Overcoming Inhibitor Resistance in the Shv Βeta

Home , Lysozyme

OVERCOMING INHIBITOR RESISTANCE IN THE SHV

ΒETA-LACTAMASE

JODI MICHELLE THOMSON

Submitted in partial fulfillment of the requirements

For the degree of Doctor of Philosophy

Dissertation Advisor: Robert A Bonomo

Department of Pharmacology

CASE WESTERN RESERVE UNIVERSITY

August, 2007

CASE WESTERN RESERVE UNIVERSITY

SCHOOL OF GRADUATE STUDIES

We hereby approve the dissertation of

______

candidate for the Ph.D. degree *.

(signed)______(chair of the committee)

______

(date) ______

*We also certify that written approval has been obtained for any proprietary material contained therein.

For Robert A. Bonomo

TABLE OF CONTENTS

Table of Contents ...... 1-2

List of Figures ...... 3-8

List of Tables...... 9

Acknowledgments...... 10-12

Abbreviations ...... 13-14

Abstract...... 15-16

Chapter 1: Introduction ...... 17-43

Body...... 17-31

Figures ...... 32-43

Chapter II: SHV Arg244: Clavulanate Resistance...... 44-81

Introduction ...... 44-46

Methods ...... 47-58

Results ...... 59-64

Discussion...... 65-69

Tables ...... 70-72

Figures ...... 73-81

Chapter III: Arg 244: Sulbactam Susceptibility ...... 82-108

Introduction ...... 82-85

Methods ...... 86-88

Results ...... 89-92

Discussion...... 93-98

Tables ...... 99-101

Figures ...... 102-108

Chapter IV: Inhibition of SHV ESBLs by Transition State

Analogs ...... 109-121

Introduction ...... 109-111

Methods ...... 112-113

Results ...... 114-115

Discussion...... 116-117

Tables ...... 118

Figures ...... 119-121

Chapter V: β-Lactamase Inhibition by BLIP ...... 122-146

Introduction ...... 122-125

Methods ...... 126-130

Results ...... 131-133

Discussion...... 134-137

Tables ...... 138-139

Figures ...... 140-146

Chapter VI: Concluding Remarks...... 147-154

Bibliography ...... 155-171

LIST OF FIGURES

1-1 Examples of the 8 generations of β-lactam antibiotics currently on the market

exemplify conservation of the lactam backbone and diversity of the side-chains

1-2 The mechanism of attachment of the penicillin binding protein (PBP) to its

natural D-Ala D-Ala substrate (top) and a β-lactam antibiotic (bottom) is strikingly

similar

1-3 The crystal structure of SHV β-lactamase (Protein Data Bank accession

number 1SHV)

1-4 The SHV β-lactamase active site pocket, highlighting important residues for

catalysis and substrate specificity

1-5 The numbering scheme for the backbone nucleus of penicillins and

cephalosporins

1-6 Structures of the β-lactamase inhibitors currently on the market. The main

difference between these and the beta-lactam antibiotics is the presence of a good

leaving group at the 1 position

1-7 The pathway of inhibition by clavulanate in the active site of class A β-

lactamases is a multi-step process with numerous intermediates

1-8 The proposed contribution of Arg244 to clavulanate inactivation of TEM-1: (A)

Direct hydrogen bonding interactions (B) Coordination of a catalytic water molecule

essential for inactivation

1-9 Structures of cephalosporins and corresponding boronic acid transition state

inhibitors (BATSIs): (A) Ceftazidime (B) Ceftazidime BATSI (C) Cefotaxime (D)

Cefotaxime BATSI (E) Cephalothin (F) Achiral Cephalothin BATSI (G) Chiral

Cephalothin BATSI

1-10 Chemical structure of novel penem inhibitors(1 and 2) developed by Wyeth

Pharmaceuticals which display large, bicyclic R1 side chains

1-11 Endo-trig cyclization of Penem 1 in the active site of a β-lactamase results in

the formation of a seven-membered ring

1-12 The results of alanine scanning mutagenesis of BLIP indicated fewer

“hotspots” of BLIP binding to SHV-1 (left) compared to TEM-1 (right). Residues

identified as critical for binding are in red (reduce binding >10 fold when mutated),

residues that, when mutated, lead to an increase in affinity are colored cyan, and

residues that had a < 10 fold effect on affinity when mutated are colored green.

(Zhang and Palzkill, JBC 279-41, 42860-42866)

2-1 Chemical structures of compounds tested in this study. The structures of

clavulanic acid and nitrocefin are labeled with the accepted ring numbering system

2-2 The proposed contribution of Arg244 to clavulanic acid inactivation of TEM-1.

The guanidinium group of Arg244 is essential for (A) hydrogen bonding

interactions with the C3 carboxylate of the inhibitor in the active site, and (B)

coordination of a proton donating water molecule essential for the saturation of the

double bond of the C2 constituent (Imtiaz et al., 1993)

2-3 Synthesis of the chiral cephalothin boronic acid transition state inhibitor (A)

(+)-Pinanediol, THF, rt; (B) (dichloromethyl)lithium, THF, –100 °C → 0 °C; (C)

lithium bis(trimethylsilyl)amide, THF, –80 °C → rt; (D) 2-thiopheneacetylchloride, 2-

thiopheneacetic acid, THF –80 → rt; (E) aqueous HCl 3N, 1h, 100 °C

2-4 Timed inactivation of SHV-1 and Arg244Ser, -Gln, -Leu, and –Glu reveal that

Arg244 variants are inactivated more rapidly by clavulanate than SHV-1. Enzymes

were incubated with KI concentrations of clavulanate and initial velocities of

nitrocefin hydrolysis measured at timepoints 0-600 seconds

2-5 (A) Deconvoluted mass spectra of SHV-1 and Arg244Ser before and after 15-

minute incubation with clavulanate show similar intermediates of inactivation.

Spectra were obtained on a Q-STAR XL quadrupole-time-of-flight mass

spectrometer equipped with a nanospray source. Eight distinct mass shifts were

identified with both enzymes. (B) Proposed intermediates in the clavulanate

inactivation pathway. The 198 Da and 173 Da adducts are represented by more

than one candidate structure (blue). The terminally inactivated 52 Da crosslinked

species is highlighted in red

2-6 The time-dependent inhibition of the chiral cephalothin boronic acid transition

state inhibitor (compound 2) is more significant with SHV-1 than with SHV Arg244

Ser, -Gln, -Leu, and -Glu. Enzymes were incubated with inhibitor at concentrations

equal to the Ki and initial velocities of nitrocefin hydrolysis assessed at 0-3600s.

Initial Ki determination was performed after a 5 minute incubation

2-7 Arg244 in relation to Val216 in the SHV-1 (1SHV) and TEM-1 (1BTL) in apo-

enzyme crystal structure representations. The distance between the backbone

carbonyl oxygen of Val216 and the nearest guanidinium nitrogen in SHV is 7.85 Å.

In contrast, the distance in TEM-1 is 5.19 Å. Shown in TEM-1 is the bridging water

molecule which makes hydrogen bonds with Val216 (2.87 Å) and Arg244 (3.08 Å).

The active site Ser70 is shown for reference

2-8 Suggested role of Arg 244 in stabilizing the β-lactam carboxylate of

penicillins, cephalosporins, and carbapenems. (A) The ring structures of ampicillin

(purple), cephalothin (yellow), and meropenem (pink) were drawn using sketcher

(Accelrys®) and energy minimized. The three structures were overlayed on the β-

lactam ring and the proposed orientation in the active site is shown. (B) The

penicillin (C) cephalosporin, and (D) carbapenem ring structures in relation to

Arg244 and Thr235 (catalytic Ser70 shown for reference). The hypothesized

proposed bonding interactions are drawn in white

3-1 The chemical structures of the clinically available β-lactamase inhibitors,

sulbactam, and clavulanic acid, as well as the novel inhibitors, penem 1 and 2.

3-2 Deconvoluted ESI-MS spectra of SHV-1 and R244S reacted with 1000:1

sulbactam:enzyme for 15 minutes. Notice that at this time point most of the SHV-1

has been recycled, while there is no unmodified R244S

3-3 Deconvoluted ESI-MS spectra of SHV-1 and R244S reacted with 500:1

Penem 1/2: enzyme for 15 minutes. At this time point there is no unreacted SHV-1

or R244S. The mass spectrometry indicates that the inhibitor covalently attaches

to the enzyme without fragmentation, as is seen with sulbactam. Interestingly there

is some evidence for the addition of multiple inhibitor molecules

3-4 SHV E166A with the trans-enamine intermediates of clavulanate (top), and

sulbactam (bottom). Important residues for interactions with R244 (green) are

shown, and the inhibitors are highlighted in yellow

3-5 Proposed endo-trig cyclization of penem 1 with SHV-1. This results in an

enzyme adduct equal to the size of the full inhibitor molecule with no fragmentation

4-1 Structures of cefotaxime (A), cefotaxime BATSI (B), ceftazidime (C),

ceftazidime BATSI (D), and a Reference Compound (E) which lacks the oxyimino-

cephalosporin R1 side chain. The arrow in (A) indicates the direction of proposed

rotation in the binding pocket as discussed in text

4-2 Progress curves of SHV-1 inhibited by ceftazidime BATSI (left) and the

cefotaxime BATSI (right) show a primarily competitive mode of inhibition

4-3 Timed inactivation experiment with the ceftazidime BATSI (left) and

cefotaxime BATSI (right) indicate a fast equilibrating inhibition without time-

dependance

5-1 The inhibition of SHV-1 (black), SHV D104K (red), SHV D104E (green), and

TEM-1 (blue) by BLIP. The D104E substitution increases the affinity for BLIP over

1000-fold, nearly to the level of TEM-1

5-2 The alignment of BLIP (colored) as it was determined in the crystal structures

of SHV-1 (cyan), TEM(orange), and SHV D104K (red). (A and B) BLIP aligns

almost identically in the active sites of SHV-1 and TEM-1. (C and D) In contrast,

the SHV D104K variant forces the F142 binding loop out of the active site pocket

of SHV

5-3 The interaction of the binding loops of BLIP with key residues of SHV-1 (A)

and SHV D104K (B). Notice the rotation of the D49 binding loop in (B) brings Y51

of BLIP into contact with Y105 of SHV. Also notable in (B) is the absence of the

BLIP F142 loop

5-4 Representation from the crystal structures of position 104 of SHV D104E

(pink), and TEM (yellow) in complex with BLIP reveals a similar geometry between

the SHV D104E:BLIP and TEM:BLIP co crystal structures. Hydrogen bonds are

pictured with the backbone at BLIP Y143, (shown in cyan for SHV and blue for

TEM) and salt bridges with BLIP K74. Figure courtesy of Melinda Hanes,

University of California-San Diego

5-5 A representation from the BLIP co-crystal structures with SHV (pink) and

SHV D104E (yellow) indicates electrostatic clash between D104 in SHV and E73

of BLIP (blue) is relieved somewhat when aspartate is mutated to glutamate in

SHV. Figure courtesy of Melinda Hanes, University of California-San Diego

5-6 The inhibition of KPC-2 by BLIP. The ultra high affinity (low picomolar) is

beyond the sensitivity of our spectrophotometer, resulting in poor curve fitting

5-7 Sequence alignment of KPC-2 with SHV-1 and TEM-1. Residues conserved

across all three enzymes are colored red. Green indicates residues conserved

between 2 of 3 enzymes. Blue indicates conservation of amino acid character

LIST OF TABLES

2-1 MICs (μg/ml) of E. coli DH10B containing SHV-1 and all 19 variants at Arg244

2-2 Kinetic properties of SHV-1 and variants at Arg244 for clavulanate, ampicillin,

and nitrocefin

2-3 Dissociation constants (Kd, or KI* μM) of β-lactam substrates and Boronic Acid

Transition State Inhibitors for SHV-1 and Arg244 variants

3-1 MICs (μg/ml) of E. coli DH10B containing SHV-1 and all 19 variants at Arg244.

Inhibitors were evaluated in the presence of 50 μg/ml ampicillin

3-2 Kinetic properties of SHV-1 and variants at R244 for sulbactam and

clavulanate

3-3 Kinetic properties of SHV-1 and variants at R244 for penem 1 and penem 2

4-1 Inhibition constants (KI) of BATSIs with SHV-1, SHV-2, SHV-5, D104K and

G238S-D104K

5-1 Inhibition constants of Class A β-lactamases, studied herein, with BLIP

5-2 r.m.s.d values between β-lactamases (blue) and BLIP molecules (yellow) in

the structures studied

ACKNOWLEDGMENTS

I owe very special thanks to so many who have helped me on this journey:

James Tracy, Ph.D. for introducing me to the world of science and giving me the experience and confidence to pursue my passion

Leslie Webster M.D. for that fateful telephone call in my senior year in college convincing me that the Case MSTP was the place for me.

Chris Bethel for teaching me all there is to know about kinetics and protein purification. You are such a patient teacher and always willing to lend a hand

Kristine Hujer for all of our late night brainstorming sessions and belly laughs. You were so essential for the maintenance of my sanity over the past three years!

Andrea Hujer for your amazing eye for detail and countless hours of proofreading.

Kimberly Reynolds, Melinda Hanes, and Tracy Handel for your rewarding and enjoyable collaborations on the BLIP study

Fabio Prati and Wyeth Pharmaceuticals for supplying me with admirably designed inhibitors for my study

Vernon Anderson for helpful insights on kinetic mechanisms

Anne Distler and Charles Hoppel for assistance with mass spectrometry

Magdalena Teracila for all your help with molecular modeling, protein purification, and crystallization. You are one of the most genuine and giving people I have ever met and I am proud to have you as a friend

Marion Helfand for all the scientific brainstorming and support you have given me

Maja Babic, Steve Marshall, Mohamed Yassin, and Brooke Decker for comic relief and emotional support

John Nilson, Krzysztof Palczewski, and the faculty and staff of the Department of Pharmacology for creating a wonderful environment for learning

Mike Maguire, John Mieyal, and George Dubyak for being a supportive and helpful thesis committee and providing such useful feedback.

Clifford Harding, Deidre Gruning, Donna MciIlwain, and Case MSTP for financial, educational and emotional support.

My parents for supporting me throughout my training, despite thinking that I was totally insane for pursuing a PhD. You two are the most amazing people I have ever met and I am so lucky to be able to call you Mom and Dad

Justine Walker, Colleen Hickey, and Matthew Lalonde for being the most incredible roommates I could have asked for. All three of you helped me through so many rough points of the past 6 years and I couldn’t have made it through without your support

Jen Collister, Janet Evans, The Cleveland Triathlon Club, and the group from PStriathlon for introducing me to another true passion of mine and being sources of inspiration for your kindness and levels of dedication unmatched by any other group of people I have ever met.

And last, but certainly not least, Robert A. Bonomo. I came to you as an orphan 4th year MSTP student without a home and you believed in me from day one. I have never met a more humble, thoughtful, and caring person than you. Not only have you taught me countless lessons about science, but also about life. I look to

you as a model for the kind of doctor I want to be someday. You have such compassion in your heart, and fire in your soul. Thank you for being a person that I will always consider my mentor, and am honored to have as a friend.

ABBREVIATIONS

EI- electron impact

bla- β-lactamase gene

BLIP- β-lactamase inhibitory protein

br- broad signal

d- doublet

E.C.- enzyme classification

EDTA- ethylenediaminetetraacetic acid

EIMS, electron impact mass spectrometry;

ESI-MS- electrospray ionization mass spectrometry

HPLC- high performance liquid chromatography

IPTG- isopropyl β-D-1-thiogalactopyranoside

IR- infrared

IRT- inhibitor resistant TEM

J- coupling constants

KPC- Klebsiella pneumoniae carbapenemase m- multiplet

MIC- minimal inhibitory concentration

Mp- melting point

NMR- nuclear magnetic resonance

PBP- penicillin binding protein

PBS- phosphate buffered saline

ppm- parts per million

q- quartet

r.m.s- root mean square

Rt- room temperature

s- singlet

SDS-PAGE- sodium dodecyl sulfate polyacrylamide gel electrophoresis

SHV- sulfhydryl Reagent Variable

t- triplet

TEM- Temoneira (patient from which the first isolate was identified)

THF- tetrahydrofuran

TMS- tetramethyl silane

WT- wild type SHV-1

δ- chemical shifts

Overcoming Inhibitor Resistance in the SHV β-lLactamase

ABSTRACT

JODI MICHELLE THOMSON

β-Lactam antibiotics have been the cornerstone of antimicrobial chemotherapy since the discovery of penicillin. Unfortunately, resistance to these life-saving agents developed as a result of drug hydrolyzing β-lactamase enzymes.

To circumvent this problem, β-lactamase inhibitors (clavulanate, sulbactam and tazobactam) were produced to target the prevalent class A β-lactamases, SHV and TEM. Single amino acid substitutions have arisen in TEM and SHV family β- lactamases that give rise to inhibitor resistance.

One amino acid position in the TEM β-lactamase that is a “hotspot” for inhibitor resistance is Arg244. Fifteen unique isolates with 6 different amino acids at this residue were identified from clinical isolates possessing TEM. Despite this, the highly similar SHV enzyme remained unchanged at this position. To anticipate the likelihood of inhibitor resistance developing in SHV, and to probe whether this is indicative of a mechanistic difference between these β-lactamase families, we performed extensive analysis of variants at Arg244 in SHV.

Our studies revealed unique properties of inhibitor resistance in SHV. As in

TEM, most substitutions at Arg244 lead to the clavulanate resistant phenotype.

However, the mechanism of resistance is unique and based solely on a reduction in inhibitor affinity. In contrast, all mutants exhibited increased susceptibility to sulbactam; this occurs despite an equal elevation in KI values. Further studies

revealed the mechanism to be highly reduced sulbactam turnover. It is interesting

that the phenotype of inhibitor resistance or susceptibility is dictated largely by the

properties of the inhibitor, despite the similarity in these compounds.

We also extended our studies to novel β-lactamase inhibitors. We first investigated penem inhibitors, which show clinical promise as broad spectrum

inactivators, but hadn’t been tested against inhibitor resistant strains. We also

tested boronic acid inhibitors, which are designed with the side chains of currently

marketed antibiotics. Not only are these high affinity transition state inhibitors, but

also great probes for active site interactions.

Finally, we examined the dynamics of β-lactamase inhibition and protein-

protein interaction of SHV with the β-lactamase inhibitor protein (BLIP). We

performed mutagenesis to construct a tight binding variant of SHV and analyzed

the inhibition of the wild type and mutant enzymes.

Chapter 1

Introduction

β-LACTAMASE OVERVIEW

The discovery of penicillin by Alexander Fleming in 1928 and its subsequent isolation and chemical characterization by Howard Florey and Ernst

Chain in the 1940’s were among the most important breakthroughs in modern medicine. Mass production of penicillin marked the beginning of the “Antibiotic Era” and saved the lives of countless troops in World War II. Subsequently, the drug served as the scaffold for the development of several classes of β-lactam antibiotics available today: the penicillins, monobactams, five “generations” of cephalosporins, and the most potent class, the carbapenems (Figure 1-1, page

32).

These life-saving drugs act by attacking key enzymes in cell wall biosynthesis, the penicillin binding proteins (PBPs). These enzymes are responsible for the final step in cell wall synthesis, a transamidation, which crosslinks growing peptidoglycan strands, stabilizing the cell wall. The β-lactam bond, the conserved component of this antibiotic class, mimics the D-Ala:D-Ala linkage that is the natural substrate of PBPs (Figure 1-2, page 33). Constrained in a four-membered ring system, the β-lactam bond is highly reactive and an efficient inhibitor of PBPs.

The effectiveness of these “front-line” antimicrobial agents, however, is constantly threatened by an expanding microbial arsenal of β-lactamases. These 17

enzymes are thought to have evolved from PBPs and comprise a very large and

diverse group of proteins (Meroueh et al., 2003). Like the PBPs, β-lactamases

attack the β-lactam bond of the antibiotic (k1 and k2); however, they are able to

efficiently hydrolyze the β-lactam (k3). This effectively renders the drugs inactive

before they can reach the PBPs and also results in regeneration of the active β- lactamase.

β-Lactamases likely evolved independently from assorted PBPs. Found in

hundreds of species of Gram-negative and Gram-positive bacteria, β-lactamases

continue to be discovered at a rapid pace. They are grouped into 4 classes based

on amino acid homology:

Class A- Serine penicillinases. Examples- TEM, SHV, KPC, CTX-M

Class B- Metallo carbapenemases. Examples- IMP, VIM

Class C- Serine cephalosporinases Examples- AmpC, P99, CMY-2

Class D- Serine Oxacillinases. Example- OXA

This body of work focuses on the Class A β-lactamases which are prevalent

in the Gram-negative bacteria Klebsiella pneumoniae (SHV and KPC). The

Escherichia coli derived TEM-1, a well studied enzyme which shares a high level

of amino acid identity with SHV-1 (68%), is often used for comparison.

The tertiary structure and key catalytic residues of Class A serine β- lactamases are highly conserved. The proteins consist of a network of α helices to the left of the active site pocket, with a stretch of five β sheets to the right. The floor of the active site pocket is comprised of a large loop structure termed the “omega loop” (Figure 1-3, page 34)

There are a number of catalytically important residues in the active site of

Class A β-lactamases as well as residues responsible for altering the enzyme substrate profile (Figure 1-4, page 35):

1. Ser70- The catalytic residue of class A β-lactamases. The hydroxyl oxygen is the nucleophile which attacks the β-lactam bond of the antibiotic and forms a covalent complex (i.e. E-S complex). Most mutants at this position are inactive, and have not been identified in nature;

2. Ser130- Donates a proton to the β-lactam nitrogen for substrate hydrolysis.

Mutants are resistant to β-lactamase inhibitors and have reduced affinity for substrates;

3. Glu166- The catalytic base essential for activating the active site water molecule

resulting in regeneration of free enzyme (deacylation). In addition, Glu166 also

serves in acylation. Mutants are deacylation deficient and have not been identified

clinically, yet have been studied extensively;

4. Gly238- This residue is located in the B3 β strand on the right side of the active site pocket. The Gly238Ser substitution results in active site pocket widening and the extended spectrum β-lactamase phenotype (ESBL), which enables the enzyme to hydrolyze oxyimino-cephalosporins;

5. Glu240- Located adjacent to Gly238 (there is no residue 239 for alignment purposes among class A β-lactamases (Jacoby, 2006)), this amino acid is postulated to be important for cephalosporin recognition. When lysine is substituted at this position, enhancement of the ESBL phenotype is noted;

6. Asp104- Located on the left side of the active site pocket, this residue is important for recognition of substrates with large/complex R1 side-chains. Thus, certain substitutions at this position increase binding of advanced generation cephalosporins, enhancing the ESBL phenotype.

7. Arg244- Located in to the right of the active site pocket in the B4 β-strand, it is implicated in recognition of the C3 carboxylate of penicillins and inhibitors and the

C4 carboxylate of cephalosporins (Figure 1-5, page 36). Substitutions at Arg244 in

TEM result in the inhibitor resistant phenotype.

β-LACTAMASE MEDIATED RESISTANCE

Mutations at residues described above, as well as others, have contributed

to the expansion to 100 unique SHV and 160 TEM clinical variants. Considering

each enzyme variant has a slightly different substrate profile, this diversity severely

compromises the use of all β-lactam antibiotics. Further compounding the problem

is the location of many of these enzymes on transferable plasmids, which increases the likelihood of intra- and inter-species spread of resistance genes.

One strategy that is employed to retain the effectiveness of β-lactam antibiotics in the face of this challenge is the discovery and use of compounds to inhibit the β-lactamases. The first inhibitor to be introduced, clavulanate, was discovered as a natural product of the bacterium Streptomyces clavigularis.

Subsequently, modifications to the structure were made in the development of two additional inhibitors, sulbactam and tazobactam (Figure 1-6, page 37).

Although these inhibitors share much structural similarity with β-lactam antibiotics, they differ in that they display a proficient leaving group at position 1 of the second, 5-membered, ring (Figure 1-5, page 36). This leads to secondary ring opening and enzyme modification through a pathway illustrated in Figure 1-7

(page 38). There has been much interest concerning which species along the inactivation pathway accounts for β-lactamase inhibition in vivo. Mass 21

spectrometry studies performed in liquid solution suggest that on a time scale of 15

minutes, a predominance of the smaller fragments (52, 70, 88 Da adducts) are

formed (Pagan-Rodriguez et al., 2004; Sulton et al., 2005). This includes what is

hypothesized to be an end product of inhibition (i.e. a crosslinked enzyme between

Ser70 and Ser130) (Figure 1-7I, page 38). In contrast, Raman crystallography and

structural determinations of deacylation deficient mutants revealed that the early,

trans-enamine intermediate (Figure 1-7D, page 38) is the dominant species

(Padayatti et al., 2005). In either case, the inhibition mechanism involves modification of the catalytic serine when these inhibitors are studied. All three inhibitors (clavulanate, sulbactam, and tazobactam) appear to follow similar reaction pathways.

The β-lactamase inhibitors are directed primarily against the class A enzymes, SHV and TEM; these are still among the most prevalent β-lactamases encountered in clinical isolates. The inhibitors do not have anti-bacterial activity themselves, as they do not inactivate most PBPs. As such, they have to be used in combination with a β-lactam antibiotic and act as shields against β-lactamase enzymes. However, as has been “the curse” for the β-lactam antibiotics, mutations in β-lactamase enzymes are contributing to the development of resistance to these important inhibitors.

To date, there have been 23 Inhibitor Resistant TEM (IRT) enzymes, and 3 inhibitor resistant SHV enzymes identified clinically. These mutations have clustered around specific residues, some of which were mentioned above. The main “hotspots” have been Ser130, Met69, Asn276, and Arg244. It was 22

hypothesized that mutations at Ser130 resulted in inhibitor resistance by

preventing the final, cross-linking step in β-lactamase inhibition (Figure 1-6, page

37) (Thomas et al., 2005). Similarly, resistance due to substitutions at Met69 is

thought to be an indirect result of a displacement of the same Ser130 residue

(Wang et al., 2002). Alternatively, mutations at the Met69 position likely distort the

oxyanion hole, an electron poor group of side chains that polarize the lactam

oxygen for nucleophillic attack. and retard inhibitor affinity (Lin et al., 1998). It is

likely that each contributes to clavulanate resistance.

Substitutions at Arg244 comprise the most numerous set of IRTs. There are

17 TEM variants identified with mutations at Arg244, including 6 different

substituted residues with diverse physical properties (Cys, Gly, His, Leu, Ser, and

Thr). Despite this mutational flexibility in TEM, position 244 has remained

completely unchanged in the highly similar SHV enzyme.

EXPLORING THE ROLE OF ARG244 IN TEM

One structural element that is essential in the design of β-lactam based

drugs is a carboxylate at position 3 on penicillins and inhibitors and position 4 of

cephalosporins (Figure 1-5, page 36). It is postulated that the C3/C4 carboxylate is essential for proper recognition and alignment in the active site of the PBPs and β- lactamases (Zafaralla et al., 1992). There have been conflicting studies concerning the nature of this recognition. Early studies suggest that hydrogen bonds in part from a conserved residue Arg244 contribute to binding (Zafaralla et al., 1992). 23

Recent studies point to a combination of hydrogen bonds along with a coordinated

water molecule held in place between Arg244 and the backbone carbonyl oxygen

of Val216 (Figure 1-8, page 39) (Diaz et al., 2003; Imtiaz et al., 1993). An

additional function has been attributed to this water molecule in the inactivation of

β-lactamases by clavulanate: the donation of a necessary proton for secondary ring opening (Imtiaz et al., 1993). This notion is supported by the discovery,

mentioned above, of several inhibitor resistant TEM (IRT) β-lactamases with mutations at position 244 (Bret et al., 1997; Jacoby, ; Lemozy et al., 1995). The

first of these to be found clinically, an Arg244Ser substitution, has been analyzed

structurally and was found to be missing this water. In addition, the rates of

inactivation by clavulanate were significantly slower in the Arg244 mutant

compared to wild type, which also argues for a mechanistic basis for resistance

(Imtiaz et al., 1993).

Mutations at position 244 in the SHV β-lactamase have not been clinically

isolated, but site-directed mutagenesis of Arg244 to Ser was shown to confer

clavulanate resistance in the highly similar OHIO-1 β-lactamase (Lin et al., 1999),

and subsequent studies showed a similar results with Arg244Ser and -Cys in SHV

(Giakkoupi et al., 1998). A detailed study of reaction intermediates in the inhibition of SHV by clavulanate with mutants at position 244 is necessary to determine the role of this residue in the inactivation pathway and clarify the important differences between these two prevalent class A enzymes.

THE CHALLENGE OF EXTENDED SPECTRUM β-LACTAMASES (ESBLs) 24

Another circumstance where single amino acid substitutions affect the

efficacy of β-lactam compounds is in the evolution of the Extended-Spectrum β-

Lactamase phenotype. Wild type SHV-1 β-lactamase confers high level resistance

to penicillins and first generation cephalosporins (Hujer et al., 2002). In an attempt

to overcome resistance mediated by these enzymes, extended spectrum

cephalosporins (i.e.”third generation”, Figure 1-1, page 32) were designed. These

β-lactam antibiotics contain bulky, rigid R1 side chains that are not easily

accommodated by the constrained active site pocket of SHV-1 or TEM-1 (Figure 1-

4, page 35). Single amino acid substitutions have subsequently been found in SHV

and TEM that allow for efficient hydrolysis of third generation, oxyimino- cephalosporins. Enzymes with this hydrolytic capacity are referred to as Extended-

Spectrum β-Lactamases (ESBLs). ESBLs are major contributors to antibiotic treatment failure and cause significant morbidity and mortality.

The residues that, when substituted, result in the ESBL phenotype are among those illustrated in Figure 1-4 (page 35). Our lab, in collaboration with Dr.

J.R. Knox, showed that a key G238S substitution opens up the binding pocket by up to 3 Ångstroms (Nukaga et al., 2003b). This primary mutation can be complemented by two other mutations outside of the binding pocket. The substitution of a lysine at position 240 (E240K), to the right of the binding pocket, enhances ceftazidime binding and hydrolysis. It has been postulated that the introduced positive charge facilitates a rotation of the aminothiazole-oxime group of ceftazidime (Huletsky et al., 1993). Interestingly, an increase in hydrolysis of 25

cefotaxime, another third generation cephalosporin is enhanced by the substitution for a lysine on the left side of the binding pocket at position 104 (Petit et al., 1995).

The exact mechanism for this increased activity is being investigated by our group.

THE ROLE OF POSITION 244 IN CEPHALOSPORIN BINDING

In addition to its role in the inactivation pathway of β-lactamase inhibitors, in vivo evidence suggests the importance of an interaction of Arg244 with cephalosporins. This comes from a study on SHV β-lactamases in an attempt to create an inhibitor resistant ESBL (Randegger and Hachler, 2001). The Arg244Ser substitution was introduced into SHV enzymes with the ESBL conferring mutations

G238S and the combination G238S / E240K. The 244Ser substitution completely abolished the ability of the ESBL enzymes to confer resistance to cefotaxime and ceftazidime in E. coli. There has recently been a report of an extended spectrum/inhibitor resistant TEM enzyme, however it required five amino acid substitutions and had diminished inhibitor resistance (Dubois et al., 2004). Kinetic study of the interaction between these mutant enzymes and cephalosporins is complicated by low turnover and auto-hydrolysis of the compounds in aqueous solutions (personal observation). However, probing this interaction may reveal additional clues as to why Arg244 has remained conserved in SHV.

BORONIC ACID TRANSITION STATE INHIBITORS: KEYS TO

UNDERSTANDING STRUCTURE ACTIVITY RELATIONSHIPS 26

With the development of resistance to advanced generation β-lactam antibiotics as well as to the β-lactamase inhibitors, much focus has turned to the discovery of novel β-lactamase inhibitors. One strategy in inhibitor design is to create a compound that closely mimics an enzyme’s natural substrate. Employing this strategy, β-lactamase inhibitors have been developed that include the R1 side chain of third generation cephalosporins (Figure 1-9, page 40)(Caselli et al., 2001)

(Powers et al., 2001). Attached to boric acid, a functionality with natural affinity for

the active site serine of β-lactamases, these compounds reversibly bind with

transition state geometry that resembles the natural substrate. Testing these

compounds against representatives of several classes of β-lactamase enzymes, a

wide range of affinities spanning from low nanomolar to high millimolar have been

found (Caselli et al., 2001; Powers et al., 2001; Wang et al., 2002).

Because of their design the boronic acid transition state inhibitors (BATSIs)

are not only effective inactivators, but also wonderful probes for active site

interactions. With these compounds, a detailed analysis of how ESBL producing mutations affect binding to the R1 side chain of advanced generation cephalosporins can be performed for the first time. For drug development purposes, it is also essential to determine how sensitive the potency of these compounds is to single amino acid substitutions within the binding pocket of β- lactamases. In addition, using boronic acid inhibitors with and without the C4 carboxylate will gain insight into the role of Arg244 in cephalosporin binding

(Thomson et al., 2006). Our studies will focus on ceftazidime and cefotaxime 27

based inhibitors for probing of ESBL defining mutations, and novel cephalothin

based inhibitors with and without a C4-like carboxylate for our studies with Arg244

(Figure 1-9, page 40).

THE PROMISE OF PENEM INHIBITORS

Another group of promising inhibitors is being developed by Wyeth

Pharmaceuticals. Carbapenems (Figure 1-1, page 32) are among the most potent

β-lactam antibiotics currently on the market. They are efficient inactivators of

PBPs, and evade hydrolysis by most β-lactamases in nature. An interesting

observation with these compounds has been that they are effective inhibitors of

many class A β-lactamases (Maveyraud, 1998). With that in mind, a series of

inhibitors has been developed with the penem backbone structure. The

methylidene penems (Figure 1-10. page 41) have several advantages over

traditional β-lactamase inhibitors. Firstly, they contain large hydrophobic R1 side chains. This side chain not only aids in cell permeability, but also adds an

additional group for affinity docking in the active site. Secondly, they have a novel

mechanism of inhibition. Rather than breaking down in the active site pocket like

clavulanate (Figure 1-7, page 38), they undergo endo-trig cyclization to form a

stable product covalently attached to the Class A, C, and D enzymes (Figure 1-11,

page 42). However, an essential property of novel β-lactamase inhibitors would be

their effectiveness against inhibitor resistant β-lactamases. This has yet to be 28

determined and will be studied herein for SHV β-lactamases with mutations at

R244.

β-LACTAMASE INHIBITORY PROTEIN (BLIP)

As mentioned above, clavulanic acid was initially discovered as a natural

product of the bacterium Streptomyces clavigularis (Reading and Cole, 1977).

Interestingly, a second novel β-lactamase inhibitor, a protein, was subsequently isolated from the same bacterium. The β-lactamase Inhibitory Protein (BLIP) is a

17.5 kDa protein cloned in 1990 (Doran et al., 1990). Its presence was detected

due to a disparity between clavulanic acid quantification and β-lactamase inhibitory

activity of S. clavigularis cell lysates. The isolated protein was found to be a very

tight binding inhibitor of the Escherichia coli TEM β-lactamase, and subsequent studies have found BLIP to be a nanomolar inhibitor of disparate Class A β- lactamases that share as little as 30% sequence identity (Zhang and Palzkill,

2004). One enigma of BLIP inhibition, however, is that it binds SHV-1 nearly 1000- fold more weakly than TEM-1. This is despite high homology between the two enzymes.

Insight into this discrepancy was gained by several structural and kinetic studies. Co-crystallization of the class A TEM-1 enzyme with BLIP, which binds with 1 nM affinity, was performed and revealed two major loops of the inhibitor in contact with the enzyme (Strynadka et al., 1996). Subsequently, a study using alanine scanning mutagenesis provided insight into the discrepancy in SHV 29

binding: inhibition of SHV, unlike the other class A enzymes, was largely unaffected by substitution of alanine into the second binding loop of BLIP (Figure

1-12, page 43) (Zhang and Palzkill, 2004). It seems that BLIP is able to inhibit SHV to micromolar affinity primarily with the interactions of a single binding loop.

Residue 104, earlier mentioned to be important for cephalosporin binding, mediates key binding to BLIP in this loop. SHV differs from the tight-binding TEM enzyme at this position, but only slightly; SHV-1 contains aspartate, while TEM-1 has a glutamate. We tested the flexibility of BLIP binding by mutating position 104 in SHV to lysine and glutamate and performing kinetic and structural studies.

The high affinity and plasticity of BLIP binding is encouraging in the search for novel β-lactamase inhibitors. As more protein-protein binding interfaces are examined, the prospect of a small peptide mimetic gets more realistic. Because of this, we next chose to study a threatening new Class A enzyme, Klebsiella pneumonia carbapenemase (KPC), with BLIP. KPC-1 was discovered in North

Carolina from a carbapenem resistant K. pneumoniae isolate (Yigit et al., 2001).

Soon after, a similar enzyme, KPC-2, differing in just 1 amino acid, was isolated from 4 isolates in Maryland (Smith Moland et al., 2003). The plasmid borne gene has since spread across the eastern half of the US and has caused major outbreaks of multi-drug resistant Klebsiella, notably in New York City (Bratu et al.,

2005; Lomaestro et al., 2006; Woodford et al., 2004). Isolates harboring KPC β- lactamases are refractory to all but the most toxic drug regimens, so discovering a novel inhibitor that effectively inactivates these dangerous enzymes is essential.

STATEMENT OF EXPERIMENTS

Resistance to β-lactamase inhibitors is growing and threatens the use β- lactam antibiotics for serious infections. To be successful, the search for novel inhibitors must include two distinct goals: Understanding the mechanism of resistance to current antimicrobials and development of novel inhibitors that are not susceptible to the same pitfalls. This study has taken aim at Arg244, an amino acid that, when mutated, results in β-lactamase inhibitor resistance in TEM-1. Our study aims to determine the role of Arg244 in SHV and characterize the resistance mechanism. Furthermore, novel small molecule inhibitors are tested and used as probes for active site chemistry. Finally, we explore the binding and affinity of BLIP and determinants of tight binding with this unusual large protein inhibitor.

Figure 1-1 Examples of the 8 generations of β-lactam antibiotics currently on the market exemplify conservation of the lactam backbone and diversity of the side- chains 32

Figure 1-2 The mechanism of attachment of the penicillin binding protein (PBP) to its natural D-Ala D-Ala substrate (top) and a β-lactam antibiotic (bottom) is strikingly similar.

Figure 1-3 The Crystal structure of SHV β-lactamase (Protein Data Bank accession number 1SHV)

Figure 1-4 The SHV β-lactamase active site pocket, highlighting important residues for catalysis and substrate specificity

Figure 1-5 The numbering scheme for the backbone nucleus of penicillins and cephalosporins

Figure 1-6 Structures of the β-lactamase inhibitors currently on the market. The main difference between these and the beta-lactam antibiotics is the presence of a good leaving group at the 1 position.

Figure 1-7 The pathway of inhibition by clavulanate in the active site of class A β- lactamases is a multi-step process with numerous intermediates 38

Figure 1-8 The proposed contribution of Arg244 to clavulanate inactivation of

TEM-1: (A) Direct hydrogen bonding interactions (B) Coordination of a catalytic water molecule essential for inactivation

Figure 1-9 Structures of cephalosporins and corresponding boronic acid transition state inhibitors (BATSIs): (A) Ceftazidime (B) Ceftazidime BATSI (C) Cefotaxime

(D) Cefotaxime BATSI (E) Cephalothin (F) Achiral Cephalothin BATSI (G) Chiral

Cephalothin BATSI

Figure 1-10 Chemical structure of novel penem inhibitors (1 and 2) developed by

Wyeth Pharmaceuticals which display large, bicyclic R1 side chains.

Figure 1-11 Endo-trig cyclization of penem 1 in the active site of a β-lactamase results in the formation of a seven-membered ring.

Figure 1-12 The results of alanine scanning mutagenesis of BLIP indicated fewer

“hotspots” of BLIP binding to SHV-1 (left) compared to TEM-1 (right). Residues

identified as critical for binding are in red (reduce binding >10 fold when mutated),

residues that, when mutated, lead to an increase in affinity are colored cyan, and

residues that had a < 10 fold effect on affinity when mutated are colored green

(Zhang and Palzkill, 2004).

CHAPTER 2

SHV Arg244- Clavulanate Resistance

Published as

PROBING ACTIVE SITE CHEMISTRY IN SHV β-LACTAMASE VARIANTS AT

AMBLER POSITION 244: UNDERSTANDING UNIQUE PROPERTIES OF

INHIBITOR RESISTANCE

Jodi M. Thomson, Anne M. Distler, Fabio Prati, and Robert A. Bonomo

The Journal of Biological Chemistry

September 8, 2006 Volume 281(36): 26734-44

INTRODUCTION

β-lactams have been the cornerstone of antibacterial chemotherapy since the

introduction of penicillin. Although this family of antibiotics now contains numerous

classes (penicillins, cephalosporins, and carbapenems, Figure 2-1, page 73), their

use is threatened by the expansion and spread of β-lactamase enzymes

(Thomson and Bonomo, 2005). These bacterial enzymes hydrolyze β-lactam

antibiotics before reaching their target, the penicillin binding proteins. In an effort to

retain the utility of several generations of life-saving β-lactam antibiotics, β- lactamase inhibitors (clavulanate, sulbactam, and tazobactam) have been developed. These mechanism-based inhibitors are β-lactam compounds that experience multi-step reactions within the active site of β-lactamase enzymes

(Helfand et al., 2003b; Pagan-Rodriguez et al., 2004; Sulton et al., 2005). Typically

the inhibitors lack antimicrobial activity, and are formulated with a β-lactam antibiotic to act as shields against β-lactamase enzymes. This allows the β-lactam to bypass the β-lactamase and inactivate the penicillin binding proteins.

Unfortunately, inhibitor resistant class A β-lactamases are emerging in the clinic and undermining the use of β-lactam/β-lactamase inhibitor therapy (Blazquez et al., 1993; Imtiaz et al., 1994b; Zhou et al., 1994). Interest has been renewed in the discovery of novel β-lactamase inactivators to circumvent these inhibitor resistant enzymes (Buynak, 2004; Morandi et al., 2003; Wang et al., 2003; Weiss et al., 2004). In this context, detailed analyses of enzymes resistant to the inhibitors are needed to minimize cross-resistance and guide discovery.

Both Ser130Gly and Met69Ile mutations conferring inhibitor resistance have been described in TEM and SHV β-lactamase families (enzymes with 68% amino acid identity) (Jacoby, 2006). A notable exception, however, is the absence of clinical variants with substitutions at Ambler position Arg244 in SHV compared to

TEM. Arg244 is situated at the periphery of the active site pocket in the B4 β strand (Ambler et al., 1991). Although 15 distinct inhibitor resistant TEM (IRT) enzymes have been reported with substitutions at this residue, mutations at

Arg244 in SHV have not been described in clinical isolates (Jacoby, 2006).

The mechanism of clavulanate resistance resulting from substitutions at position 244 in TEM was previously examined (Delaire et al., 1992; Imtiaz et al.,

1993; Zafaralla et al., 1992). Structural studies of TEM-1 show that Arg244 coordinates a water molecule with the backbone carbonyl of Val216 (Wang et al., 45

2003). This water molecule is postulated to play a key role in inhibitor affinity, and

provide the proton source essential for terminal inactivation of the enzyme (Figure

2-2, page 74) (Imtiaz et al., 1993). In addition, the guanidinium group of Arg244

contributes directly to substrate affinity through hydrogen bonding with the

conserved β-lactam carboxylate (Wang et al., 2003) (32-1 and Figure 2-2, pages

73-74). IRT β-lactamases with substitutions at Arg244 exhibit reduced rates of

inactivation (kinact) as well as diminished affinity for clavulanate and substrates

(Delaire et al., 1992; Imtiaz et al., 1993).

We hypothesized that the absence of mutations identified in SHV at this

position, despite heavy clinical drug pressure, is an indication of important

differences in active site chemistry between the two highly similar enzymes. To

test this hypothesis, we investigated the effects of multiple amino acid substitutions

by site-saturation mutagenesis and characterized select SHV β-lactamases by steady state kinetics. To compare the nature of the intermediates in the inactivation process, we inactivated SHV-1 and SHV Arg244Ser, a clavulanate resistant variant, with clavulanate and resolved the adducts by electrospray ionization mass spectrometry (ESI-MS). In addition, we studied the inhibition of several SHV β-lactamase variants at position 244 using two boronic acid transition state analogs that are chemically related to the antibiotic cephalothin. Our results reveal mechanistic differences in clavulanate inactivation between SHV-1 and

TEM-1 and suggest a new role for Arg244 in SHV.

EXPERIMENTAL PROCEDURES

Plasmids and Mutagenesis - blaSHV-1, subcloned into phagemid vector pBC SK(-)

(Stratagene, La Jolla, CA) and maintained in ElectroMAXTM DH10BTM T1R cells

(Invitrogen, Carlsbad, CA) as previously described (Rice et al., 2000) served as

the template for all mutagenesis described herein. Site-saturation mutagenesis

® was performed on the blaSHV-1 construct using the QuikChange II Site-Directed

Mutagenesis Kit (Stratagene, La Jolla, CA) and primers degenerate at Ambler

position 244. Two μL of the mutagenesis reaction was transformed into

Escherichia coli DH10B electrocompetent cells (Invitrogen) and plated on Luria

Bertani (LB) agar plates using chloramphenicol (Sigma-Aldrich, St. Louis, MO) (20

µg/ml) for selection. One-hundred colonies were screened for mutations at amino

acid position 244 by DNA sequencing with an ALF Express™ automated DNA

sequencer (GE, Piscataway, NJ). Thermo SequenaseTM fluorescent-labeled primer

cycle sequencing kit was used according to the protocol provided. Special attention was given to selecting mutants with the most common codon usage in

SHV-1.

Because the Arg244Lys, -Met, and -Phe mutants were not obtained in the initial screen, blaSHV-1(Arg244Lys), blaSHV-1(Arg244Met), and blaSHV-1(Arg244Phe) were

constructed by site-directed mutagenesis, using specific mutagenic

oligonucleotides as previously described (Hujer et al., 2002). DNA sequencing

confirmed the presence of the mutated codons as described above.

Immunoblots - E. coli DH10B cells were grown to OD600 = 0.8 and frozen.

Twenty microliters of each culture was lysed by ten minute incubation at 100º C in

sodium dodecyl sulfate loading dye buffer. Immunoblots were performed with an

anti-SHV polyclonal antibody to confirm expression of full length protein from all 20

constructs as previously described (Hujer et al., 2002).

Antibiotic Susceptibility - Minimal Inhibitory Concentrations (MICs) were

determined by the agar dilution method, using a Steer’s Replicator that delivers

104 CFU per 10 µL spot. Antibiotics tested include ampicillin, cephalothin,

piperacillin (all from Sigma-Aldrich), and lithium clavulanate (Glaxo Smith-Kline,

Surrey, England). Clavulanate susceptibility was determined in the presence of 50

μg/ml Ampicillin. MIC values reported are the most frequent number observed

(mode) in at least 3 separate experiments.

Protein Expression and Purification - E. coli DH10B cells containing the blaSHV-1 and blaSHV Arg244Ser, -Gln, -Leu, and -Glu genes in pBC SK(-) were grown overnight in SOB

medium, harvested by centrifugation at 4 °C and frozen. β-lactamase was liberated

using stringent periplasmic fractionation with 40 µg/ml lysozyme and 1mM EDTA

pH 7.8. Preparative isoelectric focusing was performed with the lysate in

Sephadex granulated gel using ampholines in the pH range of 3.5-10 (Amersham

Biosciences) as previously described (Lin et al., 1998). The protein was eluted and

dialyzed with 20 mM diethanolamine buffer, pH 8.5. Protein concentration was

assessed by Bio-Rad Protein Assay (Hercules, CA) with bovine serum albumin standards. Purity >95% was determined using 10% SDS-PAGE.

Kinetics - Kinetic constants of the β-lactamases chosen for study (SHV-1,

Arg244Ser, -Gln, -Leu, and -Glu) were measured by continuous assays at room temperature, 25 °C, using an AgilentTM 8453 diode array spectrophotometer

(Agilent, Palo Alto, CA). Each assay was performed in 20 mM phosphate-buffered saline (PBS) at pH 7.4. Measurements were obtained using nitrocefin (Becton

-1 -1 Dickinson, Franklin Lakes, NJ) (Δε482 = 17,400 M cm ) and ampicillin (Δε482 = -

-1 -1 900 M cm ). The kinetic parameters, Vmax and Km, were obtained with non-linear

least squares fit of the data (Equation 1) using Origin 7.5®.

v = (Vmax*[S])/(Km + [S]) (1)

The dissociation constants of the pre-acylation complex (Kd or KI) were determined

by direct competition with the indicator substrate nitrocefin for the substrates

(cephalothin, piperacillin, and meropenem). The KI values for the inhibitors

(clavulanic acid and the boronic acid inhibitors, compounds 1 and 2) were also

determined by direct competition since kinact/Ki was <<1.

Enzyme concentrations for these experiments were adjusted to the concentration that would result in an initial rate of nitrocefin hydrolysis between 1-

1.5 µM/s in the absence of inhibitor. Final concentrations were 7 nM (SHV-1), 10.5 nM (Arg244Gln), 21 nM (Arg244Ser, and –Leu) and 105 nM (Arg244Glu). 49

Nitrocefin concentrations equal to the Km (Table 2-2, page 71) were used for all experiments, except for Arg244Leu in which 200 µM was used due to absorbance

limitations. The data were analyzed according to Equation 2 to determine Ki :

i = [I] / ([I] + Ki (1 + ([S] / Km ))) (2)

Where i = fraction of enzyme inhibited, [S] = nitrocefin concentration, [I] =

inhibitor concentration, and Km refers to Km of the enzyme for nitrocefin (Cheng

and Prusoff, 1973).

The first-order rate constant for enzyme and inhibitor complex inactivation,

kinact, was measured directly by monitoring the reaction time courses in the

presence of clavulanate. A fixed concentration of enzyme and nitrocefin, and

increasing concentrations of clavulanate ([I]), were used in each assay. The kobs for

inactivation was determined graphically by Equation 3,

(-kobs * t) A = vf * t + (v0 - vf ) * (1 – e ) / kobs + A0 (3)

where A = absorbance, t = time (s) vf = final reaction velocity, v0 = the initial

reaction velocity in the first 5 seconds, and A0 = the initial absorbance. Each kobs

was plotted versus [I] and fit to Equation 4 to determine kinact.

kobs = kinact [I]/(Ki + [I]) (4)

The partitioning of the initial enzyme inhibitor complex between hydrolysis

and enzyme inactivation (kcat/kinact) was obtained in the following manner. First,

we incubated increasing amounts of clavulanate with a fixed concentration of β-

lactamase in a total volume of 40 µl of 20 mM phosphate-buffered saline, pH 7.4,

at room temperature. After 24 h, the sample was added to a 1 ml cuvette

containing PBS and nitrocefin (see above) and initial rates of hydrolysis were

assessed. The proportion of clavulanic acid relative to enzyme that resulted in ≥

90% inactivation after 24 h was the kcat/kinact for the enzyme.

Ki values for the boronic acid compounds were determined as described

above. However, because of time-dependant inactivation, enzyme and the

boronic acid compound 2 (Figure 2-1, page 73) were pre-incubated for 5 min in

PBS before initiating the reaction with the addition of substrate (Chen et al.,

2006; Chen et al., 2005; Wang et al., 2003).

Timed inactivation experiments were also performed with clavulanate and the boronic acid compound 2. To control for the difference in affinities, Ki concentrations of inhibitor were used in each assay. Enzyme and inhibitor were incubated in PBS for a range of time points before the reaction was initiated by nitrocefin (see above for enzyme and nitrocefin concentrations).

Mass Spectrometry - For intact protein mass spectrometry, 40 μM of SHV-1 or

Arg244Ser were incubated for 15 min with and without the addition of 40 mM

lithium clavulanate. Each reaction was terminated by the addition of 1:10 volume

1% trifluoroacetic acid, and immediately desalted and concentrated using a C4

ZipTip® (Millipore, Bedford, MA) according to the manufacturer’s protocol.

Samples were then placed on ice and analyzed within 2 h.

Spectra of the intact SHV-1 and SHV Arg244Ser proteins were generated on an Applied Biosystems (Framingham, MA) Q-STAR XL quadrupole-time-of- flight (TOF) mass spectrometer equipped with a nanospray source. Experiments were performed by diluting the protein sample with acetonitrile/ 1% formic acid to a concentration of 10 μM. This protein solution was then infused at a rate of 0.5

μL/min and data was collected for 2 minutes. Spectra were deconvoluted using the Applied Biosystems (Framingham, MA) Analyst program.

Synthesis of Boronic Acid Inhibitors – 1H and 13C nuclear magnetic resonance

(NMR) spectra were recorded on a Bruker DPX-200 or Avance 400 spectrometer;

chemical shifts (δ) are reported in parts per million (ppm) downfield from

tetramethyl silane (TMS) as internal standard (s singlet, d doublet, t triplet, q

quartet, m multiplet, br broad signal); coupling constants (J) are given in hertz.

Mass fragmentations were determined on a Finnigan MAT SSQ A mass

spectrometer (electron impact (EI), 70 eV). Optical rotations were recorded at 20

°C on a Perkin-Elmer polarimeter and specific rotations are in 10-1 deg cm2 g–1.

All reaction requiring anhydrous conditions were performed under argon using oven-dried glassware. Tetrahydrofuran (THF) was dried according to classical procedures and distilled from Na/benzophenone before use.

Chromatographic purification of the compounds was accomplished on silica gel (0.05-0.20 mm). 3-Ethoxycarbonyl-phenylboronic acid and (+)-(1S,2S,3R,5S)-

pinanediol (ee > 98%) were purchased from Aldrich. Synthesis of the achiral

cephalothin boronic acid transition state inhibitor 1 was performed as previously

described (Caselli et al., 2001). Enantioselective synthesis of the chiral

cephalothin boronic acid transition state inhibitor 2 was performed as follows

(Figure 2-3, page 75). (Morandi et al., 2003)

(+)-Pinanediol 3(ethoxycarbonyl)phenylboronate (+)-(3) –

Phenylboronic acid (1,250 g, 6.44 mM) and (+)-pinanediol (1.100 g, 6.44 mM) were dissolved in anhydrous THF (8 ml). The mixture was stirred for 1 h at room temperature (rt) and concentrated in vacuo. The crude product was purified by chromatography (light petroleum/diethyl ether 9:1) yielding the ester 3 as a white solid (2.060 g, 97%). Melting point (mp) 56-58 °C, [α]D: + 9.9 (c 0.9, CHCl3).

Infrared (IR): ν 2984, 2921, 1718.

1 H NMR (CDCl3): δ 0.88 (3H, s, pinanyl CH3), 1.20 (1H, d, J = 10.2 Hz, pinanyl

Hendo), 1.31 (3H, s, pinanyl CH3), 1.39 (3H, t, J = 7.1 Hz, O-CH2-CH3), 1.49 (3H, s,

pinanyl CH3), 1.89-2.54 (5H, m, pinanyl protons), 4.38 (2H, q, J = 7.2 Hz, O-CH2-

CH3), 4.46 (1H, dd, J = 8.6, 1.8, pinanyl CHOB), 7.44 (1H, dt, J = 0.5, 7.9 Hz, H5

arom), 7.98 (1H, dt, J = 7.4, 1.3, H4 arom), 8.13 (1H, dt, J = 7.8, 1.6, H6 arom), 8.48

(1H, br s, H2 arom).

13 C NMR (CDCl3): δ 14.4, 24.0, 26.5, 27.1, 28.7, 35.5, 38.2, 39.5, 51.4, 60.9, 78.4,

86.5, 127.7, 130.0, 132.2, 135.8, 139.1, 166.7 (C-B not seen)

Electron impact mass spectrometry (EIMS): m/z 328 (44%, M+), 313 (28%), 287

(25%), 283 (32%), 272 (20%), 259 (base peak), 245 (22%), 232 (76%), 204 (8%),

187 (15%), 177 (11%), 152 (15%), 134 (34%), 131 (20%), 109 (15%), 105 (30%),

96 (33%), 93 (13%), 83 (66%), 77 (15%), 67 (48%), 55 (25%).

Analytically calculated for C19H25BO4: C, 69,53; H, 7,68. Found: 69,66; H, 7,57.

(+)-Pinanediol (R)-(3-ethoxycarbonylphenyl)-(1,1,1,3,3,3-hexamethyl-disilazan-2-

yl)-methaneboronate (–)-4-

A solution of methylene chloride (220 μl, 3,4 mM) in THF (4 ml) was cooled at –

100 °C and treated with a 2.5 M solution of butyllithium in hexanes (1 ml, 2.56 mM)

under an argon flow and magnetic stirring: LiCHCL2 precipitated as a white

microcrystalline solid. After 30 min, a solution of the above pinanediol arylboronate

(+)-3 (700 mg, 2.13 mM) in anhydrous THF (6 ml) was added dropwise at –100 °C

over a 20 min period. The mixture was gradually allowed to reach 0 °C over 6 h

and stirred at this temperature for one additional hour. Thereafter, the solution was

cooled at –80 °C and lithium bis(trimethylsilyl)amide (1 M solution in hexane, 2.34

ml, 2.34 mM) added and the reaction mixture allowed to warm gradually overnight.

The resulting solution was partitioned between light petroleum (50 ml) and H2O (20 ml) and the organic phase was dried over MgSO4, filtered and concentrated under

reduce pressure. The residue was purified by column chromatography (light

petroleum/diethyl ether/triethylamine 70:30:2) affording 4 as a colourless viscous

oil (503 mg, 47%), [α]D: -1.7 (c 0.4, CHCl3), de ≥ 98%.

IR: ν 3404, 2936, 1720.

1 H NMR (CDCl3): δ 0.12 (18H, s, TMS), 0.87 (3H, s, pinanyl CH3), 1.29 (1H, d, J =

10.3 Hz, pinanyl Hendo), 1.33 (3H, s, pinanyl CH3), 1.40 ( 3H, t, J = 7.2, O-CH2-

CH3), 1.45 (3H, s, pinanyl CH3), 1.92-2.46 (5H, m, pinanyl protons), 4.12 (1H, br s,

CH-B), 4.37 (3H, m, O-CH2-CH3 and pinanyl CHOB), 7.34 (1H, t, J = 7.7 Hz, H5

arom), 7.71 (1H, d, J = 7.7 Hz, H4 arom), 7.84 (1H, d, J = 7.7 Hz, H6 arom), 8.24

(1H, m, H2 arom).

13 C NMR (CDCl3): δ 2.4, 14.3, 24.0, 26.5, 27.1, 28.3, 35.4, 38.2, 39.5, 46.6 (br, C-

B), 51.5, 60.6, 78.6, 86.2, 126.5, 127.5, 128.0, 129.8, 131.1, 145.5, 167.1.

+ (EIMS): m/z 339 (41%, M - N(TMS)2), 310 (8%), 294 (19%), 266 (10%), 190 (5%),

163 (100%), 161 (6%), 135 (38%), 119 (69%), 104 (10%), 91 (48%), 77 (19%), 65

(8%).

Analytically calculated for C26H44BNO4Si2: C, 62,25; H, 8,84; N, 2,79. Found: C,

62,09; H, 8,99; N, 2,64.

(+)-Pinanediol (R)-(3-ethoxycarbonylphenyl)-(2-thiophen-2-yl-acetylamino)-

methaneboronate (+)-5-

A solution of 2-thienylacetic acid (149 mg, 1.047 mM) and 2-thienylacetylchloride

(130 μl, 1.047 mM) in THF (4 ml) was slowly added to a cooled (– 80 °C) solution

of (–)-4 (500 mg, 0.997 mM) in anhydrous THF (10 ml). The resulting solution was

allowed to reach rt and to react for 15 hr; thereafter, the resulting mixture was

partitioned between n.hexane (50 ml) and H2O (20 ml), the aqueous phase was repeatedly extracted with n.hexane (2 × 20 ml). The collected organic phases were washed with saturated NaHCO3, dried over MgSO4, filtered and concentrated in

vacuo. The residue was purified by column chromatography (light

petroleum/diethyl ether 1:9) affording the amide 5 as a white solid (293 mg, 61%).

Mp 164-166 °C, [α]D: +15.5 (c 0.8, CHCl3), de ≥ 98%.

IR: ν 3173, 2979, 1721, 1612, 1276, 1201, 1090.

1 H NMR (CDCl3): δ 0.7 (3H, s, pinanyl CH3), 1.20 (1H, d, J = 10.3 Hz, pinanyl

Hendo), 1.24 (3H, s, pinanyl CH3), 1.34 (3H, s, pinanyl CH3), 1.39 ( 3H, t, J = 7.1 Hz,

OCH2CH3), 1.47-2.5 (5H, m, pinanyl protons), 4.00 (2H, s, CH2-CO), 4.12 (1H, d, J

= 2.0 Hz, CH-B), 4.21 (1H, dd, J = 8.6, 2.2 Hz, pinanyl CHOB), 4.37 (2H, q, J =

7.1, OCH2CH3), 6.68 (1H, br, NH), 6.98-7.05 (2H, m, thienyl H3 and H5), 7.23-7.40

(2H, m, phenyl H5 and H4 + thienyl H4), 7.80-7.9 (2H, m, phenyl H2 and H6).

13 C NMR (CDCl3): δ 14.3, 24.1, 26.4, 27.1, 28.7, 34.2, 36.0, 38.1, 39.8, 46.6 (br,

C-B), 52.0, 60.8, 77.3, 84.9, 126.3, 127.0, 127.5, 127.7, 128.2, 128.4, 130.6,

130.7, 133.6, 140.7, 166.6, 174.3.

EIMS: m/z 481 (62%, M+), 436 (5%), 329 (14%), 302 (5%), 285 (8%),245 (33%),

222 (9%), 178 (17%), 163 (11%), 135 (23%), 124 (33%), 107 (12.5%), 97 (100%),

84 (57%), 70 (15%), 45 (18%).

Analytically calculated for C26H32BNO5S: C, 64,87; H, 6,70; N, 2,91. Found: C,

64,77; H, 6,81; N, 2,83.

(R)-(3-Carboxyphenyl)-(2-thiophen-2-yl-acetylamino)-methaneboronic acid (–)-2.-

Compound (+)-5 (270 mg, 0.56 mM) was treated with degassed HCl 3 M (13 ml) at

100 °C for 1 hr under argon. The reaction mixture was extracted with diethyl ether

(2 × 10 ml) and the aqueous phase concentrated under reduced pressure to afford a solid residue which was crystallized from a few methanol/diethyl ether affording the desired inhibitor (–)-2 as a whitish crystalline powder. Mp 245-247 °C dec, [α]D:

-78.0 (c 0.4, CH3OH).

IR: ν 3398, 1704, 1606.

1 H NMR (CD3OD): δ 3.87 (1H, br s, CH-B), 4.18 (2H, s, CH2-CO), 7.00-7.05 (1H,

m, thienyl H4), 7.13-7.14 (1H, m, thienyl H3), 7.30-7.41 (3H, m, phenyl H5 and H6 +

thienyl H5), 7.80 (2H, m, phenyl H2 and H4).

13 C NMR (CD3OD): δ 31.0, 53.0 (br, C-B), 126.2, 127.0, 127.2, 127.4, 128.2,

128.4, 130.6, 130.8, 133.5, 141.5, 168.9, 178.5.

Elemental analysis and EIMS of the free boronic acid were not obtainable (Hall,

2005).

Structure Analysis – Protein Data Bank3 coordinates of SHV-1 and TEM-1 were

examined using ViewerLite® (Accelrys®, San Diego, CA). Research Collaboratory

for Structural Bioinformatics Protein Databank = PDB # 1SHV (SHV-1) (Kuzin et

al., 1999) and Research Collaboratory for Structural Bioinformatics Protein

Databank = PDB # 1BTL (TEM-1) (Jelsch et al., 1993) were analyzed.

Molecular Representation: ß-lactam backbones were drawn with the Sketcher module in Insight II Version 2005 Molecular Modeling System (Accelrys®)(SGI,

Octane Workstation IRIX 6.5.28). Molecules were built into 3D structures with

Converter, energy minimized, and overlayed on their β-lactam rings. ß-lactams were positioned into the active site of SHV-1 (PDB # 1SHV, see above), with only

Arg244, Thr235, and Ser70 visualized.

RESULTS

Mutagenesis and Immunoblotting- Sixteen of 19 amino acid substitutions were

obtained in the initial sequencing screen. blaSHV-1(Arg244Lys), blaSHV-1(Arg244Met), and blaSHV-1(Arg244Phe) were constructed by site-directed mutagenesis. Full length expression of all variants was confirmed by immunoblotting (data not shown).

Antibiotic Susceptibility- In order to test the effects of the single amino acid substitutions at Arg244 on in vivo β-lactam susceptibility, MICs against all 19 variants were determined for ampicillin, ampicillin/clavulanate, piperacillin, and cephalothin in E.coli DH10B cells (Table 2-1, page 70). Twelve amino acid substitutions in SHV-1 increased the MIC values for the inhibitor combination ampicillin/clavulanate by at least two dilutions (from 50/2 to > 50/8 μg/ml). MIC values for the penicillins (ampicillin and piperacillin), and the cephalosporin, cephalothin, were universally decreased.

Kinetic Behavior of Inhibitor Resistant β-lactamases With Clavulanate -

Susceptibility data served as a guide for selection of β-lactamases for further

kinetic analysis. Four clavulanate resistant enzymes were chosen: SHV

Arg244Ser, -Gln, -Leu and -Glu.

A common feature of the resistant enzymes is reduced affinity for

clavulanate, with Ki values increasing 60-1000 fold (Table 2-2, page 71).

Unexpectedly, the clavulanate resistant enzymes exhibited increased kinact values.

This is in sharp contrast to the TEM enzymes, Arg244Gln and -Thr, which 59

exhibited a 100-fold decrease in kinact (Delaire et al., 1992). At concentrations equal to the Ki for clavulanate, all SHV Arg244 variants are rapidly inactivated (Figure 2-

4, page 76). This is again much different than the analysis of Arg244Ser, -Gln, and

-Thr TEM variants at Arg244, which never achieve complete inactivation (Delaire et al., 1992; Imtiaz et al., 1993).

Despite the increased rates of inactivation, the markedly decreased affinity of these mutant enzymes for clavulanate reduces the inactivation efficiency

(kinact/Ki) by 25-500-fold (Table 2-2, page 71). Thus, decreased affinity is the primary cause for resistance to inactivation by clavulanic acid.

In 24 hour inactivation experiments, SHV-1, Arg244Ser and -Glu enzymes required approximately the same amount of clavulanate to achieve a 90% reduction in enzyme activity (kcat/kinact, or partition ratio) (Table 2-2, page 71).

Alternatively, the Arg244Gln and -Leu enzymes exhibit a 3-9 fold increase in kcat/kinact, respectively. The derived kcat values are increased for all mutants tested, most notably the Arg244Gln and -Leu mutants, and this enhanced ability to catalyze the turnover of inhibitor likely contributes to resistance.

Determining the Nature of the Intermediates: ESI Mass Spectrometry with SHV-1 and Arg244Ser- Clavulanate undergoes a multi-step reaction pathway (Sulton et al., 2005). Candidate intermediates were observed previously for SHV-1 and

Ser130Gly, also an inhibitor resistant variant (Figure 2-5, pages 77-78) (Sulton et al., 2005). Intact mass spectrometry was performed on SHV-1 and Arg244Ser to observe the covalent intermediates in the inactivation pathway. The Arg244Ser 60

variant was chosen because it is the most common mutant found in TEM at

position 244.

As seen in Figure 2-5 (pages 77-78), when SHV-1 and Arg244Ser are

incubated with clavulanate, nearly identical covalent intermediates are formed.

This includes, within experimental error, the Δ +52 adduct that is postulated to

represent the terminally inactivated, cross-linked enzyme species, and the Δ +70,

Δ +88, and Δ +155 Da adducts previously observed in TEM-1 and SHV-1 (Brown et al., 1996; Sulton et al., 2005). The only differences between the two spectra include a Δ +175 adduct seen in only SHV-1, and a Δ +194 peak visualized in

Arg244Ser only. The Δ +175 peak is very minor and may be hidden within the shoulder of the Δ +157 adduct of Arg244Ser. Alternatively, the Δ +194 peak in

Arg244Ser could be a protein modification (oxidation) which shifted the Δ +175 adduct by +19.

Kinetic Behavior of Inhibitor Resistant Enzymes With β-lactam Substrates- Kinetic parameters for SHV-1 and the inhibitor resistant mutants Arg244Ser, -Gln, -Glu, and -Leu for ampicillin and nitrocefin are reported in Table 2-2 (page 71). Affinity is reduced for all variants, with Arg244Ser and -Gln demonstrating the lowest Km values. Interestingly, kcat is universally reduced; the enzymes with the highest

kcat/kinact values for clavulanate (Arg244Gln and -Leu) display the highest kcat values for substrates among the Arg244 variants. Catalytic efficiency (kcat/Km) of all

variants was greatly reduced for ampicillin (14Æ700 fold) and nitrocefin (5Æ175

fold). 61

In order to assess the importance of the position of the C3-4 carboxylate in

substrate affinity, we next tested piperacillin, cephalothin, and meropenem (Figure

2-1, page 73). Because accurate hydrolysis of these substrates was difficult to

measure for the Arg244 variants, affinity was assessed by competition reaction

with the indicator substrate nitrocefin (Table 2-3, page 72). Again, all substitutions

at position 244 resulted in enzymes with reduced affinities for these substrates

(2→180 fold increase in Kd for piperacillin, 20→750 fold for cephalothin, and a

striking 400→ >2700 fold for meropenem). As seen above, Arg244Ser and -Gln

retain the highest affinities.

Probing the Active Site: Cephalothin Boronic Acid Transition State Analogs-

Boronic acid analogs have been developed in recent years, both as high affinity

inhibitors of β-lactamases, and as probes to study reaction mechanism. They also

serve to explore determinants of binding specificity (Chen et al., 2006; Chen et al.,

2005; Morandi et al., 2003; Powers et al., 2001; Wang et al., 2003). A majority of

these compounds have been designed as achiral molecules containing the R1

side chains of penicillins and cephalosporins. Recently, chiral compounds have

been developed to take advantage of affinity gains of the C4 carboxylate (Figure 2-

1, page 73).

To investigate the importance of Arg244 in SHV for coordination of the β- lactam carboxylate, we synthesized an achiral (compound 1) and chiral

(compound 2) boronic acid derivative of cephalothin (Figure 2-1, page 73). In particular, we reasoned that the molecular architecture of 2 would closely resemble 62

the interactions displayed by the antibiotic cephalothin (the carboxylic moiety at C4 included) with the β-lactamase.

As expected, compound 1 inhibits SHV-1 and Arg244Ser, -Gln, and -Leu with similar affinities. (Table 2-3, page 72). In contrast, the Arg244Glu enzyme demonstrates an approximately 4-fold reduction in affinity compared to wild type.

Since this analog should not interact directly with residue 244, we submit that this difference suggests a rearrangement in the tertiary structure of this variant.

Testing the chiral boronic acid transition state inhibitor, compound 2, there was a 7-fold enhanced affinity for SHV-1 compared to compound 1. Arg244Ser and -Gln both showed only modest improvements in Ki of binding to the chiral

compound (< 2 –fold), and 244Leu and -Glu have reduced affinity for 2, indicating

an unfavorable interaction of the C4 carboxylate.

As has been described for other β-lactamase families (Chen et al., 2005;

Wang et al., 2003), we observed time dependent inhibition of compound 2, which

is in contrast to the fast on/fast off time independent inhibition of other (including 1)

boronic acid compounds. To compare our analysis to others, we chose a 5 minute

pre-incubation for our studies (Chen et al., 2005; Morandi et al., 2003; Wang et al.,

2003). To ensure the suitability of the 5 min pre-incubation for reaching

equilibrium, we used the Ki concentrations measured at 5 minutes to further study the time-course of inhibition from 0-3600 sec (60 min) (Figure 2-6, page 79).

Interestingly, while all the mutant enzymes did reach steady state by 5 min, SHV-1 was increasingly inhibited with time, resulting in near complete inhibition by the

3600 sec time point. Therefore, the true steady state Ki value for SHV-1 with the chiral inhibitor is likely much lower than the 5 minute measurement of 3.8 µM.

DISCUSSION

We show that many substitutions at Ambler position 244 in SHV produce the inhibitor resistant phenotype. However, kinetic analysis of selected variants suggests that the demonstration of resistance to clavulanate in SHV is unique. The principal characteristic of resistance in SHV variants at Arg244 is a reduction in affinity for the inhibitor. This property is shared by both TEM and SHV. However, unlike TEM, SHV mutants at position 244 do not demonstrate a reduction in kinact.

These inhibitor resistant β-lactamases undergo rapid inactivation with clavulanate provided adequate inhibitor concentrations are achieved. Interestingly, both the inhibitor susceptible and inhibitor resistant SHV enzymes follow the same reaction pathway; the products of inactivation of SHV-1 and Arg244Ser by clavulanate, as determined by mass spectrometry, are identical.

What accounts for this key difference in behavior between the TEM and

SHV β-lactamases? We answer this question by examining the atomic structures of TEM-1, SHV-1 and the SHV Ser130Gly variant. Examining a representation created from the PDB coordinates available for the TEM-1 apo-enzyme, a water molecule is clearly positioned between Arg244 and the backbone carbonyl of

Val216 (Figure 2-7, page 80; Figure 2-2, page 74) (Wang et al., 2003). One explanation for the catalytic deficit in TEM mutants at position 244 is displacement of that water molecule which is essential for secondary clavulanate ring opening in the inactivation pathway. Comparing the apo-enzyme crystal structures of TEM-1

(Jelsch et al., 1993) and SHV-1 (Kuzin et al., 1999) , the distances between the 65

closest guanidinium nitrogen of Arg244 and the backbone carbonyl oxygen Val216

are significantly different (Figure 2-7, page 80). For TEM, this distance is 5.19 Å,

while in SHV the distance is 7.85 Å, a large distance for coordination of a water

molecule. Therefore it is likely that in SHV, protonation of the inhibitor is achieved

by either a coordinated water molecule elsewhere in the active site, or from bulk

water in the medium. Thus, substitutions at position 244 in SHV affect affinity, but

do not retard inhibitor turnover. This explanation is reminiscent of the observations

made in the determination of the Ser130Gly apo-enzyme structure. In Ser130Gly

the catalytic water molecule is only evident as the inhibitor is bound (Sun et al.,

2004). Our data support the general hypothesis that the active site of inhibitor

resistant SHV enzymes is rehydrated. This raises the possibility that despite

different mutations, inhibitor resistant SHV enzymes follow a common pathway to

inactivation.

Studying the dissociation constants (Kds)of penicillins, cephalosporins, and

carbapenems, provided us deeper insight into the reliance of Arg244 binding to the different spatial positions of the C3-4 carboxylate and its role in catalysis. (Figure 2-

8, page 81).

Firstly, affinity for the penicillins is the least affected by substitutions at

Arg244. The carboxylate in this case is hanging from the sp3 hybridized and S

configured C3 carbon. In our model, it is likely that Arg244 contributes one or two

weak hydrogen bonds to the carboxylate in the Michaelis complex, with additional

hydrogen bonding interactions coming from other residues in the binding pocket

(Figure 2-8B, page 81). For this reason, most inhibitor resistant variants at Arg244 66

in SHV retain weak penicillinase activity, and their emergence in the context of current β-lactam/β-lactamase inhibitor combinations, which utilize penicillins, is not excluded.

Secondly, the reductions in affinity are at the very least 20-fold (SHV

Arg244Gln) for cephalothin. Different from penicillins, the cephalosporin

2 carboxylate is directly linked to the C4 sp hybridized carbon of the six-membered ring, and therefore co-planar with C3-C4, with reduced rotational freedom due to conjugation. This likely brings the carboxylate in closer approximation to Arg244

(Figure 2-8C, page 81). The dramatic loss of cephalosporinase activity among

Arg244 variants shows that clavulanate resistant variants that arise in the clinic would be susceptible to treatment with cephalosporins. That being said, excessive use of cephalosporins in the clinical setting may mask the emergence of inhibitor resistant SHV enzymes.

Most revealing is our data for meropenem affinity. Like clavulanate, carbapenems act as very effective covalent inhibitors of class A β-lactamases4.

Interestingly, replacement of Arg244 results in total loss of affinity for meropenem.

All of the mutants bound meropenem with mM affinity, and two of them (244Glu and -Leu) had Kd values greater than 100 mM. This indicates that the carboxylate

2 linked to the C3 sp hybridized carbon of the five-membered ring makes several crucial hydrogen bonding interactions with the guanidinium group of Arg244

(Figure 2-8D, page 81). Caution should be taken in the development of future inhibitors with planar C3 carboxylates, as substitutions at residue 244 could seriously compromise their activity. 67

Lastly, we studied two boronic acid transition state analogs to further probe

the contributions of Arg244 variants to β-lactam carboxylate affinity. The achiral cephalothin boronic acid compound 1 contains just the R1 side chain of

cephalothin. In contrast, the chiral cephalothin analog 2 more closely imitates the

interactions of the natural substrate by the addition of a C4 carboxylate on a phenyl

ring (Figure 2-1, page 73). Comparing affinity of both compounds allowed us to

determine the extent to which each substitution of the β-lactamase affects binding

to the carboxylate independently.

Against SHV-1, compound 2 demonstrates significantly greater affinity for the

active site (nearly a log fold) compared to compound 1. We observed that the

Arg244Ser substitution at 244 has a modest, 40% increase in binding to 2 vs. 1

(24 µM vs. 34 µM). This supports our claim that there is another residue in the binding pocket that contributes to binding of the carboxylate, possibly Thr235

(Imtiaz et al., 1993). Alternatively, hydrophobic interactions with the phenyl ring of compound 2 could stabilize the inhibitor.

Compared to Arg244Ser, a slightly higher affinity is seen with Arg244Gln

(60% increase) for compound 2, which may imply that the Gln residue itself is able to weakly interact with the inhibitor carboxylate. On the other hand, increased Ki values for 2 with Arg244Leu and -Glu suggest unfavorable interactions with the inhibitor carboxylate, and explain the drastic affinity reductions for β-lactams.

The differences in the time dependant inactivation of compound 2 between the wildtype and mutant enzymes also provided a window into the importance of

Arg244 for turnover of substrate. The chiral compound 2 has been shown in crystal structures with TEM, CTX-M, and AmpC β-lactamases to assume the deacylation transition state (Chen et al., 2006; Chen et al., 2005; Wang et al., 2003). In TEM, although no structural rearrangements were seen upon binding to the inhibitor, there was movement of catalytic water molecules and a rearrangement of the inhibitor (Wang et al., 2003). If this reorganization is responsible for the time dependant binding of compound 2 to SHV-1, the reduction in time-dependence among the Arg244 mutants may indicate a novel substrate deacylation role for

Arg244 (Figure 2-6, page 79). Future studies to determine the microscopic rate constants k2 and k3 of SHV variants at Arg244 using stopped-flow kinetics will

provide evidence for this role. Co-crystallization of SHV-1 and Arg244 variants with

the cephalothin transition state analogs are in progress.

In conclusion, we present evidence that TEM and SHV are different in their intrinsic binding and turnover of inhibitors (Knox, 1995). Despite similarities in sequence and phenotype between TEM and SHV, the kinetic differences revealed herein suggest that these two class A β-lactamases follow unique pathways in response to antibiotic pressure (an argument that would have significant implications for the comparative study of penicillin inactivating enzymes among prokaryotes). Awareness of the subtle yet mechanistically important differences in

inactivation chemistry among class A β-lactamases could prove crucial in the

future development of β-lactamase inhibitors. 69

Table 2-1: MICs (μg/ml) of E. coli DH10B containing SHV-1 and all 19 variants at Arg244 244 Clone Ampicillin Ampicillin/Clavulanate Piperacillin Cephalothin Arg (WT) >16384 50/2 2048 64

Ala 256 50/4 64 4

Asn 256 50/4 64 4

Asp 2 50/.06 2 4

*Cys 128 50/2 64 4

Gln 1024 50/16 256 4

Glu 256 50/8 32 4

*Gly 256 50/4 64 4

*His 1024 50/8 128 4

Ile 256 50/8 32 4

*Leu 128 50/16 32 4

Lys 2048 50/8 512 8

Met 512 50/16 64 4

Phe 256 50/8-16 32 4

Pro 1 50/.06 2 4

*Ser 1024 50/8 128 4

*Thr 512 50/8 64 4

Trp 8 50/.06 4 4

Tyr 128 50/16 16 4

Val 128 50/8-16 16 4

*Substitutions found clinically in TEM Bold: Variants chosen for kinetic analysis

Table 2-2: Kinetic properties of SHV-1 and variants at Arg244 for clavulanate, ampicillin, and nitrocefin

SHV-1 Arg244Ser Arg244Gln Arg244Leu Arg244Glu

Clavulanate

Ki (μM) 1 ± .04 63 ± 3 85 ± 7 360 ± 20 1000 ± 150

-1 kinact (s ) 0.04 ± .002 0.09 ± .005 0.09 ± .007 0.11 ± .01 0.08 ± .008

kinact / Ki 0.0014 ± 0.0011 ± 0.00031 ± 0.00008 ± (μM-1s-1) 0.04 ± .003 .0001 .0001 .00003 .00001

kcat/kinact 60 ± 10 50 ± 10 180 ± 10 550 ± 30 55 ± 10

-1 kcat (s ) 2.4 ± 0.4 4.5 ± 0.9 19 ± 2 61 ± 6 4.4 ± 0.9

Ampicillin

Km (μM) 165 ± 10 255 ± 45 1190 ± 120 2240 ± 590 3500 ± 1800

-1 kcat (s ) 3700 ± 400 390 ± 50 1400 ± 160 800 ± 170 115 ± 50

kcat / Km (μM-1s-1) 22 ± 3 1.5 ± .3 1.2 ± 0.2 0.4 ± .1 0.03 ± 0.02

Nitrocefin

Km (μM) 21 ± 3 55 ± 4 110 ± 9 590 ± 30 250 ± 10

-1 kcat (s ) 290 ± 32 107 ± 11 270 ± 30 230 ± 20 20 ± 2

kcat / Km (μM-1s-1) 14 ± 2.5 2 ± 0.25 2.4 ± 0.3 0.4 ± .04 .08 ± .009

Table 2-3: Dissociation Constants (Kd, or Ki* μM) of β-lactam Substrates and Boronic Acid Transition State Inhibitors for SHV-1 and Arg244 Variants

SHV-1 Arg244Ser Arg244Gln Arg244Leu Arg244Glu

12000 ± Piperacillin 66 ± 4 134 ± 5 860 ± 50 3700 ± 400 400

56000 ± Cephalothin 74 ± 4 2000 ± 130 1500 ± 30 22000 ± 700 5000

15200 ± 17300 ± 102000 ± Meropenem 37 ± 2.5 >100000 700 700 7000

Compound 1* 30 ± 1 34 ± 2 30 ± 2 43 ± 3 116 ± 10

Compound 2* 3.9 ± 0.3 24.3 ± 0.4 18.5 ± 0.7 71 ± 7 156 ± 15

Figure 2-1 Chemical structures of compounds tested in this study. The structures of clavulanic acid and nitrocefin are labeled with the accepted ring numbering system.

Figure 2-2 The proposed contribution of Arg244 to clavulanic acid inactivation of

TEM-1. The guanidinium group of Arg244 is essential for (A) hydrogen bonding

interactions with the C3 carboxylate of the inhibitor in the active site, and (B)

coordination of a proton donating water molecule essential for the saturation of the double bond of the C2 constituent.

Figure 2-3 Synthesis of the chiral cephalothin boronic acid transition state inhibitor

(A) (+)-Pinanediol, THF, rt; (B) (dichloromethyl)lithium, THF, –100 °C → 0 °C; (C) lithium bis(trimethylsilyl)amide, THF, –80 °C → rt; (D) 2-thiopheneacetylchloride, 2- thiopheneacetic acid, THF –80 → rt; (E) aqueous HCl 3N, 1h, 100 °C.

Figure 2-4 Timed inactivation of SHV-1 and Arg244Ser, -Gln, -Leu, and –Glu reveal that Arg244 variants are inactivated more rapidly by clavulanate than SHV-

1. Enzymes were incubated with Ki concentrations of clavulanate and initial

velocities of nitrocefin hydrolysis measured at timepoints 0-600 seconds.

Figure 2-5 (A) Deconvoluted mass spectra of SHV-1 and Arg244Ser before and after 15-minute incubation with clavulanate show similar intermediates of inactivation. Spectra were obtained on a Q-STAR XL quadrupole-time-of-flight mass spectrometer equipped with a nanospray source. Eight distinct mass shifts were identified with both enzymes. (B) Proposed intermediates in the clavulanate

inactivation pathway. The 198 Da and 173 Da adducts are represented by more

than one candidate structure (blue). The terminally inactivated 52 Da crosslinked

species is highlighted in red

Figure 2-6 The time-dependent inhibition of the chiral cephalothin boronic acid

transition state inhibitor (compound 2) is more significant with SHV-1 than with

SHV Arg244 Ser, -Gln, -Leu, and -Glu. Enzymes were incubated with inhibitor at

concentrations equal to the Ki and initial velocities of nitrocefin hydrolysis assessed at 0-3600s. Initial Ki determination was performed after 5 minute incubation.

Figure 2-7 Arg244 in relation to Val216 in the SHV-1 (1SHV) and TEM-1 (1BTL) in apo-enzyme crystal structure representations. The distance between the backbone carbonyl oxygen of Val216 and the nearest guanidinium nitrogen in SHV is 7.85 Å.

In contrast, the distance in TEM-1 is 5.19 Å. Shown in TEM-1 is the bridging water molecule which makes hydrogen bonds with Val216 (2.87 Å) and Arg244 (3.08 Å).

The active site Ser70 is shown for reference.

Figure 2-8 Suggested role of Arg 244 in stabilizing the β-lactam carboxylate of penicillins, cephalosporins, and carbapenems. (A) The ring structures of ampicillin

(purple), cephalothin (yellow), and meropenem (pink) were drawn using sketcher

(Accelrys®) and energy minimized. The three structures were overlayed on the β- lactam ring and the proposed orientation in the active site is shown. (B) The penicillin (C) cephalosporin, and (D) carbapenem ring structures in relation to

Arg244 and Thr235 (catalytic Ser70 shown for reference). The hypothesized proposed bonding interactions are drawn in white.

CHAPTER 3

SHV Arg244- Sulbactam Susceptibility

Submitted as

Overcoming Resistance to β-Lactamase Inhibitors: Comparing Sulbactam to Novel

Inhibitors against Clavulanate-Resistant SHV Enzymes with Substitutions at

Ambler Position 244

Jodi M. Thomson, Anne M. Distler, and Robert A. Bonomo

Biochemistry

Submitted April, 2007

INTRODUCTION

β-Lactam antibiotics (e.g. penicillins, cephalosporins, and carbapenems)

are bactericidal agents that target penicillin binding proteins (PBPs), the essential

enzymes in cell wall biogenesis. Their use in the clinic is constantly threatened by

the ever-expanding numbers of β-lactamase enzymes (EC 3.5.2.6). Divided into 4

classes (A-D) based on amino acid homology, β-lactamases demonstrate a

diverse substrate specificity (Jacoby, 2006). The most commonly encountered β- lactamases in E. coli and K. pneumoniae (TEM and SHV, respectively) are class A enzymes (Bradford, 2001; Bush, 2001). A series of single amino acid substitutions in the TEM and SHV β-lactamases are responsible for the current shortage of

penicillins and cephalosporins active against these common enteric bacilli

(Bradford, 2001; Bush and Mobashery, 1998; Jacoby, 2006). 82

The use of β-lactamase inhibitors preserves the clinical utility of penicillins

by inactivating β-lactamases and allowing the β-lactam to reach the PBPs. Three

β-lactamase inhibitors are commercially available: clavulanate, tazobactam, and sulbactam (Figure 3-1, page 102). The β-lactam/β-lactamase formulations amoxicillin/clavulanate, ticarcillin/clavulanate, ampicillin/sulbactam, cefoperazone/sulbactam, and piperacillin/tazobactam are widely used to treat complicated infections due to β-lactam resistant organisms. Unfortunately, resistance to β-lactam/β-lactamase inhibitor combinations is a growing problem among Gram-negative pathogens harboring class A β-lactamases. Several inhibitor resistant TEM (IRT) variants have arisen with substitutions at R244

(R244S, -L, -C, -T, -H, and -G) (Jacoby, 2006). In both the TEM and SHV family of

β-lactamases, inhibitor resistant variants at M69 and S130 are also described

(Dubois et al., 2004; Helfand et al., 2003a; Prinarakis et al., 1997). Detailed kinetic and structural studies have been performed that explain the resistance observed to clavulanate among IRTs (Knox, 1995; Thomas et al., 2005; Wang et al., 2003;

Wang et al., 2002; Zafaralla et al., 1992) and inhibitor resistant SHV β-lactamases

(Pagan-Rodriguez et al., 2004; Sulton et al., 2005; Sun et al., 2004; Thomson et al., 2006).

Anticipating that inhibitor resistance phenotypes in SHV β-lactamases will arise with similar alterations as IRTs, we previously studied (see Chapter 2) the effects of substitutions at R244 in SHV β-lactamases and showed that 15 variants at R244 had increased resistance to clavulanate (Thomson et al., 2006). We also observed that the kinetic basis of the clavulanate resistant phenotype was 83

decreased affinity of the enzymes for clavulanate (60-1000 times increases in KI), presumably through loss of bonding interactions to the C3 carboxylate.

Interestingly, we also detected an increase in rates of inactivation (kinact), and, in

some cases, increases in the partition ratio (kcat/kinact) among clavulanate resistant

mutants.

In order to gain a comprehensive insight into the basis of resistance to β- lactamase inhibitors among class A enzymes, we studied the kinetic interactions of a commercially available sulfone inhibitor, sulbactam (Figure 3-1, page 102) against wildtype (WT) and three variants of SHV-1 at the 244 position (R244S,

R244Q, R244L). In addition, our analysis of the interactions of meropenem with

SHV-1 and R244 variants highlighted the importance of the R244 interaction with the C3 carboxylate, and compelled us to examine the reaction between penem

type β-lactamase inactivators (penems 1 and 2) and these clavulanate resistant

SHV type β-lactamases (Thomson et al., 2006). Penems 1 and 2 were chosen

2 because they possess a sp hybridized carboxylate at the C3 position, a bicyclic R1 side chain, and differ from clavulanate, sulbactam and tazobactam in their mechanism of inactivation of class A β-lactamases (Nukaga et al., 2003a;

Venkatesan et al., 2006). Our studies, summarized herein, reveal that sulbactam is a poor inhibitor of WT enzyme because of an extremely high partition ratio.

Substitutions at R244 reverse this phenotype, restoring some susceptibility, but not to clinically effective levels. Fortunately, β-lactam/β-lactamase inhibitor susceptibility can be restored to a clinically effective level even against inhibitor resistant SHV β-lactamases by increasing affinity via modification of the R1 side 84

chain and novel reaction chemistry.

EXPERIMENTAL PROCEDURES

Plasmids and Mutagenesis - blaSHV-1, subcloned into phagemid vector pBC SK(-)

(Stratagene, La Jolla, CA) and maintained in ElectroMAX DH10BTM cells

(Invitrogen, Carlsbad, CA) (Rice et al., 2000) served as the template for all mutagenesis at Ambler position R244 described previously (JBC).

Antibiotic Susceptibility - Minimal Inhibitory Concentrations (MICs) were determined by the agar dilution method, using a Steer’s Replicator that delivers

104 CFU per 10 µL spot. Sulbactam (Pfizer, La Jolla, CA) and penem 1 and 2

(Wyeth Pharmaceuticals, Madison, NJ) susceptibilities were determined by

increasing the concentration of inhibitor in the presence of 50 μg/ml ampicillin.

Values reported are the most frequent number observed (mode) in at least three

separate experiments.

Protein Expression and Purification - E. coli DH10B cells containing the blaSHV-1 and blaSHV R244S, -Q, and -L genes in pBC SK(-) were grown overnight in SOB medium, harvested by centrifugation at 4 °C and frozen. β-lactamase was liberated using stringent periplasmic fractionation with 40 µg/ml lysozyme and 1 mM EDTA pH

7.8. Preparative isoelectric focusing was performed with the lysate in Sephadex

granulated gel using ampholines in the pH range of 3.5-10 (Amersham

Biosciences) as previously described (Lin et al., 1998). The protein was eluted and 86

dialyzed with 20 mM diethanolamine buffer, pH 8.5. Protein concentration was

assessed by Bio-Rad Protein Assay (Hercules, CA) with bovine serum albumin standards. Purity >95% was determined using 10% SDS-PAGE.

Kinetics - Kinetic constants of the β-lactamases chosen for study (SHV-1,

R244S, -Q, and -L) were measured by continuous assays at room temperature,

25 °C, using an AgilentTM 8453 diode array spectrophotometer (Agilent, Palo

Alto, CA). The kinetic parameters were obtained with non-linear least squares fit

of the data using Origin 7.5®. Detailed descriptions of the assays performed and data analysis were previously published (Thomson et al., 2006).

Mass Spectrometry - For intact protein mass spectrometry, 40 μM of SHV-1 or

R244S were incubated for 15 min with and without the addition of 40 mM sodium sulbactam, or 20 mM penem 1 or 2. Each sulbactam reaction was terminated by

the addition of 1:10 volume 1% trifluoroacetic acid, and immediately desalted and

concentrated using a C4 ZipTip® (Millipore, Bedford, MA) according to the

manufacturer’s protocol. Reactions with penems were not terminated with the

addition of acid, as this precipitated the samples. These were desalted “as is”

using a C4 ZipTip®. Samples were then placed on ice and analyzed within 10

minutes.

Spectra of the intact SHV-1 and SHV R244S proteins were generated on an

Applied Biosystems (Framingham, MA) Q-STAR XL quadrupole-time-of-flight

mass spectrometer equipped with a nanospray source. Experiments were 87

performed by diluting the protein sample with acetonitrile / 1% formic acid to a

concentration of 10 μM. This protein solution was then infused at a rate of 0.5

μL/min and data were collected for 2 minutes. Spectra were deconvoluted using

the Applied Biosystems (Framingham, MA) Analyst program.

Structure Analysis – Protein Data Bank coordinates of SHV were examined using

ViewerLite® (Accelrys®, San Diego, CA). The following Research Collaboratory

for Structural Bioinformatics Protein Databank codes were analyzed: 2A3U (SHV

E166A with Sulbactam) (Padayatti et al., 2005) and 2A49 (SHV E166A with clavulanate) (Padayatti et al., 2005).

RESULTS

Antibiotic Susceptibility - Listed in Table 3-1 (page 99) are the results of our susceptibility testing. E. coli DH10B cells expressing blaSHV-1 from pBC SK (-) exhibit robust levels of resistance to ampicillin (> 16,000 µg/ml). High level resistance to ampicillin/sulbactam was also demonstrated (ampicillin 50 µg/ml / sulbactam 256 µg/ml). In contrast, eighteen of nineteen SHV-1 variants at position

244 expressed in this uniform genetic background were more susceptible to ampicillin/sulbactam compared to SHV-1 (Table 3-1, page 99). Only E. coli

DH10B with R244K β-lactamase (MIC = ampicillin 50 µg/ml / sulbactam 128 µg/ml) maintained resistance comparable to WT. Generally speaking, the reduction in resistance to the ampicillin/sulbactam combination was 2-4 fold. The most sulbactam susceptible mutants demonstrated a sulbactam MIC of 16 µg/mL when combined with ampicillin 50 µg/ml. This pattern of increased susceptibility to a β- lactamase inhibitor is in sharp contrast to what was observed with ampicillin/clavulanate (Thomson et al., 2006).

Compared to clavulanate and sulbactam, penem 1 is an extremely potent inhibitor of SHV-1. Penem 1 at 0.25 µg/ml restored susceptibility to ampicillin in E. coli DH10B cells expressing SHV-1 (Table 3-1, page 99). Although all variants at position 244 tested more resistant to penem 1 than WT expressed in E. coli (the pattern seen for clavulanate), the level of resistance remained much lower than ampicillin/sulbactam (MICs ranged from 0.06 to 2 μg/ml). The greatest degree of

resistance was seen with testing the R244H and R244S variants (ampicillin 50

µg/ml/penem 1, 2 µg/ml). Penem 2 combined with ampicillin was also more

effective than sulbactam with ampicillin but was less potent than penem 1 against

the R244 mutants. Fifteen of twenty variants at R244 had elevated MIC values ≥

ampicillin 50 µg/ml / penem 2 8 µg/ml. This is compared to WT of ampicillin 50

µg/ml / penem 2 2 µg/ml (Table 3-1, page 99).

Kinetic Behavior of Clavulanate Resistant β-Lactamases with Sulbactam and

Penems 1 and 2 –We chose R244S, -Q and –L for further kinetic characterization

because of their clinical importance (244S and –L) and high ampicillin/clavulanate

MICs (244Q) (Belaaouaj et al., 1994; Miro et al., 2002). In light of our

susceptibility testing, finding an elevated sulbactam KI against these three variants

was unanticipated (Table 3-2, page 100). In competition reactions with nitrocefin, we observed that the dissociation constants for the pre-acylaction complexes were increased 28-140 fold, despite the MIC result that all variants tested more susceptible to sulbactam.

Substitutions at position 244 also drastically altered the enzyme’s ability to hydrolyze sulbactam. SHV-1 exhibited a very high partition ratio (kcat/kinact) for this sulfone; in 24 hours one enzyme molecule is able to effectively hydrolyze 13,000 molecules of sulbactam. In comparison, all three R244 clavulanic acid resistant β- lactamases exhibited partition ratios of ≤ 500, the lowest being the R244S enzyme at kcat/kinact= 100. After accounting for increased kinact values among the mutants,

the kcat values for sulbactam were decreased 10-55 times. This catalytic deficiency

is likely the cause for the susceptibility increase.

The KIs of penems 1 and 2 were in the nM range for WT SHV-1 β-

lactamase (14 ± 2 and 48 ± 4 nM, respectively) (Table 3-3, page 101). Parallel to

our observations with sulbactam, we again observed a loss of affinity among the mutants at position R244. There was an 80-280 (14 nM Æ 3.9 µM) and 90-260 (48 nM Æ 12.5 µM) times increase in KI for penems 1 and 2, respectively. The most

striking feature of the penem inhibitors, however, is an extremely low turnover

when tested against the selected enzymes. Both penems 1 and 2 exhibited

partition ratios (kcat/kinact) of 2 for SHV-1 and 1 for R244S, -Q, and -L (Table 3-3,

-1 page 101). The resulting kcat values (< 0.4 s ) are lower than is seen for

clavulanate or sulbactam.

Mass Spectrometry of SHV-1 and R244S with Sulbactam and the Penem

Inhibitors- To discern the nature of the intermediates in the reaction pathway of

SHV-1 and R244S with sulbactam and the penem inhibitors, we performed

electrospray ionization mass spectrometry (ESI-MS). For sulbactam, a 1000:1

molar ratio of inhibitor, I, to enzyme, E, was prepared and the reaction proceeded

for 15 minutes before terminating with trifluoroacetic acid. In the case of SHV-1 we

observed a predominance of unmodified β-lactamase at the predicted molecular

weight of 28,872 ± 3Da (Figure 3-2, page 103). Minor peaks include adducts of Δ

+71 ± 3, Δ +90 ± 3, Δ +115 ± 3, Δ +142 ± 3, Δ +228 ± 3, most of which have been

previously documented (Helfand et al., 2003a; Helfand et al., 2003b; Sulton et al.,

2005).

With equivalent amounts of sulbactam (I/E= 1000), a markedly different result is seen studying the inactivation of R244S. The predominant intermediates are the Δ +70 ± 3 and Δ 88 ± 3 Da adducts, with minor peaks representing adducts of Δ +105 ± 3, Δ +131 ± 3, and Δ +202 ± 3 Da. We did not detect unmodified

R244S β-lactamase (28,803 ± 3Da) after 15 minute inactivation with sulbactam.

Due to the greater affinity and lower partition ratios of the penem inhibitors

for SHV-1, we reacted WT and R244S in a 500:1 I:E ratio. Unlike our observations

with sulbactam and clavulanate, we did not see evidence for inhibitor

fragmentation with the penems (Figure 3-3, pages 104-5). For both penem 1 and 2

with SHV-1 and R244S we observed adducts of Δ +307 ± 3 for penem 1, and Δ

+321 ± 3 for penem 2. In addition, uninhibited enzyme is not present after 15

minute incubation with either enzyme. Curiously, there appear to be smaller peaks

that correspond to multiple penem adducts for each enzyme (Figure 3-3, pages

104-5).

DISCUSSION

Our MIC and kinetic analyses firmly establish that R244 has a unique role in

inhibitor recognition and turnover in SHV (Thomson et al., 2006). More

importantly, our data also reveal that the specific properties of the inhibitor in the

active site largely influence the “inhibitor resistant” phenotype that is seen. The

paradoxical increased susceptibility to ampicillin/sulbactam in this panel of

ampicillin/clavulanate resistant isolates is intriguing. Eighteen of nineteen

substitutions at R244 lead to β-lactamases with increased susceptibility to

ampicillin/sulbactam by 2 or more dilutions. Equally surprising were our results

with penem 1 and, to a lesser extent, penem 2. Why are the interactions of these

inhibitors so different and what does this mean for the future of β-lactamase

inhibitor design?

To begin to understand the mechanistic basis of this phenotypic behavior

we compared the dissociation constants for the pre-acylation complexes of SHV-1,

R244S, R244Q and R244L β-lactamases against each inhibitor. It becomes readily

apparent that the principal consequence of R244 substitutions with regards to

inhibitors is increased KI values (Bret et al., 1997; Thomson et al., 2006). This observation is also common among inhibitor resistant TEM (IRT) β-lactamases as well as other SHV variants (Bermudes et al., 1999; Chaibi et al., 1998; Helfand et

al., 2003a). To further analyze the inactivation process, we next examined partition

ratios. Of critical importance is our finding that sulbactam, a sulfone, is readily

-1 hydrolyzed by SHV-1 (kcat= 730 ± 40 s ). This kcat is larger than the kcat of many β- 93

lactam substrates and explains the elevated MIC (Hujer et al., 2001). In contrast,

-1 the kcat of SHV-1 for clavulanate is 2.4 s ± 0.4. Substituting Ser for Arg at 244, we

see a 56 times reduction in kcat for sulbactam; this reduction allows for more

efficient inactivation of the enzyme (a log difference in kinact/Ki). Interestingly,

although R244 substitutions decrease hydrolysis of sulbactam, the opposite is

seen for clavulanate; R244S, R244Q, and R244L hydrolyze clavulanate from 1.9

to 25 times faster than WT. It is remarkable that the same substitutions can have

such contradictory effects on the turnover of these two inhibitors, despite their

strikingly similar reaction chemistry (Helfand et al., 2003b).

Clues into the basis of this “paradox” in SHV can be gleaned from

examining the atomic structures of the SHV E166A β-lactamase with the trans- enamine intermediates of clavulanate (Figure 3-4A, pages 106-7) and sulbactam

(Figure 3-4B, pages 106-7) trapped in the active site (Kuzin et al., 1999; Padayatti et al., 2005). Interpretation of these two structures reveals that R244 contributes to the structural stability of the β-sheet network that comprises the “right side” of the binding site pocket. This is accomplished through hydrogen bonding interactions with N276, L265 and G236 (Figure 3-4, pages 106-7). Removing these contacts

(R244S) is suspected to enlarge the active site pocket. It is also apparent that sulbactam is oriented differently in the active site pocket than clavulanate; clavulanate is slanting towards the B3 beta sheet network, while sulbactam is positioned straight out of the active site pocket (Figure 3-4, pages 106-7).

Disruptions in the beta-sheet network, which coordinate a catalytic water molecule, are more likely to have an adverse effect on sulbactam, by increasing critical 94

hydrogen bond distances even more. Thus, the loss of affinity is accompanied by

alterations in architecture that retard hydrolysis of sulbactam. Despite the resulting

MICs to ampicillin/sulbactam when R244 is mutated, the level of resistance remains too high for the combination to regain therapeutic efficacy.

We compared the trends of the kinetic constants obtained in this study with similar analyses that examined the inactivation of the TEM R244S, -C, and –H

(TEM-30, -31, and -51) β-lactamases by sulbactam (Bret et al., 1997; Imtiaz et al.,

1994a). Although these investigations also observed elevated KI values, it is

noteworthy that the R244 substitutions all lead to β-lactamases with >3 times reduction in sulbactam kinact. This contrasts to our data with SHV, which demonstrate a >2-fold increase in kinact for sulbactam among the variants studied.

These considerations, combined with our previous kinetic data for clavulanate, compel us to conclude that the “enzymatic machinery” of these two class A β- lactamases is distinctly different. This finding may have significant impact on the development of “second generation” inhibitors to target resistant β-lactamases.

Based on our studies of meropenem inactivation of SHV-1 and R244S, we

2 next explored the interactions of novel inhibitors with sp hybridization at the C3 position to the carboxylate. The two penems also possess bicyclic R1 side chains.

The clavulanate resistant variants at R244 expressed in E.coli DH10B maintain susceptibility to penem 1 and to a lesser extent penem 2 when tested with ampicillin. The most salient feature of these penem inhibitors is their extremely high affinity (nM range) for the active site of WT SHV-1. To illustrate, with penem

1, although the R244 substitutions increase KI values 80-280 times, the resulting 95

affinity was still in the low μM range, a critical factor in the maintenance of therapeutic efficacy.

Previous atomic structure determinations and molecular modeling studies lead us to hypothesize that penem 1 placed in the active site of SHV-1 makes favorable interactions between the large R1 group and Y105, and the carboxylate

of C3 with R244 (Nukaga et al., 2003a; Venkatesan et al., 2006; Weiss et al.,

2004). Although the C3 carboxylate interaction may be diminished in the R244S

variant, the R1 interactions with Y105 compensate (KI only increases to 1.1 µM).

Clavulanate and sulbactam, without R1 side chains, are primarily dependent on

interactions with the C3 carboxylate. A crystal structure of the post-acylation

complex of a related penem in SHV revealed significant stacking interactions of the

penem R1 side-chains with Y105 (Nukaga et al., 2003a). These additional

contacts help to explain the extremely high affinity of this group of inhibitors. This

interaction is also predicted to be relatively stable due to the highly conserved nature of Y105 across families of β-lactamases (Bethel et al., 2006; Shimamura et al., 2002).

In contrast to clavulanate and sulbactam, a remarkable feature of the penem inhibitors is their unique, rapid mechanism of inactivation. These inhibitors use novel, endo-trig cyclization (Figure 3-5, page 108) which results in a non- fragmented adduct during the time period studied (Nukaga et al., 2003a). The

rapid formation of a cyclic enamine within the active site may also displace the catalytic water molecule and reduce turnover (Nukaga et al., 2003a; Venkatesan et

al., 2006). This mechanism is likely to occur in SHV and the inhibitor resistant

R244S; after a 15 minute incubation all of the SHV-1 and SHV R244S exhibit

mass shifts of Δ +307 ± 3 Da for penem 1 and Δ +321± 3 Da for penem 2. There is

no evidence of inhibitor fragmentation or hydrolysis, consistent with our kinetic

analysis that indicated a remarkably low inhibitor turnover. Furthermore, our MICs

with penem 1 suggest effective cell penetration and in vivo efficacy. Combined

with impressive affinity for the active sites of WT SHV and inhibitor resistant

variants, these novel inhibitors show promise as lead compounds for future clinical

development.

The main challenges in β-lactamase inhibitor development are discovering why single amino acid changes alter inhibitor turnover and affinity, and developing novel inhibitors with activity against inhibitor resistance enzymes. Many of the inhibitor resistant defining mutations, including R244S, -L, and -Q studied herein, led to reduced affinity for these partner compounds. Therefore, a goal in the design of novel inhibitors is improving affinity against native β-lactamases, and inhibitor resistant variants. The development of an inhibitor with nM affinity for wild type enzymes, affords marked flexibility in the face of amino acid substitutions at this position. In conclusion, we continue to characterize the importance and unique role of R244 in class A β-lactamases and show that inhibitor resistant SHV variants at this position interact differently with clavulanate and sulbactam. Most importantly, this phenotype can be overcome by innovative inhibitor design that incorporates chemical properties into novel inhibitors which improve affinity by

utilizing conserved motifs and use novel reaction mechanisms that result in low partition ratios.

Table 3-1: MICs (μg/ml) of E. coli DH10B containing SHV-1 and all 19 variants at Arg244. Inhibitors were evaluated in the presence of 50 μg/ml ampicillin

Ampicillin# Clavulanate# Sulbactam Penem 1 Penem 2 Arg >16384 2 256 0.25 2 (WT) Ala 256 4 32 1 8 Asn 256 4 32 .5 8 Asp 2 N/A N/A N/A N/A *Cys 128 2 32 .5 8 Gln 1024 16 64 1 16 Glu 256 8 32 1 8 *Gly 256 4 32 .5 8 *His 1024 8 64 2 16 Ile 256 8 16 1 16 *Leu 128 16 32 1 16 Lys 2048 8 128 .25 4 Met 512 16 64 1 16 Phe 256 8-16 32 1 16 Pro 1 N/A N/A N/A N/A *Ser 1024 8 32-64 2 16 *Thr 512 8 32 1 16 Trp 8 N/A N/A N/A N/A Tyr 128 16 16 1 8 Val 128 8-16 32 1 16 *Substitutions found clinically in TEM N/A Mutants had MIC <50 μg/ml ampicillin, so could not be tested Bold: Variants chosen for kinetic analysis #From Chapter 2

Table 3-2: Kinetic properties of SHV-1 and variants at R244 for Sulbactam, and Clavulanate*

SHV-1 R244S R244Q R244L

Sulbactam

8.6 ± 0.7 240 ± 20 510 ± 34 1200 ± 100 Ki (μM) -1 kinact (s ) 0.056 ± 0.003 0.13 ± 0.01 0.14 ± 0.01 0.19 ± 0.03 0.0065 ± 0.00054 ± 0.00027 ± 0.00016 ± -1 -1 kinact / Ki (μM s ) 0.0006 0.00006 0.00003 0.00003

kcat/kinact 13,000 ± 100 100 ± 10 500 ± 20 300 ± 20

-1 kcat (s ) 730 ± 40 13 ± 2 70 ± 6 57 ± 10 Clavulanate* K (μM) i 1 ± .04 63 ± 3 85 ± 7 360 ± 20

-1 kinact (s ) 0.04 ± .002 0.09 ± .005 0.09 ± .007 0.11 ± .01 0.00031 ± -1 -1 kinact / Ki (μM s ) 0.04 ± .003 0.0014 ± .0001 0.0011 ± .0001 .00003

kcat/kinact 60 ± 10 50 ± 10 180 ± 10 550 ± 30

-1 kcat (s ) 2.4 ± 0.4 4.5 ± 0.9 19 ± 2 61 ± 6 *From Chapter 2

100

Table 3-3: Kinetic properties of SHV-1 and variants at R244 for Penem 1 and Penem 2

SHV-1 R244S R244Q R244L

Penem 1

Ki (µM) 0.014 ± .002 1.1 ± 0 .1 1.2 ± 0.1 3.9 ± 0.4

-1 kinact (s ) 0.17 ± .01 0.15 ± .01 0.14 ± .01 0.15 ± .02

-1 -1 kinact / Ki (μM s ) 12.3 ± 1.6 0.14 ± .02 0.12 ± .01 0.038 ± .007

kcat/kinact 2 ± 1 1 ± 1 1 ± 1 1 ± 1

-1 kcat (s ) 0.35 ± .17 0.15 ± .07 0.14 ± .07 0.15 ± .08

Penem 2

Ki (nM) 0.048 ± .004 4.4 ± 0.6 4.6 ± 0.3 12.5 ± 0.7

-1 kinact (s ) 0.186 ± .008 0.159 ± .007 0.140 ± .009 0.116 ± .007

-1 -1 kinact / Ki (μM s ) 3.9 ± .3 0.037 ± .002 0.031 ± .003 0.0093 ± .0008

kcat/kinact 2 ± 1 1 ± 0.5 1 ± 0.5 1 ± 0.5

-1 kcat (s ) 0.37 ± .19 0.16 ± .08 0.14 ± .07 0.12 ± .06

101

Figure 3-1 The chemical structures of the clinically available β-lactamase inhibitors, sulbactam, and clavulanic acid, as well as the novel inhibitors, penem 1 and 2.

102

Figure 3-2 Deconvoluted ESI-MS spectra of SHV-1 and R244S reacted with

1000:1 sulbactam:enzyme for 15 minutes. Notice that at this time point most of the

SHV-1 is unmodified, while all of the R244S has adducts.

103

104

Figure 3-3 Deconvoluted ESI-MS spectra of SHV-1 and R244S reacted with 500:1

Penem 1/2: enzyme for 15 minutes. At this time point there is no unreacted SHV-1 or R244S. The mass spectrometry indicates that the inhibitor covalently attaches to the enzyme without fragmentation, in contrast to the fragmentation seen with sulbactam. Interestingly there is some evidence for the addition of multiple inhibitor molecules

105

106

Figure 3-4 SHV E166A with the trans-enamine intermediates of clavulanate (top), and sulbactam (bottom). Important residues for interactions with R244 (green) are shown, and the inhibitors are highlighted in yellow.

107

Figure 3-5 Proposed endo-trig cyclization of penem 1 with SHV-1. This results in an enzyme adduct equal to the size of the full inhibitor molecule with no

fragmentation.

108

CHAPTER 4

Boronic Acid Transition State Inhibitors and the ESBL Phenotype

Adapted from

Use of Novel Boronic Acid Transition State Inhibitors To Probe Substrate Affinity in

SHV-Type Extended-Spectrum ß-Lactamases

Jodi M. Thomson, Fabio Prati, Christopher R. Bethel, and Robert A. Bonomo

Antimicrobial Agents and Chemotherapy

April, 2007, Volume 51(4): 1577-9

INTRODUCTION

Resistance to β-lactam antibiotics is a major public health threat resulting in

treatment failure of serious infections due to Gram-negative pathogens. The

principal mediator of resistance to penicillins and cephalosporins is the bacterial

production of β-lactamase enzymes. To maintain the effectiveness of β-lactam

antibiotics two strategies are employed: 1) synthesize new β-lactam compounds that resist hydrolysis; 2) develop potent β-lactamase inhibitors (Fisher et al., 2005).

The first strategy has resulted in the development of five “generations” of cephalosporins. When initially released into clinical practice, extended-spectrum cephalosporins resisted hydrolysis by the prevalent class A β-lactamases, TEM-1 and SHV-1. Unfortunately, single amino acid substitutions in these related enzymes have resulted in the emergence of the extended-spectrum β-lactamase

(ESBL) phenotype. The premier mutation responsible for this phenotype in the 109

SHV β-lactamase family is a G238S substitution (Hujer et al., 2001). The 1-3 Å binding pocket expansion caused by this substitution may be required for the bulky oxyimino-cephalosporins to bind in the active site (Hujer et al., 2002; Nukaga et al.,

2003b; Orencia et al., 2001) .

Analysis of naturally occurring TEM and SHV ESBL variants has also identified lysine residues at Ambler positions 104 and 240 as contributing to cephalosporin resistance. These substitutions (104K and 240K), studied in TEM, and recently in SHV by our laboratory, are hypothesized to increase affinity via hydrogen bonding or ionic interactions with the R1 oxyimino side chains of the β- lactam (Bethel et al., 2006; Blazquez et al., 1995; Petit et al., 1995). However, the multi-step nature of cephalosporin hydrolysis, inability to trap intermediates, and low turnover by SHV-1 limits an in-depth study of amino acid substitution effects on the ESBL phenotype.

The search for novel inhibitors of class A β-lactamases has resulted in the discovery of the boronic acid transition state inhibitors (BATSIs). These compounds are effective, reversible inhibitors of AmpC, TEM, and CTX-M β- lactamases, and also are remarkable probes for studying active site chemistry

(Chen et al., 2005; Powers et al., 2001; Powers and Shoichet, 2002). Using our work with inhibitor resistant SHV β-lactamases (Arg 244 mutants) as a guide, we reasoned that BATSIs could also be used to investigate enzyme substrate interactions that define the unique properties of the ESBL phenotype in SHV β- lactamase (Thomson et al., 2006). To this end, we tested inhibitors that contain the

R1 side chains of two extended-spectrum cephalosporins, cefotaxime and 110

ceftazidime, attached to boronic acid (Figure 4-1, page 119). We used these new

compounds to gain fresh insight into the contributions of the single amino acid

substitutions that confer resistance to the extended-spectrum cephalosporins.

Additionally, we show that these molecules have lower KI values against some

SHV enzymes displaying the ESBL phenotype.

111

EXPERIMENTAL PROCEDURES

β-Lactamase genes encoding SHV-1, -2, -5, D104K, and D104K-G238S

were cloned into pB CSK (-) as previously described (Bethel et al., 2006; Hujer et

al., 2002; Rice et al., 2000). β-Lactamases were expressed in Escherichia coli

DH10B cells and initially purified as described (Hujer et al., 2002). SHV enzymes

were quantified and purity was assessed by SDS-PAGE. The SHV-2, D104K, and

the doubly substituted D104K-G238S β-lactamases were further purified by size

exclusion chromatography using a Waters HPLC system (Bethel et al., 2006).

Kinetic measurements were performed by continuous assays at room

temperature using an Agilent 8453 Diode array spectrophotometer. Inhibition

constants, KIs, were determined by competing increasing concentrations of BATSI against the colorimetric substrate nitrocefin at 4-6 times the Km for the enzyme.

Data was analyzed using Origin® 7.5 SR2 Software and the equation:

I][ i = (Eq.1) S][ KI i 1(][ ++ ) K m

where i = fraction inhibition, [S] = nitrocefin concentration, [I] = inhibitor

concentration, and Km is the Michaelis constant for nitrocefin (Cheng and Prusoff,

1973). BATSIs were synthesized as previously described (Powers et al., 2001). In

addition, full progress curves were performed in the absence of inhibitor, and with

2 concentrations of BATSI (5µM and 10µM for ceftazidime, 10 µM and 20 µM for

cefotaxime BATSI) while varying the nitrocefin concentration to determine the

112

mode of inhibition. Results were graphed as Lineweaver-Burke plots (1/[S] vs. 1/v) and points were fit by linear regression with Origin® 7.5 SR2 software.

To determine whether these compounds exhibited time dependent inhibition, enzyme was pre-incubated with inhibitor for increasing timepoints from

0-250 minutes before initiating the reaction with the addition of nitrocefin. Initial enzyme velocities were then graphed as a function of incubation time.

113

RESULTS

In general, the cefotaxime BATSI (Figure 4-1B, page 119) demonstrated

greater affinity for the SHV ESBLs than wild type SHV-1. The inhibitor exhibited

the lowest KI value for doubly substituted SHV D104K G238S: 1.1 ± 0.2 µM, compared to the WT KI value of 8.9 ± 0.9 µM (Table 1). The greatest benefit

comes from the G238S substitution; the KI for SHV-1 (G238S) was 3.4 µM, and

affinity for SHV D104K was 5 µM.

The ceftazidime BATSI (Figure 4-1D, page 119) had a similar range of

binding affinities, but the effects of the ESBL defining substitutions are different.

Most surprising, was the relatively high affinity (2.2 ± 0.2 µM) of the ceftazidime

BATSI for the WT enzyme. Unlike with the cefotaxime BATSI, the G238S

substitution decreased binding to the ceftazidime inhibitor [6.8 ± 0.9 µM for SHV-2

(G238S), 4.5 ± 0.5 µM for SHV-5 (G238S E240K) and 4.5 ± 0.5 µM for G238S

D104K]. The only substitution that resulted in an appreciable affinity increase was

D104K alone (0.7 ± 0.1 µM).

As a control, we studied the inhibition of all enzymes with a reference

compound (Figure 4-1E, page 119) which lacks an R1 side chain. The affinity of

this compound for all enzymes was very weak (on the order of hundreds of

micromolar) confirming the importance of the R1 side chain for specificity of

binding.

We next analyzed progress curves with SHV-1 and 3 concentrations of

inhibitor (including a zero inhibitor curve) and the substrate nitrocefin to determine

the mechanism of inhibition. As shown in Figure 4-2 (page 120), both the 114

ceftazidime and cefotaxime BATSI exhibit primarily competitive inhibition, as suspected based on their design. The ceftazidime BATSI does have a somewhat mixed inhibition. However this is likely due to error caused by the higher affinity being more susceptible to a mixing effect.

As some boronic acid inhibitors exhibit marked time dependence

(compound 2, Chapter 2), we tested the effect of incubation time on inhibition of

SHV-1 by both the ceftazidime and cefotaxime BATSI. As shown in Figure 4-3

(page 121), neither BATSI exhibited any time-dependence of inhibition. Maximal inhibition was achieved by the first measurement and inhibition was stable for up to

4 hours.

115

DISCUSSION

We found that BATSIs with the R1 side chains of two extended-spectrum cephalosporins bind SHV β-lactamases with µM affinity (Table 4-1, page 118).

This was notable for both the ceftazidime and cefotaxime BATSIs. In addition, both

BATSIs inactivated SHV-1 competitively and in a time independent fashion

(Figures 4-2, 4-3, pages 120-121).

Interestingly, we found a different pattern when we compared the ceftazidime

BATSI to the cefotaxime BATSI binding to WT and ESBLs. The cefotaxime BATSI had higher affinity to enzymes with the ESBL defining G238S substitution. This increase in affinity is in line with the observation that the G238S substitution is a

“cefotaximase” specific mutation in SHV (Bethel et al., 2006; Hujer et al., 2002).

The 104K substitution also increased affinity to the cefotaxime BATSI, both alone, and in combination with the G238S substitution.

It is intriguing that the ceftazidime BATSI binds to most ESBLs more weakly than it binds SHV-1. Only the D104K substitution alone led to a substantial increase in affinity for this inhibitor. These KI measurements recall the observations made by Wang, et al. in comparing TEM-1 and TEM-52 (Glu104-->Lys, Met182--

>Thr, and Gly238-->Ser) affinity for the same BATSI (Wang et al., 2003). They observed reduction in affinity for the ESBL (3.1 µM) compared to TEM-1 (0.51µM).

This leads us to wonder whether the main impact of widening the active site cavity with the G238S substitution on class A β-lactamases is on a) precovalent

116

encounter complex, b) the formation of the high energy acylation tetrahedral

intermediate, or c) substrate reactivity post-binding. Clearly, the G→S substitution

increases the kcat of TEM-19 (G238S) and MICs of SHV-2 for ceftazidime (Hujer et

al., 2002; Raquet et al., 1994). BATSIs may be providing insight into affinity that is not possible with oxyimino-cephalosporin substrates because of poor turnover.

These data also support our previous studies of the impact of D104K on substrate hydrolysis in SHV (Bethel et al., 2006). It is now established that Lys at

104 plays a significant role in oxyimino-cephalosporin binding in SHV. In both

instances, the β-lactamase containing the 104K substitution demonstrates greater

affinity for the BATSIs. Further studies are required to investigate if the R1 side

chain rotates in the binding pocket to align a carbonyl oxygen of the side chain with

the Lys at position 104 (Figure 4-1A, page 119)) (Bethel et al., 2006).

117

Table 4-1 Inhibition constants (KI) of BATSIs with SHV-1, SHV-2,

SHV-5, D104K and G238S-D104K

KI cefotaxime KI Ceftazidime KI Reference β-lactamase BATSI (µM) BATSI (µM) Compound (µM)

SHV-1 8.9 ± 0.9 2.2 ± 0.2 300 ± 15

SHV-2 3.4 ± 0.7 6.8 ± 0.9 330 ± 10

SHV-5 5.0 ± 0.5 4.5 ± 0.5 360 ± 20

D104K 5.0 ± 0.5 0.7 ± 0.1 190 ± 20

G238S-D104K 1.1 ± 0.2 4.5 ± 0.3 300 ± 30

118

Figure 4-1 Structures of cefotaxime (A), cefotaxime BATSI (B), ceftazidime (C), ceftazidime BATSI (D), and a Reference Compound (E) which lacks the oxyimino- cephalosporin R1 side chain. The arrow in (A) indicates the direction of proposed rotation in the binding pocket as discussed in text.

119

Figure 4-2 Progress curves of SHV-1 inhibited by Ceftazidime BATSI (left) and the

Cefotaxime BATSI (right) show a primarily competitive mode of inhibition.

120

Figure 4-3 Timed inactivation experiment with the ceftazidime BATSI (left) and cefotaxime BATSI (right) indicate a fast equilibrating inhibition without time- dependence.

121

CHAPTER 5

β-Lactamase Inhibition by BLIP

Adapted in part from

Structural and Computational Characterization of the SHV-1 β-Lactamase-β-

Lactamase Inhibitor Protein Interface

The Journal of Biological Chemistry

Kimberly A. Reynolds, Jodi M. Thomson, Kevin D. Corbett, Christopher R. Bethel,

James M. Berger, Jack F. Kirsch, Robert A. Bonomo, and Tracy M. Handel

September 8, 2006, Volume 281(36):26745-53

INTRODUCTION

Clavulanic acid was initially discovered as a natural product of the

bacterium Streptomyces clavigularis (Reading and Cole, 1977). It is unknown why

a bacterium would make such an effective inhibitor of β-lactamases. The discovery

of a second β-lactamase inhibitor, a protein, was even more intriguing. The β- lactamase Inhibitory Protein (BLIP) is a 17.5 kDa protein cloned in 1990 (Doran et al., 1990). Its presence was detected due to a disparity between clavulanic acid quantification and β-lactamase inhibitory activity of S. clavigularis cell lysates. The isolated protein was found to be a stoichiometric inhibitor of the Escherichia coli pUC β-lactamase. 122

Subsequent studies have found BLIP to be a remarkably high affinity

inhibitor of disparate Class A β-lactamases, with binding affinities in the pico-molar to micro-molar range (Huang et al., 2003; Petrosino et al., 1999; Rudgers and

Palzkill, 1999; Zhang and Palzkill, 2003). This extraordinarily tight binding has spurned the search and development of several novel peptide inhibitors of β- lactamases (Huang et al., 2003; Huang et al., 1998; Huang et al., 2000). The goal is to use molecular mimicry to design a small molecule inhibitor that takes advantage of the promiscuous BLIP binding interactions.

Co-crystal structures of BLIP complexed to the TEM β-lactamase revealed two distinct binding loops of the protein interacting with the active site. The Asp49 and Phe142 BLIP binding loops both occupy the active site of TEM and approximate the binding of penicillin. Additionally, there is a very large buried surface area (>2700Å2) that would be impossible to mimic with a therapeutic agent

(Strynadka et al., 1996; Strynadka et al., 1994). Despite this, an in depth analysis

of the protein-protein interface of BLIP binding to several β-lactamases is sought to

unravel the mystery of this seemingly indiscriminate β-lactamase inhibition.

Although BLIP binds to the class A TEM-1 and SME-1 enzymes with low

nanomolar affinity, affinity for SHV-1 is 2000-times weaker (Zhang and Palzkill,

2003). This observation occurs despite a high (68%) sequence identity between

SHV and TEM. Alanine scanning mutagenesis at the TEM-1 BLIP interface revealed many residues crucial for the protein-protein binding interaction (Zhang

and Palzkill, 2004). One residue deemed critical for substrate specificity is K74 of

BLIP. In the co-crystal structure of TEM-1 with BLIP, this residue makes a 123

hypothetical salt bridge with Glu 104 of TEM-1 (Strynadka et al., 1996).

Interestingly, SHV-1 has an aspartate at this position, which may modify the interface.

To study the binding interactions of SHV with BLIP, we created both the

TEM equivalent at position 104 (AspÆ Glu), and also mutated the residue to Lys to study the effect of charge at that position. Crystal structures have been

determined for the SHV-1:BLIP, SHV D104K:BLIP, and SHV-1 D104E: BLIP

complex.

To further probe BLIP binding specificity and the utility of using this inhibitor

as a model for future antimicrobial development, we next tested its ability to inhibit

a novel family of class A β-lactamases, the Klebsiella pneumoniae

Carbapenemases, KPC. This emerging β-lactamase family was first isolated from an imipenem resistant Kelbsiella pneumoniae isolate from North Carolina in 2001.

Since that time, KPC producing isolates have been implicated in serious outbreaks of infections with carbapenem resistant gram negative bacteria in New York City, and have recently begun to appear across the globe (Deshpande et al., 2006;

Hossain et al., 2004; Villegas et al., 2007; Villegas et al., 2006; Wei et al., 2007;

Yigit et al., 2001; Yigit et al., 2003). The seriousness of the epidemic is related to the broad substrate specificity of this enzyme family. Not only can KPC enzymes hydrolyze the most potent β-lactam antibiotics, the carbapenems, they also provide bacteria with resistance to the penicillins and cephalosporins, creating near pan-β-lactam resistant isolates (Alba et al., 2005; Thomson and Bonomo,

2005). Since these genes can be found on transferable plasmids their imminent 124

spread is a serious threat to the use of β-lactam antibiotics. The discovery of a high potency inhibitor of these enzymes is essential to save this antibiotic class.

125

EXPERIMENTAL PROCEDURES

Cloning and Protein Purification—The blaSHV gene cloned into pBC SK(-) phagemid (Stratagene, La Jolla, CA) was used to construct the blaD104K and blaD104E variants of SHV-1 β-lactamase (Hujer et al., 2001; Rice et al., 2000). The

blaD104E mutation was created in pBC SK(-), whereas the blaD104K variant was

subcloned into pET- 24a(+) (Novagen, Madison, WI). Mutagenesis was then

performed at position 104 using the QuikChange® site-directed mutagenesis kit

(Stratagene). SHV-1 and SHV D104E were expressed in pBC SK(-) in E. coli

DH10B cells grown overnight, without induction. SHV D104K was expressed in E. coli BL21(DE3) cells by induction with 0.2mM isopropyl-D-thiogalactopyranoside

(IPTG) at A600 = 0.8 for 3 h at 37°C. For all enzymes, cells were harvested by

centrifugation at 4°C and frozen overnight. β-Lactamase was liberated using

stringent periplasmic fractionation with lysozyme and EDTA as previously

described (Lin et al., 1998). The blaKPC-2 construct was a kind gift from Hesna

Yigit, Centers for Disease Control and Prevention.

Preparative isoelectric focusing was performed with a Sephadex granulated gel and ampholines in the pH range of 3.5–10 (Amersham Biosciences). The protein was eluted with 20mM diethanolamine buffer, pH 8.3. An additional HPLC purification step was performed on a Waters high pressure liquid chromatograph using a Sephadex Hi Load 26/60 column (Amersham Biosciences) and eluted with phosphate- buffered saline (pH 7.4). This expression and purification scheme 126

yielded 10 mg of pure protein/liter of culture for SHV-1 and SHV D104E and 1 mg

for SHV D104K.

TEM-1 cloned into pET24a(+) with a N-terminal OmpA secretion signal was

provided by Ste´phane Gagne´ (Universite´ Laval). TEM-1 was expressed and

purified as in (Sosa-Peinado et al., 2000).

The BLIP construct was a generous gift from Susan Jensen (University of

Alberta). The BLIP bli gene was cloned into pET26b (Novagen, Madison, WI), with

the native S. clavuligeris signal sequence at the N terminus. BLIP was expressed

in E. coli BL21(DE3) cells by inducing with 1 mM isopropyl-β-D-

thiogalactopyranoside at A600 = 0.5 for 3 h at 30 °C. Cells were harvested by

centrifugation, resuspended in 20 g/100 ml sucrose, 1 mM EDTA, 30 mM Tris, pH

8.0, and incubated for 15 min. The cells were pelleted and resuspended in 5 mM

MgCl2. After centrifugation at 7,000 rpm for 10 min, the periplasmic fraction was

retained, and a Complete protease inhibitor mixture tablet (Roche Applied

Science) was added. That fraction was subsequently clarified by additional centrifugation and loaded onto a HiPrep 16/10 Q-XL anion exchange column

(Amersham Biosciences) equilibrated in 25 mM Tris, pH 8.4, 1 mM EDTA. A 250- ml gradient from 0 to 500 mM NaCl was used to isolate BLIP, which elutes at 150 mM NaCl. Fractions were pooled and concentrated to a final volume of 2 ml and subsequently passed through a HiLoad 26/60 Superdex 75 preparation grade gel filtration column (Amersham Biosciences) equilibrated in 50 mM Tris, pH 8.4, 100 mM NaCl. After purification, BLIP was concentrated to 1 mg/ml and stored at -80

127

°C. This expression and purification scheme yielded roughly 0.5 mg of pure protein/ liter of culture.

Crystallization, Data Collection, and Structure Solution— SHV-1, or SHV D104K, was mixed 1:1 with BLIP in 20mM BisTris, pH 7.25, 50 mM NaCl, concentrated to

8.7 mg/ml, and dialyzed overnight against 20 mM BisTris, pH 7.25, 50 mM NaCl.

Crystals were grown at 19 °C in microbatch format under Al’s oil (Hampton

Research, Aliso Viejo, CA) by mixing 1µL of protein with 1 µL of well solution. The

SHV-1_BLIP crystals were grown in a well solution of 60% ammonium sulfate, 50 mM cacodylate, pH 6.5; SHV D104K-BLIP crystals were grown in a well solution of

8% polyethylene glycol 8000, 50 mM cacodylate, pH 6.5. For harvesting, a cryoprotectant solution containing well solution plus 25% glycerol for SHV-1-BLIP or well solution plus 30% xylitol for SHV D104K-BLIP was added directly to the drop, and the crystals were immediately looped and flash-frozen in liquid nitrogen.

Data sets were collected on Beamline 8.3.1 at the Advanced Light Source at Lawrence Berkeley National Laboratory (MacDowell et al., 2004). A preliminary

2.1 Å data set for SHV-1_BLIP was collected on an R-Axis IV++ at the University of California Berkeley (Rigaku/ MSC, The Woodlands, TX). Data were indexed and reduced with HKL2000 (Otwinowski et al., 2003) or ELVES (Holton and Alber,

2004) using MOSFLM (Leslie and Janisch, 1992). For the 1.8Å high resolution

SHV-1-BLIP structure, molecular replacement was performed using a partially refined structure from the 2.1 Å data set. Initial maps for the SHV D104K-BLIP structure were generated by molecular replacement with PHASER (Storoni et al., 128

2004) using polyalanine models of SHV-1 (coordinates taken from Protein Data

Bank entry 1SHV) and BLIP (coordinates taken from Protein Data Bank entry

1JTG) (Kuzin et al., 1999; Strynadka et al., 1996).

The SHV D104E: BLIP crystals were made by mixing 2uL of protein solution (4.7 mg/mL 1:1 BLIP:SHV molar ratio, 30mM BisTris 50mM NaCl, pH

7.25) with 2uL well solution (40% saturated ammonium sulfate, 100mM Tris pH

8.5) in hanging drop format above 1mL of well solution. The crystals were harvested and flash frozen using 30% xylitol as a cryoprotectant. The data was collected at the APS in Chicago, and solved using the wild type SHV:BLIP structure as an initial model. The structure finished to 1.56A resolution, with final

Rwork and Rfree of 17.0% and 17.9% respectively. Spacegroup P63, and unit cell parameters a, b, and c: 127.85A, 127.85, 73.18A, and angles 90.0, 90.0, 120.0 degrees.

Manual rebuilding for all structures was carried out with O (Jones, 1997).

Refinement was carried out with Refmac (Lamzin and Wilson, 1993) using ARP to automatically place ordered waters, followed by TLS refinement (Winn et al.,

2001). Structural alignments and r.m.s. deviations were calculated with LSQMAN

(Kleywegt and Jones, 1994). All molecular figures were created with PyMOL

(DeLano, 2002).

Inhibition Assays—All kinetic determinations were performed using nitrocefin (BD

Biosciences) as the indicator substrate to measure hydrolysis rates of SHV-1, SHV

D104K, SHV D104E, KPC-2, and TEM-1 β-lactamases with and without inhibitor. 129

BLIP and β-lactamase were incubated for 2 h at room temperature in 10 mM sodium phosphate-buffered saline containing 1 mg/ml bovine serum albumin.

Initially, 7 nM enzyme was used for all assays, but this was reduced to 4 nM for

SHV D104E and KPC-2, and 2 nM for TEM-1 to obtain an accurate measurement of the low Kd values. Reactions were initiated with nitrocefin at the Km for the enzyme (25 µM for SHV-1, 15 µM for SHV D104K, 25 µM for SHV D104E, and

150 µM for TEM-1). The exception was KPC-2, which required 100 µM nitrocefim

(5.5KI) to attempt to get an accurate measurement of this very high inhibition. Final

reaction volumes were 1 ml. All measurements were performed in triplicate.

Hydrolysis rates were determined at λ = 482 nm using the extinction

coefficient, ε = 17,400 M-1 cm-1 for the hydrolyzed form of nitrocefin (O'Callaghan

et al., 1972). BLIP inhibition curves were graphed using Origin®7.5 SR software

and fit to the following equation,

2 Efree = [E0] – (([E0] + [I0] + K*d - √(([E0] + [I0] + K*d) ) – (4[E0] [I0])))/2 (Eq 1)

where Efree represents the remaining free enzyme concentration calculated based

on activity, [E0] is initial enzyme concentration, [I0] is the concentration of BLIP, and

K*d is the apparent equilibrium constant (Petrosino et al., 1999). Kd was corrected

for the presence of substrate using the following equation.

Kd = K*d / (1 + [S]/Km) (Eq 2)

130

RESULTS

Kinetics- We found a wide range of inhibition by BLIP among our panel of SHV variants (Table 5-1, page 138; Figure 5-1, page 140). As expected, BLIP had a

relatively high KI (1250 ± 50 nM) for SHV-1, compared to TEM-1 (0.32 ± .03 nM).

Surprisingly, the D104K mutant exhibited a higher affinity for BLIP than WT by

about 2-fold (580 ± 40nM). This is despite the introduction of a positive charge that

was expected to clash with K74 of BLIP. When 104 was mutated from aspartate to

glutamate, however, the binding affinity increased 1000 times. This nearly brought

SHV and TEM to similar binding affinity for BLIP (1.1 ± 0.1 nM vs 0.32 ± .03 nM).

Just as intriguing was our kinetics data with the novel class A carbapenemase KPC-2. The affinity of BLIP for KPC-2 was measured to be low picomolar (.04 ± 0.1nM). Unfortunately, the sensitivity of our spectrophotometer does not allow us to use a KPC-2 concentration less than 4nM, so the error in this measurement is inherently large. In any case, the affinity is much tighter than any

β-lactamase inhibitor that we have tested in our laboratory.

Crystallography- To date we have, along with the Handel group (University of

California-Berkeley and San Diego), crystallized BLIP with SHV-1, SHV D104K,

and SHV D104E, and Kimberly Reynolds (SHV-1:BLIP and AHV 104K:BLIP), and

Melinda Hanes (SHV D104E:BLIP) resolved the co-crystal structures.

SHV-1:BLIP 131

The SHV-1:BLIP co-crystal structure overlays with the TEM-1:BLIP structure to a

high degree (Figure 5-2, page 141). TEM-1 and SHV-1 overlay with a root mean

square (r.m.s.) difference of 0.44Å, while BLIP overlays with an r.m.s of 1.04 Å

(Table 5-2, page 139). There are two main loops of BLIP that contact SHV-1; they are referred to by the amino acid at the terminus of the loops: Asp49 and Phe-142.

Both the Asp49 and Phe-142 binding loops are present in the active site of SHV-1

as is seen in TEM-1. Absent in the structure, however, is the salt bridge between

positions 104 in SHV-1 and K74 of BLIP. There is also an additional hydrogen

bond between E104 in TEM and Y143 of BLIP that is absent in the SHV-1 co-

crystal structure.

SHV D104K:BLIP The co-crystal structure of BLIP with SHV D104K varies greatly

from the structures with the WT TEM-1 and SHV-1. The most obvious difference is

the rotation of the F142 binding loop out of the active site (Figure 5-2, page 141).

In addition, the Asp49 loop twists and re-orients in the active site of the enzyme,

with Asp49 as the “pivet point”, retaining the same contacts in both structures, to

an r.m.s. of 0.98 Å (Figure 5-3, page 142). There is one extra hydrogen bonding

interaction between Y105 of SHV D104K and Y51 of BLIP, which is not present in

the WT structure. Additionally, the rotation allows the Y50 BLIP side chain to

move into the space vacated by the F142 BLIP binding loop, which creates a

new interaction: a hydrogen bond between Y50 of BLIP and N132 of SHV.

132

SHV D104E:BLIP The structure of SHV D104E in complex with BLIP, as expected, overlay very well with the TEM-1 co-crystal structure. The crucial salt bridge between K74 and position 104 of SHV has been restored in the D104E mutant (Figure 5-4, page 143). In addition, E104 forms an additional hydrogen bond with the backbone Y143 of BLIP. Not only are positive interactions gained with the D104E substitution, but an unfavorable clash is relieved (Figure 5-5, page 144). In the SHV-1:BLIP structure there is an electrostatic clash between

D104 and E73 of BLIP. The D104E substitution partially relieves this clash.

133

DISCUSSION

Our studies have revealed important determinants of binding specificity

amongst class A β-lactamases to BLIP. A “mystery” of BLIP binding since its

discovery has been its ability to inhibit a large range of class A β-lactamases.

However, the binding of BLIP to SHV-1 is noticeably weak compared to most

enzymes studied. Alanine scanning mutagenesis of BLIP along the enzyme

binding interface revealed clues to the disparity of binding. Very few residues in

the second, F142 binding loop appeared crucial for the inhibition of SHV-1 by

BLIP. In addition to this, one of the residues in the Asp42 binding loop, K73, that

was necessary for TEM-1 affinity, didn’t appear to affect inhibition of SHV-1

(Zhang and Palzkill, 2004). From the co-crystal structure of TEM-1:BLIP, the

contribution of K73 was shown to be a crucial salt bridge with Glu104 in TEM

(Strynadka et al., 1996). This residue is not conserved among class A β-

lactamases, and is an aspartate in SHV.

To study the effect of this residue on BLIP binding to SHV-1 we created both the TEM equivalent SHV D104E, and also SHV D104K and tested the effect of these substitutions on BLIP affinity, and the structural consequences of these

changes. We were intrigued that despite a >1000 times difference in KI of BLIP binding to TEM-1 and SHV-1, the structures overlaid to a very high degree

(r.m.s.d. 0.44 Å between the enzymes). As predicted, however, the salt bridge between K74 of BLIP and position 104 of SHV was missing (Figure 5-2, page

141). 134

We anticipated that introducing a positively charged lysine at the 104

position would introduce a large electrostatic clash between K74 of BLIP and 104

of SHV. When introduced in TEM, the 104K substitution decreases affinity more than 1000 times (Petrosino et al., 1999). Unexpectedly, the substitution increases

binding of SHV to BLIP by a factor of 2-fold (Table 5-1, page 138; Figure 5-1,

page140). When examined structurally, the mechanism for this increase in

affinity is a significant movement of both binding arms of BLIP in the active site of

SHV (Figure 5-2, page 141). As mentioned above, the second, F142 binding loop

of BLIP does not contribute much to the stability of the complex, as determined

by alanine scanning mutagenesis (Zhang and Palzkill, 2004). When 104 is

mutated to lysine, this binding loop rotates completely out of the binding pocket.

This is likely due to a clash with Y143 of BLIP. The removal of this binding loop

allows the Asp49 loop to rotate and allow Y50 of BLIP to occupy the space

vacated by F142. The additional interaction that Y50 makes with N132 of SHV,

along with an extra hydrogen bond between SHV Y105 and BLIP Y51 help

explain the additional stability of the complex.

The SHV D104E substitution has a substantial effect on BLIP affinity. The

KI of BLIP for this variant (1.1 ± 0.1 nM) is more than 1000-fold less than SHV-1.

This single amino acid substitution nearly results in an enzyme with BLIP affinity

equivalent to TEM-1 (Table 5-1, page 138; Figure 5-1, page 140). When

analyzed grossly, there is little change between the SHV-1:BLIP and the SHV

D104E:BLIP structures. However, the crucial salt bridge between 104E and K74

of BLIP is present, as is an additional hydrogen bonding interaction with Y143 135

(Figure 5-4, page 143). Another consequence of the D104E substitution that may explain the extreme increase in affinity is the relief of an electrostatic clash with

E73 of BLIP (Figure 5-5, page144). The extra carbon length of Glu compared to

Asp causes a 0.5 Å movement of the carboxylate of 104 away from the negatively charged carboxylate of BLIP E73. There are earlier studies that validate this explanation. An E73A mutant of BLIP, which would relieve this clash, binds to SHV-1 with 28 times greater affinity than WT BLIP. The same

E73A mutation has little to no effect on TEM-1 KI values compared to WT BLIP

(Zhang and Palzkill, 2004).

Our studies on the mutagenesis of position 104 have shed light on the discrepancy of BLIP binding between SHV and TEM β-lactamases. Although

BLIP makes numerous contacts with both enzymes and buries >2500 Å of surface area, the length of a single carbon at amino acid 104 affects binding affinity by a 3 log difference in KI. This could have practical impact upon design of peptide inhibitors and mimetics. For example, it is possible that a non-natural amino acid with a longer side-chain could be substituted for K73 to allow for salt bridge interactions with both TEM-1 and SHV-1.

The data with SHV D104K is intriguing and may explain BLIP’s ability to inhibit such a wide range of β-lactamases. We show that BLIP is able to employ an entirely different binding conformation to adapt to the changing architecture of its binding partner. In addition, this may have important implications for molecular dynamics simulations that assume a fixed backbone. It is likely that BLIP is not unique in this binding flexibility. 136

Given what we learned about BLIP binding to SHV, we were eager to

continue to explore affinity of BLIP for KPC-2, given its novelty and clinical threat.

KPC-2 shares approximately 35-40% amino acid identity with SHV-1 and TEM-1 and its crystal structure has recently been solved to 1.85 Å resolution. The active site of KPC-2 has been shown to be more shallow than other class A enzymes that don’t have carbapenemase activity, there are numerous amino acid differences, and there is a disulfide bond between C238 and C69 (Ke et al.,

2007). When we tested the inhibition of KPC-2 by BLIP, we were amazed to see the affinity in the low picomolar range (40 ± 100 pM). This inhibition was beyond the sensitivity of our spectrophotometer, as seen by our large margin of error

(Table 5-1, page 138; Figure 5-6, page 145). Interestingly, many residues in

TEM-1 and SHV-1 that have been shown to make important contacts with BLIP are not conserved in KPC-2. Most relevant to this discussion, amino acid 104, is a proline in KPC-2 (Figure 5-7, page 146). Other key residues (See Figure 5-3,

page 142) that diverge include position 105, which is a tryptophan in KPC-2 as

opposed to a tyrosine in SHV-1 and TEM-1, and residue 244 which is an alanine

instead of arginine. A co-crystal structure is being pursued to determine the binding determinants of this unusually high inhibition. Given the results above with the 104 variants, it will be remarkable to elucidate the binding architecture of

this interaction. Continued study of BLIP and its unique mechanism of inhibition

is needed if we are to design a small molecule mimetic of this extremely adaptable inhibitor.

137

Table 5-1 Inhibition constants of Class A β-lactamases, studied herein, with BLIP.

Enzyme KI (nM)

TEM-1 .32 ± .03

SHV-1 1250 ± 50

SHV D104K 580 ± 40

SHV D104E 1.1 ± 0.1

KPC-2 .04 ± 0.1

138

Table 5-2 r.m.s.d values between β-lactamases (blue) and BLIP molecules (yellow) in the structures studied. BLIP Co-crystal Structre SHVD104E SHVD104K SHV-1 TEM

0.44 Å 0.42 Å 0.44 Å 0 TEM

0.12 Å 0.35 Å 0 1.04 Å SHV-1

N/A 0 2.24 Å 2.24 Å SHVD104K

0 N/A 0.41 Å 0.92 Å SHVD104E

139

Figure 5-1 The inhibition of SHV-1 (black), SHV D104K (red), SHV D104E

(green), and TEM-1 (blue) by BLIP. The D104E substitution increases the affinity for BLIP over 1000-fold, nearly to the level of TEM-1

140

5-8 Figure 5-2 The alignment of BLIP (colored) as it was determined in the

crystal structures of SHV-1 (cyan), TEM(orange), and SHV D104K (red). (A and B)

BLIP aligns almost identically in the active sites of SHV-1 and TEM-1. (C and D) In

contrast, the SHV D104K variant forces the F142 binding loop out of the active

site pocket of SHV

141

Figure 5-3 The interaction of the binding loops of BLIP with key residues of SHV-1

(A) and SHV D104K (B). Notice the rotation of the D49 binding loop in (B) brings

Y51 of BLIP into contact with Y105 of SHV. Also notable in (B) is the absence of the BLIP F142 loop.

142

F142 BLIP Y143 BLIP

2.89Å

D104E SHV-1

2.99Å

K74 BLIP

Figure 5-4 Representation from the crystal structures of position 104 of SHV

D104E (pink), and TEM (yellow) in complex with BLIP reveals a similar geometry between the SHV D104E:BLIP and TEM: BLIP co-crystal structures. Hydrogen bonds are pictured with the backbone at BLIP Y143, (shown in cyan for SHV and blue for TEM) and salt bridges with BLIP K74. Figure courtesy of Melinda Hanes,

University of California-San Diego

143

D104 SHV-1 D104E SHV-1

K74 BLIP 5.53Å 6.02Å

E73 BLIP

Figure 5-5 A representation from the BLIP co-crystal structures with SHV (pink) and SHV D104E (yellow) indicates electrostatic clash between D104 in SHV and

E73 of BLIP (blue) is relieved somewhat when aspartate is mutated to glutamate in

SHV. Figure courtesy of Melinda Hanes, University of California-San Diego

144

Figure 5-6 The inhibition of KPC-2 by BLIP. The ultra high affinity (low picomolar) is beyond the sensitivity of our spectrophotometer, resulting in poor curve fitting.

145

10 20 30 40 50 | | | | | SHV-1 ----MRYIRLCIISLLATLPLAVHASPQPLEQIKLSESQLSGRVGMIEMDLASGRTLTAW TEM-1 --MSIQHFRVALIPFFAAFCLPVFAHPETLVKVKDAEDQLGARVGYIELDLNSGKILESF KPC-2 MSLYRRLVLLSCLSWPLAGFSATALTNLVAEPFAKLEQDFGGSIGVYAMDTGSGATV-SY

70 80 90 100 110 | | | | | SHV-1 RADERFPMMSTFKVVLCGAVLARVDAGDEQLERKIHYRQQDLVDYSPVSEKHLADGMTVG TEM-1 RPEERFPMMSTFKVLLCGAVLSRVDAGQEQLGRRIHYSQNDLVEYSPVTEKHLTDGMTVR KPC-2 RAEERFPLCSSFKGFLAAAVLARSQQQAGLLDTPIRYGKNALVPWSPISEKYLTTGMTVA

130 140 150 160 170 | | | | | SHV-1 ELCAAAITMSDNSAANLLLATVGGPAGLTAFLRQIGDNVTRLDRWETELNEALPGDARDT TEM-1 ELCSAAITMSDNTAANLLLTTIGGPKELTAFLHNMGDHVTRLDRWEPELNEAIPNDERDT KPC-2 ELSAAAVQYSDNAAANLLLKELGGPAGLTAFMRSIGDTTFRLDRWELELNSAIPGDARDT

190 200 210 220 230 | | | | | SHV-1 TTPASMAATLRKLLTSQRLSARSQRQLLQWMVDDRVAGPLIRSVLPAGWFIADKTGA-GE TEM-1 TMPAAMATTLRKLLTGELLTLASRQQLIDWMEADKVAGPLLRSALPAGWFIADKSGA-GE KPC-2 SSPRAVTESLQKLTLGSALAAPQRQQFVDWLKGNTTGNHRIRAAVPADWAVGDKTGTCGV

250 260 270 280 290 | | | | | SHV-1 RGARGIVALLGPNNKAERIVVIYLRDTPASMAERNQQIAGIGAALIEHWQR--- TEM-1 RGSRGIIAALGPDGKPSRIVVIYTTGSQATMDERNRQIAEIGASLIKHW----- KPC-2 YGTANDYAVVWPTGRAPIVLAVYTRAPNKDDKHSEAVIAAAARLALEGLGVNGQ

Figure 5-7 Sequence alignment of KPC-2 with SHV-1 and TEM-1. Residues conserved across all three enzymes are colored red. Green indicates residues conserved between 2 of 3 enzymes. Blue indicates conservation of amino acid character. Start sites of the mature protein are bolded

146

Chapter 6

Concluding Rem arks

I have discussed herein that resistance to β-lactam antibiotics and inhibitors is rapidly developing as a result of single amino acid substitutions and the facilitated spread of β-lactamase genes on plasmids and integrons. The use of this ubiquitous class of antimicrobials can be saved by the development of novel β- lactamase inhibitors. However, discovery and testing is complicated by the expansion of β-lactamases. It is no longer suitable to test a novel β-lactam against the WT enzymes of several different β-lactamase families. The threat of inhibitor resistance looms large.

One aspect that complicates drug design is subtle differences in active site chemistry between highly similar enzymes. Here, we have shown that the TEM and SHV families of β-lactamases attain drug resistance through different mechanisms. When Arg244 is mutated, both SHV-1 and TEM-1 demonstrate resistance to clavulanate. However, in TEM the basis for this resistance is mechanistic, likely through a loss of interaction with a key catalytic water molecule, while in SHV-1 the mechanism of resistance is likely due to lack of direct hydrogen binding interactions with Arg244 in the Michaelis complex. It is suspected that the different topologies of these active sites allow for differential recruitment of water molecules, and a distinctive resistance phenotype. This becomes significant when inhibitors are designed with the purpose of catalytic residue or water replacement.

147

One example of such a design is the penem inhibitors studied in Chapter 3.

It is suspected that the stem of the R1 side chain of penems displaces the catalytic water molecule necessary for hydrolysis of β-lactam compounds from the β- lactam ase. This displacement could be a major basis for the extremely low turnover that we noted in this study. SHV-1 and TEM-1 have conserved catalytic water molecules coordinated by Glu166. However, there have recently been conflicting studies on the presence of this water molecule in the inhibitor resistant

Ser130Gly mutants of TEM-1 and SHV-1. Crystal structures clearly indicate the loss of this water molecule in the SHV Ser130Gly structure, but not in the equivalent TEM structure (Sun et al., 2004; Thomas et al., 2005). Therefore, the inhibition of SHV Ser130Gly by the penem inhibitors may not be as effective, due to the adaptation of hydrolysis through a mechanism non-reliant on the catalytic water, or from recruitment of an addition water elsewhere in the active site. Further studies are ongoing in our laboratory to answer this question. However, now that we have discovered a second location in the active site pocket that shows dissimilar phenotypes with equivalent substitutions, it is a strong indicator that as these enzymes continue to evolve, magic bullet inhibitors will become harder and harder to find.

Discovering these subtleties in β-lactamase enzyme function can be extremely complicated. Many substrates and inhibitors undergo multistep reaction cycles in the active site. With the inhibitors, the exact intermediate important for inhibition could be one of a dozen candidates in the pathway (i.e. the Michaelis-

Menton complex, the acyl-enzyme, the enamine, the imine, or even the cross- 148

linked enzyme species). Resistance or susceptibility could be defined by a slight bias towards an intermediate, like the trans-enamine, that is more or less stable to hydrolysis. Complicating this is the contrasting information obtained with the various methods utilized to study multi-step breakdown: mass spectrometry,

Raman crystallography, and x-ray crystallography.

Mass spectrometry, used heavily in this study, allows for analysis in a natural, fluid environment. In addition, the data are easy to interpret and can resolve an indefinite number of adducts. However, the output is molecular weight, absent of structural information. With Raman crystallography, conclusions about structural properties can be drawn from changes in vibrational spectra upon addition of inhibitor. Another benefit is that reactions can be followed through time.

However, the reactions are performed in a rigid crystal and analysis is complex.

Finally, with X-ray crystallography, a precise structural picture can be obtained, including the location of catalytic waters, but the location of all the atoms in the inhibitor is rarely resolvable due to the presence of multiple species within a crystal.

With those caveats, our analysis of clavulanate and sulbactam with mass spectrometry have always shown multiple species, with a predominance of the lower m olecular weight adducts (Figure 1-7GÆI) This contrasts with Raman and

X-ray data that concluded that the important species for inactivation is the trans- enamine intermediate (Figure 1-7D). Interestingly, this is an adduct that we are unable to detect with mass spectrometry. This discrepancy could be due to two different phenomenon. Ramen spectroscopy may detect earlier intermediates due 149

to the inflexibility of the enzymes in the crystal. On the other hand, the low pH environment of mass spectroscopy might artificially push the reaction to favor the smaller, more hydrolyzed adducts. In view of these observations, careful analysis of resistance defining substitutions needs to be completed with a wide variety of techniques. Currently our laboratory, along with Dr. Marion Helfand, is exploring the inhibition of Arg244 mutants using Raman spectroscopy to gain additional insight into this problem. The determination of the true inhibitory species could be of great help in the design of tighter binding inhibitors.

Analysis is also problematic for substrates that undergo complex reaction chemistry. Many cephalosporins undergo multi-step reactions which include release of the R2 side chains. This makes inferences on affinity difficult based on

Km alone, since many of these substrates have significant k2 rates. In addition, wild type SHV-1 is unable to hydrolyze third generation cephalosporins, so a direct comparison with ESBLs is not possible. For example, determining whether ESBL defining mutations affect primarily affinity versus enhanced k1, k2, or k3 rates is

impossible without a carefully chosen reference. The boronic acid transition state

inhibito rs fill an important niche in this regard. Designed with any number of R1 or

R2 side chains, these analogs can be used as probes for affinity altering

interactions among β-lactamase mutants. In this particular study we observed

great increases in affinity to boronic acid inhibitors with the D104K substitution. In contrast, affinity to the ceftazidime analog was decreased with the ESBL defining

G238S substitution. We believe that this is good evidence that the G238S substitution plays a strong role in turnover, despite the previous hypotheses that 150

the widened active site pocket results in a better “fit” of the oxyimino R1 side chain.

We have also examined the boronates with R244 variants; conclusions here are

still rudimentary; it is unknown why R244 has such an important role in the time

dependence of inactivation of the chiral boronic acid inhibitor. However, the

answer to this may hold a key to why Arg244 affects turnover of inhibitors in such

contradictory ways. Crystallographic studies are needed to guage whether or not a

structural rearrangement of the beta-sheet network occurs as the the inhibitor

binds, explaining the time dependence.

Although these inhibitors are good kinetic probes, they show tremendous

promise in the structural arena. Trapping substrates in the active site is nearly

impossible due to the speed of turnover. In the past this has been achieved by

making deacylation deficient mutants for structural analysis. However, these

structures, usually E166A mutants, not only lack catalytic water molecules, but direct comparisons to WT are complicated by the “hole” this creates in the active site. With the boronic acid inhibitors, we have the ability to design inhibitors based on what we perceive as optimal interactions, trap these in the active sites of catalytically active enzymes, and potentially get accurate pictures of molecular interactions to then optimize in further rounds of inhibitor design.

One aspect of the BATSIs that currently impedes their clinical utility is low affinity for class A β-lactamases. A useful, non-suicide inhibitor must be designed to low nanomolar affinity to be successful in vivo. This is the case with this class of inhibitors, since they are based around boronic acid, a functionality the human body is not adapted to metabolize. In light of these observations, the development 151

of a peptidomimetic is especially attractive. We have shown that the affinity of BLIP for various class A β-lactamases is low nanomolar, and in the case of KPC-2 it is low picomolar. Although there is a very large buried surface area, we have shown that the interaction of a single amino acid can change the affinity 1000-fold. Now that we have structural data on several β-lactamases, intelligently designed inhibitors can be made that capitalize on the crucial active site interactions. The main limitation with BLIP mediated inhibition, however, is that it utilizes many active site residues that are also “hotspots” for substrate modifying mutations. Five residue s shown to be important for BLIP binding have already been altered in clinical strains of class A β-lactamases: Ser130, Arg244, K234, D104, and E240.

This again speaks to the increasing difficulty of inhibiting a group of enzymes that has grown extremely large and diverse; there are currently more than 300 members of the class A β-lactamase family alone (Jacoby, 2006).

Given the inherent problems associated with the development of novel β- lactam antibiotics, as stated above, where does this leave us in the search for novel antimicrobial classes? The main problem in today’s climate is the lack of funding in the public sector and lack of interest in the private sector. Many prominent pharmaceutical companies have ceased antibiotic development and replace d the divisions with more lucrative pursuits. A Pharmaceutical company stands to gain very little by developing a highly effective antibiotic; smart prescribing practices would save these agents for highly resistant pathogens when front line agents have already failed. By the time an antimicrobial becomes conventionally prescribed, the patent on the compound has either expired, or has 152

neared its end. In contrast, a weight-loss drug may gross millions, if not billions of

dollars in its first year on the market. Until it becomes profitable to develop antibiotics, this downward trend in compound discovery will continue.

There have been a few success stories, however. One example was the

identification, by Merck, of a novel antibiotic class from a large natural product

screen. The compound, Platensimycin, originated from a soil sample collected in

South Africa. It was shown to be a strong inhibitor of Gram-positive bacteria by inhibiting a pathway of cellular lipid biosynthesis (Wang et al., 2006). Because this

pathway has not been targeted with prior antibiotics, no cross resistance has been

identified. It will be very interesting to follow the future of this compound. Discovery

is the “cheap” part of development of a drug, especially with the wide availability of sizeable compound libraries. In order to bring this to market, billions of dollars must be invested in clinical trials.

Until pharmaceutical companies have successfully marketed several new classes of antibiotics that can outperform the β-lactams, both in potency and toxicity, we will continue to develop β-lactamase inhibitors in an attempt to save the use of the dozens of β-lactams on the market. Our most promising lead compound has, without question, been penem 1. We have tested this compound not only with the wild type and inhibitor resistant class A β-lactamases, as detailed in this study, but have also shown this to be an effective inhibitor of class C and class D β-lactamases. This is an incredible feat given the genetic diversity of this set of enzymes. In the future, our laboratory hopes to begin animal studies on the

153

safety and efficacy of these compounds, and Wyeth will continue to optimize

chemical structures.

Taken together, the studies mentioned herein serve as an introduction to

the co mplex clinical and scientific questions that still need to be answered. There

are still many challenges ahead…

154

BIBLIOGRAPHY

Alba J, Ishii Y, Thomson K, Moland ES and Yamaguchi K (2005) Kinetics study of

KPC-3, a plasmid-encoded class A carbapenem-hydrolyzing beta-

lactamase. Antimicrob Agents Chemother 49(11):4760-4762.

Ambler RP, Coulson AF, Frere JM, Ghuysen JM, Joris B, Forsman M, Levesque

RC, Tiraby G and Waley SG (1991) A standard numbering scheme for the

class A beta-lactamases. Biochem J 276 ( Pt 1):269-270.

Belaaouaj A, Lapoumeroulie C, Canica MM, Vedel G, Nevot P, Krishnamoorthy R

and Paul G (1994) Nucleotide sequences of the genes coding for the TEM-

like beta-lactamases IRT-1 and IRT-2 (formerly called TRI-1 and TRI-2).

FEMS Microbiol Lett 120(1-2):75-80.

Bermudes H, Jude F, Chaibi EB, Arpin C, Bebear C, Labia R and Quentin C

(1999) Molecular characterization of TEM-59 (IRT-17), a novel inhibitor-

resistant TEM-derived beta-lactamase in a clinical isolate of Klebsiella

oxytoca. Antimicrob Agents Chemother 43(7):1657-1661.

Bethel CR, Hujer AM, Hujer KM, Thomson JM, Ruszczycky MW, Anderson VE,

Pusztai-Carey M, Taracila M, Helfand MS and Bonomo RA (2006)

DEFINING THE ROLE OF ASP104 IN THE SHV b-LACTAMASE.

Antimicrob Agents Chemother. 155

Blazquez J, Baquero MR, Canton R, Alos I and Baquero F (1993) Characterization

of a new TEM-type beta-lactamas e resistant to clavulanate, sulbactam, and

tazobactam in a clinical isolate of Escherichia coli. Antimicrob Agents

Chemother 37(10):2059-2063.

Blazquez J, Morosini MI, Negri MC, Gonzalez-Leiza M and Baquero F (1995)

Single amino acid replacements at positions altered in naturally occurring

extended-spectrum TEM beta-lactamases. Antimicrob Agents Chemother

39(1):145-149.

Bradford PA (2001) Extended-spectrum beta-lactamases in the 21st century:

characterization, epidemiology, and detection of this important resistance

threat. Clin Microbiol Rev 14(4):933-951, table of contents.

Bratu S, Landman D, Haag R, Recco R, Eramo A, Alam M and Quale J (2005)

Rapid spread of carbapenem-resistant Klebsiella pneumoniae in New York

City: a new threat to our antibiotic armamentarium. Arch Intern Med

165(12):1430-1435.

Bret L, Chaibi EB, Chanal-Claris C, Sirot D, Labia R and Sirot J (1997) Inhibitor-

resistant TEM (IRT) beta-lactamases with different substitutions at position

244. Antimicrob Agents Chemother 41(11):2547-2549.

Brown RP, Aplin RT and Schofield CJ (1996) Inhibition of TEM-2 beta-lactamase

from Escherichia coli by clavulanic acid: observation of intermediates by

electrospray ionization mass spectrometry. Biochemistry 35(38):12421-

12432.

156

Bush K (2001) New beta-lactamases in gram-negative bacteria: diversity and

impact on the selection of antimicrobial therapy. Clin Infect Dis 32(7):1085-

1089.

Bush K and Mobashery S (1998) How beta-lactamases have driven

pharmaceutical drug discovery. From mechanistic knowledge to clinical

circumvention. Adv Exp Med Biol 456:71-98.

Buynak JD (2004) The discovery and development of modified penicillin- and

cephalosporin-derived beta-lactamase inhibitors. Curr Med Chem

11(14):1951-1964.

Caselli E, Powers RA, Blasczcak LC, Wu CY, Prati F and Shoichet BK (2001)

Energetic, structural, and antimicrobial analyses of beta-lactam side chain

recognition by beta-lactamases. Chem Biol 8(1):17-31.

Chaibi EB, Peduzzi J, Farzaneh S, Barthelemy M, Sirot D and Labia R (1998)

Clinical inhibitor-resistant mutants of the beta-lactamase TEM-1 at amino-

acid position 69. Kinetic analysis and molecular modelling. Biochim Biophys

Acta 1382(1):38-46.

Chen Y, Minasov G, Roth TA, Prati F and Shoichet BK (2006) The Deacylation

Mechanism of AmpC beta-Lactamase at Ultrahigh Resolution. J Am Chem

Soc 128(9):2970-2976.

Chen Y, Shoichet B and Bonnet R (2005) Structure, function, and inhibition along

the reaction coordinate of CTX-M beta-lactamases. J Am Chem Soc

127(15):5423-5434.

157

Cheng Y and Prusoff WH (1973) Relationship between the inhibition constant (K1)

and the concentration of inhibitor which causes 50 per cent inhibition (I50)

of an enzymatic reaction. Biochem Pharmacol 22(23):3099-3108.

Delaire M, Labia R, Samama JP and Masson JM (1992) Site-directed mutagenesis

at the active site of Escherichia coli TEM-1 beta-lactamase. Suicide

inhibitor-resistant mutants reveal the role of arginine 244 and methionine 69

in catalysis. J Biol Chem 267(29):20600-20606.

DeLano WL (2002) The PyMOL Molecular Grap hics System DeLano Scientific,

San Carlos, CA.

Deshpande LM, Rhomberg PR, Sader HS and Jones RN (2006) Emergence of

serine carbapenemases (KPC and SME) among clinical strains of

Enterobacteriaceae isolated in the United States Medical Centers: report

from the MYSTIC Program (1999-2005). Diagn Microbiol Infect Dis

56(4):367-372.

Diaz N, Sordo TL, Merz Jr KM and Suarez D (2003) Insights into the acylation

mechanism of class A beta-lactamases from molecular dynamics

simulations of the TEM-1 enzyme complexed with benzylpenicillin. J Am

Chem Soc 125(3):672-684.

Doran JL, Leskiw BK, Aippersbach S and Jensen SE (1990) Isolation and

characterization of a beta-lactamase-inhibitory protein from Streptomyces

clavuligerus and cloning and analysis of the corresponding gene. J Bacteriol

172(9):4909-4918.

158

Dubois V, Poirel L, Arpin C, Coulange L, Bebear C, Nordmann P and Quentin C

(2004) SHV-49, a novel inhibitor-resistant beta-lactamase in a clinical

isolate of Klebsiella pneumoniae. Antimicrob Agents Chemother

48(11):4466-4469.

Fisher JF, Meroueh SO and Mobashery S (2005) Bacterial resistance to beta-

lactam antibiotics: compelling opportunism, compelling opportunity. Chem

Rev 105(2):395-424.

Giakkoupi P, Tzelepi E, Legakis NJ and Tzouvelekis LS (1998) Substitution of Arg-

244 by Cys or Ser in SHV-1 and SHV-5 beta-lactamases confers resistance

to mechanism-based inhibitors and reduces catalytic efficiency of the

enzymes. FEMS Microbiol Lett 160(1):49-54.

Hall DG (2005) Boronic Acids: Preparation and Applications in Organic Synthesis

and Medicine. Wiley-VCH, Weinheim.

Helfand MS, Bethel CR, Hujer AM, Hujer KM, Anderson VE and Bonomo RA

(2003a) Understanding resistance to beta-lactams and beta-lactamase

inhibitors in the SHV beta-lactamase: lessons from the mutagenesis of

SER-130. J Biol Chem 278(52):52724-52729.

Helfand MS, Totir MA, Carey MP, Hujer AM, Bonomo RA and Carey PR (2003b)

Following the reactions of mechanism-based inhibitors with beta-lactamase

by Raman crystallography. Biochemistry 42(46):13386-13392.

Holton J and Alber T (2004) Automated protein crystal structure determination

using ELVES. Proc Natl Acad Sci U S A 101(6):1537-1542.

159

Hossain A, Ferraro MJ, Pino RM, Dew RB, 3rd, Moland ES, Lockhart TJ, Thomson

KS, Goering RV and Hanson ND (2004) Plasmid-mediated carbapenem-

hydrolyzing enzyme KPC-2 in an Enterobacter sp. Antimicrob Agents

Chemother 48(11):4438-4440.

Huang W, Beharry Z, Zhang Z and Palzkill T (2003) A broad-spectrum peptide

inhibitor of beta-lactamase identified using phage display and peptide

arrays. Protein Eng 16(11):853-860.

Huang W, Petrosino J and Palzkill T (1998) Display of functional beta-lactamase

inhibitory protein on the surface of M13 bacteriophage. Antimicrob Agents

Chemother 42(11):2893-2897.

Huang W, Zhang Z and Palzkill T (2000) Design of potent beta-lactamase

inhibitors by phage display of beta-lactamase inhibitory protein. J Biol Chem

275(20):14964-14968.

Hujer AM, Hujer KM and Bonomo RA (2001) Mutagenesis of amino acid residues

in the SHV-1 beta-lactamase: the premier role of Gly238Ser in penicillin and

cephalosporin resistance. Biochim Biophys Acta 1547(1):37-50.

Hujer AM, Hujer KM, Helfand MS, Anderson VE and Bonomo RA (2002) Amino

acid substitutions at Ambler position Gly238 in the SHV-1 beta-lactamase:

exploring sequence requirements for resistance to penicillins and

cephalosporins. Antimicrob Agents Chemother 46(12):3971-3977.

Huletsky A, Knox JR and Levesque RC (1993) Role of Ser-238 and Lys-240 in the

hydrolysis of third-generation cephalosporins by SHV-type beta-lactamases

160

probed by site-directed mutagenesis and three-dimensional modeling. J

Biol Chem 268(5):3690-3697.

Imtiaz U, Billings E, Knox JR, Manavathu EK, Lerner SA and Mobashery S (1993)

Inactivation of Class A beta-Lactamases by Clavulanic Acid: The Role of

Arginine-244 in a Proposed Nonconcerted Sequence of Events. J Am

Chem Soc 115:4435-4442.

Imtiaz U, Billings EM, Knox JR and Mobashery S (1994a) A structure-based

analysis of the inhibition of class A beta-lactamases by sulbactam.

Biochemistry 33(19):5728-5738.

Imtiaz U, Manavathu EK, Mobashery S and Lerner SA (1994b) Reversal of

clavulanate resistance conferred by a Ser-244 mutant of TEM-1 beta-

lactamase as a result of a second mutation (Arg to Ser at position 164) that

enhances activity against ceftazidime. Antimicrob Agents Chemother

38(5):1134-1139.

Jacoby J, and Bush, K (2006) http://www.lahey.org/studies/webt.asp,

http://www.lahey.org/studies/webt.asp.

Jelsch C, Mourey L, Masson JM and Samama JP (1993) Crystal structure of

Escherichia coli TEM1 beta-lactamase at 1.8 A resolution. Proteins

16(4):364-383.

Jones TA, and Kjeldgaard, M. (1997) O: The Manual, Version 7.0. Uppsala

Software Factory, Uppsala, Sweden.

161

Ke W, Bethel CR, Thomson JM, Bonomo RA and Akker FV (2007) Crystal

Structure of KPC-2: Insights into Carbapenemase Activity in Class A beta-

Lactamases(,). Biochemistry.

Kleywegt GJ and Jones TA (1994) Detection, delineation, measurement and

display of cavities in macromolecular structures. Acta Crystallogr D Biol

Crystallogr 50(Pt 2):178-185.

Knox JR (1995) Extended-spectrum and inhibitor-resistant TEM-type beta-

lactamases: mutations, specificity, and three-dimensional structure.

Antimicrob Agents Chemother 39(12):2593-2601.

Kuzin AP, Nukaga M, Nukaga Y, Hujer AM, Bonomo RA and Knox JR (1999)

Structure of the SHV-1 beta-lactamase. Biochemistry 38(18):5720-5727.

Lamzin VS and Wilson KS (1993) Automated refinement of protein models. Acta

Crystallogr D Biol Crystallogr 49(Pt 1):129-147.

Lemozy J, Sirot D, Chanal C, Huc C, Labia R, Dabernat H and Sirot J (1995) First

characterization of inhibitor-resistant TEM (IRT) beta-lactamases in

Klebsiella pneumoniae strains. Antimicrob Agents Chemother 39(11):2580-

2582.

Leslie A and Janisch S (1992) Casing the joint. Health Serv J 102(5297):suppl 6,

Lin S, Thomas M, Mark S, Anderson V and Bonomo RA (1999) OHIO-1 beta-

lactamase mutants: the Arg244Ser mutant and resistance to beta-lactams

and beta-lactamase inhibitors. Biochim Biophys Acta 1432(1):125-136.

162

Lin S, Thomas M, Shlaes DM, Rudin SD, Knox JR, Anderson V and Bonomo RA

(1998) Kinetic analysis of an inhibitor-resistant variant of the OHIO-1 beta-

lactamase, an SHV-family class A enzyme. Biochem J 333 ( Pt 2):395-400.

Lomaestro BM, Tobin EH, Shang W and Gootz T (2006) The spread of Klebsiella

pneumoniae carbapenemase-producing K. pneumoniae to upstate New

York. Clin Infect Dis 43(3):e26-28.

MacDowell AA, Celestre RS, Howells M, McKinney W, Krupnick J, Cambie D,

Domning EE, Duarte RM, Kelez N, Plate DW, Cork CW, Earnest TN,

Dickert J, Meigs G, Ralston C, Holton JM, Alber T, Berger JM, Agard DA

and Padmore HA (2004) Suite of three protein crystallography beamlines

with single superconducting bend magnet as the source. J Synchrotron

Radiat 11(Pt 6):447-455.

Maveyraud L, L. Mourey, L. P. Kotra, J. Pedelacq, V. Guillet, S. Mobashery, and J.

Samama (1998) Structural basis for clinical longevity of carbapenem

antibiotics in the face of challenge by the common class A beta-lactamases

from the antibiotic-resistant bacteria. J Am Chem Soc 120:9748-9752.

Meroueh SO, Minasov G, Lee W, Shoichet BK and Mobashery S (2003) Structural

aspects for evolution of beta-lactamases from penicillin-binding proteins. J

Am Chem Soc 125(32):9612-9618.

Miro E, Navarro F, Mirelis B, Sabate M, Rivera A, Coll P and Prats G (2002)

Prevalence of clinical isolates of Escherichia coli producing inhibitor-

resistant beta-lactamases at a University Hospital in Barcelona, Spain, over

a 3-year period. Antimicrob Agents Chemother 46(12):3991-3994. 163

Morandi F, Caselli E, Morandi S, Focia PJ, Blazquez J, Shoichet BK and Prati F

(2003) Nanomolar inhibitors of AmpC beta-lactamase. J Am Chem Soc

125(3):685-695.

Nukaga M, Abe T, Venkatesan AM, Mansour TS, Bonomo RA and Knox JR

(2003a) Inhibition of class A and class C beta-lactamases by penems:

crystallographic structures of a novel 1,4-thiazepine intermediate.

Biochemistry 42(45):13152-13159.

Nukaga M, Mayama K, Hujer AM, Bonomo RA and Knox JR (2003b) Ultrahigh

resolution structure of a class A beta-lactamase: on the mechanism and

specificity of the extended-spectrum SHV-2 enzyme. J Mol Biol 328(1):289-

301.

O'Callaghan CH, Morris A, Kirby SM and Shingler AH (1972) Novel method for

detection of beta-lactamases by using a chromogenic cephalosporin

substrate. Antimicrob Agents Chemother 1(4):283-288.

Orencia MC, Yoon JS, Ness JE, Stemmer WP and Stevens RC (2001) Predicting

the emergence of antibiotic resistance by directed evolution and structural

analysis. Nat Struct Biol 8(3):238-242.

Otwinowski Z, Borek D, Majewski W and Minor W (2003) Multiparametric scaling

of diffraction intensities. Acta Crystallogr A 59(Pt 3):228-234.

Padayatti PS, Helfand MS, Totir MA, Carey MP, Carey PR, Bonomo RA and van

den Akker F (2005) High resolution crystal structures of the trans-enamine

intermediates formed by sulbactam and clavulanic acid and E166A SHV-1

{beta}-lactamase. J Biol Chem 280(41):34900-34907. 164

Pagan-Rodriguez D, Zhou X, Simmons R, Bethel CR, Hujer AM, Helfand MS, Jin

Z, Guo B, Anderson VE, Ng LM and Bonomo RA (2004) Tazobactam

inactivation of SHV-1 and the inhibitor-resistant Ser130 -->Gly SHV-1 beta-

lactamase: insights into the mechanism of inhibition. J Biol Chem

279(19):19494-19501.

Petit A, Maveyraud L, Lenfant F, Samama JP, Labia R and Masson JM (1995)

Multiple substitutions at position 104 of beta-lactamase TEM-1: assessing

the role of this residue in substrate specificity. Biochem J 305 ( Pt 1):33-40.

Petrosino J, Rudgers G, Gilbert H and Palzkill T (1999) Contributions of aspartate

49 and phenylalanine 142 residues of a tight binding inhibitory protein of

beta-lactamases. J Biol Chem 274(4):2394-2400.

Powers RA, Caselli E, Focia PJ, Prati F and Shoichet BK (2001) Structures of

ceftazidime and its transition-state analogue in complex with AmpC beta-

lactamase: implications for resistance mutations and inhibitor design.

Biochemistry 40(31):9207-9214.

Powers RA and Shoichet BK (2002) Structure-based approach for binding site

identification on AmpC beta-lactamase. J Med Chem 45(15):3222-3234.

Prinarakis EE, Miriagou V, Tzelepi E, Gazouli M and Tzouvelekis LS (1997)

Emergence of an inhibitor-resistant beta-lactamase (SHV-10) derived from

an SHV-5 variant. Antimicrob Agents Chemother 41(4):838-840.

Randegger CC and Hachler H (2001) Amino acid substitutions causing inhibitor

resistance in TEM beta-lactamases compromise the extended-spectrum

165

phenotype in SHV extended-spectrum beta-lactamases. J Antimicrob

Chemother 47(5):547-554.

Raquet X, Lamotte-Brasseur J, Fonze E, Goussard S, Courvalin P and Frere JM

(1994) TEM beta-lactamase mutants hydrolysing third-generation

cephalosporins. A kinetic and molecular modelling analysis. J Mol Biol

244(5):625-639.

Reading C and Cole M (1977) Clavulanic acid: a beta-lactamase-inhiting beta-

lactam from Streptomyces clavuligerus. Antimicrob Agents Chemother

11(5):852-857.

Rice LB, Carias LL, Hujer AM, Bonafede M, Hutton R, Hoyen C and Bonomo RA

(2000) High-level expression of chromosomally encoded SHV-1 beta-

lactamase and an outer membrane protein change confer resistance to

ceftazidime and piperacillin-tazobactam in a clinical isolate of Klebsiella

pneumoniae. Antimicrob Agents Chemother 44(2):362-367.

Rudgers GW and Palzkill T (1999) Identification of residues in beta-lactamase

critical for binding beta-lactamase inhibitory protein. J Biol Chem

274(11):6963-6971.

Shimamura T, Ibuka A, Fushinobu S, Wakagi T, Ishiguro M, Ishii Y and

Matsuzawa H (2002) Acyl-intermediate structures of the extended-spectrum

class A beta-lactamase, Toho-1, in complex with cefotaxime, cephalothin,

and benzylpenicillin. J Biol Chem 277(48):46601-46608.

Smith Moland E, Hanson ND, Herrera VL, Black JA, Lockhart TJ, Hossain A,

Johnson JA, Goering RV and Thomson KS (2003) Plasmid-mediated, 166

carbapenem-hydrolysing beta-lactamase, KPC-2, in Klebsiella pneumoniae

isolates. J Antimicrob Chemother 51(3):711-714.

Sosa-Peinado A, Mustafi D and Makinen MW (2000) Overexpression and

biosynthetic deuterium enrichment of TEM-1 beta-lactamase for structural

characterization by magnetic resonance methods. Protein Expr Purif

19(2):235-245.

Storoni LC, McCoy AJ and Read RJ (2004) Likelihood-enhanced fast rotation

functions. Acta Crystallogr D Biol Crystallogr 60(Pt 3):432-438.

Strynadka NC, Jensen SE, Alzari PM and James MN (1996) A potent new mode

of beta-lactamase inhibition revealed by the 1.7 A X-ray crystallographic

structure of the TEM-1-BLIP complex. Nat Struct Biol 3(3):290-297.

Strynadka NC, Jensen SE, Johns K, Blanchard H, Page M, Matagne A, Frere JM

and James MN (1994) Structural and kinetic characterization of a beta-

lactamase-inhibitor protein. Nature 368(6472):657-660.

Sulton D, Pagan-Rodriguez D, Zhou X, Liu Y, Hujer AM, Bethel CR, Helfand MS,

Thomson JM, Anderson VE, Buynak JD, Ng LM and Bonomo RA (2005)

Clavulanic acid inactivation of SHV-1 and the inhibitor-resistant S130G

SHV-1 beta-lactamase. Insights into the mechanism of inhibition. J Biol

Chem 280(42):35528-35536.

Sun T, Bethel CR, Bonomo RA and Knox JR (2004) Inhibitor-resistant class A

beta-lactamases: consequences of the Ser130-to-Gly mutation seen in Apo

and tazobactam structures of the SHV-1 variant. Biochemistry

43(44):14111-14117. 167

Thomas VL, Golemi-Kotra D, Kim C, Vakulenko SB, Mobashery S and Shoichet

BK (2005) Structural consequences of the inhibitor-resistant Ser130Gly

substitution in TEM beta-lactamase. Biochemistry 44(26):9330-9338.

Thomson JM and Bonomo RA (2005) The threat of antibiotic resistance in Gram-

negative pathogenic bacteria: beta-lactams in peril! Curr Opin Microbiol

8(5):518-524.

Thomson JM, Distler AM, Prati F and Bonomo RA (2006) Probing active site

chemistry in SHV beta-lactamase variants at Ambler position 244.

Understanding unique properties of inhibitor resistance. J Biol Chem

281(36):26734-26744.

Venkatesan AM, Agarwal A, Abe T, Ushirogochi H, Yamamura I, Ado M, Tsuyoshi

T, Dos Santos O, Gu Y, Sum FW, Li Z, Francisco G, Lin YI, Petersen PJ,

Yang Y, Kumagai T, Weiss WJ, Shlaes DM, Knox JR and Mansour TS

(2006) Structure-activity relationship of 6-methylidene penems bearing 6,5

bicyclic heterocycles as broad-spectrum beta-lactamase inhibitors:

evidence for 1,4-thiazepine intermediates with C7 R stereochemistry by

computational methods. J Med Chem 49(15):4623-4637.

Villegas MV, Lolans K, Correa A, Kattan JN, Lopez JA and Quinn JP (2007) First

identification of Pseudomonas aeruginosa isolates producing a KPC-type

carbapenem-hydrolyzing beta-lactamase. Antimicrob Agents Chemother.

Villegas MV, Lolans K, Correa A, Suarez CJ, Lopez JA, Vallejo M and Quinn JP

(2006) First detection of the plasmid-mediated class A carbapenemase

168

KPC-2 in clinical isolates of Klebsiella pneumoniae from South America.

Antimicrob Agents Chemother 50(8):2880-2882.

Wang J, Soisson SM, Young K, Shoop W, Kodali S, Galgoci A, Painter R,

Parthasarathy G, Tang YS, Cummings R, Ha S, Dorso K, Motyl M,

Jayasuriya H, Ondeyka J, Herath K, Zhang C, Hernandez L, Allocco J,

Basilio A, Tormo JR, Genilloud O, Vicente F, Pelaez F, Colwell L, Lee SH,

Michael B, Felcetto T, Gill C, Silver LL, Hermes JD, Bartizal K, Barrett J,

Schmatz D, Becker JW, Cully D and Singh SB (2006) Platensimycin is a

selective FabF inhibitor with potent antibiotic properties. Nature

441(7091):358-361.

Wang X, Minasov G, Blazquez J, Caselli E, Prati F and Shoichet BK (2003)

Recognition and resistance in TEM beta-lactamase. Biochemistry

42(28):8434-8444.

Wang X, Minasov G and Shoichet BK (2002) The structural bases of antibiotic

resistance in the clinically derived mutant beta-lactamases TEM-30, TEM-

32, and TEM-34. J Biol Chem 277(35):32149-32156.

Wei ZQ, Du XX, Yu YS, Shen P, Chen YG and Li LJ (2007) Plasmid-Mediated

KPC-2 in a Klebsiella pneumoniae Isolate from China. Antimicrob Agents

Chemother 51(2):763-765.

Weiss WJ, Petersen PJ, Murphy TM, Tardio L, Yang Y, Bradford PA, Venkatesan

AM, Abe T, Isoda T, Mihira A, Ushirogochi H, Takasake T, Projan S,

O'Connell J and Mansour TS (2004) In vitro and in vivo activities of novel 6-

169

methylidene penems as beta-lactamase inhibitors. Antimicrob Agents

Chemother 48(12):4589-4596.

Winn MD, Isupov MN and Murshudov GN (2001) Use of TLS parameters to model

anisotropic displacements in macromolecular refinement. Acta Crystallogr D

Biol Crystallogr 57(Pt 1):122-133.

Woodford N, Tierno PM, Jr., Young K, Tysall L, Palepou MF, Ward E, Painter RE,

Suber DF, Shungu D, Silver LL, Inglima K, Kornblum J and Livermore DM

(2004) Outbreak of Klebsiella pneumoniae producing a new carbapenem-

hydrolyzing class A beta-lactamase, KPC-3, in a New York Medical Center.

Antimicrob Agents Chemother 48(12):4793-4799.

Yigit H, Queenan AM, Anderson GJ, Domenech-Sanchez A, Biddle JW, Steward

CD, Alberti S, Bush K and Tenover FC (2001) Novel carbapenem-

hydrolyzing beta-lactamase, KPC-1, from a carbapenem-resistant strain of

Klebsiella pneumoniae. Antimicrob Agents Chemother 45(4):1151-1161.

Yigit H, Queenan AM, Rasheed JK, Biddle JW, Domenech-Sanchez A, Alberti S,

Bush K and Tenover FC (2003) Carbapenem-resistant strain of Klebsiella

oxytoca harboring carbapenem-hydrolyzing beta-lactamase KPC-2.

Antimicrob Agents Chemother 47(12):3881-3889.

Zafaralla G, Manavathu EK, Lerner SA and Mobashery S (1992) Elucidation of the

role of arginine-244 in the turnover processes of class A beta-lactamases.

Biochemistry 31(15):3847-3852.

170

Zhang Z and Palzkill T (2003) Determinants of binding affinity and specificity for

the interaction of TEM-1 and SME-1 beta-lactamase with beta-lactamase

inhibitory protein. J Biol Chem 278(46):45706-45712.

Zhang Z and Palzkill T (2004) Dissecting the protein-protein interface between

beta-lactamase inhibitory protein and class A beta-lactamases. J Biol Chem

279(41):42860-42866.

Zhou XY, Bordon F, Sirot D, Kitzis MD and Gutmann L (1994) Emergence of

clinical isolates of Escherichia coli producing TEM-1 derivatives or an OXA-

1 beta-lactamase conferring resistance to beta-lactamase inhibitors.

Antimicrob Agents Chemother 38(5):1085-1089.

171