data. The as-isolated Bla2, which we designate 1Zn-Bla2, hydrolyzed nitrocefin, -1 imipenem, cefaclor, and meropenem with kcat values ranging from 24 to 92 s and Km values ranging from 25 to 110 μM (Table 2-2). To ascertain whether Bla2 exhibits the greatest catalytic activity when containing 1 or 2 equivalents of metal, metal-free Bla2 was prepared by dialyzing the as-isolated enzyme versus EDTA, followed by exhaustive dialysis steps to remove EDTA (see Materials and Methods). ICP-AES measurements demonstrated that apo-Bla2 did not bind any detectable amounts of Zn(II), Co(II), Mn(II), or Fe. Zn(II) was added directly to the apo-enzyme to generate the 1Zn- and ZnZn-analogs of Bla2, and these analogs were used in steady state kinetic studies at pH -1 6.5 using nitrocefin as the substrate (Table 2-3). 1Zn-Bla2 exhibited a kcat of 32 s and a

Km of 28 μM; the kcat of this enzyme was ca. 30% less than that of as-isolated Bla2 (Table

2-2). ZnZn-Bla2 exhibited an increase in kcat and in Km, and the kcat did not double when the second equivalent of Zn(II) was added. The addition of more than 2 equivalents of

Zn(II) to Bla2 resulted in a drop in the kcat, suggesting that the enzyme denatures when more Zn(II) is added to the enzyme. These same studies were also conducted with Co(II), -1 and 1Co-Bla2 exhibited a kcat of 19 s and a Km of 18 μM, while CoCo-Bla2 exhibited a -1 kcat of 47 s and a Km of 30 μM. 2.3.3 UV-Vis spectroscopy To probe the structure of the active site of Bla2, apo-Bla2 was titrated with Co(II), and UV-Vis studies were conducted. The UV-Vis studies were used to determine the coordination number and ligation environment of Co(II)-substituted Bla2. The UV-Vis difference spectrum of Bla2 (spectrum of sample minus the spectrum of apo-Bla2)

containing 0.5 equivalents of Co(II) revealed the presence of a peak at 320 nm (ε320 = 608 M-1cm-1), which we assign to a Cys-S to Co(II) ligand to metal charge transfer transition (Figure 2-2). There was also a broad peak from 500 – 650 nm, which we assign to ligand field transitions of high-spin Co(II). Upon further addition of an additional 0.5 eq Co(II), and then 1.0 eq Co(II), the absorbance of each of the features increased linearly as a function of added Co(II) (Figure 2-2 inset). These results suggest Co(II) distributes

Figure 2-1: SDS-PAGE gel of purification of recombinant Bla2. Lane 1: molecular weight markers (from top to bottom: 150 kDa, 100 kDa, 75 kDa, 50 kDa, 35 kDa, 25 kDa, and 15 kDa); lane 2: boiled cell fraction of cells before induction with IPTG; lane 3: soluble fraction after Minitan ultrafiltration and dialysis against 50 mM Hepes, pH 6.5; Lane 4: purified Bla2 after SP-Sepharose chromatography.

Table 2-2: Steady-state kinetic parametersa for nitrocefin, imipenem, cefaclor, and meropenem hydrolysis by Bla2 containing 1 equivalent of Zn(II). These values represent an average of at least 3 trials and the errors are standard deviations (σ n-1).

Substrate: Substrate: Substrate: Substrate: Nitrocefin Imipenem Cefaclor Meropenem Km (μM) 25 + 4 89 + 35 67 + 6 110 + 42 -1 kcat (s ) 41 + 3 92 + 21 24 + 1 49 + 11

aKinetic assays were conducted at 25 ◦C in 50 mM Hepes, pH 6.5.

Table 2-3: Steady-state kinetic parameters for Bla2 containing 1 or 2 equivalents of Zn(II) or Co(II). These values represent an average of at least 3 trials and the errors are standard deviations (σ n-1).

1Co(II)-Bla2 CoCo(II)- 1Zn(II)-Bla2 ZnZn(II)- Bla2 Bla2 Km(μM) 18 + 2 30 + 9 28 + 4 35 + 4 -1 kcat (s ) 19 + 1 47 + 5 32 + 2 42 + 2

All experiments were conducted at 25 ◦C in 50 mM Hepes, pH 6.5. The Bla2 analogs were prepared by adding the indicated equivalents of metal to apo-Bla2 before the assays were conducted.

between the Zn1 and Zn2 sites with only a small difference in the affinities of these sites. The extinction coefficient of the ligand field transitions is 218 M-1cm-1 per Co(II), suggesting Co(II) is 4/5 coordinate in Co(II)-substituted Bla2. 2.3.4 1H NMR spectroscopy To probe the metal binding ligands in Bla2, 1H NMR spectroscopic studies were

conducted on 1Co- and CoCo-Bla2. The NMR spectrum of 1Co-Bla2 in 10% D2O shows five downfield shifted resonances between 35 and 85 ppm (data not shown). The intensities and relative broadness of the peaks are different; however since the peak shapes and relative intensities are almost identical to those of CoCo-BcII previously published by Vila and coworkers, we believe that all peaks correspond to one proton each (26). The differences in peak broadness suggest that there are two distinct Co(II) binding

modes in 1Co-Bla2. Four-coordinate Co(II) has a relatively longer T1e (electron relaxation time), and protons coupled to this metal ion yield broader NMR resonances.

On the other hand, five-coordinate Co(II) has a shorter T1e, and protons coupled to this metal ion yield sharper NMR signals. Therefore, the NMR spectrum of 1Co-Bla2 suggests two distinct binding sites for Co(II) in the 1Co-Bla2 analog. The 1H NMR spectrum of Bla2 containing two equivalents of Co(II) (CoCo-Bla2) was identical to that of 1Co-Bla2 (Figure 2-3). This result supports our conclusion that the first equivalent of Co(II) binds to both metal binding sites, generating half of the sample with fully-loaded Bla2 and half with metal-free (apo) Bla2. The spectrum of

CoCo-Bla2 in >90% D2O was also obtained, and the resonances at 78, 65, 49, and 44 ppm disappeared, indicating the presence of four histidines bound to Co(II). Based on the crystal structure of BcII, we predict that these residues are His116, His118, and His196 in

the Zn1 site and His263 in the Zn2 site. The peak at 42 ppm is not solvent-exchangeable and can be assigned to the meta proton of His118, since this histidine was shown to bind metal with the δN in BcII (26). We also tried to obtain NMR evidence for the presence of Cys coordination by Co(II) in this sample. β-Methylene protons on Co(II)-bound Cys residues typically have resonances at >175 ppm and are relatively broad. We collected spectra downfield to 400 ppm; however, we did not observe any peaks downfield of 80 ppm.

Figure 2-2: UV-Vis difference spectrum of apo-Bla2 titrated with increasing amounts of Co(II). The concentration of apo-Bla2 was 1.2 mM apo-Bla2, and the buffer was 15 mM Hepes, pH 6.5, containing 100 mM NaCl. The enzyme was titrated with 0.5, 1.0, and 2.0 equivalents of Co(II) (bottom to top at 550 nm). Inset: Absorbance changes as the equivalents of Co(II) is increased.

1 Figure 2-3: H NMR spectra of CoCo-Bla2 in 10% D2O and 90% D2O. The spectra were recorded at 500 MHz and 298 K. The enzyme concentration was 1 mM, and the buffer was 15 mM Hepes, pH 6.5, containing 100 mM NaCl.

2.3.5 EPR spectroscopy EPR spectra from 1Co-Bla2 and CoCo-Bla2 were recorded under non-saturating, saturating, and rapid-passage conditions and are shown in Figure 2-4. The spectra from 1Co-Bla2 (Figure 2-4A) and CoCo(II)-Bla2 (Figure 2-4B) recorded under non-saturating conditions exhibited no resolvable rhombicity or 59Co hyperfine structure and were characteristic of 5- or 6-coordinate Co(II) with at least one water ligand. Spectra recorded 3 at high power and low temperature (Figure 2-4D) did not reveal any MS = |± /2〉 signals that would indicate tetrahedral Co(II). Slight differences were observed between the non- saturated spectra of 1Co-Bla2 and CoCo-Bla2. The difference spectrum (Figure 2-4C) revealed the presence of a second Co(II) species in CoCo-Bla2, indicating at least some preference for one site over the other, though this could not be quantitated due to the lack of resolved features in the spectra from which individual single EPR species could be deconvoluted. Rapid-passage spectra were informative; the rapid passage spectra of 1Co- Bla2 and CoCo-Bla2 (Figure 2-4E) were indistinguishable in form and differed only in intensity, by a factor of 2. This indicates that the species that is responsible for this spectrum is the same in both 1Co-Bla2 and CoCo-Bla2 and that twice as much of it is present in CoCo-Bla2 than 1Co-Bla2. These data are consistent with either a positive cooperative binding mechanism or a random mechanism with similar affinities for the sites but are completely inconsistent with a sequential mechanism. Comparison of the normal, non-saturated spectrum of 1Co-Bla2 and the (CoCo-Bla2 - 1Co-Bla2) difference spectrum (Figure 2-4C) shows that the spectrum of 1Co-Bla2 is sharper in the g⊥ region, around 2000 G (200 mT). The derivative of the rapid passage spectrum (of either 1Co- Bla2 or CoCo-Bla2) is, in turn, sharper still than that of 1Co-Bla2; the former more likely represents a single chemical species and provides more evidence that the signal from 1Co-Bla2 contains more than one species; again, this is inconsistent with strictly sequential binding. EXAFS spectroscopy Comparing the EXAFS of 1Co-Bla2 and CoCo-Bla2 (Figure 2-5A) shows little quantifiable difference, consistent with the optical, NMR and EPR studies described above. This similarity is also reflected in the curve fitting results, summarized in Table 1; best fits are shown in Figure 2-5B and 2-5C. The distributed binding indicated above

Figure 2-4: EPR spectra from Co(II)-containing Bla2 (A) 1Co(II)-Bla2, 2 mW, 11 K; (B) CoCo(II)-Bla2, 2 mW, 11 K; (C, solid line) = (B) – (A); (C, dashed line) 1Co(II)- Bla2, 2 mW, 11 K; (D, solid line) CoCo(II)-Bla2 × 0.5, 50 mW, 7 K; (D, dashed line) 1Co(II)-Bla2, 50 mW, 7 K; (E, solid line) CoCo(II)-Bla2 × 0.5, 100 mW, 7 K, rapid passage; (E, dashed line) 1Co(II)-Bla2, 100 mW, 7 K, rapid passage; (F, solid line) derivative of rapid passage spectrum of CoCo(II)-Bla2; (F, dashed line) 1Co(II)-Bla2, 2 mW, 11 K. Rapid passage spectra were recorded using second-derivative quadrature phase-sensitive detection. The intensities of spectra shown in A – C are correct relative to each other. Intensities of pairs of spectra in D – F are arbitrary, but within each pair the intensities are correct when the multiplication factors given are taken into account.

leads to the expectation that both 1Co- and CoCo-Bla2 would reflect an average of the

Zn1 and Zn2 coordination spheres, with 4 N/O and 0.5 S making up the primary coordination sphere. However, in neither case does inclusion of a partial sulfur scatterer improve the quality of the first shell fits. This is consistent with our earlier studies of the B1 MβL CcrA, where the sulfur contribution could not be resolved (17). Also in accord with our previous studies of CcrA, we see no evidence of a Co-Co interaction in CoCo- Bla2, suggesting that the aquo bridge is not present. In contrast to the distributed binding observed with Co, the binding of Zn(II) to Bla2 appears to be sequential. The Fourier transformed EXAFS for 1Zn-Bla2 shows much higher outer shell intensity than its ZnZn counterpart, which shows much better resolution in its outer shell scattering (Figure 2-6A). Qualitatively, this is consistent with

binding of the first equivalent of Zn(II) in the Zn1 site, which carries three His ligands.

Addition of the second equivalent then populates the Zn2 site, which is expected to (1) broaden the first shell with addition of the sulfur ligand, (2) lower the overall outer shell scattering, as the average number of His ligands is reduced from 3 to 2.5 and (3) sharpen the outer shell features as the dinuclear site becomes better organized. This description is fully consistent with the curve fitting results (Figure 2-6B and 2-6C), which show the Zn(II) in 1Zn-Bla2 is best represented as coordinated by 4 N/O donors, including 3 His ligands, while the average Zn(II) in ZnZn-Bla2 is coordinated by 4 N/O (including 2.5 His ligands) and 0.5 S donors. The inclusion of a Zn-Zn interaction leads to a ca. 25 % improvement in the fit residual, and the refined distance of 3.44 Å is the same as that seen in other resting state dinuclear MβLs. 2.3.7 Stopped-flow UV-Vis kinetic studies To probe the kinetic mechanism for the hydrolysis of β-lactams by Bla2, stopped- flow UV-Vis kinetic studies were conducted. All studies were conducted with 1Zn-Bla2, because this is most likely the biologically-relevant form of the enzyme (12, 27) and because ZnZn-Bla2 is unstable. Three β–lactam containing substrates were used in the pre–steady state kinetic studies: nitrocefin, cefaclor, and imipenem. In stopped-flow, diode array UV-Vis kinetic studies with 20 μM Bla2 and nitrocefin concentrations, ranging from 10 μM to 50 μM, only two distinct absorbance features

Figure 2-5: Fourier transformed EXAFS spectra of Co(II)-substituted Bla2. (A) Direct comparison of 1Co- (black line) and CoCo-Bla2 (gray line). (B) Best fit for 1Co-Bla2. (C) Best fit for CoCo-Bla2. See Table 2-1 for fit details.

Figure 2-6: Fourier transformed EXAFS spectra of Zn(II)-substituted Bla2. (A) Direct comparison of 1Zn- (black line) and ZnZn-Bla2 (gray line). (B) Best fit for 1Zn-Bla2. (C) Best fit for ZnZn-Bla2. See Table 2-1 for fit details.

were present in the absorbance versus wavelengths plots: a feature at 390 nm, which corresponds to substrate that decreased during the course of the reaction, and a feature at 485 nm, which corresponds to product that increased during the course of the reaction. There was no feature at 665 nm, corresponding to an intermediate that was observed in reactions of nitrocefin with CcrA and L1 but not with BcII (22, 24, 28). Bla2 was reacted with 25 μM and 50 μM nitrocefin, and progress curves at 390 nm and 485 nm were obtained and fitted with single exponential equations to obtain kobs. The kobs versus substrate concentration plot was hyperbolic (data not shown), suggesting a rapid- equilibrium, two-step binding mechanism (Scheme 2-1). Nitrocefin (25 μM and 50 μM ) was reacted with

Scheme 2-1:

K1 k2 k3 E+S ES EX E+P

k-2

20 μM Bla2, and the resulting progress curves were simulated using KINSIM, the mechanism in Scheme 2-1, and the rate constants in Table 2-4 (Figure 2-7). The King

Altman method was used to determine the theoretical expressions for kcat and Km, based

on the mechanism in Scheme 2-1, and these theoretical steady-state kinetic constants (kcat -1 = 7.4 s and Km = 62 μM) were similar to those determined in steady-state kinetic studies (Table 2-2). Stopped-flow kinetic studies were also used to probe the reaction of 20 μM Bla2 with imipenem and cefaclor. Stopped-flow diode-array UV-Vis studies with Bla2 and imipenem or cefaclor showed that the only UV-Vis detectable species in these reactions were the substrates. The resulting progress curves were fitted with single exponential

equations, and the kobs values were plotted versus substrate concentration. The kobs versus imipenem or cefaclor concentration plots were linear (data not shown), suggesting one- step binding mechanisms for both substrates (Scheme 2-2). Imipenem and cefaclor

Scheme 2-2

K1 k2 E+S ES E+P

(10, 25, and 50μM) were reacted with 20 μM Bla2, and the resulting progress curves were simulated using KINSIM, the mechanism in Scheme 2-2, and the rate constants in Table 2-4 (Figures 2-8 and 2-9). The King-Altman method yielded the theoretical expressions for kcat and Km. The theoretical values of kcat and Km for imipenem (kcat -1 -1 theoretical = 35 s and Km theoretical = 194 μM) and cefaclor (kcat theoretical = 15 s and Km theoretical = 150 μM) are similar to those determined by steady-state kinetic studies (Table 2-2). 2.4 Discussion Mβl BcII has previously been characterized by crystallographic, computational, kinetic, and structural studies in order to determine if mono or dinuclear BcII has higher catalytic activity. The initial crystal structure of BcII showed a mononuclear Zn(II) binding enzyme, with Zn(II) binding in the consensus Zn1 site (29). However, a second crystal structure (and subsequent structures) showed that BcII contains a dinuclear Zn(II) binding site (12, 30). The conflicting crystallographic data were explained by different

Zn(II) binding affinities to the Zn1 and Zn2 sites in BcII. Currently, the Zn1 and Zn2 sites in BcII are being studied to determine the roles of each metal site. Page and coworkers used pH dependent, kinetic studies with Co(II)-substituted BcII to evaluate the activity of the dinuclear metal ion containing analog. This group asserts that the dinuclear Co(II)- containing (and Zn(II)-containing, by analogy) analog is the only catalytically-active form of the enzyme (31, 32). In contrast, Vila, de Seny, and Wommer report that both the mononuclear and dinuclear Zn(II)-containing analogs of BcII are catalytically-active (27, 33-36); however, these researchers differ on a number of significant points. For example,

Vila contended that the Zn1 site is essential in early work (35, 36); however, the most recent work from his lab suggests that the Zn2 site is the vital metal binding site in the mononuclear Zn(II)-containing form of BcII (26, 34). Wommer and coworkers concluded

Table 2-4: Kinetic constants used in KINSIM simulations. ProK software (nonlinear Marquardt-Levenberg algorithm) was used to determine the error in the rate constants (22, 25).

Kinetic rate constants used in KINSIM simulations

constant value used in simulation value used in simulation value used in simulation for nitrocefin for imipenem for cefaclor

K1 0.12 ± 0.01 0.13 ± 0.01 2.2 + 0.1 k2 10 ± 1 35 ± 1 15 ± 1 k-2 4 ± 1 Set to 0 Set to 0 k3 40 ± 1 - - k-3 Set to 0 - -

Figure 2-7: Progress curves of the reaction of nitrocefin and Bla2 containing 1 eq. Zn(II) at 4 oC. The concentration of product increased over time using 25 (□) and 50 (○) μM nitrocefin, while substrate concentrations decreased over time using 25 (◊) and 50 (Δ) μM nitrocefin. The solid progress curves were generated by KINSIM, using the mechanism in Scheme 1 and the kinetic constants in Table 2-4.

Figure 2-8: Progress curves of the reaction of imipenem and Bla2 containing 1 eq. Zn(II) at 4 oC. The concentrations of substrate were (Δ) 50 μM, (□) 25 μM, and (○) 10 μM, and the concentration of Bla2 was 20 μM. The solid lines were generated by using KINSIM, the mechanism in Scheme 2, and the kinetic constants in Table 2-3.

Figure 2-9: Progress curves of the reaction of cefaclor and Bla2 containing 1 eq. Zn(II) at 4 oC. The concentrations of substrate were (Δ) 50 μM, (□) 25 μM, and (○) 10 μM, and the concentration of Bla2 was 20 μM. The solid lines were generated by using KINSIM, the mechanism in Scheme 2, and the kinetic constants in Table 2-4.

53 that BcII has a high affinity and a low affinity Zn(II) site and speculated that the tight binding site is the Zn1 site, based on earlier crystallographic studies (27). De Seny et al. reported that BcII exhibits negative cooperative Zn(II) binding and that there are not high and low affinity Zn(II) binding sites in BcII (33). They further contended that the first equivalent of Zn(II) distributes between the two sites and hinted that the binding of substrate may lead to Zn(II) preferentially binding the Zn2 site in the active enzyme (33, 34). Subsequent studies by Vila on Co(II)-substituted BcII suggest that the first equivalent of Co(II) distributes in both sites of BcII (26). It is clear that the issue of metal binding to BcII is far from being understood. In an effort to shed light on metal binding to BcII, we conducted structural and mechanistic studies on Bla2, which shares 89% amino acid sequence identity and 92% amino acid sequence homology with BcII. In order to conduct these studies, large amounts of Bla2 were required. Previously, Palzkill and coworkers reported the over-expression and purification of recombinant Bla2 (11). In this protocol, the ompA gene was fused to the gene for Bla2, and this construct was designed to export Bla2 into the culture growth medium. Palzkill used cation exchange and Sepharose S-200 chromatographies to obtain 0.4 mg of purified Bla2 per liter of growth culture. To improve yield, we modified the published purification protocol for recombinant Bla2 (11). Since Bla2 is exported into the growth medium, we used a Minitan ultrafiltration system to concentrate the crude protein solution to 50 mL. A similar protocol was used to purify recombinant CcrA, which is also exported into the growth medium during over-expression (13). We then purified Bla2 using a single SP- Sepharose chromatographic step, yielding 17-23 mg of purified Bla2 per liter of growth medium. SDS-PAGE was used to verify the purity of Bla2, and MALDI-TOF MS was used to verify the identity of the recombinant enzyme. The initial characterization of Bla2 involved metal binding and steady-state kinetic studies. The as-isolated Bla2 was shown to bind 1.0 + 0.2 equivalents of Zn(II), which is lower than the amount reported for as-isolated BcII (1.4 – 1.8 equivalents per protein (26)). To determine if metal dissociated from Bla2 during purification, excess Zn(II) was added to the as-isolated enzyme, and the mixture was extensively dialyzed against metal-free buffer. The resulting enzyme was shown to bind 1.4 + 0.1 equivalents of Zn(II). In similar studies, BcII has been reported to tightly bind 2.0 + 0.1 equivalents

54 of Zn(II) (37). It is not clear why there is a difference in the metal content of BcII and Bla2; however, it is possible that the quantitation of the enzymes may explain some of the differences. The extinction coefficient of Bla2 is 25,400 M-1cm-1, as estimated by amino acid analyses, and the extinction coefficient of BcII is 30,500 M-1cm-1, estimated by amino acid analyses (37). The larger extinction coefficient for BcII would result in smaller calculated concentrations for BcII and larger metal:enzyme stoichiometries. In fact, the use of the extinction coefficient for BcII to determine the concentration of Bla2 would result in 1.7 equivalents of Zn(II) tightly bound to Bla2. Since the amino acid sequences of the enzymes are so similar, it is hard to explain why the two enzymes have extinction coefficients that vary by 17%. Steady-state kinetic studies on as-isolated Bla2 containing 1 equivalent of Zn(II) were conducted using several different substrates, and the data show that the as-isolated enzyme is catalytically active (Table 2-2). Previously, Palzkill reported that recombinant -1 Bla2 (Zn(II) content unknown) exhibited Km of 75 + 5 μM and a kcat 313 + 21 s , when using nitrocefin as substrate (11). In our hands, recombinant Bla2 containing 1 equivalent -1 of Zn(II) exhibited a Km of 25 + 4 μM and a kcat of 41 + 3 s when using nitrocefin as substrate. Our data compare favorably with previous steady-state kinetic data on BcII, -1 which exhibits a Km of 9.8 μM and a kcat of 30 s . The difference in steady-state kinetic behavior of the two recombinant Bla2 preparations could possibly be explained by the enzymes having different equivalents of Zn(II); however, the large differences in kcat cannot be explained completely by differing metal content. In addition, our assays were conducted at 25 oC, while those of Palzkill and coworkers were conducted at 30 oC (11). The differences in temperature are expected to result in only a 1.25-fold difference in the reported kcat values. The differences can also be explained by differing conditions used in the steady-state kinetic studies. Our initial efforts to use the identical conditions (buffer of 50 mM Hepes, pH 7.5, containing 50 μM Zn(II) and 20 μg of bovine serum albumin (BSA) lead to irreproducible results (11). We attributed the irreproducible results on enzyme instability in the reaction conditions. The predicted pI of Bla2 is 7.6; therefore at pH 7.5, the enzyme would be expected to be predominantly zwitterionic and insoluble. We therefore lowered the pH of our assay buffer to 6.5, and the kinetic assays were much more reproducible. It is not clear why the Bla2 in Palzkill’s studies was stable. Another

55 big difference in our assay conditions and Palzkill’s was the inclusion of Zn(II) in the assay buffer. Our studies on Zn(II) being added to apo-Bla2 suggest that the addition of Zn(II) to Bla2 causes protein denaturation in steady-state kinetic experiments (Table 2-3). Again, it is not clear why Palzkill’s Bla2 was stable to the addition of extra Zn(II). We will need to purify Bla2 using the identical protocol used by Palzkill to rule out any differences in the enzymes caused by different purification protocols. Clearly, the enzymes are different as Palzkill reported that his recombinant Bla2 does not hydrolyze -1 imipenem (11), and our recombinant Bla2 exhibits a Km of 89 μM and a kcat of 92 s when using imipenem as a substrate (Table 2-2).

Structural studies were conducted to probe the Zn1 and Zn2 metal binding sites in Bla2. Co(II)-substituted Bla2 was prepared by direct addition of Co(II) to apo-Bla2. The structural studies determined whether Zn(II) and Co(II) bind in a sequential (metal binds to the Zn1 first than the Zn2 binding sites), random (metal distributes among the Zn1 and

Zn2 binding sites) or cooperative (the binding of the first equivalent of metal results in a higher affinity for the binding of a second 2 equivalent of metal) mode. The UV-Vis, 1H NMR, and EPR studies were conducted with Co(II)-substituted Bla2 due to native Zn(II) being diamagnetic resulting in spectroscopically-silent data. Steady-state kinetic studies revealed that Bla2 containing 1 equivalent of Co(II) is catalytically-active. The addition of the second equivalent of Co(II) to generate CoCo-Bla2 resulted in a ca. a two-fold increase in kcat (Table 2-3). This result suggests that Co(II) binding to Bla2 is random and that both metal ions bind with similar KD values or that Bla2 exhibits positive cooperative binding of Co(II). The issue of metal binding KD(s) to BcII is controversial. Early work suggested that metal ions bind much more tightly to the Zn1 site than to the Zn2 site, perhaps due to the presence of a Arg at position 121 (27, 29, 35, 36). However, Vila generated R121C mutants and showed that this Arg did not affect metal binding (36).

Other studies have revealed that the metal binding sites in BcII have very similar KD values (26). Since there is no consensus on the KD values of metal binding to the Zn1 and

Zn2 sites in BcII that is almost identical to Bla2, it is not possible to determine whether the binding of Co(II) to Bla2 is random or positive cooperative. In contrast, ZnZn-Bla2 is not twice as active as 1Zn-Bla2, suggesting that Zn(II) binding is sequential; although, we cannot rule out the possibility that addition of a

56 second equivalent of Zn(II) to 1Zn-Bla2 results in some protein denaturation with steady- state kinetics alone. Since the EXAFS results demonstrate Zn(II) binding to be sequential in Bla2 (Figure 2-6), it suggests that the addition of the second equivalent of Zn(II) does not greatly affect catalysis and that the first site of Zn(II) binding is more important kinetically (33, 34, 36). Lastly, the steady-state kinetic buffers contain ca. 100 nM Zn(II), even if the buffers are Chelex-treated, and the enzyme concentration is 1-10 nM. The exact metal content of the Bla2 analogs shown in Table 2-3 is unknown. We therefore conducted stopped-flow kinetic studies to address whether 1Zn-Bla2 is active. UV-Vis studies on Co(II)-substituted Bla2 confirmed that Co(II) binding to apo- Bla2 is random or positive cooperative. A cysteine sulfur to Co(II) LMCT was observed in the spectrum of Bla2 containing only 0.5 equivalents of Co(II), and the intensity of this peak was maximized only after 2 equivalents of Co(II) were present. We were unable to obtain the UV-Vis spectrum of Bla2 containing 3 equivalents of Co(II) because the protein quickly precipitated. A similar coincidence occurred during the Co(II) titrations of CcrA by Wang et al. (13). 1H NMR spectra of Co(II)-substituted Bla2 further confirm the random binding or positive cooperative binding of Co(II) to apo-Bla2. Five paramagnetically-shifted 1H resonances were observed in the spectrum of 1Co-Bla2, and no additional peaks were observed in the spectrum of CoCo-Bla2. Four of the peaks were assigned to solvent-exchangeable NH protons on Co(II) bound histidines, and the other peak was assigned to a meta proton on a Co(II) bound histidine. The 1H NMR spectra of Bla2 are very similar to those previously reported by Vila (26) with one exception. Despite considerable effort, we were never able to detect peaks between 100 and 400 ppm assignable to the β-CH2 protons on Co(II) bound cysteine. It is possible that our procedure to prepare the 1H NMR samples resulted in oxidation of Cys221, and similar oxidation has been reported for CcrA (17). In addition, oxidized Cys221 has been observed in crystal structures of BcII and VIM-2 (17). The EPR spectra of Bla2 containing 1 and 2 equivalents of Co(II) were obtained to further probe Co(II) binding to Bla2. The EPR spectra were consistent with both Co(II) ions being 5- or 6-coordinate with at least one water ligand, as there was no 3 observable MS = |± /2〉 signal at low temperature and high power indicative of tetrahedral Co(II). This result along with the UV-Vis studies demonstrates that both Co(II)’s in Bla2

57 are 5-coordinate. The EPR studies also confirm that Co(II) binding to Bla2 is not sequential, but these studies cannot unambiguously determine whether Co(II) binding is random or positive cooperative. Spectra at non-saturating conditions demonstrate a slight, but non-quantifiable, preference for Co(II) binding to one site over the other; however, spectra obtained at rapid passage conditions show identical signals for the 1Co- and

CoCo-Bla2 samples, suggesting a random binding mechanism with metal binding KD’s similar for both sites. EXAFS was used to probe the binding of Zn(II) and Co(II), to verify that Zn(II) has the same metal binding mechanism as Co(II). The studies with Co(II) substituted Bla2 showed a cooperative binding mechanism; Cys 221 was not observed with Co(II) substituted Bla2, and no Co-Co interaction was detected, consistent with previously reported studies of Co(II) substituted CcrA (17). The binding of Zn(II) by Bla2 does not follow the same pathway as Co(II) binding by Bla2. Studies of Zn(II)-Bla2 show Zn(II) binds to the Zn1 site first (Cys221 was not observed), while the addition of the second equivalent of Zn(II) leads to observation of a Zn-S interaction, consistent with Cys221 coordination, and metal bound to the Zn2 site. The EXAFS studies clearly show that Zn(II) follows a sequential binding mechanism, while Co(II) follows a cooperative binding mechanism (Figure 2-10). The EXAFS studies further support the steady-state kinetic data with the metal analogs of Bla2. To further address the activity of Bla2 metal analogs, stopped-flow UV-Vis kinetic studies were conducted. We decided to use 1Zn-Bla2 in these studies because we wanted to probe the activity of the mononuclear Zn(II) containing analog, which has

Zn(II) in the Zn1 site (Figure 2-10). In addition, our steady-state kinetic studies suggested that Zn(II) may cause precipitation of Bla2. The stopped-flow studies demonstrate that mononuclear Zn(II)-containing Bla2 is catalytically-active. This result contrasts recent reports on BcII by Page and coworkers, who argue that the dinuclear metal ion containing forms of BcII are the only catalytically active analogs (31, 32). The rapid kinetic studies also showed that 1Zn-Bla2 does not stabilize the nitrocefin-derived reaction intermediate observed with CcrA and L1 (13, 22, 28). This result is not surprising because BcII does not stabilize this intermediate either, and previous studies on L1 showed that the metal

Figure 2-10: The proposed active site of Bla2 after the addition of 1 or 2 equivalents of Zn(II) or Co(II) to apo-Bla2.

59 ion in the Zn2 site is required for stabilization of the intermediate (12, 24). Analysis of the kinetic data demonstrates that Bla2 utilizes a rapid-equilibrium mechanism to hydrolyze nitrocefin, and the rate constants suggest that β-lactam C-N bond cleavage is rate- limiting. Vila and coworkers proposed the same rate-limiting step for BcII (12, 24). Stopped-flow studies with imipenem and cefaclor showed that Bla2 utilizes a one-step mechanism to hydrolyze these substrates (23). Previous studies on L1 and CcrA have shown that these enzymes utilize a different kinetic mechanism to hydrolyze nitrocefin than other substrates. Diode-array spectra of the reactions of Bla2 with imipenem or cefaclor did not reveal any reaction intermediates that absorb UV-Vis radiation. Taken together, our data demonstrate that 1Zn-Bla2 is catalytically-active and

EXAFS data on this analog demonstrate that Zn(II) binds in the consensus Zn1 site. The EXAFS data also show that Zn(II) binding is sequential, similar to CcrA and L1 (17, 38). In contrast, all of the spectroscopic and kinetic data on Co(II)-containing Bla2 strongly suggest that Co(II) binding is either random or positive cooperative. This result suggests that caution should be exercised when extrapolating results on Co(II)-substituted proteins to the native Zn(II)-containing forms of Bla2, and possibly BcII. A summary scheme showing how Zn(II) and Co(II) bind to Bla2 is shown in Figure 2-10. Further studies are required to address why Zn(II) and Co(II) behave differently in Bla2, and X-ray crystal structures of both analogs may yield information about the different metal binding properties. Since Bla2 and BcII are almost identical proteins in terms of amino acid sequence, a crystal structure of Bla2 may shed light on why these enzymes appear to have different biochemical properties.

2.5 References 1. Jernigan, J. A., Stephens, D. S., Ashford, D. A., Omenaca, C., Topiel, M. S., Galbraith, M., Tapper, M., Fisk, T. L., Zaki, S., Popovic, T., Meyer, R. F., Quinn, C. P., Harper, S. A., Fridkin, S. K., Sejvar, J. J., Shepard, C. W., McConnell, M., Guarner, J., Shieh, W. J., Malecki, J. M., Gerberding, J. L., Hughes, J. M., Perkins, B. A., and Anthrax Bioterrorism, I. (2001) Bioterrorism-related inhalational anthrax: The first 10 cases reported in the United States, Emerg. Infect. Dis. 7, 933-944. 2. Bohn, K., Mears, B., and Fiegel, E. (2008) U.S. Officials declare researcher is anthrax killer, in CNN, Atlanta, GA. 3. Page, M. I., and Laws, A. P. (1998) The mechanism of catalysis and the inhibition of beta-lactamases, Chem. Commun., 1609-1617. 4. Bush, K. (1989) Characterization of β-lactamases, Antimicro. Agents. Chemo. 33, 259-263. 5. Bush, K. (1989) Classification of β-lactamases - Group-1, Group-2A, Group-2B, and Group-2B', Antimicro. Agents. Chemo. 33, 264-270. 6. Hall, B. G., and Barlow, M. (2005) Revised Ambler classification of beta- lactamases, J. Antimicro. Chemo. 55, 1050-1051. 7. Bush, K. (1998) Metallo-β-lactamases: A class apart, Clin. Infect. Dis. 27, S48- S53. 8. Galleni, M., Lamotte-Brasseur, J., Rossolini, G. M., Spencer, J., Dideberg, O., Frere, J. M., and Metallo-β-Lactamase Working, G. (2001) Standard numbering scheme for class B β-lactamases, Antimicro. Agents. Chemo. 45, 660-663. 9. Moran-Barrio, J., Gonzalez, J. M., Lisa, M. N., Costello, A. L., Dal Peraro, M., Carloni, P., Bennett, B., Tierney, D. L., Limansky, A. S., Viale, A. M., and Vila, A. J. (2007) The metallo-β-lactamase GOB is a mono-Zn(II) enzyme with a novel active site, J. Biol. Chem. 282, 18286-18293. 10. Chen, Y. H., Succi, J., Tenover, F. C., and Koehler, T. M. (2003) β-lactamase genes of the penicillin-susceptible Bacillus anthracis Sterne strain, J. Bacteriol. 185, 823-830.

11. Materon, I. C., Queenan, A. M., Koehler, T. M., Bush, K., and Palzkill, T. (2003) Biochemical characterization of β-lactamases Bla1 and Bla2 from Bacillus anthracis, Antimicro. Agents. Chemo. 47, 2040-2042. 12. Crowder, M. W., Spencer, J., and Vila, A. J. (2006) Metallo-β-lactamases: Novel weaponry for antibiotic resistance in bacteria, Acc. Chem. Res. 39, 721-728. 13. Wang, Z. G., and Benkovic, S. J. (1998) Purification, characterization, and kinetic studies of a soluble Bacteroides fragilis metallo-β-lactamase that provides multiple antibiotic resistance, J. Biol. Chem. 273, 22402-22408. 14. Crowder, M. W., Walsh, T. R., Banovic, L., Pettit, M., and Spencer, J. (1998) Overexpression, purification, and characterization of the cloned metallo-β- lactamase L1 from Stenotrophomonas maltophilia, Antimicro. Agents. Chemo.42, 921-926. 15. Yang, K. W., and Crowder, M. W. (2004) A method for removing ethylenediaminetetraacetic acid from apo-proteins, Anal. Biochem. 329, 342-344. 16. Crawford, P. A., Yang, K. W., Sharma, N., Bennett, B., and Crowder, M. W. (2005) Spectroscopic studies on cobalt(II)-substituted metallo-β-lactamase ImiS from Aeromonas veronii bv. sobria, Biochemistry 44, 5168-5176. 17. Periyannan, G. R., Costello, A. L., Tierney, D. L., Yang, K. W., Bennett, B., and Crowder, M. W. (2006) Sequential binding of cobalt(II) to metallo-β-lactamase CcrA, Biochemistry 45, 1313-1320. 18. Wang, D. M., and Hanson, G. R. (1995) A New Method for Simulating Randomly Oriented Powder Spectra in Magnetic-Resonance - the Sydney-Opera-House (Sophe) Method, J. Mag. Res. A 117, 1-8. 19. Thomas, P. W., Stone, E. M., Costello, A. L., Tierney, D. L., and Fast, W. (2005) The quorum-quenching lactonase from Bacillus thuringiensis is a metalloprotein, Biochemistry 44, 7559-7569. 20. Newville, M. (2001) IFEFFIT: interactive EXAFS analysis and FEFF fitting, Synchrotron Rad 8, 322-324. 21. Ankudinov, A. L., Ravel, B., Rehr, J. J., and Conradson, S. D. (1998) Real-space multiple-scattering calculation and interpretation of X-ray-absorption near-edge structure, Phys. Rev. B 58, 7565-7576.

22. McManus-Munoz, S., and Crowder, M. W. (1999) Kinetic mechanism of metallo- β-lactamase L1 from Stenotrophomonas maltophilia, Biochemistry 38, 1547- 1553. 23. Spencer, J., Clarke, A. R., and Walsh, T. R. (2001) Novel mechanism of hydrolysis of therapeutic β-lactams by Stenotrophomonas maltophilia L1 metallo- β-lactamase, J. Biol. Chem. 276, 33638-33644. 24. Rasia, R. M., and Vila, A. J. (2003) Mechanistic study of the hydrolysis of nitrocefin mediated by B. cereus metallo-β-lactamase, ARKIVOC, 507-516. 25. Sharma, N. P., Hajdin, C., Chandrasekar, S., Bennett, B., Yang, K. W., and Crowder, M. W. (2006) Mechanistic studies on the mononuclear Zn-II-containing metallo-β-lactamase ImiS from Aeromonas sobria, Biochemistry 45, 10729- 10738. 26. Llarrull, L. I., Tioni, M. F., Kowalski, J., Bennett, B., and Vila, A. J. (2007) Evidence for a dinuclear active site in the metallo-β-lactamase BcII with substoichiometric Co(II) - A new model for metal uptake, J. Biol. Chem. 282, 30586-30595. 27. Wommer, S., Rival, S., Heinz, U., Galleni, M., Frere, J. M., Franceschini, N., Amicosante, G., Rasmussen, B., Bauer, R., and Adolph, H. W. (2002) Substrate- activated zinc binding of metallo-β-lactamases - Physiological importance of the mononuclear enzymes, J. Biol. Chem. 277, 24142-24147. 28. Wang, Z. G., Fast, W., and Benkovic, S. J. (1999) On the mechanism of the metallo-β-lactamase from Bacteroides fragilis, Biochemistry 38, 10013-10023. 29. Carfi, A., Pares, S., Duee, E., Galleni, M., Duez, C., Frere, J. M., and Dideberg, O. (1995) The 3D structure of a zinc metallo-β-lactamase from Bacillus cereus reveals a new type of protein fold, Embo J. 14, 4914-4921. 30. Fabiane, S. M., Sohi, M. K., Wan, T., Payne, D. J., Bateson, J. H., Mitchell, T., and Sutton, B. J. (1998) Crystal structure of the zinc-dependent β-lactamase from Bacillus cereus at 1.9 angstrom resolution: Binuclear active site with features of a mononuclear enzyme, Biochemistry 37, 12404-12411.

31. Badarau, A., Damblon, C., and Page, M. I. (2007) The activity of the dinuclear cobalt-β-lactamase from Bacillus cereus in catalysing the hydrolysis of β-lactams, Biochem. J. 401, 197-203. 32. Badarau, A., and Page, M. I. (2006) The variation of catalytic efficiency of Bacillus cereus metallo-β-lactamase with different active site metal ions, Biochemistry 45, 10654-10666. 33. de Seny, D., Heinz, U., Wommer, S., Kiefer, M., Meyer-Klaucke, W., Galleni, M., Frere, J. M., Bauer, R., and Adolph, H. W. (2001) Metal ion binding and coordination geometry for wild type and mutants of metallo-β-lactamase from Bacillus cereus 569/H/9 (BcII) - A combined thermodynamic, kinetic, and spectroscopic approach, J. Biol. Chem. 276, 45065-45078. 34. Llarrull, L. I., Fabiane, S. M., Kowalski, J. M., Bennett, B., Sutton, B. J., and

Vila, A. J. (2007) Asp-120 locates Zn2 for optimal metallo-β-lactamase activity, J. Biol. Chem. 282, 18276-18285. 35. Orellano, E. G., Girardini, J. E., Cricco, J. A., Ceccarelli, E. A., and Vila, A. J. (1998) Spectroscopic characterization of a binuclear metal site in Bacillus cereus β-lactamase II, Biochemistry 37, 10173-10180. 36. Rasia, R. M., and Vila, A. J. (2002) Exploring the role and the binding affinity of a second zinc equivalent in B. cereus metallo-β-lactamase, Biochemistry 41, 1853-1860. 37. Paul-Soto, R., Zeppezauer, M., Adolph, H. W., Galleni, M., Frere, J. M., Carfi, A., Dideberg, O., Wouters, J., Hemmingsen, L., and Bauer, R. (1999) Preference of Cd(II) and Zn(II) for the two metal sites in Bacillus cereus β-lactamase II: A perturbed angular correlation of gamma-rays spectroscopic study, Biochemistry 38, 16500-16506. 38. Costello, A., Periyannan, G., Yang, K. W., Crowder, M. W., and Tierney, D. L. (2006) Site-selective binding of Zn(II) to metallo-β-lactamase L1 from Stenotrophomonas maltophilia, J. Biol. Inorg. Chem. 11, 351-358.

Chapter 3

Conclusion 3.1 Conclusion

Howard, Florey and Ernst Chain’s contributions to society resulted in the production of a miracle drug called penicillin. Penicillin was first clinically utilized in 1942 during World War II to ward off bacterial infections found in wounded World War II soldiers. Clinics, which used penicillin to treat staph infections, discovered that certain strains of staph were not responding to penicillin by 1945 (1). The rapid emergence of antibiotic resistance is currently a major clinical and public health concern. One of the most common strategies of antibiotic resistance is the production of β-lactamases by bacteria. The rapid emergence of distinct β-lactamases has resulted in different derivatives of β-lactam antibiotics being ineffective. The most common reason for the emergence of antibiotic resistance is the misuse of antibiotics by patients who demand that physicians prescribe an antibiotic to treat viral infections or who do not follow the recommended dosage regimen. The normal clinical response to infections caused by bacteria that are resistant to one antibiotic is to prescribe a different, usually more potent, antibiotic. If this antibiotic does not work, the physician will prescribe yet another antibiotic, and so on. In an effort to stop this cycle, physicians are currently being advised not to prescribe an antibiotic until the patient has an infection for three weeks (2-4), which suggests that the patient has a bacterial rather than a viral infection. Another way to combat antibiotic resistance is to use drug/inhibitor combinations, which are made up of an antibiotic to kill the bacterium and an inhibitor to inactivate the enzyme/protein that the bacterium uses to achieve the resistance phenotype. The most successful drug/inhibitor combination is Augmentin, which is made up of amoxicillin, a penicillin, and clavulanic acid, a serine β-lactamase inhibitor. Augmentin is not effective at treating infections caused by bacteria that produce metallo-β- lactamases. Dr. Crowder’s group has structurally and mechanistically characterized multiple metallo-β-lactamases in different sub-groups, including CcrA (subgroup Ba), ImiS (subgroup Bb), and L1 (subgroup Bc). The goal by Dr. Crowder’s group is to find common structural and mechanistic features, which can be targeted for the development 65

of a universal metallo-β-lactamase inhibitor. The characterization of Bla2 (subgroup Ba), a metallo-β-lactamase from Bacillius anthrasis (anthrax), is of interest because β-lactams are one of the drugs of choice to treat anthrax infections (5,6). The appearance of a metallo-β-lactamase from Bacillius anthrasis was discovered by Koehler and coworkers in 2003 (7). Koehler and coworkers reported that Bla2 shares an 89% amino acid sequence identity and 92% amino acid sequence similarity to BcII from Bacillus cereus (sub-group Ba) (6,7). However, the initial kinetic data on Bla2 were significantly different than that reported for BcII (6). In addition, the published data on BcII to describe structural or mechanistic aspects of Bla2 is complicated by conflicting data reported on the enzyme (6, 8, 9, 12). Therefore, our characterization of Bla2 may lead to information that could lead to an inhibitor of Bla2 and also help resolve the extensive controversy about BcII. The characterization of Bla2 required the use of several structural and mechanistic approaches. This thesis describes the preparation of Co(II)-substituted Bla2 for UV-Vis, 1H NMR, EPR, and EXAFS spectroscopic studies and mechanistic results on 1Zn-Bla2 with different β-lactam containing antibiotics. Spectroscopic studies were used to probe

the Zn1 and Zn2 sites in Bla2 and to evaluate how Co(II) and Zn(II) bind to the enzyme. It was determined from EXAFS studies that the binding of Zn(II) is different than the binding of Co(II). Specifically, EXAFS studies indicated that Zn(II) binds sequentially. On the other hand, Co(II) binding to Bla2 follows a random or positive cooperative binding mechanism. To our knowledge, this is the first example of different metal binding mechanisms for Zn(II)/Co(II) binding to a Zn(II) metalloenzyme. The mechanistic and spectroscopic characterization of Bla2 demonstrates the

catalytic importance of the Zn1 site. The EXAFS studies showed that Zn(II) sequentially

binds initially to the Zn1 site. The spectroscopic results on Bla2 are similar to the initial spectroscopic findings for BcII (8, 9). Initial crystallographic results on BcII indicate that

Zn(II) has a higher binding affinity for the Zn1 site than for the Zn2 site (10, 11). Current publications from Page are not in agreement with the studies on Bla2 reported in this thesis. Page’s recent work on BcII indicates that the dinuclear Zn(II)-containing analog is the only catalytically active form and that the mono-Zn(II) analog is not active (12). Steady-state kinetic results on Bla2 indicate that the mono-Zn(II) analog of Bla2 is active,

while the addition of a second Zn(II) to mono-Zn(II) Bla2 results in denaturation of the enzyme. Based on the steady-state kinetic and structural studies on Bla2, 1Zn(II)-Bla2 was used in stopped-flow kinetic experiments. The proposed mechanism for Bla2 with one

equivalent of Zn(II) in the Zn1 site suggests that the rate-limiting step in the hydrolysis reaction of imipenem or cefaclor is β-lactam C-N bond cleavage. Mechanistic studies on L1 (subgroup Bc) and ImiS (subgroup Bb) showed that these enzymes share the same rate-limiting step (13, 14). The discovery of this common rate-limiting step is significant because inhibitors designed to structurally mimic the substrate during the transition state of the rate-limiting step could be excellent, tight-binding inhibitors of all metallo-β- lactamases (15, 16). The transition state of this rate-limiting step is most likely a tetrahedral species (at carbon), and similar transition states have been proposed for many peptidase-catalyzed reactions. Bartlett and coworkers reported that phosphonate and phosphinate analogs of the peptide substrates have been shown to be very tight-binding -15 inhibitors, with one inhibitor of carboxypeptidase exhibiting a Ki value of 10 M (17). Following a similar strategy, a penicillin or cephalosporin derivative with a phosphonate/phosphinate group at the β-lactam carbonyl position is hypothesized to be an excellent and universal inhibitor of the metallo-β-lactamases (15, 16) (Figure 3-1). Unfortunately, we know of no existing compounds that resemble this “designed” inhibitor, and we do not know how difficult it would be to prepare such a compound. The next step in our strategy to prepare a universal inhibitor involves varying the X and Y substituents to maximize enzyme-inhibitor contacts. Buynak and co-workers showed

that penicillin-derived inhibitors with small X and Y substituent resulted in low IC50 values for several β-lactamases (18). Our inhibitor “redesign” approach involves screening a number of inhibitors versus metallo-β-lactamases from each of the distinct subclasses, and solving the crystal structures of the enzymes bound to the tightest binding inhibitors. These studies are underway in the Crowder lab now. The work presented in this thesis enhances our current understanding of Bla2 compared with metallo-β-lactamases from other subgroups. It was found that Co(II) binds to Bla2 in a random or positive cooperative manner. EXAFS and kinetic studies on

Bla2 demonstrate that the Zn1 site is catalytically important, and this characteristic of Bla2 represents a significant difference between Bla2 and amino acid sequence-similar

BcII. Since Co(II) binding is different than Zn(II) binding to Bla2, spectroscopic studies on Co(II)-substituted Bla2 cannot be used to offer insight on inhibitor/substrate binding or mechanism. Despite the differences in metal binding, Bla2 shares a common mechanistic feature found in other metallo-β-lactamases, and this feature may serve as an excellent target for the generation of the first universal inhibitor of these enzymes.

Figure 3-1: A penicillin derivative inhibitor for metallo-β-lactamases with a phosphinate group at the β-lactam carbonyl position.

3.2 References 1. Drews, J. (2000) Drug Discovery: A historical perspective, Science 287, 1960- 1964. 2. Walsh, C. (2000) Molecular mechanisms that confer antibacterial drug resistance, Nature 406, 775-781. 3. Neu, H. C. (1992) The crisis in Antibiotic-Resistance, Science 257, 1064-1073. 4. Walsh, F. M., and Amyes, S. G. B. (2004) Microbiology and drug resistance mechanisms of fully resistant pathogens, Curr. Opin. Microbiolo.7, 439-444. 5. Jernigan, J. A., Stephens, D. S., Ashford, D. A., Omenaca, C., Topiel, M. S., Galbraith, M., Tapper, M., Fisk, T. L., Zaki, S., Popovic, T., Meyer, R. F., Quinn, C. P., Harper, S. A., Fridkin, S. K., Sejvar, J. J., Shepard, C. W., McConnell, M., Guarner, J., Shieh, W. J., Malecki, J. M., Gerberding, J. L., Hughes, J. M., and Perkins, B. A. (2001) Bioterrorism-related inhalational anthrax: The first 10 cases reported in the United States, Emerg. Infec. Dis. 7, 933-944. 6. Materon, I. C., Queenan, A. M., Koehler, T. M., Bush, K., and Palzkill, T. (2003) Biochemical characterization of β-lactamases Bla1 and Bla2 from Bacillus anthracis, Antimicro. Agents Chemo. 47, 2040-2042. 7. Chen, Y. H., Succi, J., Tenover, F. C., and Koehler, T. M. (2003) β -lactamase genes of the penicillin-susceptible Bacillus anthracis sterne strain, J. Bacteriol. 185, 823-830. 8. Llarrull, L. I., Tioni, M. F., Kowalski, J., Bennett, B., and Vila, A. J. (2007) Evidence for a dinuclear active site in the metallo-β-lactamase BcII with substoichiometric Co(II) - A new model for metal uptake, J. Biol. Chem. 282, 30586-30595. 9. Orellano, E. G., Girardini, J. E., Cricco, J. A., Ceccarelli, E. A., and Vila, A. J. (1998) Spectroscopic characterization of a binuclear metal site in Bacillus cereus β-lactamase II, Biochemistry 37, 10173-10180. 10. Carfi, A., Pares, S., Duee, E., Galleni, M., Duez, C., Frere, J. M., and Dideberg, O. (1995) The 3-D structure of a zinc Metallo-β-Lactamase from Bacillus cereus reveals a new-type of protein fold, EMBO J. 14, 4914-4921.

11. Fabiane, S. M., Sohi, M. K., Wan, T., Payne, D. J., Bateson, J. H., Mitchell, T., and Sutton, B. J. (1998) Crystal structure of the zinc-dependent β-lactamase from Bacillus cereus at 1.9 angstrom resolution: Binuclear active site with features of a mononuclear enzyme, Biochemistry 37, 12404-12411. 12. Badarau, A., and Page, M. I. (2008) Loss of enzyme activity during turnover of the Bacillus cereus β-lactamase catalysed hydrolysis of β-lactams due to loss of zinc ion, J. Biol. Inorg. Chem. 13, 919-928. 13. McManus-Munoz, S., and Crowder, M. W. (1999) Kinetic mechanism of metallo- β-lactamase L1 from Stenotrophomonas maltophilia, Biochemistry 38, 1547- 1553. 14. Sharma, N. P., Hajdin, C., Chandrasekar, S., Bennett, B., Yang, K. W., and Crowder, M. W. (2006) Mechanistic studies on the mononuclear Zn-II-containing metallo-β-lactamase ImiS from Aeromonas sobria, Biochemistry 45, 10729- 10738. 15. Funke, T., Healy-Fried, M. L., Han, H., Alberg, D. G., Bartlett, P. A., and Schonbrunn, E. (2007) Differential inhibition of class I and class II 5- enolpyruvylshikimate-3-phosphate Synthases by tetrahedral reaction intermediate analogues, Biochemistry 46, 13344-13351. 16. Priestman, M. A., Healy, M. L., Becker, A., Alberg, D. G., Bartlett, P. A., Lushington, G. H., and Schonbrunn, E. (2005) Interaction of phosphonate analogues of the tetrahedral reaction intermediate with 5-enolpyruvylshikimate-3- phosphate synthase in atomic detail, Biochemistry 44, 3241-3248. 17. Kaplan, A. P., and Bartlett, P. A. (1991) Synthesis and Evaluation of an Inhibitor of Carboxypeptidase-a with a Ki Value in the Femtomolar Range, Biochemistry 30, 8165-8170. 18. Buynak, J. D., Chen, H. S., Vogeti, L., Gadhachanda, V. R., Buchanan, C. A., Palzkill, T., Shaw, R. W., Spencer, J., and Walsh, T. R. (2004) Penicillin-derived inhibitors that simultaneously target both metallo- and serine-β-lactamases, Bioorg. Med. Chem. Letters 14, 1299-1304.