THE ROLE OF THE DSB SYSTEM IN

ANTIMICROBIAL RESISTANCE

PhD Thesis

Submitted to the Department of Life Sciences, Imperial College London

in partial fulfilment of the requirements for the degree of

Doctor of Philosophy

Supervisors: Dr Despoina Mavridou and Professor Alain Filloux

NIKOL KADEŘÁBKOVÁ MRC CMBI, Imperial College London

September 2020

2

COPYRIGHT DECLARATION AND DECLARATION OF

ORIGINALITY

The copyright of this thesis rests with the author. Unless otherwise indicated, its contents are licensed

under a Creative Commons Attribution-Non Commercial-No Derivatives 4.0 International Licence.

Researchers may copy and redistribute the thesis in any medium or format on the condition that they

credit the author, that they do not use it for commercial purposes and that they do not distribute modified

versions of the work. When reusing or sharing this work, researchers must ensure that the licence terms

are clear to others by naming the licence and linking to the licence text. Researchers must seek

permission from the copyright holder for uses of this work that are not included in this licence or

permitted under UK Copyright Law.

I hereby declare that all work presented in this thesis, unless detailed below or referenced appropriately

in the text, is my own.

Figure 3.3 A, Figure 3.4 A, Figure 3.5, Figure 3.7, Figure 3.9, Figure 5.2 A, and Figure 5.3 A –

Experiments performed by Dr R. Christopher D. Furniss.

Figure 3.11 – Experiments performed by Evgenia Maslova and Dr Ronan R. McCarthy.

The MG1655, MG1655 acrA, MG1655 tolC strains and pSLTS plasmid were from Dr Jessica M. A.

Blair. The pCB112 plasmid was a kind gift from Professor Thomas G. Bernhardt (Harvard Medical

School). The Pseudomonas and Stenotrophomonas clinical isolates were a kind gift of Dr Laurent

Dortet (Bicêtre Hospital, Le Kremlin-Bicêtre). Antibodies against DsbA, AcrA, and TolC were the

from Professor Jonathan R. Beckwith (Harvard Medical School), Dr Felicity Alcock (Newcastle

University) and Professor Vassilis Koronakis (University of Cambridge), respectively.

3

LIST OF PUBLICATIONS ARISING FROM THIS WORK

Furniss, R.C.D.*, Kadeřábková, N.*, Barker, D., Bernal, P., Maslova, E., Antwi, A.A.A., McNeil,

H.E., Pugh, H.L., Dortet, L., Blair, J.M.A., Larrouy-Maumus, G., McCarthy, R.R., Gonzales, D.,

Mavridou, D.A.I., Breaking antimicrobial resistance by disrupting extracytoplasmic protein folding.

eLife (in revision) - included in this thesis as Chapter 10 - APPENDIX II.

*Equally-contributing first authors

Kadeřábková, N., Furniss, R.C.D., Maslova, E., Bernal, P., Filloux, A., Gonzales, D., McCarthy, R.R.,

Mavridou, D.A.I., Breaking species-specific antimicrobial resistance in Gram-negative pathogens by

targeting disulfide bond formation. (manuscript written)

4

ACKNOWLEDGEMENTS

First and foremost, I would like to thank my main supervisor, Dr Despoina Mavridou. I came to you as

green a microbiologist as one could be, thank you for guiding me, letting me learn and develop at my

own pace. This thesis would not have been written without your trust in me, your endless patience, and

your support. I have learnt so much from you both inside and outside the lab over the past three years

and have grown as a scientist and as a person thanks to all your effort and dedication. Thank you.

I would also like to acknowledge my second supervisor Professor Alain Filloux, for his guidance,

expertise, and advice throughout my PhD. Thank you for believing in me at the beginning.

My deepest thank you goes to Dr Chris Furniss for being there all the way, for teaching me, supporting

me, and letting me make my own mistakes. You answered every single little question and then some,

offered advice and listened to my thoughts and ideas. I cannot express enough how lucky I was to have

such a kind and helpful guide, lab partner and friend. Thank you for everything.

The end of my three years would not have been what they were without the help, support, and hard

work of Dr Alex McCarthy. You gave up your lab time, your office and even your amazing scones to

make my final four months as stress free as you could. Thank you.

My PhD would not be possible without the following people:

Dr Diego Gonzales, for expertise, in silico studies and advice.

Dr Patricia Bernal, for experimental advice, expertise, and guidance in the field of genetic

manipulation of Pseudomonas species.

Helen McNeil, Hannah Pugh, and Dr Jessica Blair, for strains and expertise on efflux.

Evgenia Maslova and Dr Ronan McCarthy, for running my Galleria mellonella in vivo models.

Dr Sabrina Slater, for her expertise and magician’s touch with computer software.

5

A special thank you goes to Amanda Antwi and Declan Barker for their friendship and

companionship at various times in the lab and beyond. Last, but not least thank you to all the

members of the CMBI Level 5 and the Filloux group on CMBI Level 1 for friendship, support, and

feedback over the years.

In the life beyond the lab, there are many who deserve a thank you from me, and I would like to give

special mention to the following:

Suja Moore, Anna Cooke, Magdalena Lemanczyk, Sophie Bennet, Fabia Borrmann, Hannah Moore,

Emma Lambe and Stacey Smiley-Carr, our time together has been short but sweet and I feel I have

known you my entire life. Thank you for getting me out of the lab and into the ‘real’ world on a

regular basis.

Dang Quoc Anh, Charlie, I still cannot pronounce your name despite the 13 years of friendship. This

is a heartfelt thank you for your never-ending snarky humour and support – you have been making

me laugh and despair at the same time for over a decade and your spot-on humour was never more

needed than now. Thank you my friend.

Tony Emmerson and Dr Nicola Howarth, you were there to spark and light the fire for my love of

research and this thesis would never had been written had it not been for you believing in me, pushing

me and challenging me right at the start. Thank you.

Alasdair Keith, thank you for your support and understanding over the many years. I would not have

embarked on so many adventures without you. Thank you.

My brother Filip, who let it be known that there is only one ‘right doctor’, your high regard of me is

humbling and I am grateful for your belief in me. Thank you.

My parents, and my grandparents, thank you for being there throughout the years and opening so

many doors at the start. No price was high enough for you in supporting me and words cannot express

how grateful I am for the opportunities you have given me. Thank you.

This thesis is dedicated to Mirka and Vaclav Kaderabkovi for their unwavering faith in me.

6

ABSTRACT

Extensive use of antibiotics in medicine and agriculture has led to increasing emergence of

antimicrobial resistance in bacterial populations. Dwindling resources in the discovery of novel active

compound leads and the increasing demands for safety and efficacy of new drugs mean that we are now

faced with treatment failures due to multi-drug resistant pathogens. In the quest for new targets that will

enable us to counter antibiotic resistance, it is often ignored that many resistance mechanisms precede

the clinical use of antibiotics. Instead, the ability to adapt, survive and bypass the toxicity of many

chemical compounds is wired within the bacterial genome. Continuous inter-strain and inter-species

competition have given microorganisms tools to thrive under conditions of chemical warfare.

Recognising this is important when characterising mechanisms underpinning bacterial antimicrobial

resistance, as it can lead to novel strategies that can help us bypass it.

The work described here explores the connection between the disulfide bond formation system, a key

oxidative protein folding pathway in the cell envelope of Gram-negative bacteria, and two widespread

antimicrobial resistance mechanisms, -lactamase catalysed hydrolysis of -lactam antibiotics and

efflux-mediated drug expulsion. It is demonstrated that oxidative-protein-folding-mediated proteostasis

is crucial for both resistance mechanisms, and its inhibition can sensitise multidrug-resistant pathogens

to existing antibiotics. Preliminary results from an experimental evolution approach, set the scene for

future exploration of the importance of disulfide linkages for the capacity of -lactamase enzymes to

evolve under selective pressure. Together, these findings aim to address the mechanistic basis of a new

avenue for antibiotic adjuvant therapy, whereby targeting a non-essential process would allow us to

potentiate existing antibiotics towards previously resistant bacterial strains. With novel essential targets

against bacteria being scarce, adjuvant approaches like this one could prolong the use and efficacy of

existing drugs against some of the most resistant Gram-negative pathogens.

7

TABLE OF CONTENTS

Copyright Declaration and Declaration of Originality 2

List of Publications Arising From This Work 3

Acknowledgements 4

Abstract 6

Table of Contents 7

List of Figures and Tables 11

List of Abbreviations 15

1 Introduction 17

1.1 Bacteria, the Causative Agents of Disease 18

1.2 Brief Overview of Antibiotic Development 20

1.3 Antimicrobial Resistance and Its Spread in Bacterial Communities 21

1.3.1 Intrinsic resistance 21

1.3.2 Acquired resistance 28

1.3.3 Foreign DNA acquisition 28

1.3.4 Mutational resistance 34

1.4 Overcoming Antibiotic Resistance By Using Antibiotic Adjuvants 36

1.4.1 Inhibition of antibiotic modification 36

1.4.2 Membrane permeabilising compounds 38

1.4.3 Inhibition of drug efflux 40

1.4.4 Inhibiting the spread of antimicrobial genes 41

1.5 The Bacterial Cell Envelope 44

1.5.1 Protein folding and transport into the cell envelope 45

1.5.2 Protein transport across the inner membrane of Gram-negative bacteria 46

1.6 Oxidative Protein Folding Pathways 49

1.6.1 Disulfide bond formation in eukaryotes 49

1.6.2 Disulfide bond formation in prokaryotes 53

1.6.3 Polymorphisms of the Gram-negative DSB system 74

1.6.4 Gram-positive bacteria 75

1.6.5 Targeting bacterial pathogenicity through inhibition of the DSB system and oxidative

folding 76

1.7 Aims of This Work 81

2 Materials and methods 82

8

2.1 Reagents and Bacterial Growth Conditions 82

2.2 Genetic Manipulation Techniques 82

2.2.1 Genomic DNA extraction, purification of plasmid DNA and PCR products of genes 82

2.2.2 PCR amplification 83

2.2.3 Agarose gel electrophoresis 84

2.2.6 Restriction digestion 84

2.2.7 Ligation 85

2.2.4 Site-directed mutagenesis 85

2.2.5 DNA sequencing 86

2.2.8 Preparation and transformation of chemically competent cells 86

2.2.9 Preparation and transformation of electrocompetent cells 87

2.3 Bacterial Strains and Plasmids 87

2.3.1 Cloning of -lactamase genes 93

2.3.2 Generation of E. coli dsbA, degP and marR mutants 93

2.3.3 Generation of P. aeruginosa dsbA1 mutants 94

2.3.4 Generation of S. maltophilia dsbA1 dsbL1 mutant 94

2.3.5 Triparental conjugation of P. aeruginosa and S. maltophilia 95

2.3.6 Complementation of E. coli MG1655 dsbA 95

2.4 Minimum Inhibitory Concentration (MIC) Assays 96

2.5 SDS-PAGE Analysis and Immunoblotting 97

2.6 -lactam Hydrolysis Assay 98

2.7 NPN Uptake Assay 99

2.8 PI Uptake Assay 99

2.9 CPRG Cell Envelope Integrity Assay 99

2.10 Motility Assay 100

2.11 AMS labelling 100

2.12 Bacterial Growth Assay – dsbA Mutant 100

2.13 Bacterial Growth Assay – DSB System Chemical Inhibitor 101

2.14 In Vivo Clearance Assay 101

2.15 Statistical Analysis of Experimental Data 102

3 The Importance of Disulfide Bond Formation for the Function of Mobile Class D -

lactamases Enzymes of Pseudomonas aeruginosa 103

3.1 Introduction 103

3.2 Results 106

9

3.2.1 Deletion of dsbA substantially decreases -lactamase mediated antibiotic resistance in E.

coli MC1000 106

3.2.2 Deletion of dsbA does not affect the integrity of the cell envelope in E. coli MC1000 109

3.2.3 Deletion of dsbA deletion does not affect the viability of E. coli MC1000 111

3.2.4 Class D -lactamases misfold in absence of DsbA 112

3.2.5 DsbA is a tractable target for class D -lactamases 114

3.3 Discussion 122

4 The Importance of Disulfide Bond Formation for the Function of Chromosomally-resident

-lactamase enzymes 124

4.1 Introduction 124

4.2 Results 127

4.2.1 The activity of cysteine-containing chromosomally encoded β-lactamase enzymes is

dependent on DsbA 127

4.2.2 Chromosomally encoded β-lactamase enzymes degrade or misfold in the absence of DsbA

130

4.2.3 Deletion of dsbA1 compromises the function of the intrinsic β-lactamase OXA-50 in P.

aeruginosa laboratory strains and clinical isolates 132

4.2.4 Deletion of dsbA1 results in sensitization of P. aeruginosa clinical isolates to existing β-

lactam antibiotics 134

4.2.5 Deletion of dsbA1 and dsbL1 results in increased sensitivity of a S. maltophilia clinical

isolate to ceftazidime 136

4.3 Discussion 138

5 The Importance of Disulfide Bond Formation for the Function of Resistance-Nodulation-

Division Efflux Pumps 140

5.1 Introduction 140

5.2 Results 144

5.2.1 Deletion of dsbA in E. coli MG1655 does not affect the outer or the inner membrane

permeability 144

5.2.2 Deletion of dsbA in E. coli MG1655 causes only minor decreases in bacterial viability 146

5.2.3 RND efflux pump function is compromised in the absence of DsbA 147

5.2.4 Compromised function of RND efflux pumps is due to altered periplasmic proteostasis

148

5.2.5 DsbA as a tractable target for RND efflux pumps 150

5.3 Discussion 152

10

6 The Importance Of Disulfide Bond Formation For The Expansion Of The Hydrolytic

Spectrum Of -Lactamase Enzymes 153

6.1 Introduction 153

6.2 Experimental Design 157

6.2.1 Deletion of dsbA does not severely impact the resistance to b-lactams conferred by the

narrow-spectrum b-lactamases SHV-1 and TEM-1 157

6.2.2 Setup of the experimental evolution experiment 158

6.3 SHV-1 Pilot Study Results 163

6.3.1 Absence of DsbA decreases the potential for evolution of antibiotic resistance to

ceftazidime upon exposure to increasing antibiotic concentrations 163

6.3.2 Characterisation of evolved SHV-1 expressing strains 164

6.3.3 Increase of the hydrolytic spectrum of SHV-1 does not affect bacterial fitness 167

6.4 Discussion 169

7 Discussion and Future Work 171

8 References 176

9 APPENDIX I 208

10 APPENDIX II 217

11

LIST OF FIGURES AND TABLES

Table 1 Comparison of the features of the bacterial cell-envelope of the Gram-positive and Gram-

negative bacterial species. 19

Figure 1.1 Examples of the five main efflux pump superfamilies. 24

Figure 1.2 Schematic of cell wall synthesis and the role of -lactam antibiotics in the inhibition of PBP-

catalysed peptidoglycan cross-linking. 25

Figure 1.3 Clinically available -lactamase inhibitors target class A, C and D b-lactamases. 37

Figure 1.4 The first identified oxidative folding catalyst PDI1 carries a classical Trx fold with a

conserved disulfide bond in a CXXC motif, a common feature in all thiol-redox enzymes.

51

Figure 1.5 The oxidative pathway of the disulfide bond formation system. 54

Figure 1.6 Crystal structure of the primary oxidase of the E. coli DSB system, EcDsbA. 55

Figure 1.7 Differential binding of DsbA to substrate peptide or DsbB is directed by the hydrophobic

surfaces surrounding the active site of DsbA and the histidine residue, His32. 57

Figure 1.8 Crystal structure of EcDsbB, the membrane partner protein of DsbA. 61

Figure 1.9 The proposed mechanism of DsbA oxidation by DsbB. 62

Figure 1.10 The isomerase pathway of the DSB system. 65

Figure 1.11 Crystal structure of the primary isomerase of the E. coli DSB system, EcDsbC. 66

Figure 1.12 Crystal structure of E. coli DsbG, EcDsbG, rendered in cartoon representation. 70

Figure 1.13 Crystal structure of the reduced state of the N-terminal domain of E. coli, nDsbD, rendered

in cartoon representation. 71

Figure 1.14 Crystal structures of EcDsbC - nDsbD and cDsbD - nDsbD complexes elucidate the

mechanism behind the transfer of reductive potential through the periplasmic subunits of

DsbD. 72

Figure 1.15 Fragment based screening identified the first active inhibitors of the DSB system. 77

Figure 1.16 Peptidomimetic library screening identified EcDsbA inhibitors. 79

Figure 1.17 Inhibitors of the DsbA partner proteins, DsbB and VKOR. 80

Table 2 Bacterial strains used in this thesis. 87

Table 3 Plasmids used in this thesis. 89

Table 4 Oligonucleotide primers used in this thesis. 91

Figure 3.1 Antimicrobial resistance mediated by OXA-type -lactamases depends on disulfide bond

formation. 107

12

Figure 3.2 Complementation of dsbA restores the β-lactam MIC values for E. coli MC1000 expressing

class D β-lactamases. 108

Figure 3.3 Deletion of dsbA has no effect on outer membrane permeability in E. coli MC1000. 110

Figure 3.4 Deletion of dsbA does not result in damage to the bacterial inner membrane cell envelope.

111

Figure 3.5 Deletion of dsbA does not have drastic effects on the growth of E. coli MC1000. 112

Figure 3.6 Class D -lactamase enzyme levels remain unaffected by the absence of DsbA. 113

Table 5 The hydrolytic activities of tested β-lactamase enzymes are significantly decreased in the

absence of DsbA. 114

Figure 3.7 Chemical inhibition of the DSB system impedes DsbA re-oxidation and flagellar motility

in E. coli MC1000. 116

Figure 3.8 Chemical inhibition of the DSB system phenocopies the β-lactam MIC changes observed

using E. coli MC1000 dsbA mutant. 117

Figure 3.9 Chemical inhibition of the DSB system has no effect on the growth of E. coli MC1000.

118

Table 6 Chemical inhibition of the DSB system via DsbB shows no effects on multidrug-resistant

P. aeruginosa clinical isolates. 119

Figure 3.10 Absence of DsbA1, the principal pseudomonal DsbA analogue, sensitizes multidrug-

resistant clinical P. aeruginosa isolates to first-line and last-resort -lactam antibiotics. 120

Figure 3.11 Absence of DsbA1 from a P. aeruginosa clinical isolate expressing OXA-198 allows it to

be cleared from infected G. mellonella larvae by piperacillin. 121

Figure 4.1 Antimicrobial resistance mediated by chromosomally resident -lactamases depends on

disulfide bond formation. 128

Figure 4.2 Complementation of dsbA restores the β-lactam MIC values for E. coli MC1000 expressing

β-lactamases. 130

Figure 4.3 The majority of tested β-lactamase enzymes become unstable in the absence of DsbA. 131

Table 7 The hydrolytic activities of tested β-lactamase enzymes are significantly decreased in the

absence of DsbA. 132

Figure 4.4 Absence of DsbA1, the principal pseudomonal DsbA analogue, from P. aeruginosa

laboratory strains and clinical isolates expressing OXA-50 results in a two-fold decrease in

their β-lactam MIC values. 133

Figure 4.5 Absence of DsbA1, the principal pseudomonal DsbA analogue, sensitises P. aeruginosa

clinical isolates expressing AIM-1 to penicillins and cephalosporins. 135

13

Figure 4.6 Absence of DsbA1 and of its analogue DsbL1 significantly decreases the MIC of the S.

maltophilia GUE clinical isolate, expressing L2-1 and L1-1, to ceftazidime. 137

Figure 5.1 Structure of the E. coli AcrAB-TolC efflux pump. 141

Figure 5.2 Deletion of dsbA has no effect on the membrane permeability or on the outer membrane

integrity of E. coli MG1655. 144

Figure 5.3 Deletion of dsbA does not result in damage to the bacterial cell envelope. 145

Figure 5.4 Deletion of dsbA causes a small defect in the growth of E. coli MG1655. 146

Figure 5.5 Antimicrobial resistance mediated by a tripartite efflux pump, AcrAB-TolC, of E. coli

MG1655 is affected in absence of DsbA. 147

Figure 5.6 Complementation of dsbA restores efflux-pump substrate MIC values for E. coli MG1655.

148

Figure 5.7 RND efflux pump function is impaired in the absence of DsbA due to accumulation of

unfolded AcrA resulting from insufficient DegP activity. 149

Figure 5.8 Deletion of dsbA sensitizes the efflux-active E. coli MG1655 strain to chloramphenicol.

150

Figure 5.9 Deletion of marR results in increased expression of the AcrAB pump. 151

Table 8 An overview of commonly occurring amino acid substitutions in TEM-1 and SHV-1 -

lactamases that mediate expansion of their hydrolytic activity. 155

Figure 6.1 Antimicrobial activity of narrow-spectrum enzymes, SHV-1 and TEM-1, does not

dependent on DsbA. 158

Table 9 E. coli MC1000 strains constitutively expressing SHV-1 and TEM-1 -lactamases as well

as the single-cysteine variants of these enzymes were plated on ceftazidime containing

plates (0.5-64x MIC). 160

Table 10 Bacterial suspensions resulting from overnight growth of E. coli MC1000 strains

constitutively expressing wild-type SHV-1 and TEM-1 -lactamases. 161

Figure 6.2 Schematic of the experimental evolution method to be used for strains expressing TEM-1

and SHV-1 -lactamases. 162

Table 11 E. coli MC1000 strains expressing SHV-1 and its single-cysteine variant, develop resistance

upon exposure to increasing ceftazidime concentrations. 164

Figure 6.3 Determination of -lactam MIC values (µg/mL) of three isolates of E. coli MC1000 pDM2-

blaSHV-1 obtained during passage II from plates containing 32 g/mL of ceftazidime. 165

14

Figure 6.4 Determination of -lactam MIC values (µg/mL) of three isolates of E. coli MC1000 dsbA

pDM2-blaSHV-1 obtained during passage II from plates containing 32 g/mL of ceftazidime.

166

Figure 6.5 Determination of -lactam MIC values (µg/mL) of three isolates of E. coli MC1000 pDM2-

blaSHV-1 C54A obtained during passage II from plates containing 32 g/mL of ceftazidime.

167

Figure 6.6 The experimental evolution process does not affect the fitness of any of the evolved strains.

168

Supplementary Table 1 Overview of the β-lactamase enzymes investigated in this thesis. 208

Supplementary Table 2 Deletion of dsbA lowers the β-lactam MIC values for E. coli MC1000

expressing diverse β-lactamases. 209

Supplementary Table 3 Chemical inhibition of the DSB system reduces the MIC values of

representative -lactam antibiotics for E. coli MC1000 expressing disulfide-

bond-containing class D β-lactamases in a similar manner to the deletion of

dsbA. 211

Supplementary Table 4 Antibiotic resistance profiles (MIC values in µg/mL) of the clinical isolates

and laboratory strains tested in this study for -lactam compounds. 212

Supplementary Table 5 Antibiotic resistance profiles (MIC values in µg/mL) of the clinical isolates

and laboratory strains tested in this study for a range of commonly used non-

-lactam antibiotics. 213

Supplementary Table 6 Deletion of dsbA does not decrease the β-lactam MIC values for E. coli

MC1000 expressing the narrow-spectrum β-lactamases TEM-1 and SHV-1 at

either 37°C or 42°C. 214

Supplementary Table 7 β-lactam MIC values and MIC fold changes (FC) recorded in evolved and

original backgrounds, after experimental evolution of E. coli MC1000 strains

expressing the narrow-spectrum β-lactamase SHV-1. 215

15

LIST OF ABBREVIATIONS

AAC – aminoglycoside N-acetyltransferase

ABC – ATP-binding cassette efflux pump

AC – amoxicillin

AM – ampicillin

AMR – antimicrobial resistance

AMS – 4-acetamido-4'-maleimidylstilbene-

2,2'-disulfonic acid

ANT – aminoglycoside O-

nucleotidyltransferase

AP – alkaline phosphatase

APH – aminoglycoside O-

phosphotransferase

AT – aztreonam

ATP – adenosine triphosphate

ASST – aryl-sulfate sulfotransferase

BMD – broth microdilution

bp – base pair

cDsbD – C-terminal domain of DsbD

CCCP – carbonyl cyanide m-chlorophenyl

hydrazone

CFU – colony forming unit

CI – ciprofloxacin

CO – colistin

CPRG – chlorophenyl red-β-D-

galactopyranoside

DSB – disulfide bond formation system

DMSO – dimethyl sulfoxide

DNA – deoxyribonucleic acid

DTT – dithiothreitol

EcDsbA/DsbB/DsbC/DsbG – Escherichia

coli DsbA/DsbB/DsbC/DsbG

EDTA – Ethylenediaminetetraacetic acid

EPI – efflux pump inhibitor

ERO1p – ER oxidoreductin 1

Erv1 and Erv2 – protein essential for

respiration and vegetative growth

ER – endoplasmic reticulum

ESBL – extended-spectrum -lactamase

FAD – flavin adenine dinucleotide

FC – fold change

GFP – green fluorescent protein

GM – gentamicin

GTP – guanosine-5'-triphosphate

HEPES – N-2-hydroxyethylpiperazine-N-

ethanesulfonic acid

HGT – horizontal gene transfer

HRP – horseradish peroxidase

Ig – immunoglobulin

IMS – intermembrane space

IPTG – isopropyl β-D-1-

thiogalactopyranoside

IP – imipenem

IR – inverted repeat

IS – insertion sequence

ITS/MISS – intermembrane space targeting

signals

L-Ara4N – 4-amino-4-deoxy-L-arabinose

LB – Lysogeny broth

LPS – lipopolysaccharide

MATE – multidrug and toxic compound

extrusion efflux pump

MCS – multiple cloning site

16

MES – 2-(N-morpholino)ethanesulfonic

acid

MFS – major facilitator superfamily of

efflux pump

MH – Mueller-Hinton

MIC – minimum inhibitory concentration

MOPS – 3-(N-morpholino) propanesulfonic

acid

MQ – menaquinone

mRNA – messenger ribonucleic acid

MtDsbA – Mycobacterium tuberculosis

DsbA

nDsbD – N-terminal domain of DsbD

NmDsbA1/DsbA2/DsbA3 – Neisseria

meningitidis DsbA1/DsbA2/DsbA3

NPN – 1-N-phenylnaphthylamine

NS – narrow-spectrum -lactamase

OD – optical density

ORI – origin of replication

P – periplasmic

PBP – penicillin binding proteins

PBS – phosphate buffered saline

PCR – polymerase chain reaction

PDI – protein disulfide isomerase

PI – propidium iodide

PP – piperacillin

PmDsbA – Proteus mirabilis DsbA

PT – piperacillin/tazobactam combination

QSOX – quiescin sulfhydryl oxidase

RNA – ribonucleic acid

RND – resistance-nodulation-division efflux

pump

RPM – revolutions per minute

SAR – structure-activity relationship

SDS PAGE – sodium dodecyl sulphate-

polyacrylamide gel electrophoresis

Sec – general secretory pathway

SeDsbA – Salmonella enterica serovar

typhimurium DsbA

SMR – small multidrug resistance efflux

pump

SOB – super optimal broth

Tat – twin arginine pathway

TBS – tris-buffered saline

TBS-T – tris-buffered saline – Tween20

TINS – a target-immobilized NMR

screening

TM – transmembrane

tmDsbD – transmembrane domain of DsbD

TS – trimethoprim/sulfamethoxazole

combination

TR – trimethoprim

Trx – thioredoxin

TZ – ceftazidime

UPEC – uropathogenic Escherichia coli

UQ – ubiquinone

VKOR – vitamin K epoxide reductase-like

protein

X-Gal – 5-bromo-4-chloro-3-indolyl--d-

galactopyranoside

XM – cefuroxime

17

1 INTRODUCTION

Part A

Antimicrobial Resistance in Bacteria

18

1.1 BACTERIA, THE CAUSATIVE AGENTS OF DISEASE

Since the discovery of bacteria by Anton van Leeuwenhoek in the late 1600s, human understanding of

infections and disease has evolved significantly. The role of pathogenic microorganisms in disease, as

described by the ‘Germ Theory’, is now seen as common knowledge and is widely accepted by modern

society.1,2 It is thus remarkable to consider that, despite the scientific observations by Leeuwenhoek, it

was not until the mid to late 1800s that the ideological shift from the ‘Miasma theory’ to ‘Germ theory’

begun in Europe, thanks to the work of John Snow, Louis Pasteur and Robert Koch.3 Interestingly,

literature sources show that the first notions of Germ Theory appeared in cultures outside the European

sphere of influence much earlier, dating as far back as ancient Greece, Rome or India, or recently

appearing in accounts of Islamic medicine from the late Middle Ages.2,4,5

Bacteria represent a large fraction of prokaryotic microorganisms that are found in most habitats on

Earth, from temperate soil and water to acidic hot springs.6,7 Their survival, even in the most unlikely

of places, is supported by their fast replication, high gene mutational rates, flexible genomes capable of

quickly incorporating or removing genetic elements, and their ability to form both symbiotic and

parasitic relationships with other species in their vicinity. This remarkable diversity and complexity

mean that, despite decades of intensive research, only a small fraction of bacteria has been characterised

to date.

Bacterial cells show a general lack of intracellular organisation, in comparison to eukaryots, and thus,

structurally, their cytoplasmic composition is fairly uniform. Most distinct differences can be observed

at the membrane and extracellular level, the so called “bacterial cell envelope”. This protective structure

has been intensively studied, using biochemical and biophysical methods, to further understand its

protective effects that allow bacteria to grow and survive in hostile environments. Further, differences

in the envelope can be used to classify bacteria using the Gram stain and morphological appearance

(Table 1). Discovered in late 19th century by Hans Christian Gram, the Gram stain is composed of a

crystal violet stain and a safranin counterstain and it exploits the inherent differences in the bacterial

cell wall compositions to separate the bacterial domain into two distinct groups, the Gram-positive

(stain purple) and Gram-negative (stain red) species (Table 1). Additional classification is then afforded

by cell morphology, with the two most common being the separation between spherical cocci and rod-

shaped bacilli; this gives rise to the commonly used four-group classification system.8 Smaller groups

19

can also be recognised, such as the comma-shaped Vibrio, the spiral-shaped Spirilla, or the tightly-

coiled Spirochetes although their use in bacterial classification is not as prevalent.8

Gram-negative bacteria, such as the Escherichia, Pseudomonas or Neisseria spp, are surrounded by a

well organised three-layered bacterial cell envelope composed of the outer membrane, the cell wall, and

the inner membrane.9 The glycolipid outer membrane, built from lipopolysaccharide (LPS) chains,

protects Gram-negative bacteria against a wide range of toxic compounds; it is this layer that is

recognised by the human immune system.10 Attached to the inner leaflet of the outer membrane is a

thin, cross-linked peptidoglycan layer which forms the rigid cell wall that determines the bacterial

shape, prevents Gram staining and provides physical protection of the cell interior.11 In between this

outer layer and the inner membrane lies the aqueous and highly oxidative periplasmic space. For

bacteria that cause disease, the periplasm is home not only to housekeeping pathways, but also virulence

factors and antimicrobial resistance (AMR) determinants.

Table 1 Comparison of the features of the bacterial cell-envelope of the Gram-positive and Gram-negative bacterial

species.12

Gram-negative bacteria Gram-positive bacteria

Peptidoglycan layer thin (single-layered) thick (multi-layered)

Teichoic acids absent present in many

Periplasmic space present absent

Outer membrane present absent

Lipopolysaccharide (LPS) content high none

Lipid /lipoprotein content

high

(presence of outer

membrane)

Low

(acid-fast bacteria have lipids linked

to peptidoglycan)

Flagellar structure four rings in basal body two rings in basal body

Toxins production primarily endotoxins primarily exotoxins

Resistance to physical disruption low high

Susceptibility to anionic detergents low high

Resistance to sodium azide low high

Resistance to drying low high

Cell wall composition

70-120 Å thick; two layers.

Lipid content is 20-30%

(high), Murein content is 10-

20% (low)

100-120 Å thick; single layer.

Lipid content is low,

Murein content is 70-80% (high).

Intrinsic antibiotic resistance more resistant more susceptible

20

In contrast to Gram-negative species, Gram-positive bacteria, such as Staphylococci or Bacilli, lack an

outer membrane. Instead a thick cell wall composed of many layers of peptidyglycan-techoic/mycolic

acids is anchored onto their cytoplasmic membrane. This permits the entrance of more chemicals,

including the Gram stain, and makes these bacteria more susceptible to several antimicrobial

compounds that Gram-negative species are resistant to.

1.2 BRIEF OVERVIEW OF ANTIBIOTIC DEVELOPMENT

The discovery and successful isolation of the antibiotic compound penicillin in the late 1920s and early

1940s, respectively, has long been considered the major milestone of antibiotic discovery. Historically,

however, the appearance of antibiotic compounds dates much further back, with traces of tetracycline,

a broad-spectrum antibiotic produced by Streptomyces spp., found in human remains as far back as

A.D. 350-550 in line, perhaps, with the first seeds of germ theory described by the ancient Greek,

Roman, and Middle Eastern cultures.2,13,14

Interestingly, the first “modern” hospital use of antibiotics also predates the discovery of penicillin by

over three decades. The extracts of Pyocyanase from Pseudomonas aeruginosa, prepared by Emmerlich

and Löw, were used in 1899 to treat a range of infections.13,15 Even though the practice and the outcomes

of these applications were questionable at best, they sparked interest in the discovery and development

of other antimicrobial compounds.13,15 By the end of the 19th century, Ehrlich and his group had started

to work on large-scale synthesis and screening of hundreds of compounds in an attempt to find a cure

for syphilis.13,16 This was a resounding success; not only did their experiments lead to the identification

of Salvarsan and Neosalvarsan, they also provided the methodological basis for future identification of

other active compounds.13,16 Using a similar systematic approach, the sulfonamide Protonsil was

identified a few decades later.13,17 These antibiotics, in combination with penicillin have formed the

cornerstone of pharmaceutical drug discovery. In the following decades, several additional antibiotic

classes were discovered and introduced to the hospital setting, until progress tapered off in 1960s.18

Despite the major advances in technology in the early 20th century, drug discovery timescales have

remained long and little progress has been made in developing clinically approved novel-target, broad-

spectrum antimicrobials.

Broadly, antibiotic compounds can be separated into bactericidal (killing bacteria) and bacteriostatic

(inhibiting the growth of bacteria). They target a wide range of cellular processes, including, but not

21

limited to, cell-wall synthesis (-lactams, glycopeptides), transcription (quinolones, rifampicin) and

translation (aminoglycosides, macrolides). Although hundreds of active compounds are now available

in clinics, they all belong to a relatively small group of approximately 15 antibiotic classes. Taking into

consideration the ever-growing number of resistance mechanisms, this collection is quickly becoming

insufficient and the treatment of multidrug-resistant infections, for example those caused by P.

aeruginosa, Acinetobacter baumannii or Klebsiella pneumoniae, is becoming increasingly difficult.

Indeed, the World Health Organisation has issued a warning on the rise of multi-drug resistant

pathogens and composed a list of priority ‘ESKAPE’ pathogens for which novel treatments are urgently

required.19,20

1.3 ANTIMICROBIAL RESISTANCE AND ITS SPREAD IN BACTERIAL

COMMUNITIES

Antimicrobial resistance is a multifaceted concept, examples of which can be found for any cellular

process that can be affected by a chemical compound exerting selective pressure. It can arise through

drug-target modifications, decreased cell-wall permeability, increased efflux rate, pathway variation, or

antibiotic inactivation; this list is representative but certainly not comprehensive. Generally, these

resistance mechanisms can be divided into two types, intrinsic and acquired. Intrinsic resistance is

linked to the natural ability of bacteria to survive and thrive in the presence of other organisms that

produce antimicrobial compounds in shared environments, such as soil or water habitats. As expected,

antibiotic-producing species are themselves naturally resistant to the antibiotics they produce,

broadening the pool of resistant organisms.21 By contrast, acquired resistance emerges in previously

susceptible strains under the selective pressure applied by an antimicrobial agent.

1.3.1 Intrinsic resistance

Complex microbial communities, containing bacteria as well as eukaryotic organisms like fungi, are

major sources of antimicrobial compounds. These communities are essentially a cocktail of microbial

species, metabolic by-products, signalling molecules and toxic compounds that form highly complex

environments in which bacteria have co-evolved to survive. Natural fitness advantages required for

bacterial survival in the presence of a specific antibiotic and environmental stressors are termed intrinsic

resistance mechanisms, have developed independently of previous antibiotic exposure without

22

horizontal gene transfer, and are usually present across the entire species.22 Even though these intrinsic

resistance mechanisms have not been evolved in clinical settings, they can contribute to antibiotic

treatment failure, not only because they confer resistance to several opportunistic pathogens but also

because bacteria that harbour such mechanisms can act as reservoirs from which resistance genes can

be mobilised and spread to other species (discussed further in section 1.3.2).21,23 Although this is not a

comprehensive review, the most important of these mechanisms are discussed below.

1.3.1.1 Membrane permeability

The uptake of nutrients such as amino acids, ions, or sugars is essential for bacterial survival. In Gram-

negative species, transfer of these molecules across the outer membrane is mediated by protein

channels, called porins.24 In addition, to allowing important nutrients through, these structures can be

used by some antibiotics, such as -lactams, aminoglycosides or fluoroquinolones, to cross the outer

membrane barrier and access the interior of the cell.24–27 The simplest way for bacteria to protect

themselves from these antimicrobials is to limit their membrane permeability. There are many examples

of this in nature with Pseudomonas spp. offering some of the most impressive ones. P. aeruginosa , an

increasingly critical Gram-negative pathogen, has an intrinsically less permeable outer membrane (up

to 100-fold less permeable than Escherichia coli), which makes it naturally less susceptible to harmful

substances.25 In addition, several Pseudomonas species exhibit deficiency in OprD, a small but specific

porin, which has been shown to be the major route for imipenem uptake; imipenem is a -lactam of

last-resort.28,29 The formation of biofilms, complex structures where the cells are encapsulated by a

robust extracellular matrix, also decreases bacterial permeability as it prevents drugs from reaching the

cell membrane; P. aeruginosa biofilms, especially in cystic fibrosis patients, are known to cause severe

problems during antibiotic treatment.25,30,31

1.3.1.2 Removal of toxic substances

In cases where antibiotics enter the bacterial cell, other mechanisms are in place to ensure survival. One

of the most widespread resistance mechanisms is the expulsion of chemicals from the periplasm and

the cytoplasm through the action of efflux pumps.32 These complex, and often multi-protein, molecular

machines are present in both prokaryotic and eukaryotic organisms and ensure the removal of metabolic

side products and toxic molecules from the cellular interior using either ATP hydrolysis or chemical

gradients for energy.33–35 Their broad substrate scope means that efflux pumps can carry out removal of

23

a wide range of antibiotic classes, which leads to multidrug-resistant phenotypes and can result in

treatment failure.32 Further, overexpression of efflux pumps is commonly observed in pathogenic

bacteria, including P. aeruginosa and Mycobacterium tuberculosis, which further exacerbates antibiotic

resistance.36,37

To date, efflux pumps have been categorised into five main super-families: the ATP-binding cassette

(ABC) family, the major facilitator superfamily (MFS), the multidrug and toxic compound extrusion

(MATE) familyi, the small multidrug resistance (SMR) family and the resistance-nodulation-division

(RND) family (Figure 1.1).24,38 Most efflux pumps are chromosomally encoded; however specialised

machineries have been found on large AMR plasmids, often in conjunction with other AMR

determinants.37,39

The most studied efflux system in bacteria is the tripartite E. coli RND pump AcrAB-TolC, composed

of the periplasmic component AcrA, the cytoplasmic membrane protein AcrB, and the outer membrane

β-barrel TolC; all three components are essential for pump function.49 Its overexpression is commonly

observed as a direct response to antibiotic stress and leads to increased levels of intrinsic resistance.50,51

This prevents the use of entire antibiotic classes, as observed by the major contribution of AcrAB-TolC

to macrolide resistance in E. coli.21,52 Similar phenotypes are observed in other species, such as P.

aeruginosa, which expresses the MexAB-OprM RND pump or K. pneumoniae, which expresses the

AcrAB and KexEF systems.32,53,54 Notably, many organisms encode more than one efflux pump; their

function can be specific, whilst at the same time, redundancy between efflux systems can also be in

play.

i The MATE family of efflux pumps can be subdivided further into DinF, NorM and eukaryotic pump subfamilies that attain the same overall

fold but exhibit great variety of substrate binding sites.157

24

Figure 1.1 Examples of the five main efflux pump superfamilies. (A) The cytoplasmic ABC pump Sav1866 from

Staphylococcus aureus, whose activity is stimulated by doxorubicin and vinblastine (anti-cancer therapeutics).40 PDB code:

2HYD. (B) The tripartite ABC pump MacAB-TolC from E. coli, responsible for resistance to macrolide antibiotics38; it is

composed of a TolC trimer, MacA hexamer and MacB dimer. Removal of MacAB from Salmonella typhimurium leads to

increased susceptibility to oxidative stress and loss of virulence in mice.38,41 PDB code: 5NIK.42 (C) The MFS efflux pump

MdfA from E. coli, shown to remove chloramphenicol.38,43 Another the E. coli MFS pump, EntS, has been linked to secretion

of the siderophore enterobactin.38,43,44 The MFS group is the largest and most diverse family of efflux pumps.38 PDB code:

4ZOW. (D) The MATE pump NorM-VC from Vibrio cholerae, conferring resistance to tigecycline, a last resort antibiotic for

methicillin- and vancomycin-resistant S. aureus.45 PDB code:3MKT. (E) The SMR pump EmrE from E. coli, responsible for

resistance against lipophilic cations.46,47 PDB code: 3B5D. (F) The tripartite RND efflux pump MexAB-OprM from P.

aeruginosa; it is composed of TolC trimer, MexA hexamer and MexB trimer subunits. Its structure is homologous to the model

RND pump AcrAB-TolC of E. coli.24,48 PDB Code: 6TA6. Figure adapted from Du et al.38

1.3.1.3 Inherent pathway variation

Intrinsic resistance mechanisms are commonly due to natural target or pathway variations in some

bacterial species. This abrogates the bacteriostatic or bactericidal effects of drugs despite active

antibiotic uptake. Pathway differentiation resulting in polymyxin B and colistin resistance has been

studied in detail for several species, such as Proteus spp., Serratia spp., and Burkholderia cepacia.7,55–

57 This type of resistance originates from variations in the LPS of the bacterial cell wall of these species.

For example, modification of the lipid A moiety of the LPS through the addition of 4-amino-4-deoxy-

25

L-arabinose (L-Ara4N) results in an overall positively charged cell envelope which leads to decreased

polymyxin binding and antibiotic resistance.57 This modification is a consequence of variation in

specific two-component systems, like PhoP/PhoQ or PmrA/PmrB, that lead to constitutive

overexpression of LPS modifying genes.57–61

In addition to outer membrane modifications, antibiotic resistance at the outer membrane can occur

through expression of diverse penicillin binding proteins (PBPs), membrane associated acyltransferases

that act as catalysts for the final crosslinking reaction in cell wall synthesis (Figure 1.2).62,63 Many

bacterial species express enzymes that introduce functional modifications to PBP substrates. These

variations are reflected in the PBP active site and can lead to abrogation -lactam antibiotic recognition

and binding. For example, the Gram-positive Streptococci express murM and murN, which catalyse the

addition of short dipeptides onto the muropeptide backbone and result in branched peptidoglycan and

penicillin resistance.63–66

Figure 1.2 Schematic of cell wall synthesis and the role of -lactam antibiotics in the inhibition of PBP-catalysed

peptidoglycan cross-linking. The bacterial cell wall is composed of alternating N-acetylmuramic/N-acetyglucosamine sugar

subunits which are cross-linked at the terminal D-alanine residue. The cross-linking reaction is catalysed by penicillin binding

proteins (PBPs) that bind the D-Ala-D-Ala moiety of the nascent NAM subunit and the diamino residue of a neighbouring

polypeptide.63 This residue (represented by ‘X’) can exhibit extensive species-specific variation; with glycine present in S.

aureus, lysine in Streptococcus pneumoniae and a diaminopimelate group in E. coli.63 -lactam antibiotics emulate the D-Ala-

26

D-Ala motif (shown in red) and covalently inhibit PBPs at their active site serine residue. Figure adapted from Zeng & Lin

and Macheboeuf et al.63,72

1.3.1.4 Chemical modification of the key pharmacophore

A large sub-set of antibiotic resistance mechanisms is driven by the binding, break down or modification

of active antibiotic compounds to effectively lower their cellular concentrations. Some of the most

extensively studied examples include the -lactamase or aminoglycoside-modifying enzymes.5 The

former is elaborated upon as an example below.

-lactamase enzymes catalyse the hydrolysis of a strained -lactam peptide bond that is key for the

activity of -lactam antibiotics. These drugs include some of the most commonly used therapeutics due

to their relatively low toxicity levels and activity against both Gram-positive and Gram-negative

species.67 Continuous research has resulted in the development and deployment of numerous -lactam

derivatives that can broadly be separated into four categories, the simplest 1st generation penicillins

(amoxicillin, ampicillin), followed by the increasingly specific cephalosporins (cefuroxime,

ceftazidime), the broad and higher-activity carbapenems (imipenem, meropenem), as well as the

narrow-spectrum monobactam compounds (aztreonam).

Despite their diversity, all -lactam antibiotics disrupt the last step of bacterial cell wall formation by

emulating the highly specific D-Ala-D-Ala moiety of the nascent peptidoglycan chain terminus. This

structural similarity results in competitive binding and interaction with the native PBPs.62 Interaction of

the strained -lactam ring with the catalytic serine of these DD-transpeptidases results in the formation

of an irreversible covalent bond which blocks the entry of native substrates, the nascent N-

acetylmuramic Acid / N-acetylglucosamine peptide subunits.5,68–70 In addition to stopping the synthesis

of new cell wall and thus preventing cellular division, the activity of -lactam antibiotics leads to

accumulation of un-crosslinked precursors prompting increased expression of autolytic hydrolases.5,68–

70 These enzymes degrade both the precursor molecules and the existing peptidoglycan layer, which

enhances the antibiotic effects and lead to further cell swelling and eventual lysis.5,68–70 Notably, the

inclusion of monobactam compounds in the list of classical -lactam antibiotics has recently been called

into question, due to their divergent monocyclic core and evidence suggesting inhibition of a different

enzyme in the cell wall biosynthesis process which, among other things, explains their extensive activity

against P. aeruginosa infections.71

27

Inactivation of -lactam compounds by bacteria poses a threat to many currently available therapies.

Although treatment failure can also arise from species-specific differences in the amounts or affinities

of PBPs, usually resistance to these compounds arises from -lactamase-mediated hydrolysis of the -

lactam moiety. Encoded on the bacterial chromosome, as well as on plasmids and transposable

elements, thousands of -lactamase enzymes have spread rapidly through bacterial populations with

their high mutational rates complicating the development of agents mitigating their activity (more detail

in section 1.3.2).67

-lactamases, first described in the 1940s in Balantidium coli by Fleming as well as Abraham & Chain,

predate the first clinical use of -lactam drugs.13,73 Since then, they have been shown to be produced by

both Gram-positive (secreted into extracellular space) and Gram-negative (excreted into the periplasmic

space) bacteria and their high diversity has given rise to the amino-acid-sequence-based Ambler

classification system (class A-D).67,74 Classes A, C, and D comprise serine-based lactamases which have

a conserved catalytic serine (Ser70) residue in their active site and are characterised by structural

similarity to the naturally-encoded PBPs.67 By contrast, class B lactamases have evolved along a

distinctly different evolutionary pathway from metallo-hydrolase enzymes and depend on the presence

of a catalytically active Zn2+ in their active site.62,67,75,76

Within the Ambler classification system, -lactamases can be further separated by their activity into

narrow-spectrum, extended-spectrum, or carbapenem-hydrolysing enzymes. Narrow-spectrum -

lactamases only hydrolyse penicillins and early generation cephalosporins, while extended-spectrum

enzymes act on most -lactam antibiotics, except for carbapenem drugs. Although most ‘modern’ -

lactamase enzymes are located on mobile genetic elements and can break down even the last-generation

β-lactams, they originate from narrow-spectrum chromosomally-resident species which were acquired

by pathogens through horizontal gene transfer and have mutated into broader-spectrum hydrolases.

Examples of such archetypical chromosomally-resident enzymes include SHV family of enzymes from

K. pneumoniae, capable of hydrolysing penicillin and ampicillin or AmpC from P. aeruginosa which

hydrolyses early-generation cephalosporins.24,77,78

Notably, intrinsic resistance mechanisms are not strictly limited to inherently-resistant species, and

many of them can be attained through mutations (for example porin gene mutations) or acquired via

horizontal gene transfer (discussed further in section 1.3.3). As a matter of fact, the ensemble of intrinsic

resistance mechanisms creates a basis for the multi-factorial way in which antimicrobial resistance can

28

spread and evolve, as can be seen by the multitude of resistance determinants encoded by clinical

isolates of highly-resistant organisms like P. aeruginosa or Stenotrophomonas maltophilia.28,79–81

1.3.2 Acquired resistance

Acquired antimicrobial resistance develops in susceptible strains over time and in response to exposure

to an antimicrobial agent. Although resistance to most antibiotic classes is becoming increasingly

prevalent, the rate of development depends strongly on the target of the antibiotic compound. For

example, resistance to a single-target antibiotic, such as rifampicin, occurs more easily than mechanisms

of resistance to antimicrobial compounds with greater scope. In all cases, however, resistance can arise

through two main routes: a) incorporation of external genetic material from an already-resistant strain,

and b) mutation(s) within the bacterial genome.

1.3.3 Foreign DNA acquisition

Apart from cell division, horizontal gene transfer (HGT) is the most important mechanism for

acquisition of DNA both within and in between bacterial species.82 The majority of DNA acquired

through this route has either no or negative effect on the receiving organism. However, a small

percentage of the acquired material results in gain of beneficial traits, which subsequently become

vertically propagated within a bacterial population.22 Most retained genes drive a simple function, like

the production of a single antimicrobial resistance determinant (for example a -lactamase enzyme),

which may come with its own set of regulatory components to ensure expression.22,83,84 Successful DNA

incorporation events rarely involve central cellular pathways, like transcription or translation, as these

vary substantially between species and any “swaps” would be detrimental to bacterial fitness.83 There

are three main mechanisms driving HGT, which commonly occur in environments where toxic-

compound-producing organisms co-exist with non-producing species: a) transformation, b)

transduction and c) conjugation.

1.3.3.1 Transformation

Natural transformation is a mechanism for external DNA uptake in many bacterial species, although its

contribution to antimicrobial resistance in clinically-relevant bacteria is limited.22 Overall,

29

transformation occurs when a competent receiving strain uptakes DNA from its environment under

normal growth conditions, and subsequently integrates the acquired genetic material into its own

chromosome.83 Unlike Neisseria gonorrhoeae, whose competence appears to be constitutively active,

the majority of bacteria develop competence transiently to achieve this process.22,85 This usually occurs

in response to starvation to specific nutrients, increased cell density or environmental pressures like

temperature.86 This mechanism of HGT is remarkably widespread in the bacterial world, including

Gram-positive (Staphylococcus spp., Streptococcus spp.) as well as Gram-negative (Helicobacter spp.,

Pseudomonas spp.) human pathogens, and it is clear that it contributes greatly to bacterial evolution.22,86

The primary requirement for transformation is the presence of extracellular DNA in the surrounding

environment. Active excretionii, from species such as Acinetobacter calcoaceticus, Bacillus subtilis87,88

or P. aeruginosa87,89 and passive release from decomposing or disrupted cells and virus particles, can

lead to a highly varied mixture of genetic information, ranging from short linear fragments to fully

circularised plasmids. Although the majority of the DNA is likely to be damaged or decomposed by

environmental enzymes, it has been shown that large plasmids and chromosomal DNA can remain

intact for several hours and small plasmids and linear fragments for even longer time periods.90 Non-

covalent binding of DNA to cell surface receptors promotes translocation across the membrane in a

sequence-dependent (Haemophilus influenzae or N. gonorrhoeae) or independent (B. subtilis or S.

pneumoniae) manner.83,86,91 During the DNA import process the double-helix structure of DNA needs

to unwind; only a single strand of DNA is imported.92 This means that plasmid DNA needs to be either

integrated into the chromosome or reconstituted to generate a form that would successfully replicate.

Due to this, more than one copy of plasmid DNA usually needs to be taken up to achieve successful

transformation.83 While plasmid DNA is often self-sufficient, linear fragments of chromosomal DNA

must be incorporated into the chromosome of the recipient cell to confer any function. Mechanisms that

allow this integration will be discussed in more detail in section 1.3.3.4.

ii While active secretion of DNA into the environment may eventually lead to transformation, its role has been implicated in other processes,

such as the formation in biofilms in P. aeruginosa.30

30

1.3.3.2 Transduction

A less common route for DNA acquisition is mediated by phages; these are bacteria-specific viruses

which can be divided into two types, virulent and temperate.82,93,94 The maturation and release of virulent

phages causes host-cell lysis leading to DNA fragmentation.95 Some of this DNA can then be integrated

into the phage head and passed on to a new host during the next infection cycle.93,95,96 This process is

called generalised transduction, reported for example in the case of the P1 phage in E. coli, and results

in a non-specific dissemination of genetic material.95 By contrast, temperate phages insert into the

bacterial chromosome without killing their host. In this case, host DNA incorporation into the virus

occurs due to incorrect excision of the prophage from the bacterial chromosome.93 This results in the

inclusion of surrounding bacterial genes, as observed in phage λ in E. coli which has been shown to

package only genes from the neighbouring galactose or biotin metabolism pathways.97,98

Useful genetic determinants increasing virulence or conferring antimicrobial resistance can be acquired

through transduction. For example, pathogenicity islands in S. aureus have been shown to be excised,

replicated, and integrated into phage particles upon phage infection.93,99 This leads to dissemination to

a range of new host cells. In V. cholerae a key virulence determinant, the cholera toxin, is encoded in

the genome of a CTXΦ phage.100 Transduction has been documented in many environmental settings,

including but not limited to the soil, marine, and fresh water environments, suggesting that it plays an

important role in gene dissemination and evolution.93,98

1.3.3.3 Conjugation

Conjugation is a highly efficient route for DNA transfer occurring at high rates in the human

gastrointestinal tract under antibiotic treatment.22 It is the most prevalent route of HGT responsible for

the spread of antimicrobial resistance and is believed to be the primary mechanism through which

development of hospital-acquired resistance occurs.22,101 In contrast to the previous two modes of HGT,

this mechanism requires two metabolically-active cells to establish cell-to-cell contact and form a

junction that allows them to successfully transfer genetic material from a donor to a receiver cell.83,86

Direct transfer between chromosomes has rarely been reported for bacterial conjugation.102 Instead

mobile genetic elements such as insertion sequences (IS), transposons, genomic islands, plasmids,

integrative and conjugative elements and miniature inverted repeat (IR) transposable elements are

efficiently transferred.22,103–105 The wide range of mobile genetic elements that are transferred through

31

conjugation, allows for gene expression without the need for plasmid reconstruction or homology-based

chromosomal integration in the recipient cells. Some of these mobile genetic elements will be discussed

further below.

Insertion sequences are the simplest and most common examples of mobile genetic elements. Their

inter-species recurrence promotes homology-based matching and efficient insertion into the

chromosome of the receiving cell.83 Classically, they encode their own transposition proteins allowing

them to move easily both within and in between bacterial chromosomes.106 With only minor exceptions,

the open reading frame is flanked by a short terminal IRs that double as transposase binding domains

and strand cleavage/transfer sites.106 The transposase enzyme catalyses the generation of a free plasmid-

like copy containing the IR and the DNA to be transposed by joining together the IS ends and forming

a transient tnpA promoter to increase insertion frequency.83 If two identical IS are present on either side

of a DNA region, mobility of the entire genetic segment is enabled leading to transposon

formation.106,107 This can lead to beneficial or detrimental effects, depending on the final insertion

location and the gene carried. For example, a mobile antibiotic resistance gene cassette can introduce a

survival advantage upon transposition and thus have beneficial effects. On the other hand, insertion of

an IS element into a gene sequence or its promoter region is likely to cause disruptions in protein

expression that may negatively affect bacterial fitness.106 Overall, these mobile genetic elements offer

a convenient route of DNA mobilization. Their effects can be seen in the dissemination of AIM-1, a

newly emerging -lactamase isolated from P. aeruginosa clinical isolates.108 Class B3 -lactamase

enzymes, such as AIM-1, have long been believed to be immobile environmental enzymes. The

characterisation of AIM-1 has demonstrated the presence of ISCR stability elements that have enabled

this enzyme to be mobilised from the genome and transferred between P. aeruginosa strains.108,109 This

case demonstrates how powerful HGT is for the evolution of bacterial populations and the spread of

antimicrobial resistance.

Plasmids are circular mobile genetic elements that are commonly gained through natural transformation

and conjugative transfer and often encode virulence, detoxification or antibiotic resistance genes which

can function autonomouslyiii.110,111 Plasmids transferred via conjugation are acquired fully functional,

bypassing the need for their reconstitution.83 All plasmids contain a replicon region which encodes the

iii Plasmids can also be incorporated into the genome; however, they tend not to be maintained as such for long.110,111

32

origin of replication (ORI) as well as the relevant control elements. The ORI is an AT-rich DNA

sequence from which replication ensues; it determines the copy number, the stability and the

compatibility of the plasmid with its host.110 Despite their ‘independence’, plasmids rely on host

machinery for replication. This means that the resources for replication are limited and any genes

encoded will generally have to be beneficial for both the plasmid and the host in order to be maintained

over time.110,112 Bacterial cells can take up and maintain more than one plasmid at a time if the genes

encoded confer a strong selective advantage to make up for any decrease in fitness caused by their

carriage. However, sharing of the available resources also means that plasmids utilising the same

replication system cannot co-exist in the same cell. Over time, the propagation of one plasmid leads to

the dilution of the other until one is lost from the population.110 Notably, plasmids responsible for

spreading antimicrobial resistance genes often encode their own conjugative machinery, something that

makes their replication and maintenance more robust.101,113

1.3.3.4 DNA integration into the chromosome

The spread of antimicrobial resistance via HGT requires DNA stabilisation (plasmid reconstitution) or

its incorporation into the chromosome of the host (linear fragments of DNA cannot replicate on their

own). This can be achieved in a homology-dependent or independent manner. Homologous

recombination can occur in cases where the donor and the recipient share high sequence identity, as is

the case between members of the same genus or species.83 Studies show that the average maximum

divergence for successful recombination through this mechanism is approximately 25%.83 This route,

unlike many others, ensures that the size of the recombined region remains unchanged and, if integrated

successfully, fully functional.83 In stark contrast, homology-independent recombination, observed for

example in E. coli, occurs via double-strand DNA breaks enabling risky fragment insertions into the

genome.114 As this procedure requires blunt-end joining of DNA fragments it can lead to detrimental

effects due to incorrect recombination. These two mechanisms can also come together in the case of

additive integration. In this instance, two circular DNA molecules, or one circular molecule or linear

fragment and the chromosome integrate together.83,115,116 For this to happen short, high-similarity

regions overlap enabling integration. However, total similarity remains low and leads to new DNA

addition, exchange or even to host DNA deletions. This allows the acquisition and integration of DNA

from phylogenetically distant species.

33

Finally, DNA integration can be achieved by integrons, site-specific recombination systems encoding

an integrase enzyme responsible for incorporation of external DNA into its own binding site.103 An

upstream promoter ensures gene expression, which allows the bacterium to ‘test’ the new DNA for

function.103 In addition, many of these systems carry a LexAiv-binding site nearby, thus allowing the

SOS pathway of the host to influence genome mobility and DNA integration frequency.117 As such,

under stress conditions gene mobility and integration may be upregulated in an attempt to increase the

chance of beneficial gene acquisition. Integrons have classically been divided into two types,

chromosomal ones responsible for a build-up of long-term genome complexity, and mobile ones

associated with integration of plasmids and transposons.118 A gene cassette can be recruited into an

integron of any of the two types; self-mobilisation then makes it dependent on other mobile genetic

elements and associated transposase genesv. Nonetheless, it is the mobile integrons that are mostly

linked to inter-species genome penetration and development of modern antibiotic resistance in

clinically-relevant pathogenic strainsvi.22,118

1.3.3.5 Gene maintenance

The acquisition of a potential resistance gene is only the beginning of an on-going fight for bacterial

survival in the presence of antimicrobial agents. Numerous adaptive changes need to happen in the

recipient cell, and where relevant, in the donor cell to ensure proper gene expression, successful AMR

determinant maturation and future gene propagation. Overall, only a minor proportion of externally

acquired DNA remains actively maintained in the genome across multiple generations, as vertical gene

transfer is strongly dependent on gene stability as well as the net effect associated with the costs and

benefits of its upkeep. Considering this, in the absence of selective pressure large plasmids or integrons

carrying many gene cassettes are less likely to be maintained in the long term.

Though highly adaptable, the ability of bacteria to offset the fitness costs originating from the

acquisition of AMR determinants varies extensively between species. It depends, among other things,

iv LexA is a transcriptional repressor responsible for the repression of SOS genes linked to DNA repair. Its autolytic cleavage leads to the SOS

response, activated among others by the presence of -lactam antibiotics, and to the expression of the integron integrase IntI.117,118

v It can be easily imagined how the integrons’ ability to incorporate DNA can become the springboard for the development of new mobile

elements such as plasmids carrying multiple resistance genes.

vi It has been noted that the ability of integrons to cross horizontally between genomes is somewhat limited to the environmental origin of the

species as specific integron sequences vary between soil and marine environments.118

34

on habitat, nutrient availability at any given timepoint during evolution and even gene expression level

for the AMR gene of interest. While intrinsic resistance mechanisms are likely to be constitutively

active in order to afford protection in the natural environment of the resistant bacterium, acquired genes

may be more prone to stress-mediated induction and might require adaptations to cellular pathways in

order to be expressed.39 Successful development of antimicrobial resistance extends beyond the

genome’s capacity to integrate, stabilise and express the relevant genes. Key additional considerations

include the ability of cells to correctly fold and translocate the synthesised products to appropriate

cellular locations. While exact details of the processes involved may vary between species and gene

products, they will inevitably depend on the same general principles guiding protein synthesis, folding

and transport; these are discussed further in section 1.5.1.

1.3.4 Mutational resistance

The emergence and accumulation of mutational changes in genomes is at the heart of evolution and

natural selection. The high replication rates of most bacterial species allow the generation of a

considerable number of mutations that result in phenotypic changes over a short period of time. In

general, these mutations lead to similar resistance mechanisms as described above for naturally-resistant

bacteria i.e. drug target modifications, decrease in drug uptake through porin mutations, efflux

upregulation or changes to metabolic pathways and their associated regulatory networks.22 Such

adaptations can be caused by single nucleotide and amino acid changes or through more extensive

sequence deletions and insertions.119 This process often generates substitutions which decrease bacterial

fitness under normal conditions, but offer a significant survival advantage under antibiotic pressure. In

this case, clearing of susceptible cells by antibiotics will ensure the propagation of the resistant species,

and maintenance of the given mutation through vertical gene transfer.22 This is particularly relevant in

clinical environments where repetitive treatment with multiple antibiotics can exacerbate the

development of multidrug-resistant strains and lead to poor patient outcomes. Indeed, M. tuberculosis

or P. aeruginosa have been shown to be particularly efficient in developing antibiotic resistance in this

way.119 In such cases, combinatorial therapies are often employed to exploit the different modes of

action of antibiotic compounds in order to decrease the chances of mutations.

In M. tuberculosis, RNA polymerase mutations have been shown to quickly develop during rifampicin

exposure. Discovered in mid-1960s in the soil bacterium Amycolatopsis rifamycinica, rifampicin was

reported as a wide-spectrum antibiotic with strong selectivity for prokaryotic enzymes. The drug targets

35

the prokaryotic DNA-dependent RNA polymerase, RpoB, by binding to its -subunit and sterically

preventing the access of nascent RNA into its binding site, blocking RNA chain extension.120,121 In E.

coli up to 95% of the commonly observed mutations have been shown to be point mutations in the N-

terminus of RpoB, between amino acids 505 and 537, leading to decrease in binding affinity.121–123

Similarly, in M. tuberculosis all mutations occur in nucleotides encoding amino acids 419-451.121,124

Point mutants emerge at high frequencies of up to 10-8 per bacterium per cell division, and their

occurrence leads to the development of varying levels of resistance that is vertically transmitted and

can be enhanced upon further exposure to the antibiotic through accumulation.121,123 Often, the only

option to bypass this development is to use rifampicin, and its derivative compounds, only in

combination therapies, as is done in its clinical use against M. tuberculosis.

Other examples of mutational resistance are common in bacteria. For example, mutations in -

lactamase genes increasing their hydrolytic activity or the aforementioned OprD porin deficiency of P.

aeruginosa are well-reported.24,27 Of particular interest to our work is the OXA family of mobile -

lactamases consisting of over 750 different enzymes. All members reported so far exhibit distinct

hydrolytic profiles originating from a variety of amino acid mutations. Examples include the Ser73Asn

substitution, conferring extended-spectrum activity to enzymes in the OXA-7 family and the Gly157Asp

substitution which leads to high-level of resistance to ceftazidime.125

In summary, bacterial exposure to different stressors and environmental factors results in the activation

of variety of stress-response cascades which drive the differential induction of a wide range of survival

mechanisms.39 These can lead to the development of varying levels of drug resistance, as seen in the

cases of RpoB-mediated rifampicin resistance in M. tuberculosis or the non-enzymatic aminoglycoside

resistance in P. aeruginosa.126 The exact level of resistance to antibiotics and the location of the bacterial

infection will ultimately determine the choice and efficiency of antibiotic therapy, as even resistant

strains may be cleared by a drug that they are resistant to if the bacterial load and the bioavailability

that can be achieved in the infection location of interest are favourable. Susceptibility breakpoints tables

have been developed to correlate these in vivo considerations with the results of in vitro studies to guide

clinicians and scientists in determining the best strategies for antibiotic therapy.22

36

1.4 OVERCOMING ANTIBIOTIC RESISTANCE BY USING ANTIBIOTIC

ADJUVANTS

The development and spread of antimicrobial resistance have been exacerbated by extensive, and

sometimes inappropriate, use of antibiotic compounds in clinical and agricultural settings. Common

antibiotic strategies that target essential bacterial pathways generate strong selective pressure and lead

to rapid emergence of antibiotic-resistant strains.18 Combined with the limited success in the

development of novel antibiotics during the past 60 years, this increase in antimicrobial resistance

creates an urgent need for the discovery of next-generation antibacterial strategies, as well as approaches

that sensitise resistant pathogens to existing antibiotic compounds. With the hope of exerting minimal

selective pressure, an increasingly adopted approach is to target pathways which are not essential for

bacterial survival in the absence of the antibiotic compound of interest. Several examples of such

adjuvant approaches are currently in consideration, whilst some are already in use in clinical therapy.

These can be broadly divided into four types, based on the use of a) resistance enzyme inhibitors, b)

membrane-permeabilising compounds, c) efflux pump inhibitors, and d) horizontal gene transfer

blockers.37,101,127

1.4.1 Inhibition of antibiotic modification

Due to the importance of -lactam antibiotics, routes to disable -lactamase activity have been

extensively investigated, and -lactamase inhibitors form the biggest class of AMR breakers to

date.67,75,76 The development of such compounds is challenging because of the vast chemical and

structural diversity of known -lactamase enzymes. Nonetheless, several inhibitor compounds have

been deployed and can achieve bacterial clearance when co-administered with -lactam antibiotics,

such as clavulanic acid which is routinely co-administered with amoxicillin.62

Clavulanic acid, the first clinically used -lactam inhibitor, is a -lactam derivative that, on its own,

exhibits poor bactericidal properties (Figure 1.3).128,129 It is commonly used in combination with

amoxicillin or ticarcillin to irreversibly inhibit serine -lactamases; its -lactam moiety forms a stable

acyl-enzyme complex with the catalytic Ser70 and blocks the hydrolytic action of the enzyme.37

Although the use of clavulanic acid is a definite success, the high mutational capacity of -lactamase

genes resulted in rapid emergence of adjuvant-resistant strains that prompted the search for additional

37

inhibitors.130 Modification of the oxazolidine in the heterocyclic core of clavulanic acid gave birth to

the sulfone-based sulbactam and tazobactam inhibitors, and expanded our capacity to inhibit -

lactamase activity in highly resistant Gram-negative pathogens, such as Pseudomonas, Acinetobacter,

or Klebsiella spp.37,131 It is interesting to note that both compounds show very distinct activity profiles

despite their similar structure. Sulbactam, for example, exhibits species-specific inhibition of

Acinetobacter spp., while piperacillin-tazobactam combination shows more extensive inhibitory

activity of more Gram-negative pathogens than the cefoperazone-sulbactam combination.37,131,132

Despite the apparent mechanistic similarity of class A and C serine--lactamases, class C enzymes are

particularly resilient to inhibition by the aforementioned classical -lactamase inhibitors. This is due to

an active site variation that enables them to hydrolyse the ‘inhibitory’ serine peptide bond with these

compounds.133,134 A novel non--lactam inhibitor, avibactam, has been shown to substantially decrease

the resistance levels of class C -lactamase-carrying bacteria, and ceftazidime/aztreonam are approved

for use in complicated intra-abdominal and urinary tract infections in combination with this adjuvant

(Figure 1.3).62,133,135,136 Unlike classical inhibitors, avibactam, acts as a semi-reversible inhibitor by

stabilising the carbamoyl complex at the active site and promoting re-cyclisation rather than hydrolytic

release from the active site.134,137 This means, that the inhibitor concentration is not depleted over time

allowing on-going competitive inhibition of the -lactamase enzymes.

Figure 1.3 Clinically available -lactamase inhibitors target class A, C and D -lactamases. Clavulanic acid, sulbactam

and tazobactam emulate the key -lactam ring of -lactam antibiotics and are effective, primarily, against class A and D -

lactamase enzymes. Novel inhibitors avibactam and vaborbactam have been developed to target class C and KPC-type -

lactamases, respectively; these enzymes therapeutically challenging targets.37

38

Another novel inhibitor, vaborbactam, a boronic acid transition state compound, was also recently

approved for use in complicated urinary tract infections as a co-treatment with meropenem.37,138,139 Its

novel mode of action makes it particularly suited for use against KPC-producing Enterobacteriaceae,

but unsuitable for use with A. baumannii or P. aeruginosa.138

Despite the non-lethality of these inhibitors, clinical use in combination with antibiotics has already

been shown to cause resistance development, seen for example with the emergence of mutated KPC-3

(Asp179Tyr/Thr243Met) -lactamase in K. pneumoniae clinical isolates.140 Further, class B metallo--

lactamase enzymes remain unaffected by inhibitor compounds, due to their unique and divergent

hydrolytic mechanism that depends on the use of Zn2+ ions; they are thus of increasing clinical

concern.62

As -lactamase resistance is widespread across pathogenic species, some of the most worrying

representatives of bacterial pathogens, such as K. pneumoniae, Enterobacter cloacae or M. tuberculosis,

are treated with aminoglycoside compounds such as streptomycin, gentamicin or tobramycin.141

Aminoglycoside antibiotics inhibit protein synthesis by targeting the ribosome and strong synergistic

behaviour with other drug classes has been reported.141,142 Structural modification to these antibiotics is

carried out by the often mobile aminoglycoside N-acetyltransferase (AACs), aminoglycoside O-

nucleotidyltransferase (ANTs) or aminoglycoside O-phosphotransferase (APHs) enzymes.143 These

minimise the aminoglycoside binding by respectively acetylating, adenylating, or phosphorylating the

hydroxyl or amine groups.141,143,144 Several strategies for targeting these modifying enzymes have been

suggested and include the repurposing of current protein kinase inhibitors or the synthesis of substrate

and peptide-based analogues. However, no compounds have reached the clinic to date.37

1.4.2 Membrane permeabilising compounds

Decreased membrane permeability affects the treatment of many Gram-negative infections. In addition

to other antibiotic resistance mechanisms, pathogens, such as P. aeruginosa, commonly mutate or

downregulate the expression of transmembrane -barrel porin structures to further non-selectively

exclude small molecules.28,145

39

Polymyxin compounds, like colistin, are key molecules, used to disrupt the Gram-negative outer

membrane. These compounds interact with LPS-associated cations and permeabilise the outer

membrane to facilitate the entrance of small molecules, including antibiotics, into the cell.146 This

strategy has enabled the sensitisation of several ESKAPE pathogens to previously ineffective drugs,

like azithromycin.37,147 The polymyxin colistin is a special case in AMR inhibitors, as it exhibits

antibiotic function and is also used as an antibiotic drug on its own. Despite significant nephrotoxicity,

colistin use is on the rise in an attempt to counteract the loss of other antibiotic classes due to

resistance.148,149

Interestingly, truncated versions of polymyxin antibiotics retain their permeabilising activities and these

compounds could make suitable antibiotic adjuvants.37,150The polymyxin B nonapeptide synthesised by

Vaara et al. was shown to sensitize resistant K. pneumoniae, P. aeruginosa and S. typhimurium strains

to different classes of antibiotic compounds, including the -lactam ampicillin, and the Gram-positive

drug vancomycin.37,151 Research into less cytotoxic derivatives of these cationic peptides has yielded

several adjuvants that are currently in clinical trials and might open new avenues for antimicrobial

peptide strategies.37

Antimicrobial peptides, natural or synthetic, encompass a large family of multipurpose scaffolds that

exhibit promising in vitro activities, while synergising with many clinically used compounds.152,153 As

such, strategies for the use of organic acids, cholic acids, and polyethyleneimines have been proposed;

their clinical use, however, is complicated by a number of factors including lack of in vivo activity, low

specificity and high cytotoxicity.154,155

Overall permeabilising compounds are a promising strategy that could help increase the intracellular

concentrations of existing antibiotics while also potentially exhibiting antibiotic function when used in

isolation. Most importantly, targeting the outer membrane may open the use of compounds previously

unsuitable for the treatment of Gram-negative species due to their exclusion by the cell wall, as is the

case with the antibiotic vancomycin.

40

1.4.3 Inhibition of drug efflux

Inhibition of prokaryotic efflux has long been attempted to prevent antibiotic removal, following the

passage of drugs through the outer membrane. However, development of both active and specific

compounds against efflux is challenging due to the innate role of efflux pumps in the expulsion of

metabolic side-products, which is conserved in both prokaryotic and eukaryotic species.156 Strategies

proposed to date involve energy decoupling, steric inhibition or manipulation of pump expression.37,156

Small-molecule and peptide-based inhibitors designed to enable steric blocking of efflux pumps are

collectively known as EPIs, efflux pump inhibitors.156 Numerous alkaloid, flavonoid, polyphenol,

quinolone, aryl and heterocyclic derivatives have been developed and tested for activity, and while

many promising hits have been identified, cytotoxicity or specificity issues have, in most cases,

prevented their clinical use.37,156 A notable exception is verapamil, a small molecule channel blocker

used in the treatment of hypertension which has been approved for the use in M. tuberculosis infections,

where it competitively inhibits the activity of the multidrug and toxin extrusion pumps.156–159

Similar competitive inhibitors to verapamil have been extensively investigated and yielded several

promising avenues of research. Two representative compounds are the carbonyl cyanide-m-

chlorophenylhydrazone (CCCP) that disrupts the proton motive force and PAN, which competitively

inhibits the efflux activity of RND pumps.160,161 Through different modes of action, both compounds

result in sensitisation of critical Gram-negative pathogens to previously ineffective antibiotics; CCCP

sensitises Klebsiella spp. to tetracycline, whilst PAN allows the use of erythromycin and

chloramphenicol on resistant P. aeruginosa.161,162 Additionally, CCCP exhibits synergistic effects with

a range of other antibiotics, including carbapenems, resulting in metabolic arrest.156,160,163

A slightly different approach for targeting efflux resistance uses peptide mimics of nucleic acids and

synthetic DNA molecules to downregulate or block the expression of essential pump components. This

has led to the sensitisation of Campylobacter jejuni strains to ciprofloxacin and erythromycin.37,164

While successful, this method requires a thorough understanding of the transcriptional and translational

regulation behind efflux pump expression, which is likely to be species and even strain specific. As

such, although in principle effective, it is unlikely to be suitable for the development of broad-acting

inhibitor compounds.

41

Overall, broad-acting breakers of antimicrobial resistance, such as -lactamase or efflux inhibitors

represent attractive treatment. However, given the low percentage of successful clinically used

inhibitors, in comparison to the number of compounds in development, new strategies may need to be

considered for the future. The use of repurposed drugs or modified antibiotic substrates and covalent

inhibitors may allow the exploitation of this approach to generate next-generation broad-acting

resistance breakers.

1.4.4 Inhibiting the spread of antimicrobial genes

Conjugation of AMR-gene-carrying plasmids is the most prevalent mechanism for the spread of

resistance determinants in bacterial populations.101,165 Out of the 28 conjugative plasmid types

characterised to date, four families have been strongly linked to the transmission of AMR genes.101,165

Their presence is particularly notable in Enterobacteriaceae, where these large, low-copy and self-

transmitting mobile elements encode multiple resistance determinants including carbapenemases and

extended-spectrum -lactamases (ESBLs).101,127 Further, unlike small-sized mobile elements that are

easily lost in the absence of selective pressure, these ‘AMR plasmids’ appear to be maintained for

extended time periods in E. coli.127,166,167 Thus, inhibition of the spread or maintenance of these gene

pools is a potentially promising approach to mitigate the spread of antimicrobial resistance.82 Inhibition

of plasmid conjugation can, theoretically, be achieved at different levels: a) through inhibition of

pathways in the recipient cell or donor cell, and b) through inhibition of the conjugation machinery.

The generation of inhibitors targeting either the recipient or donor cells is challenging, due to the variety

of recipient-donor combinations possible. However, inhibitors targeting plasmid maintenance

mechanisms open the possibility of clearing donor cells of pre-existing plasmids or, alternatively,

preventing the stabilisation and maintenance of newly acquired plasmids in the recipient cells.127,167,168

These approaches are reminiscent of the current practice of perioperative antibiotics use. Targeted

design of conjugative ‘interference plasmids’ that exploit the ori incompatibility of many closely related

plasmids resulting in plasmid loss in the population, has been successfully applied in vitro and in mouse

models by Kamruzzaman et al.167 While clinical use of these plasmids is unlikely, it provides an

interesting tool for laboratory use and acts as a proof of principle. More relevant is the identification of

two already existing anti-HIV drugs, abacavir and azidothymine, in a medium throughput screen; these

prevented the transmission of two AMR plasmids isolated from E. coli and K. pneumoniae.127 While

42

these two compounds offer species- and plasmid- specific protection, further development may improve

their scope.127

Finally, conjugation can be interrupted through the inhibition of the conjugative machinery, such as the

relaxosome and the type IV secretion system. Several strategies have been employed in the literature,

like the use of intracellular antibodies to target the relaxase or ATP inhibitors against the secretion

system or its chaperones.101,169–172 These anti-HGT approaches are similar to the concepts driving the

development of novel anti-virulence strategies that target non-essential bacterial pathways responsible

for pathogenicity in an attempt to minimise selective pressure while potentially allowing bacterial

clearance. In addition to stringent and rational therapy design, these approaches can help maximise

treatment success and minimise chances of resistance development for challenging resistant pathogens.

43

Part B

Protein Folding in the Cell Envelope

44

1.5 THE BACTERIAL CELL ENVELOPE

The bacterial cell envelope is a complex cellular compartment that protects the cytoplasm from external

stresses. Its composition depends on the bacterial species and its natural environment, nonetheless two

main structural arrangements are largely recognised: the Gram-positive cell envelope, mostly composed

of a thick peptidoglycan layer, and the Gram-negative cell envelope, where an aqueous compartment

along with a thin peptidoglycan cell wall are sandwiched between the inner and outer membranes.

Most of the Gram-negative cell envelope is taken up by the highly oxidative periplasm. This soluble

fraction houses a variety of pathways that safeguard the overall stability of the envelope and ensure the

optimal function of its components.173 Periplasmic chaperones and proteases play an essential role in

managing proteostasis in this part of the cell, either by directing proteins to their final destinations or

by managing misfolding and aggregation events.174–177 Although the activity of the three major

representatives of these folding catalysts, SurA, Skp and DegP, shows a level of functional redundancy,

combined losses (SurA/Skp or SurA/DegP) are not tolerated and result in deleterious effects and

decreased bacterial survival. The widely conserved DegP protease/chaperone, for example, is

responsible for the degradation of a range of membrane-associated and oxidatively-damaged proteins.

It exhibits a temperature-sensitive protease/chaperone activity, with its chaperone activity prevalent at

low temperatures (28°C) and its protease activity appearing at higher temperatures and becoming

essential for survival at temperatures above 42°C.178,179

At the interface between the periplasm and the outer membrane resides a protective layer of rigid

peptidoglycan that is linked to the outer membrane by murein proteins, and that determines the shape

of the cell.5,11 Finally, the outer membrane itself is composed of a lipid bilayer, where outward facing

glycolipids act as a charge exclusion barrier and size-restricting -barrel porin structures, such as OmpF

or OmpC, control the transport of most biomolecules.12 An additional single layer of protein, the S-

layer, is present in some species and encapsulates the cell offering further protection. Overall, the cell

envelope evolved to allow the selective passage of nutrients and waste to and from the cytoplasm,

respectively, to safeguard the cytoplasmic content while ensuring structural integrity of the cell.

45

1.5.1 Protein folding and transport into the cell envelope

Protein folding refers to the ensemble of cellular processes by which proteins attain their biologically

stable and functional three-dimensional structures. As mRNA molecules are translated into nascent

polypeptides by ribosomes in the cytoplasm, co-translational folding driven by non-covalent

intramolecular forces, such as hydrophilic, hydrophobic and ionic interactions or hydrogen bonds,

begins at the N-terminus.180 Protein folding is generally an energetically favourable process, which

often occurs spontaneously; the majority of initial folding interactions are both enthalpically and

entropically favourable and follow the thermodynamic gradient towards the lowest energy

conformation. The three-dimensional topology of approximately 70% of all protein amino acids is

governed by hydrogen bonding and consists of highly regular α-helices, β-sheets, and, less commonly,

- or -loops. Interplay between these discrete structural elements leads to the collapse of hydrophobic

regions to the interior of the protein molecule and the rearrangement of hydrophilic regions to coat the

external surface of the forming structure. Interactions between non-polar protein regions and their

surrounding aqueous environment are, thus, minimised to decrease the total energy of the structure.

Achievement of the correct folding intermediate is critical for the protein’s continued existence at this

stage; misfolding in the secondary and super-secondary structure often leads to costly protease-

mediated degradation to prevent toxicity caused by accumulation and aggregation of misfolded

proteins.180–182

Although most protein amino acids can spontaneously self-arrange into highly ordered states, a lack of

strong secondary structure is generally observed in catalytically active sites. As enzymatic activity is

often controlled by amino acids that are polar or charged in nature, their close spatial proximity causes

steric strain resulting in loss of structural regularity.181,182 Thus, unusual kinks and angles in amino acid

backbones result in above-average representation of random coils in the active-site regions.181,183,184 In

addition, key residues are commonly located within, or in proximity to, hydrophobic pockets to

maximise productive van der Waals surface interactions between an enzyme and its substrate. Given

the exposure of active sites to the surrounding aqueous environment, interactions at these non-polar

surfaces increase the total energy of the system.181 Therefore, high thermodynamic favourability in the

rest of the protein fold is essential to offset these inherently unfavourable interactions and achieve

overall protein stability and functionality. This delicate balance can be disrupted even by a single change

in the primary amino acid sequence giving rise to stability-function trade-offs. Trade-offs of this nature

underpin the ability of enzymes to evolve improved functions as seen, for example, with the effects of

46

single amino acid substitutions on the stability or hydrolytic activity of several class D OXA-family -

lactamase enzymes.181,183,185

While all proteins are synthesised in the cytoplasm, acquisition of their native folds may not be achieved

until their target destination has been reached. Often, protein folding depends on the action of specific

post-translational adaptation machinery, incorporation into a membrane, formation of a higher-order

protein complex, or simply translocation across the cytoplasmic membrane. A key role in all these cases

is performed by cytoplasmic chaperones that slow down or entirely prevent spontaneous protein folding

during and immediately after mRNA translation. These helper molecules trap partially folded protein

intermediates, prevent their misfolding and aggregation or help mediate their post-translational

transport. In Gram-negative bacteria, approximately 20-30% of all expressed gene products is destined

to perform key functions in the bacterial cell envelope, such as swimming and twitching motility,

nutrient transport, or protection of the cell against mechanical stresses and antimicrobial agents. It is

thus essential that all of these protein components efficiently traverse the inner membrane through co-

or post-translational protein translocation pathways.186,187

1.5.2 Protein transport across the inner membrane of Gram-negative bacteria

The bacterial inner membrane serves as a selective barrier that separates the cytoplasm from the

periplasm. In the absence of defined organelle-like structures in bacteria, this divider provides a

platform for all membrane-associated reactions and functions, including energy production, lipid

biosynthesis, and protein secretion and transport. Given the harsh conditions of the extra-cytoplasmic

environment, and the segregation of potentially harmful degradative enzymes such as RNAses or

phosphatases in the intermembrane space of the cell envelope, membrane transport is a tightly

controlled process.188 Polypeptides that are required to cross the inner membrane carry a characteristic

24 to 30 amino-acid-long N-terminal sequence; this signal peptide is composed of a positively-charged

N-terminus, a hydrophobic -helical core (h-region), and a C-terminal domain with a signal peptidase

cleavage site (exported proteins) or a hydrophilic region (transmembrane proteins).188,189 The nature of

the signal peptide along with the post-translational folding state of the protein, determine the

translocation pathway that mediates the peptide’s transport outside the cytoplasm.

The general secretory pathway (Sec) selectively transports unfolded proteins across the inner membrane

during translation, or integrates them into the inner membrane, immediately after they are translated.188

47

Most transmembrane proteins are integrated into the membrane in a co-translational fashion. This helps

prevent their release from the ribosome and minimises unfavourable interactions with the aqueous

cytoplasmic environment that could result either in aggregation or in errors in signal peptide

recognition. Ribosome-bound signal peptide and its signal recognition particle associate with the signal

receptor at the membrane. In turn, the signal receptor hydrolyses GTP to transfer the signal peptide into

the SecYEG translocase channel.187,190–193 This process depends heavily on the hydrophobic h-region in

the signal peptide which drives the efficient interaction of the signal peptide with its signal recognition

particle.187,190–193 Continuing polypeptide synthesis from the ribosome, pushes the unfolded polypeptide

further into the SecYEG channel. The subsequent integration of the protein into the inner membrane

has been shown to be mediated by the membrane chaperone YidC and the ensemble of

hydrophobic/hydrophilic interactions between the h-region of the signal peptide, the transmembrane

helices of the folding protein, the hydrophobic domains of the Sec translocase, and the membrane

itself.187,189,194–197

In contrast, soluble periplasmic proteins are usually transported across the inner membrane post-

translationally. In E. coli, the cytoplasmic chaperone SecB has been shown to direct substrates to the

translocation channel through its interaction with the Zn2+ coordination site of SecA.187,188,198 SecA, a

membrane associated ATPase, drives polypeptide translocation across the membrane. Notably, SecA

interacts with both the C- terminal and the N-terminal parts of the SecYEG channel and this binding is

thought to be responsible for the opening of the channel to the polypeptide chain.187,199 The signal

peptide is looped through the channel such that the N-terminus of the polypeptide remains at the

cytoplasmic side of the membrane while the C-terminus enters into the periplasm.187,188,200 Consecutive

ATP-powered insertion and de-insertion cycles of SecA generate a push-and-slide movement of the

polypeptide across the membrane.187,188,200 Throughout this process, the permeability of the membrane

remains unperturbed due to the polypeptide plugging the SecYFG channel.187,201 To this end, the

association of SecYEG with the SecDF complex has been shown to prevent polypeptide backsliding

and increase transport efficiency in vivo.187 Interestingly, the prokaryotic Sec translocation pathway

shares many similarities with the Sec61/62/63 membrane complex and the associated BiP ATPase of

the endoplasmic reticulum (ER) in eukaryotes, and has only rarely been retained in mitochondria.188,202

The second most common inner membrane protein translocation system in prokaryotic cells is the twin-

arginine translocation (Tat) pathway.203 In E. coli this pathway has specialised to transport folded and

co-factor binding proteins that are too large to pass through the Sec channel and require a larger complex

48

assembly to transverse the inner membrane.186,189,204 Although these proteins represent a smaller fraction

of exported polypeptides, they are often essential in energy metabolism as they form parts of the

respiratory and photosynthetic electron transport chains.186,189,204

In E. coli proteins are targeted to the TatABC translocase-docking complex by a 30 amino acid signal

peptide adapted to avoid the Sec machinery through the inclusion of an essential twin arginine (RR)

motif, a less hydrophobic h-region and a positively-charged Sec-avoidance signal prior to the C-

terminal cleavage site.189,203,204 Recognition of the signal peptide by TatC and deep insertion of the

signal peptide into the protein, leads to interaction of the h-domain of the signal with TatB and exposure

of the cleavage site to the peptidase on the other side of the inner membrane.186 Notably, evidence

suggests that assembly of several TatBC complexes at the membrane allows simultaneous binding of

multiple substrate proteins.186

TatB docking and its interaction with TatC and the signal peptide drives the recruitment of TatA to the

membrane.186,203,205 TatA oligomerisation leads to the formation of large (100 to 500 kDa) multimeric

structures that are thought to translocate proteins across the membrane via two possible mechanisms,

the formation of a translocation pore or through the weakening of the membrane.186,206,207 In both

potential mechanisms, TatA assembly and protein transport are driven by the proton motive force; the

existence of an antiporter mechanism coupled to this process has also been suggested.186 Interestingly,

the Tat system is able to recognise misfolded proteins, prevent their export, and direct them to be

degraded in the cytoplasm.186,203,208–210 While the mechanism behind this activity is not fully understood,

it has been linked to co-factor insertion and thus potentially to conformational stability of the

translocated protein.186,189,204 In the case of B. subtilis, the WprA protease was shown to be essential for

protein translocation by directly interacting with the TatAyCy system; this suggests that Tat

translocases can be linked to specific localised degradation systems.186,203,211

An essential step in achieving full protein export through either the Sec or the Tat pathway is the

exposure of and cleavage at the C-terminal cleavage site when the polypeptide has reached the

periplasm. This liberates the mature protein which is then ready for further post-translational

modifications, folding and transport in the cell envelope.186,188,189,203

Sec and Tat translocation machineries exhibit distinct substrate profiles. However, protein substrates

belonging to the same protein superfamily are not necessarily transported exclusively by one of the two

49

systems. For example, at least 89 signal peptides have been identified in -lactamase enzymes to date,

showcasing the existing diversity in bacterial protein export.189 With PBPs invariably undergoing Sec

transport, the evolutionary-related -lactamases were also initially thought to be only transported via

the Sec pathway.189,212 However, examples from M. tuberculosis (BlaC, BlaS), S. maltophilia (L2-1),

and even Pseudomonas luteola (LUT-1, (SignalP 5.0 likelihood scores: Sec/SPI = 0.0572, Tat/SPI =

0.9312, Sec/SPII (lipoprotein) = 0.0087, other = 0.0029)) have proven that the export of some of these

hydrolases can be Tat dependent.189,212,213 These discrepancies are likely linked to post-translational

processing requirements of some -lactamases. For example, some enzymes undergo Sec-mediated

transport to gain access to periplasmic post-translational modification machinery, like the disulfide

bond formation system, which helps them achieve their biologically active structure. Others, like the

representative enzymes from M. tuberculosis, S. maltophilia or P. luteola do not require such

modifications and are transported into the periplasm in their fully functional state.189,213

1.6 OXIDATIVE PROTEIN FOLDING PATHWAYS

The stability of proteins located in the cell envelope is key for bacterial survival. A plethora of

chaperones and proteases work against the harsh conditions of the extra-cytoplasmic environment (for

example low pH or high salt content) to safeguard the integrity of the cell envelope proteome.214 In

addition, the formation of disulfide bonds in many extra-cytoplasmic proteins improves their stability

by reinforcing their non-covalent intra- and inter-molecular interactions, like van der Waals or

hydrophobic interactions and ionic or hydrogen bonds. The biochemical processes resulting in the

formation and isomerisation of covalent bonds between two spatially proximal cysteine amino acids,

are collectively known as oxidative protein folding and occur post-translationally. The process of

disulfide bond formation, despite the simplicity of the oxidation reaction that entails the removal of two

protons and two electrons from the thiol side groups of the cysteines, is catalysed by dedicated protein

systems in both prokaryotic and eukaryotic species.215,216 A brief overview of oxidative protein folding,

with a focus on Gram-negative bacteria, is given below.

1.6.1 Disulfide bond formation in eukaryotes

The highly compartmentalised nature of eukaryotic cells offers many surfaces where oxidative protein

folding can take place, and as such more than one organelle structure plays a role in disulfide bond

50

formation. Two major and distinct oxidative pathways have been identified in eukaryotes, the protein

disulfide isomerase pathway in the endoplasmic reticulum (ER) and the MIA pathway in the

mitochondrial intermembrane space (IMS).217 Oxidative folding also takes place in thylakoid

membranes of plant chloroplasts, although this field is highly understudied and will not be discussed

further here.218

The existence of a catalyst responsible for oxidative protein folding was first noted in 1963 in the

process of re-oxidation of reduced RNase obtained from rat liver,; this process can occur spontaneously,

but only at a very slow rate and under highly specific conditions.219,220 The rate of RNase oxidation was

shown to accelerate by addition of crude rat liver homogenate. Characterisation of this homogenate by

fractionation let to the discovery of the protein disulfide isomerase (PDI) enzyme.219,221 Since then, the

function and mechanism behind PDI catalysis has been extensively studied in Saccharomyces

cerevisiae, where its isomerase activity was shown to be essential for cell viability, and three key players

were identified, PDI1 (PDI family member), ER oxidoreductin 1 (ERO1p) and a protein essential for

respiration and vegetative growth (Erv2p).220,222–224 Interestingly, although ERO1p and Erv2 do not

share sequence similarity, they both contain two active cysteine pairs, bind a flavin adenine dinucleotide

(FAD) co-factor within a 4-helix core, and carry out similar functions.218,223,225

PDI1 is a soluble, V-shaped enzyme with a classical thioredoxin (Trx) fold that has a highly conserved

CGHC functional motif (Figure 1.4).218,226 In its oxidised state, the protein acts as a disulfide bond donor

and forms a mixed-disulfide bond with nascent polypeptides upon their translocation into the ER

lumen.217 Nucleophilic attack by a secondary thiol leads to the release of the oxidised substrate and

generation of reduced PDI1. In turn, the latter can either act as an isomerase or be re-oxidised by

membrane associated ERO1p/Erv2p.217 These two proteins shuttle electrons through a FAD co-factor

to the respiratory chain.217,223,225 The formation of FAD2H and its subsequent reaction with molecular

oxygen yields hydrogen peroxide and regenerates ERO1p/Erv2p.217,223,225

51

Figure 1.4 The first identified oxidative folding catalyst PDI1 carries a classical Trx fold with a conserved disulfide

bond in a CXXC motif, a common feature in all thiol-redox enzymes. (A) A schematic of the classical thioredoxin fold;

the N-terminal domain carries the catalytically active CXXC motif. An helix connects the N-terminal domain to the C-

terminal domain, where a conserved cis-proline is located; this is a key residue for substrate binding and release. Adapted from

Shouldice et al.218 (B) A crystal structure of the PDI1 enzyme from S. cerevisiae, showing a V-shaped fold composed of four

Trx domains. The N- and C-terminal domains contain a conserved pair of cysteine residues; one of the conserved thioredoxin

domains containing a disulfide bond (Trx domain 1) is highlighted. In the second cysteine-containing Trx domain the cysteines

are in their reduced form. PDB code: 2B5E.66,224

Eukaryotic PDI proteins are remarkably diverse, with five different yeast and over 20 distinct human

homologues.218 It is, thus, not surprising that in more complex mammalian species, additional routes

for PDI re-oxidation have been identified and include numerous peroxidase enzymes or vitamin K

epoxide reductase-like proteins (VKOR is discussed further in section 1.6.3).217,227,228 For example,

peroxiredoxin IV mediates PDI oxidation alongside the ERO-1 pathway in mammals, using hydrogen

peroxide as the electron sink.217,228 Its absence leads to impaired disulfide formation and growth.217,228

Other peroxidase enzymes, such as the glutathione peroxidases 7 and 8, have also been shown to oxidise

members of the PDI family and likely act in concert with the main PDI pathway in order to utilise the

hydrogen-peroxide-to-water conversion and maximising the use of reduced oxygen.217,228

In addition to the oxidative protein folding pathways in the ER, disulfide bond formation also occurs in

the mitochondria where it mitigates the effects of restrictive protein import across the mitochondrial

membrane. As the mitochondrial structure closely copies that of its prokaryotic predecessors, with an

outer membrane, an intermembrane space, an inner membrane, and a matrix, the passage across its outer

membrane is restricted to reduced and unfolded polypeptides.202,218,229 The introduction of disulfide

bonds into these protein precursors ensures their entrapment within the organelle and their functional

activation.218,229 The mitochondrial MIA pathway is composed of two folding catalysts, the

52

oxidoreductase Mia40 and a FAD dependent sulfyhydryl oxidase Erv1, both of which have similar

redox interactions as the prokaryotic DsbA-DsbB system (discussed further in section 1.6.2.1.1).229,230

Mia40 is a highly conserved protein with an -hairpin core that, despite its redox activity, does not

contain the classical Trx redox fold.231,232 Instead, catalytic activity is driven by a CPC functional motif,

which is supported by a hydrophilic binding cleft and stabilised by two oxidised N-terminal CX9C

moieties.229,231 Its substrates include polypeptides from the respiration and protein biogenesis pathways,

such as the Tim family of chaperones, and commonly include twin CX9C or CX3C motifs.229,231,233

Substrate oxidation is thought to occur through the formation of a mixed disulfide between the second

Mia40 cysteine and its substrate.229 Reduced Mia40 is re-oxidised by Erv1 and electrons are passed

through its FAD domain to cytochrome c, and ultimately to the respiratory chain where they generate

water molecules.229,230

Mia40 appears to be able to ‘selectively’ introduce correct disulfide linkages in the presence of more

than two cysteine residues, likely through the recognition and folding of a nine-amino-acid-long IMS-

targeting signal (ITS/MISSvii).229,231,233,234 The presence of proofreading mechanisms has been proposed,

including the utilisation of reduced glutathione or direct Mia40-facilitated disulfide isomerisation.233,234

Additionally, in a fashion that is not too dissimilar from the dual role of the prokaryotic isomerase

DsbC, Mia40 has also been shown to act as a chaperone for numerous cysteine-rich proteins, including

the mitochondrial protease Atp23.229,234

The similar functions of the thiol-oxidase and thiol-oxidoreductase enzymes of the PDI and MIA

pathways result in the formation of the majority of eukaryotic disulfide bonds. However, despite the

mechanistic similarities, their respective substrate profiles are entirely distinct; polypeptides folded in

the ER are targeted to all cellular and extra-cellular locations, while substrates of the MIA pathway are

oxidised for exclusive use in the mitochondria.218

Last but not least, it should be noted that in addition to the PDI and MIA pathways, mammalian cells

are also able to oxidise nascent polypeptides through the action of the quiescin sulfhydryl oxidase

vii Research by Sideris et al. suggests that the IMS-targeting sequence overlaps with the mitochondria IMS-sorting signals previously observed

by Milenkovic et al. in some Mia40 substrates. For ease of reference and limited understanding these targeting sequences are now commonly

and jointly referred to as the ITS/MISS sequences.229,233,459

53

(QSOX).217,235 Identified by Thorpe and Coppock, this enzyme is present in the ER membranes, the

Golgi apparatus and the extra-cellular matrix and given that it contains both an FAD co-factor and a

Trx domain, it has been suggested that its role might be to form disulfide bonds in the extra-cellular

environment.217

1.6.2 Disulfide bond formation in prokaryotes

1.6.2.1 Gram-negative bacteria – the archetypical E. coli system

Bacterial cells lack the clearly organised organelle structures that provide the surface for oxidative

protein folding reactions in eukaryotes. Instead, in Gram-negative bacteria, these processes take place

in the oxidative environment of the outer leaflet of the inner membrane, in the periplasm. The first

bacterial oxidative folding mechanism was discovered in E. coli, and this archetypical system has since

been characterised in depth.236–240 It is composed of two complementary pathways, the oxidative and

the isomerase pathways, which are respectively responsible for the introduction and rearrangement of

disulfide bonds in newly translocated polypeptide chains.

1.6.2.1.1 The oxidative pathway

The oxidative pathway introduces consecutive disulfide bonds into polypeptides during, or immediately

after, their periplasmic translocation through the Sec translocase. As cysteine residues enter the

periplasmic space, they are trapped in a mixed-disulfide bond by the soluble thiol oxidase, DsbA. The

resolution of this intermediate complex occurs via disulfide transfer to the polypeptide substrate.

Reduced DsbA is then recycled by its membrane-bound partner DsbB, a unique quinone reductase

located in the inner membrane (Figure 1.5).

54

Figure 1.5 The oxidative pathway of the disulfide bond formation system. The oxidative pathway introduces disulfide

bonds into polypeptide chains containing two or more cysteine residues, through the activity of the primary oxidase DsbA,

and immediately after protein translocation through the Sec pathway. The removal of two electrons from two consecutive

cysteine residues leads to substrate oxidation and DsbA reduction. DsbA is re-oxidised by its partner, the membrane-embedded

protein DsbB. Two electrons are shuttled from DsbA to DsbB through an intermolecular disulfide. A thiol-disulfide exchange

transfers the acquired charge onto ubiquinone or menaquinone carriers and from there to the electron transport chain. Bond

lengths not to scale. Figure adapted from Messens & Collett.241

1.6.2.1.1.1 DsbA

DsbA is a 21 kDa soluble monomeric protein with a conserved active-site disulfide bond that is

available for transfer to substrates with two or more cysteine residues in their primary sequence.242 The

protein core comprises an extended Trx-like fold that is disrupted by an -helical domain (Figure 1.6

A).242 The classical Trx fold is composed of two sub-domains, the N-terminal and the

C-terminal domains, connected by a single -helix (Figure 1.6 A).218,243 In DsbA, the loop linking the

-helix to the -sheet is broken up by an insertion of a compact helical unit of three clustered -

helices (−) flanked by single -helices ( at the N-terminal end of , Figure 1.6 A).218,242

cytoplasm

periplasm

55

Figure 1.6 Crystal structure of the primary oxidase of the E. coli DSB system, EcDsbA. (A) Structure of EcDsbA rendered

in a cartoon representation. The protein is composed of a Trx core, 1/1/2-2- 3/4/3, disrupted by a compact -helical

domain insertion which is composed of three clustered -helices, 2-4, followed by single -helix, 5.218,242 The active site

disulfide bond lies on the N-terminal side of the helix and is depicted by yellow spheres.242,245,246 (B) A close-up view of

the active site of EcDsbA. The catalytically active motif, Cys30-Pro-His-Cys33, and the substrate binding loop, Gln145-Leu-

Arg-Gly-Val150, are shown in cyan stick representation. Highlighted are the Pro31-His32 dipeptide residues responsible for the

protein’s strong oxidative potential and the conserved cPro151 residue critical to successful substrate release. (C) The

electrostatic surface of EcDsbA generated using PyMol with a normalised hydrophobicity scale developed by Eisenberg et al.;

a sliding colour scale shows the hydrophilic regions in light grey/pink and hydrophobic residues in darker red.248 The active

site residues are shown in cyan (Pro31-His32 dipeptide) and yellow (Cys30 and Cys33). On the left, the hydrophobic patch

responsible for the recognition and binding of substrate polypeptides is prominent in dark red colour to the left of and above

the active site. The DsbB binding groove (Groove 1) runs below the active site.242,246,247 On the right, a side-on view shows

the presence of a hydrophilic groove (Groove 2) of unknown function.242 Information for this figure compiled from Guddat et

al., Shouldice et al. and Martin et al.; PDB code: 1FVK.66,218,242,248–250

56

Interestingly, a similar but much less close-packed insertion is observed in the eukaryotic glutathione

peroxidase enzyme responsible for the formation of disulfide bonds in glutathione.242 The inserted

cluster creates a flexible -turn allowing hinge-like motion to the two DsbA sub-domains.244 At the

interface of the N-terminal end of the Trx domain and the -helical domain lies the active site.242,245,246

The C30XXC33 functional motif (CPHC in E. coli) is surrounded by highly hydrophobic amino acids

responsible for recognition and binding of substrate polypeptides and the partner protein DsbB (Figure

1.6 C).242,246,247 Specifically, a loop connecting the and forms a hydrophobic patch just above the

active site (Figure 1.6 A and C).242,243,246 The C-terminal end of this loop carries a conserved cis-proline

(cPro151, Figure 1.6 B) residue which extends into a deep hydrophobic groove running just below the

active site.242,243,246 An additional prominent groove, lined with polar acidic residues, is located on the

opposite side of the protein, but its function remains unclear.242

DsbA binds and oxidizes a plethora of polypeptides, including virulence factors such as toxins, pili or

secretion system components.216,218,237,251–253 With the exception of an observed bias for an even number

of cysteines, these substrates show no clear conservation of amino acid sequences or structural motifs

proximal to their cysteine residues.246 Instead, substrate recognition seems to be occurring at the DsbA

level; a crystal structure of DsbA in complex with a peptide originating from a native substrate shows

that binding interactions occur in the hydrophobic patch directly above the active site (Figure 1.7

A).246,251 Access to this region in E. coli DsbA (EcDsbA) is dictated by the protein’s oxidation state and

controlled by a histidine residue (His32, Figure 1.6 B, Figure 1.7 C) located in the C30PHC33 dipeptide

motif.246 In particular, in oxidised DsbA, the histidine is positioned across the hydrophobic groove

located under the active site allowing access of the substrate to the hydrophobic patch (Figure 1.7).242,246

57

Figure 1.7 Differential binding of DsbA to substrate peptide or DsbB is directed by the hydrophobic surfaces

surrounding the active site of DsbA and the histidine residue, His32. DsbA is shown either shown in cartoon representation

coloured pink or in an electrostatic surface representation using the normalised hydrophobicity scale developed by Eisenberg

et al.; a sliding colour scale shows the hydrophilic regions in light grey/pink and hydrophobic residues in darker red.248 The

DsbA active site cysteines are coloured in yellow or shown as yellow spheres. The conserved loop of DsbA (QLRGV) known

to affect substrate binding is shown in cyan sticks.246,251 (A) Top panel: DsbA in complex with a substrate polypeptide (dark

blue). Substrate binding is promoted by surface-surface interactions with the hydrophobic patch above the active site. The

active site dipeptide moiety is shown in cyan. Bottom panel: Stick representation of the DsbA-substrate binding site showing

the key interactions with the Gln145- Val150 loop.243,246,251 PDB code: 3DKS. (B) Top panel: DsbA in complex with the P2

periplasmic loop of DsbB (dark blue); DsbB active site cysteines are coloured in orange. The DsbA active site dipeptide moiety

is shown in cyan. Binding is achieved by interaction with both the hydrophobic patch above and the hydrophobic groove below

the active site. Bottom panel: Stick representation of the DsbA-DsbB showing the P2 (Pro100-Phe106) loop in the hydrophobic

groove.247,250,252,262,265 (C) Binding of the substrate or DsbB is directed by the position of the His32 residue which changes its

conformation depending on the oxidative state of DsbA. When the substrate is bound in the active site, the aromatic ring of

His32 (dark blue) blocks the DsbB binding groove. Upon DsbA reduction, a shift of the His32 thiolate to the cyan conformation

results in the opening of the hydrophobic groove to allow DsbB binding.242,246 PDB code: 2HI7. Figure adapted from Paxman

et al. and Inaba et al.66,246,247

DsbA-polypeptide binding occurs primarily through backbone hydrogen bonding with only few direct

surface-surface interactions observed.246 This weak binding is common to many Trx species and is in

)

58

line with the enzyme’s broad substrate scope.246,254 A Gln145-Leu-Arg-Gly-Val150 motif of the Trx loop

was identified as key for substrate recognition (Figure 1.6 B, Figure 1.7).243,246,251 While mutations that

alter these residues do not completely abrogate the enzyme’s disulfide-forming ability, they result in

alteration of the substrate specificity and the redox potential.243,246,251 For example, the single amino

acid substitution Val150Gly leads to complex formation defects for a range of substrate proteins.246,251

Substitution of the entire native E. coli motif (QLRGV, Figure 1.6 B) with the equivalent sequences

from the Neisseria meningitidisviii (QIDGT of DsbA1 or QISGT of DsbA2) affects the ability of DsbA

to confer resistance to the reducing agent dithiothreitol (DTT).246,251 This suggests that additional

interactions and conformational changes may be required to promote correct substrate binding.249

Immediately following the substrate recognition motif is a conserved cis-proline residue (cPro151, Figure

1.6 B) which plays a key role in DsbA stability and its ability to release substrates.251,252 Substitutions

in this residue result in significant loss of activity, likely due to its key position at the end of the Trx

loop and its spatial positioning within the hydrophobic groove.251,252 For example, substitution for

alanine or threonine increases the stability of the reaction intermediates resulting in the accumulation

of the DsbA-polypeptide complex.246,249,252 By contrast, substitution for serine propagates extensive

conformational changes through the Trx loop and results in the accumulation of the DsbA-DsbB

complex.251,252

The process of disulfide bond formation is catalysed by the interaction of the oxidised Cys30XXCys33

motif of DsbA with the cysteine residues of the substrate proteins.242,249 The cysteine pair of DsbA is

strictly conserved across all DsbA homologues, but minor variations are observed in the enclosed

dipeptide moiety.218 While EcDsbA harbours the very common Pro31-His32 combination, DsbA3 of N.

meningitidis, for example, contains a Val-His sequence.218,242,255 In general, His32 is often conserved in

thiol-oxidases, including in the human PDI, as it increases the oxidative power of the enzymes through

stabilisation of their reduced states.256,249 This is due to the residue’s inherently destabilising nature

caused by unfavourable interactions with the nearby -helix dipole in oxidised DsbA.256,249 -

galactosidase functional assays in E. coli by Grauschopf et al. confirm that the dipeptide motif is

viii Multiple homologues of DSB system components have been observed in pathogenic species, such as N. meningitidis. This can result in

more differentiated catalytic profiles (see section 1.6.3 for further discussion of the DSB system polymorphism).

59

responsible for the enzyme’s oxidative properties.257 Substitutions in the CPHC motif in EcDsbA to

CPPC, CPLC or CTRC markedly reduces the protein’s oxidative powerix.250,257

The transfer of the disulfide bond from DsbA to its substrate begins with a nucleophilic attack from the

first, deprotonated cysteine residue of the substrate at the N-terminally located Cys30 of

DsbA.242,246,251,258,259 A stabilising mixed-disulfide bond forms between DsbA and the substrate

polypeptide chain.254 The resultant complex is then resolved by a second nucleophilic attack from the

substrate cysteine to Cys30 of DsbA that leads to the release of the oxidised substrate protein and reduced

DsbA.241,245,251,260,261 DsbA reduction is a favourable process due to the formation of the thiolate anion

of the highly acidic Cys33.218,242,249 Structural comparison of the two redox states of DsbA by Kortemme

et al. and Guddat et al. shows that this thiolate anion forms a favourable electrostatic interaction with

the positive dipole moment of the proximal -helix and a hydrogen bond with the side chain of

His32.249,250,256,257 Jointly, these interactions stabilise reduced DsbA.249,250,256,257 In addition, the thiolate-

His32 interaction causes a conformational change in His32 that results in the opening of the hydrophobic

groove of DsbA in preparation for DsbB binding (Figure 1.7 C).242,246 Unlike substrate binding,

specificity during DsbA-DsbB complex formation is achieved by a great number of interactions

between the two proteins localised at both hydrophobic surfaces that are proximal to the DsbA active

site, the patch above it and the groove underneath it (Figure 1.7 B).246,247,251,254 DsbA is re-oxidised

through a disulfide exchange cascade with DsbB. Mechanistic understanding of this process has been

achieved with the help of four structures of reaction intermediates (three crystal structures and one NMR

structure), which are discussed in more detail in the next section.247,262–264

1.6.2.1.1.2 DsbB

Re-oxidation of DsbA following its catalytic cycle is carried out by a dedicated transmembrane partner

protein, DsbB, which generates the DsbA disulfide de novo. A series of redox reactions allows DsbB

to remove two electrons from DsbA and to re-instate it in its active oxidised state. Electrons removed

from DsbA are then passed from DsbB to terminal oxidases via ubiquinone (UQ, aerobic conditions)

or menaquinone (MQ, anaerobic conditions).241,266 Notably, in comparison to DsbA, DsbB has a low

ix More recent quantum-mechanical simulations of the active site by Carvalho et al. confirmed that highly localised changes in the hydrogen

bonding network affect the protein’s redox activity.260 Reductase species, for example, were observed in the presence of less stabilising

residues, such as glycine or tyrosine.260

60

redox potential, and as a result prevention of electron backflow through the DsbA-DsbB complex is

achieved by substantial conformational changes in the DsbB backbone.247,267

DsbB is a 21 kDA integral membrane protein with an -helical transmembrane domain, two flexible

periplasmic domains and two conserved disulfide-bonded cysteine pairs (Figure 1.8).265,268,269 Four

transmembrane -helices (TM1-4) traverse the inner membrane such that the N- and C-termini of the

protein are located at the cytoplasmic side, whilst the two flexible loops that connect the transmembrane

segments are on the periplasmic side.247 The shorter of the two loops (P1, Gln33-Arg48) links TM1 and

TM2, and the longer loop (P2, Tyr97-Gln144) connects TM3 and TM4 (Figure 1.8).247 Each of the loops

carries a conserved disulfide bond. Cys41-Cys44 is present in the P1 loop and interacts with the quinone

molecule, which acts as an electron sink.240,263–265 Cys104-Cys130 is located in the P2 loop and interacts

with the DsbA thiols as an electron acceptor.240,263–265 The P2 loop is associated with the periplasmic

leaflet of the inner membrane via a short and conserved amphipathic -helix (Leu116-Val120, assignment

is structure-dependent, Figure 1.8).247,263–265 This places the two cysteine residues of the Cys104-Cys130

bond to different protein regions with distinct mobilities. Mutational studies show that this feature

underpins the structural flexibility of DsbB which is central to its ability to oxidise DsbA, despite the

aforementioned unfavourable redox potential.247,263,264 The flexibility of the P2 region makes

crystallisation of isolated DsbB challenging, therefore the protein’s mode of action has mostly been

studied using trapped complexes between DsbA and DsbB cysteine mutants.247,262

61

Figure 1.8 Crystal structure of EcDsbB, the membrane partner protein of DsbA. The left structure shows a side view of

DsbB Cys130Ser variant in complex with UQ, while the structure on the right shows a top view of DsbB. Four helical

transmembrane domains (TM1-4) dock in the inner membrane with the N- and C-termini emerging in the cytoplasm. Flexible

periplasmic loops carry two catalytically active disulfide bonds, Cys41-Cys44 (P1 loop) and Cys104-Cys130 (P2 loop, Cys130Ser

mutant is used in this crystal structure). The second half of the P2 loop, connecting the amphipathic -helix to the TM4 domain,

is very flexible and its position was not determined in the structures above. Thus the Ser130 residue is not shown.240,263–265

Cysteines are shown in blue stick representation, the sulfur atoms are represented by yellow spheres. The ubiquinone (UQ)

co-factor is shown in black sticks. PDB code: 2HI7. Figure adapted from Inaba et al.66,247

DsbA re-oxidation depends on the formation of a DsbA-DsbB complex, which is held together through

interactions of DsbB with the hydrophobic surfaces surrounding the DsbA active site.247 A crystal

structure of DsbA Cys33Ala/DsbB Cys130Ser (Figure 1.7 B, Figure 1.9 A) shows that DsbB interacts

with the hydrophobic patch above the DsbA active site in manner similar to DsbA substrates. In

addition, the specificity of DsbB for DsbA is dictated by an additional interaction with the hydrophobic

groove below the DsbA active site.246,247 Groove binding occurs through the N-terminal end of the P2

periplasmic loop (DsbB, Pro100-Phe106)x and depends primarily on hydrogen bonds, salt bridges and

surface-surface interactions with conserved DsbA residues, such as the cis-proline (cPro151, section

1.6.2.1.1.1) or the Pro63-Phe174 sequence (Figure 1.7 B).247,250,252,262,265 Additional close contacts between

the Ala126-Ala131 sequence of DsbB and DsbA were identified by NMR.263–265 Van der Waals surface

interactions between the active site His32(DsbA) and Ala102-Thr103(DsbB), as well as the formation of a short,

antiparallel -sheet between Arg148-Val150(DsbA) and Cys104-Phe106(DsbB) initiate DsbA re-oxidation

placing Cys104(DsbB) in spatial proximity of Cys30(DsbA) (Figure 1.7).246,247

x Crystal structure by Inaba et al. has shown notable interactions with the Pro163, Gln164, Thr168,Met171 and Phe174.247

62

Figure 1.9 The proposed mechanism of DsbA oxidation by DsbB. DsbA is shown in pink, DsbB is shown in orange. (A)

Crystal structure of DsbA Cys33Ala - DsbB Cys130Ser complex prior to disulfide rearrangement. DsbB binds into the

hydrophobic groove of DsbA via the P2 loop and forms an intermolecular disulfide bond (Cys30(DsbA)-Cys104(DsbB), cyan and

blue, respectively).246,247 The Cys41-Cys44 disulfide bond in the P1 loop of DsbB remains intact (blue stick representation).262–

265 PDB code: 2ZUP.247 (B) Crystal structure of DsbA Cys33Ala - DsbB Cys130Ser after disulfide rearrangement.

63

Conformational shift in the amphipathic -helix leads to the formation of an intramolecular disulfide Cys130-Cys41 (not shown

as Cys130Ser mutant is used). This results in formation of a Cys44 thiolate which interacts with the bound UQ. PDB code:

3E9J.262 (C) NMR structure of the DsbB Cys44Ser/Cys104Ser mutant following DsbA dissociation. A transient intramolecular

Cys130-Cys41 disulfide is shown, trapped by the use of the double cysteine mutant. PDB code: 2K74.263 (D) Crystal structure

of DsbB Cys41Ser mutant showing a Cys104-Cys130 disulfide where full reoxidation has been achieved; the protein can now

undergo another round of DsbA oxidation. PDB code: 2ZUQ.264 Figure adapted from Bushweller et al.66,265

Oxidation of DsbA begins with a nucleophilic attack from the Cys30(DsbA) on Cys104(DsbB); the latter is

located in the P2 loop where also UQ is bound.247,265 This leads to formation of a mixed Cys30(DsbA)-

Cys104(DsbB) disulfide bond leaving Cys130(DsbB) in a thiolate anion form (Figure 1.9 A).247,265,267 At this

point, the C-terminal amphipathic -helix (Figure 1.8) facilitates a conformational shift in the N-

terminus of the P2 loop and increases the distance between the Cys130(DsbB) thiolate and the mixed

disulfide bond to ~8Å.262–265 This prevents the backflow of electrons to DsbA, and places the thiolate in

the vicinity of the Cys41-Cys44 disulfide bond in the P1 loop of DsbB (Figure 1.9 B).262–265

The exact sequence of steps leading to the resolution of this DsbA-DsbB-UQ complex has yet to be

determined, but it invariantly includes a nucleophilic attack on the DsbA-DsbB intermolecular disulfide

bond, rearrangement of DsbB, and an ensuing thiol-disulfide exchange cascade.265 Two non-exclusive

pathways have been described in literature.267 In the rapid pathway, the Cys30(DsbA)-Cys104(DsbB) is

resolved by an immediate attack of Cys33(DsbA) (Figure 1.9 B, C), resulting in release of oxidised DsbA

and a hemi-oxidised state of DsbBxi (Figure 1.9 C, D).262,264,267 By contrast, in the slow pathway, the

DsbA-DsbB covalent bond is retained until the removed electrons have been passed on to the electron

transport chain.267 In both proposed mechanisms, the formation of an interdomain disulfide (Cys130-

Cys41) in DsbB is favoured by the strongly electron-withdrawing Cys41, and this in turns promotes the

interaction of a Cys44 thiolate with UQ (Figure 1.9 B).262–265

While the DsbB-UQ binding site is yet to be fully characterised, available structures suggest that face-

on interactions occur between the quinone ring, the TM1 C-terminus and the TM2 N-terminus, with the

UQ isoprenyl chains extending into a TM1-TM4 groove (Figure 1.8, Figure 1.9).262,263,265,270–272 These

contacts are stabilised by the guanidinium group of an essential Arg48 (P1 loop) and the TM2 helix

xi Cys104-SH, Cys130-SH, Cys41-Cys44 / Cys104-SH, Cys130-Cys41, Cys44-SH / Cys104-Cys130, Cys41-SH, Cys44-SH

64

dipole moment.262 Generally, the proximity of the DsbB Cys41-Cys44 disulfide to the quinone ring

promotes a nucleophilic attack between the Cys44 thiolate and UQ and results in the formation of the

DsbB-UQ charge transfer complex that provides a direct electron gradient between DsbA and the

electron transport chain (Figure 1.9 B).247,262,264,265,270,273

The mechanism behind de novo disulfide bond formation depends on the binding state of DsbB. The

slow pathway proceeds via a cascade of nucleophilic substitution reactions that start with the attack of

Cys33(DsbA) and terminate with the attack of Cys41(DsbB) on the Cys44-UQ complex.267 Conversely, the

rapid pathway (where DsbA is released prior to P2 loop rearrangement) depends on the re-instatement

of the Cys41-Cys44 disulfide mediated by the Cys44-UQ adduct (Figure 1.9 D).267 Experimental evidence

by Inaba et al. suggests that the rapid pathway is likely preferred, in part, due to the transient tethering

of Cys130(DsbB) in an intermediate location between both Cys104 and Cys41, which would facilitate the

electron exchange.247,267 Irrespective of the release pathway mechanism, the DsbB disulfide cascade

results in the release of oxidised DsbA which enables further oxidative protein folding steps to take

place in the cell envelope maintaining its protein homeostasis.

1.6.2.1.2 The isomerase pathway

The oxidative pathway of the DSB system rapidly introduces disulfide bonds between consecutive

cysteine residues of cell envelope proteins as the unfolded polypeptides transverse the inner membrane.

Due to the non-discriminatory nature of DsbA-catalysed cysteine coupling, in cases where a protein has

more than two cysteine amino acids the formation of non-native disulfide bonds can occur. The

detection and correction of oxidatively misfolded proteins is carried out by the isomerase pathway of

the DSB system, which is composed of the freely diffusible isomerase, DsbC, its analogue DsbG, and

their membrane-bound partner protein, DsbD (Figure 1.10).274 The rearrangement of incorrect disulfide

linkages is crucial in preventing protein aggregation which could eventually intoxicate the bacterial

cell.241,275

65

Figure 1.10 The isomerase pathway of the DSB system. The isomerase pathway corrects proteins misfolded by DsbA

through the activity of the isomerase DsbC. The misfolded substrate binds in the wide cleft formed by the N-terminal

dimerisation domains of DsbC. Thiol-disulfide exchange reactions between the reduced cystine pairs of the C-terminal

domains of DsbC and the misfolded substrate results in the release of a reduced substrate and oxidised DsbC. The

transmembrane partner protein DsbD ensures that DsbC is present in its active, reduced form. DsbD itself is kept reduced by

cytoplasmic thioredoxins. The isomerase pathway also encompasses the DsbC homologue, DsbG. The function of this protein

is to protect single thiols from irreversible oxidation. Figure adapted from Messens & Collett and Katzen & Beckwith.241,276

1.6.2.1.2.1 DsbC

DsbC is a 46 kDa V-shaped homo-dimeric protein composed of an N-terminal dimerization domain and

a C-terminal catalytic domain containing two conserved disulfide bonds formed by two conserved pairs

of cysteine residues (Figure 1.11).274 Two monomeric DsbC subunits are held together by hydrogen

bond interactions between their N-terminal domains, each comprising a six-strand antiparallel -sheet

(−) and a single N-terminal -helix () (Figure 1.11 A).274 Interaction and binding of the two

domains is mediated by the exchange of the 4 strand of monomer 1 and the 5’ strand of monomer 2,

cytoplasm

periplasm

66

such that strands 1-4 of monomer 1 overlap with strands 5’-6’ of monomer 2 and vice versa (Figure

1.11 B).274 A cis-proline (Pro50) residue disrupts the -sheet planarity at the C-terminus and, along with

the proximal flexible -helix linker, ensures that the catalytic domain extends away from the

dimerization domain.274 This creates the opposing arms of the protein and minimises undesirable inter-

C-domain interactions between the two monomers (Figure 1.11 A and C).274 A localised 13Å twist

around one of the -helices of the catalytic domain exposes one of the C-terminal active sites to the

periplasmic environment.277

Figure 1.11 Crystal structure of the primary isomerase of the E. coli DSB system, EcDsbC. (A) (top) Cartoon

representation of the V-shaped DsbC homodimer held together by the N-terminal dimerization domains, -−. The catalytic

site is located in the extended Trx core (- −-) at the distal end of the C-terminus which is extended away from the

dimerization domain via the helix.274 (bottom) The structure of a DsbC monomer in cartoon representation. The active

cysteine residues of the C98-GY-C101 motif are shown as sticks in cyan with the sulfur atoms represented by yellow spheres.

(B) Close up view of the N-terminal dimerization domain of EcDsbC. Hydrogen bond interactions occur through the exchange

of strand of monomer 1 (pink) and the ' strand of monomer 2 (dark red); 1-4 of monomer 1 overlap with strands ’−’ of

monomer 2 and vice versa.274 (C) The electrostatic surface of EcDsbC generated using PyMol (top) and using PyMol with a

normalised hydrophobicity scale developed by Eisenberg et al. (bottom; a sliding colour scale representation shows the

hydrophilic regions in light grey/pink and hydrophobic residues in darker red.66,248 Hydrophobic residues line the wide binding

cleft formed by the N-terminal dimerization domains.274,281 PDB code: 1EEJ. Figure adapted from McCarthy et al.274

67

The cysteine-containing C-terminal domains are, in their monomeric form, structurally similar to DsbA

(Figure 1.6 A, Figure 1.11 A).274 Interestingly, early studies by Rietsch et al. and Missiakas, Schwager

& Raina showed that DsbC can partially complement DsbA activity under suppression of DsbD and in

the absence of DsbA.239,245,278 This behaviour is not observed under standard conditions, due to kinetic

and structural restrictions introduced by the dimerization domain and the -helix linker ()xii.279–281

The C-terminal domains are formed from an extended Trx core ( − and Figure 1.11 A), which

is disrupted by an -helical insertion () in the − connecting loop.242,274 This -helical insertion

is smaller in comparison to that of DsbA, something that likely prevents the formation of the

hydrophobic groove observed under the active site of DsbA.274 Instead, the wide hydrophobic cleft

formed by the -sheets of the dimerization domain enables binding of misfolded substrates, while the

two extended C-terminal domains jointly interact with their partner protein DsbD.274,281 The great

conformational flexibility in DsbC promotes interactions with a wide range of partially folded and

misfolded polypeptides and additionally supports the secondary role of DsbC as a periplasmic

chaperone.245,253,274,278,281 In further contrast to DsbA, the catalytically active disulfide bond C98XXC101

(CGYC in E. coli) of DsbC is reduced to enable interactions with disulfide bonds rather than with free

thiols.279 Stabilisation of the reduced form of DsbC has been shown to be partially supported by

interactions between the Cys98 and Thr182 sidechains.282 An additional disulfide bond, located at Cys141-

Cys163, is catalytically inactive and its loss causes increased DsbC sensitivity to denaturation, thus

showing a role in protein stability.277,283

The isomerisation process begins by substrate binding into the uncharged hydrophobic cleft of DsbC,

likely mediated by the loop residues of the Trx fold, Thr182-cPro183, which are located in an equivalent

position as the loop residues shown to affect DsbA-substrate/DsbB and DsbC/DsbD complex formation

(see section 1.6.2.1.1.1).243,246,249,252,274,281 This results in the exposure of the misfolded disulfide to the

C-terminal active site and subsequent nucleophilic attack by the essential and reactive N-terminal Cys98

of DsbC forming an intermolecular disulfide bond .245,274,278,279,283 While the mechanism of the resultant

DsbC-substrate complex resolution has yet to be conclusively established, two mutually non-exclusive

xii Research by Rozhkova et al. showed that native DsbA-DsbB or DsbC-DsbD interactions are kinetically 3-7 orders of magnitude more

favourable than cross-pathway interactions of DsbA-DsbD or DsbC-DsbB.277,292 Steric hindrance and lack of a hydrophobic groove under the

DsbC active site prevents DsbC association with DsbB, while the -helical domain of DsbA is believed to prevent interactions with

DsbD.245,247,280,295 Segatori et al. show that the absence of the -helix linker () in DsbC also results in the loss of resistance to DsbB.245,274,280

68

pathways have been proposed in literature. In similar manner to PDI isomerisation in eukaryotic cells,

resolution could occur via a nucleophilic attack from a second free substrate cysteine to the

intermolecular disulfide, thus resulting in the release of the isomerised substrate and reduced DsbC.226

Alternatively, in the second proposed mechanism, the mixed disulfide undergoes a nucleophilic attack

by Cys101(DsbC) to release a fully reduced substrate and oxidised DsbC. A second round of DsbA mediated

oxidation reaction would then be required to introduce new disulfide bonds to the now semi-folded

substrate, while oxidised DsbC would be reduced by DsbD.245,284

Experimental evidence suggests that the second, stepwise process is more likely to occur as DsbC

accumulates in its oxidised form in the absence of DsbD.279 Further, DsbC cysteine substitution studies

by Liu and Wang and Rietsch et al. suggest that both active site cysteines, Cys98 and Cys101, are involved

in the isomerisation process; the loss of Cys101 causes extensive defects in the folding of a model

urokinase substrate, a eukaryotic protease with several non-consecutive disulfide bonds.245,279,283,284 As

expected, while variations in the Cys98-XX-Cys101 motif greatly affect the catalytic activity of DsbC,

they have no effect on its stability or its role as a periplasmic chaperone.277,283

Further supporting evidence comes from evolution experiments on antibiotic resistance, where an

artificial variant of an isomerisation-dependent TEM -lactamase was studied in dsbC mutants of

Legionella pneumophila.245,285 Presence of multiple non-consecutive disulfide bonds in the TEM -

lactamase enzyme resulted in overexpression of DsbA under antibiotic selection pressure.245,285 Limited

availability of DsbB, in comparison to the amount of DsbA produced, led to accumulation of the

reduced form of DsbA in the periplasm and enabled this non-oxidising form to interact with and reduce

misfolded disulfide bonds in the -lactamase.245,285,286 In this case, isomerisation of the TEM variant

was achieved in a stepwise fashion and, given the structural similarity between monomeric DsbC and

DsbA, it is likely that the classical E. coli isomerisation route follows a similar mechanism.

Irrespective of its mechanism of action, the ability of DsbC to break incorrectly formed disulfide bonds

plays an essential role in processes even beyond new protein synthesis and folding, like copper

detoxification; the presence of copper in the periplasm results in fast and uncontrolled catalysis of

disulfide bond formation.278,287 Although the ability of DsbC to identify misfolded substrates is not yet

fully understood, its activity is key to maintaining proper oxidative folding of numerous components of

the cell envelope.

69

1.6.2.1.2.2 DsbG

In addition to correcting wrongly-formed disulfide linkages catalysed by indiscriminate DsbA folding,

the isomerase pathway plays a key role in bacterial survival under oxidative stress conditions.287,288 The

presence of reactive oxygen species catalyses the formation of unwanted disulfide bonds as well as the

oxidation of thiols into sulfenic, sulfinic and sulfonic acids.288 While DsbC reverses the former, its

ability to reduce fully oxidised cysteine residues is limited. As the conversion into sulfinic and sulfonic

acids results in terminal protein misfolding and necessitates protease-mediated degradation, it is

important that the intermediate sulfenic acids are promptly reduced to thiols.245 Reduction of these side

chains is driven by DsbG, whose action safeguards key periplasmic mechanisms such as the cell wall

synthesis; DsbG prevents the loss of catalytic single cysteine residues in several transpeptidase

enzymes, such as YbiS and ErfK.241,287,288

DsbG, a V-shaped homo-dimeric protein with monomers composed of an N-terminal dimerization

domain and C-terminal catalytic domain, exhibits many structural similarities to DsbC (Figure 1.12).282

Primary differences arise from variation in size, as well as the polarity of the wide cleft which is formed

by the interacting N-terminal dimerization domains.282 In DsbG the substrate binding cleft is lined with

several acidic residues and an additional highly conserved, charged groove is formed by the −

sheets; together these polar traits promote the binding of the oxidised single cysteine residuesxiii.282 The

Trx-like C-terminal domains of DsbG are extended further away from the dimerization domain, in

comparison to DsbC.282 They carry the catalytically active Cys109-XX-Cys112 motif (Cys109-Pro-Tyr-

Cys112 in E. coli) and it has been shown that substitutions of the two amino acid dipeptide between the

cysteines result in stabilisation of the DsbG-substrate complex which prevents its resolution and he

release of reduced substrate (Figure 1.12).241,288

xiii This contrasts with DsbC, where the binding cleft is purely hydrophobic in nature.

70

Figure 1.12 Crystal structure of E. coli DsbG, EcDsbG, rendered in cartoon representation. Two homodimers form a V-

shaped homo-dimeric protein through interactions at the -sheets of the N-terminal dimerization domains (−) in a EcDsbC-

like fashion. Yellow spheres represent the sulfur atoms of the active cysteine residues. PDB code: 1V58. Figure adapted from

Heras et al.66,282

Together the isomerase proteins of the DSB system exhibit enough structural and functional flexibility

to allow a comprehensive protection of periplasmic cysteine residues under oxidative stress conditions.

Their contribution to a several periplasmic protein modifications/processes is in stark contrast to the

more restricted DsbA-DsbB oxidative pathway, which only catalyses introduction of disulfide bonds,

albeit to more substrates than the isomerase pathway.

1.6.2.1.2.3 DsbD

In the highly oxidative environment of the periplasm, disulfide bond isomerisation depends on the

activity of the reduced forms of DsbC or DsbG. This is ensured by a transmembrane protein, DsbD,

which was initially identified for its role in the reduction of CcmGxiv and which transfers reductive

equivalents across the inner membrane.265,289 DsbD is a 59 kDa protein composed of three domains.

Two of these domains are in the periplasm and are formed by the N- and C-terminal ends of the protein

(nDsbD and cDsbD, Figure 1.13 and Figure 1.14 D). These two soluble domains flank a third

transmembrane domain (tmDsbD).290,291 Each domain carries an essential pair of cysteine residues that,

xiv CcmG is an enzyme responsible for cytochrome c maturation.

71

Figure 1.13 Crystal structure of the reduced state of N-terminal domain of E. coli, nDsbD, rendered in cartoon

representation. Cysteine residues are shown as blue sticks or yellow spheres. (A) The Ig fold of nDsbD is formed by two

antiparallel -sheets, 1/2/8/5 and 4/9/12, with the active-site cysteines located in the randomly coiled region perpendicular to the

Ig fold. (B) A close-up view of the active site of nDsbD. The catalytically active cysteines are protected by a Glu69-Phe-Tyr-

Gly-Lys73, and particularly by the Phe70 and Tyr71 aromatic sidechains (blue). On the right a partially rendered electrostatic

surface of the cap residues shows the steric hindrance effects caused by the protective aromatic ring moieties which limit

access to the active site in absence of DsbC/DsbG. PDB code: 3PFU.66,297

72

Figure 1.14 Crystal structures of EcDsbC - nDsbD and cDsbD - nDsbD complexes elucidate the mechanism behind the

transfer of reductive potential through the periplasmic subunits of DsbD. (A) Crystal structure of the DsbC Cys101Ser –

nDsbD Cys103Ala complex; DsbC is shown in pink with reactive sulfur shown as yellow spheres and nDsbD is shown in orange

with its reactive sulfur moiety shown in orange spheres. Extensive conformational changes in the cap-loop region of nDsbD,

enable Cys98(DsbC) and Cys109(nDsbD) to come into contact with the cysteine residues of substrate proteins like DsbC/G, in order

to transfer reductive potential; this eventually leads to the formation of an intermolecular disulfide in nDsbD. (B) A close-up

view of the active site of nDsbD Cys103Ser - cDsbD Cys464Ser complex showing the proximity of Cys109(nDsbD) and Cys98(DsbC)

prior to formation of the disulfide bond. PDB code: 1JZD. (C) Crystal structure of nDsbD Cys103Ser - cDsbD Cys464Ser shows

the electron transfer step between the two periplasmic domains occurring through an interdomain disulfide between Cys109

and Cys461 (orange spheres). PDB code: 1SE1. (D) Crystal structure of reduced cDsbD in cartoon representation. PDB code:

2FWH. Figure adapted from Rozhkova et al., Haebel et al., and Stirnimann et al.66,281,292,299

73

through progressively increasing redox potentials, mediate the transfer of reductant from the cytoplasm

to the periplasm via a series of thiol-disulfide exchange reactionsxv.265,292 Co-operation between the three

domains of DsbD enables the movement of electrons from cytoplasmic thioredoxins to periplasmic

substrates.293 The exact mechanism of electron transfer is not fully understood due to the lack of

structural data on tmDsbD.276 Nonetheless, crystal structures of the two periplasmic domains along with

other biochemical evidence suggest that significant conformational changes occur both in DsbD and its

partners, DsbC and DsbG, in order for this electron cascade to take place.265,291,294

nDsbD has an immunoglobulin-like (Ig) fold composed of two antiparallel -sheets ( and 4/9/12),

something quite uncommon in redox-active enzymes (Figure 1.13 A).281,295–297 The catalytic C103-X5-

C109 motif is located in a -sheet perpendicular to the Ig fold ( ) and is protected by a cap

structure formed by the flexible Glu69-Phe-Tyr-Gly-Lys73 loop, and especially by the sidechains of Phe70

and Tyr71 (Figure 1.13 B).281,292 The interaction of oxidised DsbC/G with reduced nDsbD leads to

extensive conformational changes in the cap-loop region of nDsbD, enabling electrostatic and hydrogen

bonding interactions with the cPro183(DsbC) that lead to the opening of the cap loop and exposure of the

active site of nDsbD (Figure 1.13 A vs Figure 1.14 A).281

Binding and recognition of oxidised DsbC/G in turn leads to the formation of an intermolecular disulfide

bridge between Cys98(DsbC) and Cys109(nDsbD) (Figure 1.14 A and B).281,295,296 The DsbC-DsbD complex is

rapidly resolved by a second nucleophilic attack from Cys103(nDsbD) and reduced DsbC is promptly

released.293,296 Electrons from the nDsbD Cys103-Cys109 disulfide bond are then transferred to the Trx-

like cDsbD domain (Figure 1.14 C and D). Rapid formation and breakdown of an inter-domain Cys109-

Cys461 complex shows the favourability of this disulfide exchange reaction and results in a Cys461-Cys464

disulfide in cDsbD (Figure 1.14 C).292,298 Interestingly, direct reduction of DsbC by cDsbD can also

occur, but happens almost five times slower than the nDsbD catalysed process.292

The last steps of the cascade require the interaction of cDsbD with tmDsbD followed by the interaction

of tmDsbD with cytoplasmic thioredoxins. tmDsbD, predicted to be composed of 8 membrane-inserted

helices, carries two conserved cysteine residues, Cys163 (TM1) and Cys285 (TM4).276,292,294,300,301

Formation of the Cys163-Cys285 disulfide is counteracted by productive interaction with cytoplasmic

xv nDsbD is also known as DsbD, cDsbD is also known as DsbD, tmDsbD is also known as DsbD

74

thioredoxin (TrxA) which leaves Cys163 and Cys285 in their reduced thiol form.290,293,296 These steps

depend extensively on conformational changes that are yet to be fully elucidated.291 Characterisation of

a functional homolog of DsbD, CcdA, which comprises six transmembrane helices, homologous to

TM1-6 of DsbD, provides an insight into this process. Recently, an NMR structure of Thermus

thermophilus CcdA has suggested that this protein uses an elevator-type transport mechanism.265,302,303

Rotational movement of the transmembrane helices results in conformational shift between an inward

and an outward state protecting the inner membrane integrity and enabling the vertical movement of

the active cysteines across the inner membrane.265,302 Interaction with the cytoplasmic thioredoxins is

mediated by the Cys20(CcdA) of TM1 (equivalent to Cys163(DsbD)) while Cys127(CcdA) (equivalent to

Cys285(DdsbD)) interacts with periplasmic substrates.265,290,302,304,305

Along with glutathione, DsbD is the only other source of reductant for the E. coli cell envelope. Its

ability to transfer reductant obtained from the cytoplasmic compartment to periplasmic substrates, is

key to several periplasmic processes.275,290 For example, the loss of DsbD results in a pleiotropic

phenotype, which includes increased susceptibility to benzylpenicillin and temperature sensitivity.239

Overall, the role of the DSB isomerase pathway extends beyond the isomerisation of disulfide bonds

performed by DsbC, to the synthesis and maturation of c-type cytochromes via DsbD/CcmG, or the

protection of single-cysteine residues from oxidation via DsbG.275,290

1.6.3 Polymorphisms of the Gram-negative DSB system

The prototypical E. coli DSB system described here, was long thought to be conserved across Gram-

negative bacterial species. However, bioinformatic studies by several groups unveiled great diversity in

the DSB protein players.215,216,306 Interestingly, DSB polymorphisms are often found in bacterial species

that are human pathogens, such as N. meningitidis, P. aeruginosa or Salmonella enterica.307–309 Many

of these species encode an extended set of DSB proteins, some of which are thought to fold specialised

substrates.215,216

The archetypical E. coli DsbA (EcDsbA) is a promiscuous catalyst that introduces disulfide bonds into

any polypeptide with more than one cysteine residue. However, significant substrate specificity has

been observed in the additional DsbA components of species with extended DSB systems, like the

oxidative pathway of the causative agent of cerebrospinal meningitis, N. meningitidis. N. meningitidis

encodes three DsbA proteins, two of which are membrane bound lipoproteins (NmDsbA1 and 2) and a

75

third one that is a soluble periplasmic protein (NmDsbA3).216,255,310 NmDsbA1 and 2 are responsible for

the formation of pili through the folding of PilQ, which in turn ensures successful binding and uptake

of extracellular DNA.255 While both of these proteins can complement the activity of EcDsbA to a

certain degree, NmDsbA2 has been shown to be unable to restore strain motility suggesting that the two

proteins have distinct catalytic profiles.218,246,255

Another example of extended DSB systems has been described for E. coli strains associated with extra-

intestinal infections, such as uropathogenic E. coli (UPEC), which often encode a second pair of

functional redox enzymes, EcDsbL (EcDsbA-like) and EcDsbI (EcDsbB-like), alongside the classical

oxidative pathway.218,311 The DsbL/I pair is responsible for the folding of a single periplasmic substrate,

arylsulfate sulfotransferase (ASST), an enzyme linked to urinary tract colonisation.216,311 Notably, DSB

system variations are not limited to chromosomally resident enzymes. S. enterica serovars, for example,

carry a EcDsbA-like enzyme, SeSrgA, encoded on their virulence plasmid in addition to the classical

DSB proteins and the EcDsbL-EcDsbI pair.216,312 This DsbA-like protein is responsible for the folding

of structural subunits of PefA fimbrae.216,312

1.6.4 Gram-positive bacteria

While disulfide bonds are formed in the periplasmic space of Gram-negative bacteria, Gram-positive

species lack this cellular compartment. In its absence, the peptidoglycan-teichoic/mycolic-acid cell wall

of Gram-positive bacteria creates a periplasm-like environment suitable for disulfide bond

formation.11,236,313,314 Unlike the archetypical DSB system of E. coli, the formation of these bonds in

Gram-positive organisms, is not carried out by a conserved protein pathway but is rather species-

dependent.

A well-studied disulfide bond formation pathway in Gram-positive bacteria is that of the MtDsbA-

VKOR oxidation system in M. tuberculosis, where a membrane-anchored MtDsbA catalyses disulfide

formation.236,315 Re-oxidation of this oxidase is mediated by an analogue of human vitamin K epoxide

reductase, VKOR, often encoded directly next or fused to its MtDsbA protein partner.236,315 This is

interesting in light of the fact that unlike VKOR, DsbB is specific to prokaryotic cells.236 In addition, a

secondary pair of oxidoreductase enzymes, DsbE and DsbF, has been identified in M.

tuberculosis.236,316–318 However, unlike the primary pathway, these proteins are not essential for M.

tuberculosis survival and their function remains unclear.236,316–318

76

Other examples in Gram-positive bacteria include the MdbA-VKOR functional pairs identified in

Actinomyces and Corynebacterium diphtheriae or the BsBdb proteins of B. subtilis.236 Notably, the

four-protein Bdb system in B. subtilis closely resembles the oxidative pathway of E. coli, with BsBdbA

and BsBdbD being EcDsbA-like and BsBdbB and BsBdbC being similar to EcDsbB-like proteins.236,319

In other species, such as S. aureus or Listeria monocytogenes, disulfide formation occurs through

EcDsbA orthologues that appear to be functionally independent of other enzymes.216,320 Despite

numerous studies of disulfide bond formation in Gram-positive species, no isomerisation pathway has

been identified to date, pointing to a more limited disulfide bond formation system or to a lack of

proteins with more than two cysteines in these organisms.

1.6.5 Targeting bacterial pathogenicity through inhibition of the DSB system and oxidative

folding

While polymorphisms in the DSB proteins across the bacterial phylogeny are not yet fully characterised,

the conserved function of archetypical DSB proteins, which fold hundreds of proteins in the cell

envelope, has been well established. In the case of pathogenic species, DsbA is essential for the stability

and folding of numerous virulence factors, including secretion systems, toxins and pili that are required

for bacterial pathogenicity.215,216,236,246,251 Most notably, some human pathogens, such as Shigella

flexnerii or P. aeruginosa, depend on oxidative protein folding for intracellular survival, pointing to a

more generalized role of the DSB system in the survival of bacteria in their environmental niches.309,321

These observations open the avenue of abrogating the dangers posed by bacterial pathogens through the

use of the DSB system as a target for broad-acting anti-virulence strategies that do not impair bacterial

viability.216,236,306,322–324

Several strategies for targeting bacterial virulence through inhibiting the DSB system have been

proposed in the literature. These include both DsbA and DsbB as targets, and can be broadly divided

into three approaches: a) fragment-based discovery for targeting DsbA or DsbB, b) development of

peptide and peptidomimetic inhibitors of DsbA and c) high-throughput screening methods for DsbB

inhibitors.322,325–327 Whilst several chemical inhibitors have been identified for both DsbA and DsbB,

no clinically suitable compounds have been developed to date.

77

1.6.5.1 Fragment-based discovery of small molecule inhibitors against DsbA and DsbB

Fragment-based drug discovery is a commonly used technique that allows the screening of a library of

compounds against a pharmaceutical target of interest. Historically, membrane-embedded proteins such

as EcDsbB were unsuitable for this technique as it required large quantities of purified and solubilised

target protein.322,326,328 Früh et al. have bypassed these limitations by employing a target-immobilized

NMR screening (TINS) strategy. With this technique, EcDsbB was solubilised in detergent micelles

prior to its immobilisation onto a resin support. Normalisation of the results enabled the identification

of several candidate compounds with two distinct modes of action, compounds that were perturbing the

DsbB-UQ binding or molecules that were disrupting the binding of EcDsbB to both UQ and EcDsbA

(Figure 1.15).326,328 The work of Früh et al., and in particular Compound 2 (Figure 1.15 A), was used

for structure-activity-relationship (SAR)-based optimisation studies by Halili et al. and resulted in the

development of Compound 19.326,328,329 These molecules represent the first active inhibitors against the

DSB system. Though, the exact mode of action of Compound 19 is unconfirmed, Halili et al. show that

this UQ derivative binds a reduced cysteine residue and covalently inhibits either EcDsbA and/or

EcDsbB.329 Further, and more importantly, this work confirmed the possibility of developing

prokaryote-specific anti-DSB compounds and gave rise to further DSB anti-virulence efforts.329

Figure 1.15 Fragment based screening identified the first active inhibitors of the DSB system.326,328,330 Compound 2 and

19 by identified by Halili et al. and Adams et al. target the transmembrane EcDsbB while compound 40 acts as a competitive

inhibitor of EcDsbA.326,328,330 These small molecules confirmed the possibility of inhibiting the DSB system without affecting

cell viability and provide the a platform for further drug design.

78

Simultaneously to Halili et al., Adams et al. used fragment-based discovery to identify EcDsbA specific

inhibitors.330 Positive hits were narrowed down to a single halogen-substituted phenyl thiazole

compound which was investigated further using SAR studies and resulted in the isolation of a

competitive inhibitor binding within the hydrophobic groove below the EcDsbA active site (Compound

40, Figure 1.15).330 Chemical treatment using this molecule resulted in the inhibition of cell motility but

did not affect bacterial viability, thus confirming the second key consideration of anti-virulence

therapies.324,326,330

More recently Totsika et al. expanded on this work to consider DSB system inhibition in pathogens

with multiple copies of DsbA.322 Use of phenylthiophene- and phenoxyphenol-based inhibitors of

EcDsbA resulted in more variable and less pronounced motility effects in UPEC (EcDsbA and EcDsbL)

and S. typhimurium (SeDsbA and SeSrgA) strains.322 This highlighted the potential challenges ahead

for the development of broad-acting inhibitors against the DSB pathway.

1.6.5.2 Designing peptide and peptidomimetic inhibitors of DsbA

A different approach for achieving DsbA inhibition was taken by Duprez et al. who took advantage of

our in-depth structural understanding of the mechanism leading to DsbA-DsbB complex formation,

supported by high-resolution structures, in an attempt to develop peptide-based DsbA inhibitors.327,331

A synthetic peptide, ‘Pro-Phe-Ala-Thr-Cys-Asp-Ser’, that mimics the P2 loop of DsbB was used to

further the understanding behind DsbA-DsbB complex formation and was confirmed to bind in the

hydrophobic groove of DsbA.327 Targeted design led to the development of the DsbA-specific peptide

‘Pro-Trp-Ala-Thr-Cys-Asp-Ser’.327 SAR studies showed that the cysteine residue was critical for

binding suggesting that covalent peptides may be suitable and potent candidates for DSB inhibition.327

Interestingly, this peptide was also active against the Proteus mirabilis DsbA (PmDsbA).332

Virtual screening of a peptidomimetic library, using the ‘Pro-Trp-Ala-Thr-Cys-Asp-Ser’ as a template,

yielded additional scaffolds suitable for DsbA inhibition. In comparison to the rational drug design, like

the strategy described above, this method identified several small molecule fragments that acted as weak

non-covalent inhibitors (Figure 1.16).331 Though not investigated further, these peptide-like scaffolds

could benefit from mimicking the now known protein-protein interactions, while evading the issues

associated with many peptide-based therapeutics, including poor oral bioavailability, low stability or

high toxicity.326,331

79

Figure 1.16 Peptidomimetic library screening identified EcDsbA inhibitors. Scaffold design took advantage of our in-

depth knowledge of the DsbA-DsbB interaction, gleaned from multiple high-resolution structures, and is based on a ‘Pro-Trp-

Ala-Thr-Cys-Asp-Ser’ template. These preliminary inhibitors open up the possibility of SAR modifications that could results

in improved specificity, bioavailability, and stability - issues commonly observed with peptide-based therapeutic

candidates.326,331

1.6.5.3 High throughput screening for small molecule inhibitors of DsbB

Cell-based high-throughput screening methods have allowed the identification of in vivo DSB system

inhibitors from large libraries of compounds. 250, 000 potential inhibitor compounds were screened

using a chromogenic -galactosidase assay.325,333 Here, the usually cytoplasmic -galactosidase enzyme

was targeted to the periplasm where it became oxidised by DsbA.325 Introduction of a non-native

disulfide bond inhibits enzymatic function and prevents X-Gal hydrolysis.325 Absence, or inhibition of

the DSB system, rescues -galactosidase activity and results in blue pigment formation.325

A first study resulted in the identification of several EcDsbB inhibitors, and their further development

led to the identification of Compound 12 (Figure 1.17).325 The activity of this compound was specific

to Gram-negative bacteria, and inhibition of the VKOR of M. tuberculosis or the human PDI were not

observed.325 Notably, Compound 12 displayed different inhibition levels for DsbB enzymes from

several Gram-negative pathogens such as A. baumannii, S. enterica, or P. aeruginosa.325 Further

investigation as part a second study showed that species-specific inhibition of P. aeruginosa and M.

tuberculosis is possible through targeted development, and resulted in Compounds PA1 and MT17,

respectively (Figure 1.17).333

80

Figure 1.17 Inhibitors of the DsbA partner proteins, DsbB and VKOR. Described in two studies by Landeta et al., inhibitor

Compounds 12 and PA1 are specific to the Gram-negative EcDsbB and PaDsbB1, while MT17 showed selective activity

against the Gram-positive MtVKOR.325,333 In all cases, no cross-activity was observed with the human PDI enzyme showing

the potential for the development of prokaryotic and species-specific inhibitors of the DSB system.325,333

It should be noted that despite the development of functional DsbB inhibitors by Landeta et al.,

successful abrogation of oxidative folding in vivo will most likely require inhibitors targeting

DsbA.325,327,333 This arises from the fact that DsbA re-oxidation due to molecular oxygen or the dipeptide

cystine, which is abundant in most biological fluids, can occur spontaneously, albeit slowly, in the

absence of DsbB, and thus partially mitigate the inhibitory effects of DsbB inhibition.334 This would be

evaded through development of efficient DsbA inhibitors, as these exogenous oxidants cannot

efficiently oxidise the substrates of DsbA.

Therapeutic approaches targeting non-essential bacterial pathways are believed to be useful in

decreasing the chance of resistance development as well as minimise deleterious effects on the host’s

natural microbiota.82,216,236,306,322–324 To date, several strategies targeting bacterial virulence have been

proposed but were limited by narrow activities and high target specificities and their use would likely

depend on the development of complex drug cocktails.82A desirable alternative would be to inhibit a

conserved broad-acting pathway, such as the DSB system in Gram-negative bacteria. The combined

work of Totsika et al., Kurth et al. and Landeta et al. identified several inhibitor compounds with activity

against both EcDsbA and EcDsbB-like proteins.322,325,327,329–333 Their work confirms the potential behind

the use of the DSB system as a target for both narrow and broad-spectrum anti-virulence agents that do

not inhibit bacterial viability, and at the same time highlights several key challenges behind this

81

approach. It is clear that to target the DSB system with the purpose of abrogating bacterial virulence, a

good understanding of its inherent polymorphisms, which are particularly prominent in pathogenic

organisms, is required.216 In addition, if species-specific DSB inhibitors are more promising, their

deployment into clinical practice will hinge on the ability to rapidly detect and identify the bacterial

species responsible for the infections of interest.

1.7 AIMS OF THIS WORK

The Gram-negative DSB system has long been perceived as a non-essential house-keeping-only system.

More recently, this system has been implicated in bacterial pathogenesis due to its role in folding and

safeguarding the stability and function of many key virulence factors across different bacterial species.

For example, oxidative folding, catalysed by the primary oxidase DsbA, has been shown to be essential

for the function of V. cholerae toxins, the type III secretion in S. enterica and P. aeruginosa as well as

the formation of fimbriae in uropathogenic E. coli.309,335–337

While the role of the DSB system proteins in bacterial virulence is now broadly accepted, and chemical

inhibitors of the DSB proteins have even been generated with the hope to abrogate virulence, very little

is known about the contribution of the DSB proteins to mechanisms of antimicrobial resistance, despite

the fact that several of these mechanisms are protein-based, contain more than one cysteine residue, and

are located in the cell envelope. Thus, the aims of this work are to:

1. Investigate the effects of loss of DsbA on antibiotic resistance mediated by cysteine-containing

mobile class D OXA-type -lactamase enzymes.

2. Investigate the role of DsbA in intrinsic antibiotic resistance conferred by cysteine-containing

chromosomally-resident -lactamase enzymes.

3. Characterise the relationship between efflux-pump mediated resistance and cell envelope

protein homeostasis safeguarded by the DSB proteins.

4. Understand the role of disulfide bonds in the mutational evolution capacity of extended-

spectrum -lactamase enzymes.

82

2 MATERIALS AND METHODS

2.1 REAGENTS AND BACTERIAL GROWTH CONDITIONS

Unless otherwise stated, chemicals and reagents were acquired from Sigma Aldrich, growth media were

purchased from Oxoid and antibiotics were obtained from Melford Laboratories. Lysogeny broth (LB)

(10 g/L peptone, 5 g/L yeast extract, 10 g/L NaCl) and agar (1.5% w/v) was used for routine growth of

all organisms at 37°C with shaking at 220 RPM, except for S. maltophilia strains which were routinely

grown in low-salt LB (Luria protocol, 10 g/L peptone, 5 g/L yeast extract, 0.5 g/L NaCl).

Mueller-Hinton (MH, 300 g/L beef dehydrate infusion, 17.5 g/L casein hydrolysate, 1.5 g/mL starch)

broth and agar (1.5% w/v) were used for Minimum Inhibitory Concentration (MIC) assays. Growth

media were supplemented with the following, as required: 0.25 mM Isopropyl β-D-1-

thiogalactopyranoside (IPTG) (for strains harbouring β-lactamase-encoding pDM1 plasmids), 12.5

g/mL tetracycline, 33 g/mL chloramphenicol, 100 g/mL ampicillin, 50 g/mL kanamycin, 10

g/mL gentamicin, 50 g/mL streptomycin (for cloning purposes), 2000-6000 g/mL streptomycin (for

the construction of P. aeruginosa mutants), and 6000 g/mL streptomycin (for the construction of a S.

maltophilia mutant).

2.2 GENETIC MANIPULATION TECHNIQUES

2.2.1 Genomic DNA extraction, purification of plasmid DNA and of PCR products

Extraction of genomic DNA was carried out from a single bacterial colony harbouring the β-lactamase

gene of interest. The colony was resuspended in 100 L of molecular-biology-grade H2O and boiled at

100°C for 10 minutes. The lysed cells were centrifuged on a tabletop centrifuge (13 500 rpm, 3 min)

and the DNA-containing supernatant was used as template for PCR amplification.

Purification of plasmid DNA was carried out using the QIAprep Spin Miniprep Kit (Qiagen), according

to the manufacturer’s instructions. Briefly, overnight cultures were centrifuged to remove the culture

media and resultant cell pellets were resuspended in the P1 Resuspension Buffer containing a RNase I.

83

Subsequently, the alkaline P2 Lysis Buffer was added to lyse the cells. Addition of the N3 Neutralisation

Buffer followed by a 10-minute centrifugation at 13, 000 rpm was used to separate the cellular debris,

including membranes, proteins and genomic DNA, from the plasmid DNA. Supernatants were applied

to QIAquick silica membrane spin columns allowing binding of plasmid DNA. Bound plasmid DNA

was washed with Buffer PB and ethanol-based Buffer PE and then eluted into molecular-biology-grade

H2O.

Purification of PCR products was carried out using the QIAquick PCR Purification Kit PCR according

to the manufacturer’s instructions. Briefly, amplified DNA was mixed with Buffer PB, in a 1:5 ratio,

and applied to a QIAquick silica membrane spin column. The column was then washed with Buffer PE

and eluted in molecular-biology-grade H2O.

Purification of restriction digestion products was carried out using the QIAquick Gel Extraction Kit

(Qiagen) according to the manufacturer’s instructions. Briefly, DNA molecules were separated using

agarose gel electrophoresis and the fragment of interest was excised using a clean scalpel. The gel piece

was dissolved in Buffer QG in 1:3 w/v ratio at 50°C and mixed with 1 gel volume of isopropanol. The

solution was applied to a QIAquick silicon membrane spin column which was washed with Buffers PB

and PE and the DNA was eluted using molecular-biology-grade H2O.

2.2.2 PCR amplification

KOD DNA polymerase (Merck) was used for all the cloning steps described in this thesis. Generally,

the PCR reactions were carried out using the following reaction mixture composition: 5 L 10X buffer,

5 L dNTP mix (2mM of each base) mix, 3 L 2mM MgCl2, 1 L KOD DNA polymerase, 1.5 L of

each primer (in an appropriate dilution), 2 L genomic DNA or 1 L plasmid DNA. The reaction

mixture was made up to 50 L with molecular-biology-grade H2O. For cloning of genes with high %GC

content 5 L DMSO was included in the reaction mixture. For cloning using genomic DNA from

Pseudomonas or Stenotrophomonas isolates, 5 L betaine was added.

84

Standard thermo-cycling conditions used were as follows:

1. Hot start denaturation - 95°C, 2 minutes, 1 cycle

2. Denaturation - 95°C, 30 seconds, 30 cycles

3. Annealing - 50-65°C, 20 seconds, 30 cycles

4. Extension - 70°C, 20 seconds per amplified kb, 30 cycles

5. Final extension - 95°C, 5-10 minutes, 1 cycle

6. Storage - 4°C, up to 14 hours

OneTaq DNA polymerase (New England BioLabs) was standardly used for PCR amplification during

colony screening for the construction of dsbA1 and dsbA1 dsbL1 mutants of P. aeruginosa and S.

maltophilia clinical isolates, respectively. These reactions were carried out using the following reaction

mixture composition: 7.5 L OneTaq DNA Polymerase, 2.5 L betaine, 0.75 L of each primer (in

appropriate dilution), 1 L genomic DNA or 1 colony. The reaction mixture was made up to 25 L with

molecular-biology-grade H2O. Standard thermo-cycling conditions were used as described above,

except for with increased extension time of 1 minute per amplified kb.

For both PCR reaction setups, the annealing temperature and time, as well as the extension time were

adapted as required for specific genes and primer combinations.

2.2.3 Agarose gel electrophoresis

Agarose gel slabs were made in TAE buffer (40 mM Tris-acetate (pH 8.0), 1 mM EDTA) by adding 1

% agarose and 50 nL/ml SybrSafe while DNA samples were added into 6 x gel loading dye (10 mM

EDTA, 50 % v/v glycerol, 0.5 % bromophenol blue). Electrophoresis was performed in TAE buffer at

a constant current of 70 mA. DNA bands were revealed by SybrSafe fluorescence under UV light.

2.2.4 Restriction digestion

DNA restriction digestions were standardly performed with high-fidelity restriction endonuclease

enzymes (New England Biolabs) and the following reaction mixture composition: 5 L CutSmart®

buffer, 1000 ng DNA template, 1 L of each restriction digestion enzyme. The reaction mixture was

made up to 50 L with molecular-biology-grade H2O. Reactions were incubated at 37°C for 2 hours,

85

analysed using agarose gel electrophoresis and purified before further use. Where high fidelity enzymes

were not available, stepwise digestion with standard restriction digestion enzymes followed by gel

extraction purification after every step was performed.

2.2.5 Ligation

DNA ligations were standardly performed at 16°C overnight using T4 DNA ligase (New England

Biolabs) and the following reaction mixture composition: 5 L 1x T4 DNA Ligase Reaction Buffer, 0.2

L T4 DNA ligase, 50 ng vector, DNA insert at a 5:1 ratio to the vector. The reaction mixture was made

up to 20 L with molecular-biology-grade H2O. Ligation reactions were stored at 4°C and used directly

for transformation into competent cells.

2.2.6 Site-directed mutagenesis

Site-directed mutagenesis was performed using the QuickChange site-directed mutagenesis method

(Stratagene) as per the manufacturer’s instruction manual. Briefly, KOD DNA polymerase was used to

amplify plasmid DNA using QuickChange primers and the following reaction mixture composition: 5

L 10X buffer, 5 L dNTP mix (2mM of each base), 3 L 2mM MgCl2, 1 L KOD DNA polymerase,

125 ng per primer, 1/5/10 ng plasmid DNA. The reaction mixture was made up to 50 L with molecular-

biology-grade H2O.

Standard thermo-cycling conditions used were as follows:

1. Hot start - 95°C, 2 minutes, 1 cycle

2. Denaturation - 95°C, 30 seconds, 18 cycles

3. Annealing - 55°C, 20 seconds, 18 cycles

4. Extension - 70°C, 3-5 minutes, 18 cycles

5. Final extension - 95°C, 10 minutes, 1 cycle

6. Storage - 4°C, up to 14 hours

Following thermal cycling, the methylated template DNA was digested by addition of 1 L Dpn I (New

England Biolabs) and one-hour incubation at 37°C. Resultant amplified plasmids were directly

86

transformed into chemically competent E. coli DH5, selected on antibiotic-supplemented LB, and

confirmed by DNA sequencing.

2.2.7 DNA sequencing

All the plasmids used in this work were sequenced to confirm that only the desired genes had been

incorporated. Plasmid DNA Sanger sequencing was carried out externally by Eurofins Genomics.

2.2.8 Preparation and transformation of chemically competent cells

E. coli strains to be made chemically competent were grown in Super Optimal Broth (SOB, 20 g/L

peptone, 5 g/L yeast extract, 584 mg/L NaCl, 184 mg/L KCl, 2.03 g/L MgCl2, 2.44 g/L MgSO4)

overnight, sub-cultured 1:100 in SOB and grown to optical density at 600 nm (OD600) 0.45. The cultures

were incubated on ice for 30 minutes before centrifugation at 4°C and 2 000 x g for 15 minutes. Cell

pellet was resuspended in 66 mL of Buffer RF1 (12 g/L RbCl, 9.9 g/L MnCl2•H2O, 2.95 g/L K(OAc),

1.5 g/L CaCl2•H2O, 150 g/L glycerol; pH 5.8 with AcOH; filter sterilised) and incubated on ice for 1

hour. Cell pellets obtained from a second centrifugation step were resuspended in 16 mL of Buffer RF2

(2.09 g/L, MOPS, 1.2 g/L RbCl, 11 g/L CaCl2•H2O, 150 g/L glycerol; pH 6.8 with NaOH; filter

sterilised) and incubated on ice for 15 minutes. The cell suspension was aliquoted in 220 L aliquots

and frozen on liquid N2.

E. coli DH5 chemically competent cells were routinely used for transformations for cloning and site-

directed mutagenesis purposes. E. coli MC1000, MC1000 dsbA, and MC1000 dsbA atTn7 dsbA

competent cells were used for transformation with plasmids carrying -lactamase genes for experiments

described in Chapters 3, 4, and 6. In both cases the following steps were performed: 10 L of ligation

reaction mixture / 1 L of plasmid DNA / 50 ng linear DNA were added to 50 L of cells and incubated

on ice for 15 minutes. Cells were then subjected to a 30-45 second heat shock at 42 °C and cooled on

ice for 2 minutes. After addition of 900 L SOB, a one-hour recovery step was carried out at 37°C

before transformants were selected on antibiotic-supplemented LB agar overnight.

87

2.2.9 Preparation and transformation of electrocompetent cells

Overnight cultures of cells to become electrocompetent were centrifuged at 4°C and 4000 x g for 10

minutes in order to remove culture media. Cell pellets were then washed three times with ice-cold sterile

MQ H2O, resuspended in 100 L ice-cold sterile MQ H2O and used immediately. 2 L of plasmid or

linear DNA were routinely electroporated into E. coli MG1655, Pseudomonas or Stenotrophomonas

strains in 1mm electroporation cuvettes at 25 FD, 200 , and 1.5 V. Cells were recovered in 900 L

SOB at 30°C for 3 hours or at 37°C for 1 hour as required. Transformants were selected for on antibiotic-

supplemented LB agar overnight.

2.3 BACTERIAL STRAINS AND PLASMIDS

Bacterial strains, plasmids and oligonucleotides described in this thesis are listed in Table 2, Table 3

and Table 4, respectively.

Table 2. Bacterial strains used in this thesis. All listed strains in the “Clinical isolates / laboratory strains” section are clinical

strains except for P. aeruginosa PAO1LA and P. aeruginosa PAO1LD, which are laboratory strains. FNRCAR refers to the

French National Reference Centre for Antibiotic Resistance in Le Kremlin-Bicêtre, France. CRBIP stands for Centre de

Ressources Biologiques de l’Institut Pasteur, France.

Name Description Source

Escherichia coli

DH5α F– endA1 glnV44 thi-1 recA1 relA1

gyrA96 deoR nupG purB20

φ80dlacZ∆M15 ∆(lacZYA-argF)U169

hsdR17(rK–mK

+) λ–

344

CC118λpir araD Δ(ara, leu) ΔlacZ74 phoA20 galK

thi-1 rspE rpoB argE recA1 λpir 345

HB101 supE44 hsdS20 recA13 ara-14 proA2

lacY1 galK2 rpsL20 xyl-5 mtl-1 346

MC1000 araD139 ∆(ara, leu)7697 ∆lacX74

galU galK strA 347

MC1000 dsbA dsbA::aphA, KanR 251

MC1000 dsbA attTn7::Ptac-dsbA dsbA::aphA attTn7::dsbA, KanR 338

MG1655 K-12 F– λ– ilvG– rfb-50 rph-1 348

MG1655 dsbA dsbA::aphA, KanR This study

MG1655 dsbA attTn7::Ptac-dsbA dsbA::aphA attTn7::dsbA, KanR This study

MG1655 acrA acrA 338

88

MG1655 tolC tolC 338

MG1655 degP degP::strAB, StrR This study

MG1655 marR marR::accC, GentR This study

MG1655 dsbA marR dsbA::aphA marR::accC, KanR, GentR This study

Clinical isolates / laboratory strains

Pseudomonas aeruginosa SOF-1 blaOXA-4 349

Pseudomonas aeruginosa PU21 blaOXA-10 350

Pseudomonas aeruginosa PAe191 blaOXA-19 351

Pseudomonas aeruginosa PA43417 blaOXA-198 352

Pseudomonas aeruginosa PA43417 dsbA1 dsbA1 blaOXA-198 This study

Pseudomonas aeruginosa 51170 blaBEL-1 353

Pseudomonas luteola CIP 102067 blaLUT-1 CRBIP

Pseudomonas aeruginosa G4R7 blaAIM-1 FNRCAR

Pseudomonas aeruginosa G4R7 dsbA1 dsbA1 blaAIM-1 This study

Pseudomonas aeruginosa G6R7 blaAIM-1 FNRCAR

Pseudomonas aeruginosa G6R7 dsbA1 dsbA1 blaAIM-1 This study

Stenotrophomonas maltophilia GUE blaL2-1 blaL1-1 354

Stenotrophomonas maltophilia GUE dsbA dsbL dsbA dsbL blaL2-1 blaL1-1 This study

Pseudomonas otitidis CIP 109236T blaPOM-1 CRBIP

Pseudomonas aeruginosa PA14 blaOXA-50 355

Pseudomonas aeruginosa PA14 dsbA1 dsbA1 blaOXA-50 This study

Pseudomonas aeruginosa PAO1 LA blaOXA-50 356

Pseudomonas aeruginosa PAO1 LA dsbA1 dsbA1 blaOXA-50 This study

Pseudomonas aeruginosa PAO1 LD blaOXA-50 350

Pseudomonas aeruginosa PAO1 LD dsbA1 dsbA1 blaOXA-50 This study

89

Table 3. Plasmids used in this thesis.

Name Description Source

pDM1 pDM1 vector (GenBank MN128719), p15A ori, Ptac

promoter, lacI, MCS, TetR

Mavridou

lab

pDM2 pDM1 derivative, p15A ori, BioFab promoter, MCS, TetR Mavridou

lab

pDM1-blaL2-1 blaL2-1 cloned into pDM1, TetR 338

pDM1-blaOXA-4 blaOXA-4 cloned into pDM1, TetR This study

pDM1-blaOXA-10 blaOXA-10 cloned into pDM1, TetR This study

pDM1-blaOXA-198 blaOXA-198 cloned into pDM1, TetR This study

pDM1-blaBEL-1 blaBEL-1 cloned into pDM1, TetR This study

pDM1-blaBPS-1m blaBPS-1m cloned into pDM1, TetR This study

pDM1-blaCARB-2 blaCARB-2 cloned into pDM1, TetR This study

pDM1-blaFTU-1 blaFTU-1 cloned into pDM1, TetR This study

pDM1-blaLUT-1 blaLUT-1 cloned into pDM1, TetR This study

pDM1-blaAIM-1 blaAIM-1 cloned into pDM1, TetR This study

pDM1-blaPOM-1 blaPOM-1 cloned into pDM1, TetR This study

pDM1-blaSMB-1 blaSMB-1 cloned into pDM1, TetR This study

pDM1-blaL1-1 blaL1-1 cloned into pDM1, TetR 338

pDM1-blaOXA-50 blaOXA-50 cloned into pDM1, TetR This study

pDM1-StrepII-blaOXA-4 blaOXA-4 encoding OXA-4 with an N-terminal StrepII tag

cloned into pDM1, TetR This study

pDM1-blaOXA-10-StrepII blaOXA-10 encoding OXA-10 with a C-terminal StrepII tag

cloned into pDM1, TetR This study

pDM1-blaOXA-198-StrepII blaOXA-198 encoding OXA-198 with a C-terminal StrepII

tag cloned into pDM1, TetR This study

pDM1-blaBEL-1-StrepII blaBEL-1 encoding BEL-1 with a C-terminal StrepII tag

cloned into pDM1, TetR This study

pDM1-blaBPS-1m-StrepII blaBPS-1m encoding BPS-1m with a C-terminal StrepII tag

cloned into pDM1, TetR This study

pDM1-blaCARB-2-StrepII blaCARB-2 encoding CARB-2 with a C-terminal StrepII tag

cloned into pDM1, TetR This study

pDM1-blaFTU-1-StrepII blaFTU-1 encoding FTU-1 with a C-terminal StrepII tag

cloned into pDM1, TetR This study

pDM1-blaLUT-1-StrepII blaLUT-1 encoding LUT-1 with a C-terminal StrepII tag

cloned into pDM1, TetR This study

pDM1-blaAIM-1-StrepII blaAIM-1 encoding AIM-1 with a C-terminal StrepII tag

cloned into pDM1, TetR This study

pDM1-blaPOM-1-StrepII blaPOM-1 encoding POM-1 with a C-terminal StrepII tag

cloned into pDM1, TetR This study

pDM1-blaSMB-1-StrepII blaSMB-1 encoding SMB-1 with a C-terminal StrepII tag

cloned into pDM1, TetR This study

pDM1-blaL2-1-StrepII blaL2-1 encoding L2-1 with a C-terminal StrepII tag cloned

into pDM1, TetR 338

pDM1-blaL1-1-StrepII blaL1-1 encoding L1-1 with a C-terminal StrepII tag cloned

into pDM1, TetR 338

90

pDM1-blaOXA-50-StrepII blaOXA-50 encoding OXA-50 with a C-terminal StrepII tag

cloned into pDM1, TetR This study

pDM2-blaSHV-1 blaSHV-1 cloned into pDM2, TetR This study

pDM2-blaSHV-1 C54A blaSHV-1 C54A cloned into pDM2, TetR This study

pDM2-blaTEM-1 blaTEM-1 cloned into pDM2, TetR This study

pDM2-blaTEM-1 C86A blaTEM-1 C86A cloned into pDM2, TetR This study

pGRG25-Ptac::dsbA

Ptac::dsbA fragment cloned within the Tn7 of pGRG25;

when inserted into the chromosome and the plasmid

cured, the strain expresses DsbA upon IPTG induction,

AmpR

338

pSLTS Thermosensitive pSC101ori, ParaB for λ-Red, PtetR for

I-SceI, AmpR 339

pUltraGFP-GM

Constitutive sfGFP expression from a strong Biofab

promoter, p15A ori, (template for the accC cassette),

GentR

357

pKNG101 Gene replacement suicide vector, oriR6K, oriTRK2,

sacB, (template for the strAB cassette), StrR 342

pKNG101-dsbA-PA

PCR fragment containing the regions upstream and

downstream P. aeruginosa dsbA1 cloned in pKNG101;

when inserted into the chromosome the strain is a

merodiploid for dsbA1 mutant, StrR

338

pKNG101-dsbA dsbL-SM GUE

PCR fragment containing the regions upstream and

downstream S. maltophilia GUE dsbA and dsbL genes

cloned in pKNG101; when inserted into the chromosome

the strain is a merodiploid for dsbA dsbL mutant, StrR

This study

pRK600 Helper plasmid, ColE1 ori, mobRK2, traRK2, ClR 358

pCB112 CPRG membrane integrity assay vector, lacIq Plac ::lacZ ,

MCS, ClR 359

pMK-RQ bps-1m GeneArt® cloning vector containing bps-1m, ColE1 ori,

(template for bps-1m), KanR This study

pMK-RQ carb-2 GeneArt® cloning vector containing carb-2, ColE1 ori,

(template for carb-2), KanR This study

pMK-T ftu-1 GeneArt® cloning vector containing ftu-1, ColE1 ori,

(template for ftu-1), KanR This study

pMK-RQ smb-1 GeneArt® cloning vector containing smb-1, ColE1 ori,

(template for smb-1), KanR This study

pMQ-RQ shv-1 GeneArt® cloning vector containing shv-1, ColE1 ori,

(template for shv-1), KanR This study

pMK-T tem-1 GeneArt® cloning vector containing tem-1, ColE1 ori,

(template for tem-1), KanR This study

91

Table 4. Oligonucleotide primers used in this thesis. The “Brief description” column provides basic information on the

primer design (restriction enzyme used for cloning, encoded protein or gene replaced by antibiotic resistance cassette, forward

or reverse orientation of the primer (F or R); QC stands for QuickChange primers and SQ stands for sequencing primers).

Number Brief description Sequence (5ˊ-3ˊ)

P1 SacI.OXA-4.F ctggagctcaaaaacacaatacatatcaacttcgc

P2 KpnI.OXA-4.R cagggtaccttataaatttagtgtgtttagaatggtg

P3 SacI.OXA-10.F ctggagctcaaaacatttgccgcatatgtaattatcgc

P4 KpnI.OXA-10.R cagggtaccttagccaccaatgatgccctc

P5 SacI.OXA-18.F ctggagctccaacggagcctgtccatga

P6 KpnI.OXA-18.R cagggtacctcagaagttttccgacagggc

P7 NdeI.OXA-198.F actgcatatgcataaacacatgagtaagctcttc

P8 KpnI.OXA-198.R ctgggtaccttattcgatcccctttgctt

P9 SacI.BEL-1.F ctggagctcaaactgctctacccgttattgc

P10 PstI.BEL-1.R cagctgcagtcagtgaacatattgacgtgc

P11 SacI.BPS-1m.F ctggagctcaatcattctccgttgcgccgctc

P12 XmaI.BPS-1m.R caacccgggtcaggcgaacgcccgcgcg

P13 SacI.CARB-2.F ctggagctcaagtttttattggcattttcgc

P14 KpnI.CARB-2.R cagggtacctcagcgcgactgtgatgta

P15 SacI.FTU-1.F ctggagctccgtctattagttacaactttatc

P16 XmaI.FTU-1.R ctgcccgggttatttataagtgttagtcagatc

P17 SacI.LUT-1.F ctggagctcaatgtcatcctgaaccgtcga

P18 PstI.LUT-1.R cagctgcagtcagcctgtcacccattcag

P19 SacI.AIM-1.F ctggagctcaaacgtcgcttcaccctgg

P20 KpnI.AIM-1.R ctgggtacctcaaggccgcgcgccgctg

P21 SacI.POM-1.F ctggagctccgtaccctgaccctcg

P22 KpnI.POM-1.R cagggtaccttatgcgtcatcagagacctc

P23 NdeI.SMB-1.F cagctccatatgaaaatcatcgcttccctgatcc

P24 XmaI.SMB-1.R ctgcccgggtcagcgtttctcgctggcca

P25 SacI.OXA-50.F ctggagctccgccctctcttcagtg

P26 KpnI.OXA-50.R cagggtacctcagggcagtatcccgagag

P27 QC.BPS-1m.F gaattcgcccttctgcccgggtcagcgtttc

P28 QC.BPS-1m.R gaaacgctgacccgggcagaagggcgaattc

P29 QC.FTU-1.F acccgggctatttttcaaattgcggatggctcc

P30 QC.FTU-1.R ggagccatccgcaatttgaaaaatagcccgggt

P31 QC.LUT-1.F ttgggtgacatgctggctgcggataa

P32 QC.LUT-1.R ttatccgcagccagcatgtcacccaa

P33 KpnI.StrepII.OXA-4.R cagggtaccttatttttcaaattgcggatggctccaagcgct

ccctaaatttagtgtgtttagaatggtgatc

92

P34 signalseq.StrepII.OXA-4.body.R tgcaacagtagagatatctgttgatttttcaaattgcggatggctccaagcgctcc

ctgcactggcgctgctgta

P35 OXA-4.body.F tcaacagatatctctactgttgca

P36 KpnI.StrepII.OXA-10.R cagggtaccttatttttcaaattgcggatggctccaagcgctcccgccaccaatg

atgccctcacttg

P37 KpnI.StrepII.OXA-18.R cagggtaccttatttttcaaattgcggatggctccaagcgctcccgaagttttccg

acagggcc

P38 KpnI.StrepII.OXA-198.R ctgggtaccttatttttcaaattgcggatggctccaagcgctcccttcgatcccctt

tgcttg

P39 PstI.StrepII.BEL-1.R cagctgcagttatttttcaaattgcggatggctccaagcgctcccgtgaacatatt

gacgtgctaac

P40 XmaI.StrepII.BPS-1m.R ctgcccgggctatttttcaaattgcggatggctccaagcgctcccggcgaacgc

ccgcgcggcg

P41 KpnI.StrepII.CARB-2.R cagggtaccttatttttcaaattgcggatggctccaagcgctcccgcgcgactgt

gatgtataa

P42 XmaI.StrepII.FTU-1.R ctgcccgggctatttttcaaattgcggatggctccaagcgctccctttataagtgtt

agtcagatcattag

P43 KpnI.StrepII.AIM-1.R cagggtaccttatttttcaaattgcggatggctccaagcgctcccaggccgcgc

gccgctggag

P44 KpnI.StrepII.POM-1.R cagggtaccttatttttcaaattgcggatggctccaagcgctcccgccgcgctgc

ttc

P45 XmaI.StrepII.SMB-1.R ctgcccgggctatttttcaaattgcggatggctccaagcgctcccgcgtttctcgc

tggccag

P46 KpnI.StrepII.OXA-50.R cagggtaccttatttttcaaattgcggatggctccaagcgctcccgggcagtatc

ccgagagcc

P47 QC.SHV-1.C54A.F cacgtgccagaactgcaccagccagaacaactttaaaggtg

P48 QC.SHV-1.C54A.R cacctttaaagttgttctggctggtgcagttctggcacgtg

P49 QC.TEM-1.C86A.F cggctcagaactgcaccagccagcagaactttaaagg

P50 QC.TEM-1.C86.R cctttaaagttctgctggctggtgcagttctgagccg

P51 NotI.Ptac.EcDsbA.F

ctggcggccgctgacaattaatcatcggctcgtataatgtgtggaattgtgacta

gtcgaggtccaggacctcggatcgctaagataggatgattgtatgaaaaagattt

ggctggc

P52 XhoI.EcDsbA.R ctgctcgagttattttttctcggacagatatttc

P53 EcdsbA::aphA.F atgaaaaagatttggctggcgctggctggtttagttttagcgtttagcgcgtgtag

gctggagctgcttc

P54 EcdsbA::aphA.R ttattttttctcggacagatatttcactgtatcagcatactgctgaacaagggaatta

gccatggtccat

P55 EcdegP::strAB.F atgaaaaaaaccacattagcactgagtgcactggctctgagtttaggtttggaac

tgcacattcgggatatttctc

P56 EcdegP::strAB.R ttactgcattaacaggtagatggtgctgtcgccgcgctgaatgttgagtgccagg

ccggatctagatatctagtatga

P57 EcmarR::accC.F atggttaatcagaagaaagatcgcctgcttaacgagtatctgtctccgctggtga

agttcctatactttctagagaataggaacttcaagatcccctg

P58 EcmarR::accC.R ttacggcaggactttcttaagcaaatactcaagtgttgccacttcgtccgcgaagt

tcctattctctagaaagtataggaacttcacttactcaatggaattctagatcg

P59 SQ.dsbA1.Paeruginosa.F tacctgctcaagcagatgcatg

P60 SQ.dsbA1.Paeruginosa.R ggtgttcatgtcgcccatca

P61 XbaI.dsbA1.Paeruginosa F ggttcctctagagcctacttcgccagccagaa

P62 dsbA1.Paeruginosa.body.R ctacttcttgttacgcatcgttcactc

P63 dsbA1.Paeruginosa.body.F atgcgtaacaagaagtaggcaaggtga

93

P64 BamHI.dsbA1.Paeruginosa.R aattaaggatcctcatcactaccaccagcgcg

P65 SQ.dsbAdsbL.Smaltophilia.F atggtgccgttcgtgcaga

P66 SQ.dsbAdsbL.Smaltophilia.R acagcacctgcatttccgg

P67 XbaI.dsbAdsbL.Smaltophilia.F ggttcctctagatcttctggtacagcacctgcatttccg

P68 dsbAdsbL.Smaltophilia.body.R tgcgtgtcgatgaggttggctcactga

P69 dsbAdsbL.Smaltophilia.body.F tctcttggatcagtgagccaacctcat

P70 BamHI.dsbAdsbL.Smaltophilia.R aattaaggatcctcgctggaggtggatttcagcaagacc

P71 SQ.pKNG101.vector.upstream ccctggatttcactgatgag

P72 SQ.pKNG1010.vector.downstream catatcacaacgtgcgtgga

2.3.1 Cloning of -lactamase genes

Genes for β-lactamase enzymes were amplified from genomic DNA extracted from clinical isolates,

with the exception of bps-1, carb-2, ftu-1, smb-1, shv-1, and tem-1, which were synthesized by GeneArt

Gene Synthesis (ThermoFisher Scientific). All β-lactamase genes, except shv-1 and tem-1, were cloned

into the IPTG-inducible plasmid pDM1 using primers P1-26 (Table 4).338 All StrepII-tag fusions of

these β-lactamases (constructed using primers P37-46, Table 4) bear a C-terminal StrepII tag

(GSAWSHPQFEK), except OXA-4. The latter has an N-terminal StrepII tag inserted between the Sec

signal peptide and the body of the protein using the primers P33-36 (Table 4). Synthesized shv-1 and

tem-1 genes were designed to include a StrepII tag sequence which was placed after the STOP codon.

The STOP codon was preceded by an AvrII restriction site and followed by a LLAGAVLCGAVLLSX

linker sequence. In this way, if needed, single restriction digestion and re-ligation steps on each

construct would generate plasmids expressing StrepII tagged versions of the enzymes. The synthesised

sequences were digested out of their respective cloning vectors and ligated into pDM2 (using SacI/XmaI

sites), whilst point mutants of each gene were obtained by QuickChange mutagenesis (using primers

P47-50, Table 4).

2.3.2 Generation of E. coli dsbA, degP and marR mutants

E. coli MG1655 dsbA, degP, and marR gene mutants were constructed using a pSLTS vector and a

modified lambda-Red recombination method, as previously described.339 Briefly, the pSLTS plasmid

was transformed into chemically competent E. coli and 10 mL of LB were innoculated with 100 L of

an E. coli pSLTS overnight and grown at 30°C for 1 hour. Lambda-Red recombinase expression was

induced using L-arabinose at final concentration of 2 mM and the culture was grown to OD600 0.7-0.9.

94

Cells were harvested by centrifugation (4500 x g), washed twice with ice-cold MQ H2O, and

resuspended in 100 L of ice cold MQ H2O. 50-bp DNA fragments upstream and downstream of the

dsbA, degP, marR genes were amplified from E. coli gDNA (primers P53-58, Table 4). 100 ng of the

linear mutation cassettes were electroporated into the prepared electro-competent E. coli pSLTS cells

and after 1 hour of incubation at 30°C the outgrowth was plated on kanamycin- and ampicillin-

supplemented LB agar for overnight incubation at 30°C. Western blot analysis was used to screen single

colonies and identify/confirm protein loss and gene deletion.

2.3.3 Generation of P. aeruginosa dsbA1 mutants

The dsbA1 mutants of P. aeruginosa laboratory and clinical strains (Table 2) were constructed by allelic

exchange, as previously described.341 Briefly, the dsbA1 gene area of all P. aeruginosa strains (including

the dsbA1 gene and 600 bp on either side of this gene) was amplified (primers P59-60, Table 4) and the

obtained DNA was sequenced to allow for accurate primer design for the ensuing cloning step.

Subsequently, 500-bp DNA fragments upstream and downstream of the dsbA1 gene were amplified

using P. aeruginosa PA43417 genomic DNA (primers P61-62 (upstream) and P63-64 (downstream),

Table 4). A fragment containing both of these regions was obtained by overlapping PCR (primers P61

and P64, Table 4) and inserted into the XbaI/BamHI sites of the pKNG101 vector. The suicide vector

pKNG101342 is not replicative in P. aeruginosa; it was maintained in E. coli CC118λpir and mobilized

into the P. aeruginosa strains by triparental conjugation (see section 2.3.5).

2.3.4 Generation of the S. maltophilia dsbA1 dsbL1 mutant

The dsbA1 dsbL1 mutant of the S. maltophilia GUE clinical isolate was constructed by allelic exchange,

as previously described.341,343 Briefly, the dsbA dsbL gene area of S. maltophilia strains (including the

dsbA dsbL genes and 600 bp on either side of these genes) was amplified (primers P65-66, Table 4) and

the obtained DNA was sequenced to allow for accurate primer design for the ensuing cloning step.

Subsequently, 600-800-bp DNA fragments upstream and downstream of the dsbA dsbL genes were

amplified using S. maltophilia genomic DNA (primers P67-68 (upstream) and P69-70 (downstream),

Table 4). A fragment containing both of these regions was obtained by overlapping PCR (primers P67

and P70) and inserted into the XbaI/BamHI sites of pKNG101. The suicide vector pKNG101342 is not

replicative in S. maltophilia; it was maintained in E. coli CC118λpir and mobilized into S. maltophilia

strains by triparental conjugation (see section 2.35) on low salt LB (Luria protocol).

95

2.3.5 Triparental conjugation of P. aeruginosa and S. maltophilia

Triparental conjugations of Pseudomonas or Stenotrophomonas clinical isolates was carried out using

an E. coli HB101 helper strain and an E. coli CC118pir donor strain transformed with the helper

pRK600 plasmid and the mutation-cassette-harbouring pKNG101 plasmid, respectively. 20 L of the

helper and donor strains were spotted on LB agar, both separately and on top of one another, and

incubated at 37°C for 2 hours. Concurrently, the recipient isolate to be conjugated was cultured at 43°C.

20 L of the recipient strain was then spotted on top of the combined helper + donor spot, as well as on

its own. The plate was incubated facing up at 37°C overnight. The donor + helper + recipient spot and

the individual spots were scraped using a sterile loop and resuspended into 1 mL of sterile PBS. Then

successful conjugants were selected on streptomycin-supplemented VBM (2% technical agar

supplemented with 200 mg/L MgSO4.7H2O; 2 g/L anhydrous citric acid; 10 g/L K2HPO4; 3.5 g/L

H5NNaO4P.4H2O; pH 7.0; filter-sterilised) media at 37°C for 24 -72 hours. Single colonies were

patched on LB agar supplemented with 20% w/v sucrose and LB agar supplemented with streptomycin

and incubated at room temperature for 24-48 hours. Colonies susceptible to sucrose, but resistant to

streptomycin, were screened by PCR (primers P61, 64, 67, 70, 71, 72; Table 4). PCR reactions could

yield two distinct band sizes, each representative of the direction in which the pKNG101 mutator

plasmid had incorporated into the bacterial chromosome. A colony where the pKNG101 mutator

plasmid was incorporated was selected and the less common orientation was selected and incubated at

30°C until turbidity was observed. 100 L was then plated on LB agar supplemented with 20% sucrose

and incubated at room temperature for 24-72 hours. Single colonies were patched on LB agar

supplemented with streptomycin, LB and Pseudomonas Isolation Agar (1.4 g/L MgCl2, 20 g/L peptic

digest of animal tissue, 10 g/L K2SO4, 0.025 g/L triclosan, 10 mL/L glycerol, 13.6 g/L agar), and

incubated at 37°C overnight. Colonies with good growth on LB and PIA with susceptibility to

streptomycin were screened for the absence of dsbA1 and dsbA1 dsbL1 (primers P61, 64, 67, 70; Table

4) and mutants were confirmed by DNA sequencing.

2.3.6 Complementation of E. coli MG1655 dsbA

To complement the E. coli dsbA mutant, dsbA was reintroduced into the E. coli chromosome at the

attTn7 site, as previously described using the pGRG25 plasmid carrying dsbA under a Ptac promoter

(Table 3).340 Briefly, pGRG25 plasmid was transformed into E. coli dsbA and a single transformant was

grown in LB supplemented with 0.3% w/v L-arabinose and ampicillin at 30°C overnight. The overnight

96

culture was diluted to 2 x 10-6, 20 L of this dilution was plated on LB agar and incubated at 42°C

overnight. Single colonies were sub-cultured on ampicillin-supplemented LB to check for plasmid

absence. Western blot analysis was then used to confirm dsbA complementation.

2.4 MINIMUM INHIBITORY CONCENTRATION (MIC) ASSAYS

Unless otherwise stated, β-lactam MIC assays were carried out in accordance with the EUCAST

recommendations using E-test strips (BioMérieux). Briefly, overnight cultures of each strain to be tested

were standardized to OD600 0.063 in 0.85% NaCl (equivalent to McFarland standard 0.5) and distributed

evenly across the surface of MH agar plates. E-test strips were placed on the surface of the plates, evenly

spaced, and the plates were incubated for 18-24 hours at 37°C. MICs were read according to the

manufacturer’s instructions. β-lactam and vancomycin MICs were also determined using the Broth

dilution (BD) method, as required. Briefly, a series of antibiotic concentrations was prepared by two-

fold serial dilution in MH broth in a clear-bottomed 96-well microtiter plate (Corning) or 7 mL bijou

containers. The strain to be tested was added to the wells/containers at approximately 5 x 105 colony

forming units (CFU) per well and incubated for 18-24 hours at 37°C. When 7mL bijou containers were

used, shaking of the cultures was performed at 220 RPM. The MIC was defined as the lowest antibiotic

concentration with no visible bacterial growth. All colistin sulphate (Acros Organics) MIC assays were

also performed using the BMD method as described above using MH broth supplemented with CaCl2

at a final concentration of 0.223mM. When used, tazobactam was included in broth or agar at a fixed

concentration of 4 g/mL, in accordance with the EUCAST guidelines.

The covalent DsbB inhibitor 4,5-dichloro-2-(2-chlorobenzyl)pyridazin-3-one325 was used to chemically

impair the function of the DSB system. Inactivation of DsbB results in abrogation of DsbA function360

only in media free of small-molecule oxidants, such as the dipeptide cystine.238 Therefore, MIC assays

involving chemical inhibition of the DSB system were performed using M63 broth (15.1 mM

(NH4)2SO4, 100 mM KH2PO4, 1.8 mM FeSO4.7H2O, adjusted to pH 7.4 with KOH) and agar (1.5%

w/v) supplemented with 1 mM MgSO4, 0.02% w/v glucose, 0.005% w/v thiamine, 31 µM FeCl3.6H2O,

6.2 μM ZnCl2, 0.76 µM CuCl2.2H2O, 1.62 µM H3BO3, 0.081 µM MnCl2.4H2O, 84.5 mg/L alanine, 19.5

mg/L arginine, 91 mg/L aspartic acid, 65 mg/L glutamic acid, 78 mg/L glycine, 6.5 mg/L histidine, 26

mg/L isoleucine, 52 mg/L leucine, 56.34 mg/L lysine, 19.5 mg/L methionine, 26 mg/L phenylalanine,

26 mg/L proline, 26 mg/L serine, 6.5 mg/L threonine, 19.5 mg/L tyrosine, 56.34 mg/L valine, 26 mg/L

tryptophan, 26 mg/L asparagine and 26 mg/L glutamine for E. coli strains. For P. aeruginosa strains

97

MOPS medium (40 mM 3-(N-morpholino)propanesulfonic acid, 4 mM Tricine, 9.53 mM NH4Cl, 0.5

μM CaCl2, 0.52 mM MgCl2.7H2O, 50 mM NaCl, 0.01 mM FeSO4.7H2O and 1.32 mM K2HPO4,

adjusted to pH 7.2 with KOH) and agar (1.5%) supplemented with 0.02% w/v glucose, 500 mg/L L-

glutamine, 0.03 μM (NH4)6Mo7O24.4H2O, 4 μM H3BO3, 0.3 μM CoCl2, 0.01 μM CuSO4, 0.081 μM

MnCl2.4H2O, 0.01 μM ZnSO4, and 4.2 μM NiCL2.6H2O were used. Either DMSO (vehicle control) or

the covalent DsbB inhibitor 4,5-dichloro-2-(2-chlorobenzyl)pyridazin-3-one (final concentration of 50

μM) (Enamine)325 were added to the M63 or MOPS medium, as required. The strain to be tested was

added at an inoculum that recapitulated the MH medium MIC values obtained for each strain.

2.5 SDS-PAGE ANALYSIS AND IMMUNOBLOTTING

All samples for immunoblotting were prepared as follows with the exception of samples for strains

expressing FTU-1. Strains to be tested were grown on LB agar plates as lawns in the same manner as

for MIC assays described above. Bacteria were collected using an inoculation loop and resuspended in

LB to OD600 2.0 (except for strains expressing OXA-4, where OD600 6.0 was used). The cell suspensions

were spun at 10,000 x g for 10 minutes and bacterial pellets were lysed by addition of BugBuster Master

Mix (Merck Millipore) for 25 minutes at room temperature with gentle agitation. Subsequently, lysates

were spun at 10,000 x g for 10 minutes at 4 °C and the supernatant was added to 4 x Laemmli buffer.

Samples were boiled for 5 minutes before separation by SDS-PAGE. Strains expressing FTU-1 were

grown in LB broth supplemented with 12.5 μg/mL tetracycline until OD600 0.4. FTU-1 -lactamase

expression was induced with 0.5 mM IPTG for 4 hours. Bacteria were then diluted in LB to OD600 2

and spun at 10,000 x g for 10 minutes. Bacterial pellets were lysed by addition of BugBuster Master

Mix (Merck Millipore) for 5 minutes at room temperature with gentle agitation. Subsequently, lysates

were added to 4 x Laemmli buffer and separated by SDS-PAGE.

Unless otherwise stated, SDS-PAGE analysis was carried out using 10% BisTris NuPAGE gels

(ThermoFisher Scientific) using MES/SDS running buffer prepared according to the manufacturer’s

instructions and including pre-stained protein markers (SeeBlue Plus 2, ThermoFisher Scientific).

Proteins were transferred to Amersham Protran nitrocellulose membranes (0.45 µm pore size, GE Life

Sciences) using a Trans-Blot Turbo transfer system (Bio-Rad) before blocking in 3% w/v Bovine Serum

Albumin (BSA)/TBS-T (0.1 % v/v Tween 20) or 5% w/v skimmed milk/TBS-T and addition of primary

and secondary antibodies. The following primary antibodies were used in this study: Strep-Tactin-HRP

conjugate (Iba Lifesciences) (dilution 1:3,000 in 3 w/v % BSA/TBS-T), Strep-Tactin-AP conjugate (Iba

98

Lifesciences) (dilution 1:3,000 in 3 w/v % BSA/TBS-T), rabbit anti-DsbA antibody (dilution 1:1,000

in 5 w/v % skimmed milk/TBS-T), rabbit anti-AcrA antibody (dilution 1:10,000 in 5 w/v % skimmed

milk/TBS-T), rabbit anti-TolC antibody (dilution 1:5,000 in 5 w/v % skimmed milk/TBS-T), rabbit

anti-HtrA1 (DegP) antibody (Abcam) (dilution 1:1,000 in 5 w/v % skimmed milk/TBS-T) and mouse

anti-DnaK 8E2/2 antibody (Enzo Life Sciences) (dilution 1:10,000 in 5% w/v skimmed milk/TBS-T).

The following secondary antibodies were used in this study: goat anti-rabbit IgG-AP conjugate (Sigma

Aldrich) (dilution 1:6,000 in 5% w/v skimmed milk/TBS-T), goat anti-rabbit IgG-HRP conjugate

(Sigma Aldrich) (dilution 1:6,000 in 5% w/v skimmed milk/TBS-T), goat anti-mouse IgG-AP conjugate

(Sigma Aldrich) (dilution 1:6,000 in 5% w/v skimmed milk/TBS-T) and goat anti-mouse IgG-HRP

conjugate (Sigma Aldrich) (dilution 1:6,000 in 5% w/v skimmed milk/TBS-T). Membranes were

washed three times for 5 minutes with TBS-T prior to development. Development for AP conjugates

was carried out using a SigmaFast BCIP/NBT tablet, while HRP conjugates were visualized with the

Novex ECL HRP chemiluminescent substrate reagent kit (ThermoFisher Scientific) or the Luminata

Crescendo chemiluminescent reagent (Merck) using a Gel Doc XR+ Imager (Bio-Rad).

2.6 -LACTAM HYDROLYSIS ASSAY

β-lactam hydrolysis measurements were carried out using the chromogenic β-lactam nitrocefin

(Abcam). Briefly, overnight cultures of strains to be tested were centrifugated, pellets were weighed

and resuspended in 150 L of 100 mM sodium phosphate buffer (pH 7) per 1 mg of wet-cell pellet, and

cells were lysed by sonication. Lysates were transferred into clear-bottomed 96-well microtiter plates

(Corning) and used at the following loadings: strains harbouring pDM1, pDM1-blaL2-1, pDM1-blaL1-1,

pDM1-blaOXA-10 and pDM1-blaOXA-50 (lysates equivalent to 0.34 mg of cell pellet); pDM1-blaOXA-4

(lysates equivalent to 0.2 mg of cell pellet); pDM1-blaBEL-1, pDM1-blaAIM-1 and pDM1-blaSMB-1 (lysates

equivalent to 0.17 mg of cell pellet); pDM1-blaPOM-1 (lysates equivalent to 0.1 mg of cell pellet); pDM1-

blaOXA-198 (lysates equivalent to 0.015 mg of cell pellet); pDM1-blaBPS-1m (lysates equivalent to 0.07 mg

of cell pellet); and pDM1-blaCARB-2 (lysates equivalent to 0.03 mg of cell pellet). In all cases, nitrocefin

was added at a final concentration of 400 M, and the final reaction volume was made up to 100 L

using 100 mM sodium phosphate buffer (pH 7). Nitrocefin hydrolysis was monitored at 25°C by

recording absorbance at 490 nm at 60-second intervals for 15 minutes using an Infinite M200 Pro

microplate reader (Tecan). The amount of nitrocefin hydrolysed by each lysate in 15 minutes was

calculated using a standard curve generated by acid hydrolysis of nitrocefin standards.

99

2.7 NPN UPTAKE ASSAY338

1-N-phenylnaphthylamine (NPN) (Arcos Organics) uptake assays were performed as described by

Helander & Mattila-Sandholm.361 Briefly, mid-log phase cultures of strains to be tested were diluted to

OD600 0.5 in 5 mM HEPES (pH 7.2) before transfer to clear-bottomed 96-well microtiter plates

(Corning) and addition of NPN at a final concentration of 10 M. Colistin sulphate (Arcos Organics)

was included at a final concentration of 0.5 g/mL, as required. Immediately after the addition of NPN,

fluorescence was measured at 60-second intervals for 10 minutes using an Infinite M200 Pro microplate

reader (Tecan); the excitation wavelength was set to 355 nm and emission was recorded at 405 nm.

2.8 PI UPTAKE ASSAY338

Exponentially growing (OD600 0.4) E. coli strains harbouring pUltraGFP-GM357 were diluted to OD600

0.1 in phosphate buffered saline (PBS) (pH 7.4) and cecropin A was added to a final concentration of

20 M, as required. Cell suspensions were incubated at room temperature for 30 minutes before

centrifugation and resuspension of the pellets in PBS. Propidium iodide (PI) was then added at a final

concentration of 3 M. Suspensions were incubated for 10 minutes at room temperature and analysed

on a two-laser, four colour BD FACSCalibur flow cytometer (BD Biosciences). 50,000 events were

collected for each sample and data were analysed using FlowJo v.10.0.6 (Treestar).

2.9 CPRG CELL ENVELOPE INTEGRITY ASSAY

Exponentially growing (OD600 0.4) E. coli MG1655 and MC1000 pCB112 strains were diluted to 1:105

in MH broth and plated on MH agar containing CPRG and IPTG at final concentrations of 20 g/mL

and 50 M, respectively. Plates were incubated at 37°C for 18 hours. Plates images were analysed using

Adobe Photoshop CS4 (Adobe).362 Briefly, plate images were converted to CMYK colour space format,

colonies were manually selected using consistent tolerance (26, anti-alias, contiguous) and edge

refinement (32 px, 100% contrast). Magenta colour was quantified for each image and used to assess

changes in the cell envelope integrity of the tested strains.

100

2.10 MOTILITY ASSAY338

500 μL of overnight culture of each strain to be tested were centrifuged and the pellets were washed in

M63 broth before resuspension in the same medium to achieve a final volume of 25 L. Bacterial

motility was assessed by growth in M63 medium containing 0.25% w/v agar supplemented as described

above. DMSO (vehicle control) or the covalent DsbB inhibitor 4,5-dichloro-2-(2-

chlorobenzyl)pyridazin-3-one (final concentration of 50 M) (Enamine) were added to the medium, as

required. 1 L of the washed cell suspension was inoculated into the centre of a 90 mm diameter agar

plate, just below the surface of the semi-solid medium. Plates were incubated at 37 °C in a humidified

environment for 16-18 hours and growth halo diameters were measured.

2.11 AMS LABELLING338

Bacterial strains to be tested were grown for 18 hours in M63 broth supplemented as described above.

DMSO (vehicle control) or the covalent DsbB inhibitor 4,5-dichloro-2-(2-chlorobenzyl)pyridazin-3-

one (final concentration of 50 μM) (Enamine) were added to the medium, as required. Cultures were

standardized to OD600 2.0 in M63 broth, were spun at 10,000 x g for 10 minutes and bacterial pellets

were lysed by addition of BugBuster Master Mix (Merck Millipore) for 25 minutes at room temperature

with gentle agitation. Subsequently, lysates were spun at 10,000 x g for 10 minutes at 4 °C prior to

reaction with 4-acetamido-4ˊ-maleimidyl-stilbene-2,2ˊ-disulfonic acid (AMS) (ThermoFisher

Scientific). AMS alkylation was performed by vortexing the lysates in 15 mM AMS, 50 mM Tris-HCl,

3% w/v SDS and 3 mM EDTA (pH 8.0) for 30 minutes at 25°C, followed by incubation at 37°C for 10

minutes. SDS-PAGE analysis and immunoblotting was carried out as described above, except that 12%

BisTris NuPAGE gels (ThermoFisher Scientific) and MOPS/SDS running buffer were used. DsbA was

detected using a rabbit anti-DsbA primary antibody and an AP-conjugated secondary antibody, as

described above.

2.12 BACTERIAL GROWTH ASSAY – DSBA MUTANT

Overnight cultures of the strains to be tested were centrifuged and the pellets were washed in LB broth

before transfer to clear-bottomed 96-well microtiter plates (Corning) at approximately 5 x 107 CFU/well

(starting OD600 ~ 0.03). LB broth, supplemented with antibiotics as required, was used as a growth

101

medium. Plates were incubated at 37 °C with orbital shaking (amplitude 3 mm, equivalent to ~ 220

RPM) and OD600 was measured at 900-second intervals for 18 hours using an Infinite M200 Pro

microplate reader (Tecan).

2.13 BACTERIAL GROWTH ASSAY – DSB SYSTEM CHEMICAL INHIBITOR338

Overnight cultures of the strains to be tested were centrifuged and the pellets were washed in M63 broth

before transfer to clear-bottomed 96-well microtiter plates (Corning) at approximately 5 x 107 CFU/well

(starting OD600 ~ 0.03). M63 broth supplemented as described above was used as a growth medium.

DMSO (vehicle control) or the covalent DsbB inhibitor 4,5-dichloro-2-(2-chlorobenzyl)pyridazin-3-

one (final concentration of 50 M) (Enamine) were added to the medium, as required. Plates were

incubated at 37°C with orbital shaking (amplitude 3 mm, equivalent to ~ 220 RPM) and OD600 was

measured at 900-second intervals for 18 hours using an Infinite M200 Pro microplate reader (Tecan).

2.14 IN VIVO CLEARANCE ASSAY338

The wax moth model Galleria mellonella was used for in vivo clearance assays.363 Briefly, overnight

cultures of the strains to be tested were standardized to OD600 1.0. Suspensions were centrifuged and

the pellets were washed three times in PBS and serially diluted. 10 l of a 10–5 dilution of each bacterial

suspension was injected into the last right abdominal proleg of 5 G. mellonella larvae per condition; a

second, equal-size group of larvae were injected with PBS as negative control. 3 hours after infection,

larvae were injected with 13 l of piperacillin to a final concentration of 12 g/mL in the last left

abdominal proleg. 24 hours after infection larvae were euthanized and macerated individually in 1 ml

of PBS by vortexing for 15 minutes. The larval suspension was then serially diluted and 20 l of each

dilution plated on Pseudomonas Isolation Agar. Plates were incubated at 37°C for 16 hours before CFU

counting.

102

2.15 STATISTICAL ANALYSIS OF EXPERIMENTAL DATA

For MIC assays, all recorded values were plotted; bar charts were predominantly used to display MIC

results. For experiments where 3 independent experiments were performed, the bar indicates the median

of the recorded values, whilst for experiments where 2 independent experiments were performed the

bar indicates the most conservative of the two recorded values.

For all other assays, statistical analysis was performed in GraphPad PRISM v8.3.1 using an unpaired

T-test with Welch’s correction, a one-way ANOVA with correction for multiple comparisons, or a

Kruskal-Wallis test with correction for multiple comparisons, as appropriate. Statistical significance

was defined as p < 0.05. Outliers were defined as any technical repeat >2 SD away from the average of

the other technical repeats within the same biological experiment. Such data were excluded, and all

remaining data were included in the analysis. Detailed information for each analysis is provided in the

relevant figures.

103

3 THE IMPORTANCE OF DISULFIDE BOND FORMATION FOR

THE FUNCTION OF MOBILE CLASS D -LACTAMASE

ENZYMES OF PSEUDOMONAS AERUGINOSA

3.1 INTRODUCTION

Resistance determinants encoded on mobile elements are a common trait of multi-drug resistant

organisms. The dissemination potential of such genes contributes to the increase in resistance of

pathogenic bacteria and creates a great challenge for current antibiotic therapies. Mobilizable -

lactamases are easily transferred, both among clinically significant species and across environmental

strains. Several families of these enzymes have become prevalent in some of the most concerning

bacterial pathogens, such as K. pneumoniae, P. aeruginosa or A. baumannii. These include the SHV,

CTX-M and KPC (class A -lactamases), the NDM and VIM (class B -lactamases) or the OXA (class

D -lactamases) families of enzymes.62

P. aeruginosa is a highly resistant opportunistic pathogen, primarily affecting immunocompromised

individuals. Infections result in a variety of nosocomial diseases that often have poor clinical outcomes,

including sepsis, pneumonia, urinary tract and soft-tissue infections.28,75 In addition to its virulence

traits, P. aeruginosa strains have multiple resistance mechanisms, including, for example the reduction

of membrane permeability via loss of OprD proteins, structural modifications of topoisomerase

enzymes, overexpression of efflux pumps or production of many endogenous and acquired -lactamase

enzymes belonging to different Ambler classes.28 Due to a highly flexible and diverse genome, P.

aeruginosa is uniquely adapted to retain its essential genes while it incorporates new ones that are

typically acquired in hostile environments, such as hospital surfaces.20,81 For these reasons P.

aeruginosa is considered a high priority organism for novel intervention therapies.20,81

Although a comprehensive overview of the overall prevalence of the OXA -lactamase family in P.

aeruginosa has not been carried out, several country-specific studies have noted the appearance of these

oxacillinase enzymes in hospital infections. For example, samples from 12 clinical laboratories across

South Korea in 2005 identified 252 P. aeruginosa strains with enzymes from the OXA-7 (OXA-10,

104

17), OXA-36 (OXA-2) and OXA-224 (OXA-1, 4) families.364 A more recent study in Warsaw, Poland

noted that 110 out of 900 tested strains of the pathogen were phenotypically-positive for extended-

spectrum -lactamases, with many expressing members from the OXA-7 and OXA-36 families.365 It is

notable that the enzymes identified in these two epidemiological studies originate from several

phylogenetic families, confirming their wide dissemination and importance in antibiotic resistance.

Predominantly encoded near mobilizable elements or on plasmids, class D OXA enzymes are active

serine -lactamases with a conserved catalytic serine (Ser70) and a critical carboxylated lysine residue

in their active site.366,367 With some exceptions, for example OXA-198, they can be inhibited by

clinically available compounds, such as clavulanic acid or tazobactam.62 The OXA-family enzymes

form the largest group of class D -lactamases with over 750 members, sub-divided into over 50

phylogenetic sub-families that display a diverse range of hydrolytic profiles. These often originate from

single of amino acid variations that result in high levels of resistance to all classes of -lactam

antibiotics. Commonly observed mutations include the Ser73Asn replacement, responsible for extended-

spectrum activity in the OXA-7 sub-family, or the Gly157Asp substitution leading to ceftazidime

resistance.28,125

A characteristic trait of these serine hydrolases is the Tyr144-Gly-Asn146 or Phe144-Gly-Asn146 functional

motif.352,368 This structural element, denoted as the loop, is composed of a pair of NH hydrogen-bond

donors located next to an acyl group carbonyl oxygen. The arrangement of the hydrogen-bonding

partners is known as the “oxyanion hole” and has been shown to play a critical role in stabilising the

formation of enzymatic intermediates, while it is also believed to affect the hydrolytic efficiency of the

catalytic site.352,366,369 Computed molecular simulations by Szarecka et al. and Simakov et al. suggest

that further stabilisation in this region could stem from the presence of a disulfide bridge which shapes

the highly defined loop conformational arrangement in OXA-198 or OXA-10-like enzymes.366,368

Further, they suggest that the stabilising effect of this disulfide bond could drive the evolution of new

hydrolytic functions in these enzymes.181,366

Previous work in the Mavridou lab has shown a link between DsbA activity and the ability of disulfide-

bond containing class A -lactamases, and one class B enzyme, to confer resistance to a range of -

lactam antibiotics. The importance of oxidative protein folding for class D OXA enzymes, many of

which have highly conserved disulfide bridges, was not investigated. The wide dissemination of these

-lactamases in the accessory genomes of clinically isolated P. aeruginosa strains, along with the fact

105

that resistance to classical inhibitors is often observed, offered an incentive to examine their dependence

on DsbA-catalysed folding.364,365 Four enzymes, OXA-4, 10, 18, and 198, commonly encoded by P.

aeruginosa strains causing hospital-acquired infections, were selected as representative members of

phylogenetically distinct sub-families of the OXA group of enzymes. All of these proteins have a

cysteine pair at different positions in their primary sequence, and exhibit varied hydrolytic profiles

(Supplementary Table 1).364,365 Testing the importance of DsbA for their activity, thus offers a

comprehensive overview of the role of disulfide bond formation for cysteine-containing members of

the class D OXA family.

106

3.2 RESULTS

3.2.1 Deletion of dsbA substantially decreases -lactamase mediated antibiotic resistance

in E. coli MC1000

The selected -lactamase enzymes, OXA-4, 10, 18, and 198, were expressed in the E. coli MC1000 K-

12 strain and its isogenic dsbA mutant. A panel of -lactam antibiotics was selected for each enzyme

according to its known hydrolytic capabilities, and MIC values were recorded. An aminoglycoside

antibiotic, gentamicin, which cannot be neutralised by -lactamase enzymes, as well as strains

harbouring the pDM1 empty vector or expressing the class A -lactamase L2-1, were included as

negative controls. The class A β-lactamase L2-1 from S. maltophilia contains three cysteine residues,

but lacks a disulfide bond (PDB ID: 5NE1).137 The reason for this is that it is transported to the periplasm

pre-folded via the Tat pathway, rather than by the Sec system, and thus is not a DsbA substrate.189

The absence of DsbA resulted in a substantial decrease in MIC values for all tested enzymes across

multiple -lactam compounds (Figure 3.1). In addition, the observed effects were specific to the tested

resistance proteins and their interaction with DsbA and not a result of a general inability of the dsbA

mutant to resist antibiotic stress; no decreases in MIC values were recorded for the aminoglycoside

antibiotic gentamicin nor strains carrying the disulfide-free -lactamase L2-1 or the pDM1 empty vector

(Figure 3.1).

107

Figure 3.1 Antimicrobial resistance mediated by OXA-type -lactamases depends on disulfide bond formation. β-lactam

minimum inhibitory concentration (MIC) values for E. coli MC1000 expressing class D disulfide-bond-containing β-

lactamases from P. aeruginosa are substantially reduced in the absence of DsbAxvi. No changes in MIC values are observed

for the aminoglycoside antibiotic gentamicin (white bars, MIC fold changes: < 2) confirming that absence of DsbA does not

compromise the general ability of this strain to resist antibiotic stress. Further, no changes in MIC values are observed for

strains harbouring the empty vector control (pDM1) or those expressing a class A β-lactamase L2-1, which contains three

cysteines but no disulfide bond (PDB ID: 5NE1; top row). Graphs show MIC fold changes (fold change is defined as MC1000

MIC (µg/mL) / MC1000 dsbA (µg/mL),) for β-lactamase-expressing E. coli MC1000 and its dsbA mutant and represent three

independent experiments; black dotted lines indicate an MIC fold change of 2.

It should be noted that although the overall fold changes in MIC values recorded for the extended-

spectrum -lactamase OXA-18 were not higher than 2, the cut off below which we considered observed

effects insignificant, MIC values for this enzyme in the absence of DsbA decreased by 500 or even

1000 MIC points for the compounds tested (Supplementary Table 2). This suggests that its hydrolytic

activity is affected in the absence of disulfide bond formation. Nonetheless, this -lactamase is

extremely efficient in hydrolysing all of the tested -lactam antibiotics (MIC values of 1000-2000

xvi Where broth dilution experiments were required (pDM1-blaOXA-4, pDM1-blaOXA-10, pDM1-blaOXA-18, pDM1-blaOXA-198,) Ampicillin was used

instead of amoxicillin. Both antibiotics are 2nd generation amino-benzyl penicillins but have varying salt-solubility properties. Amoxicillin hydrate is insoluble at the concentrations required for broth dilution experiments. The more soluble amoxicillin sodium salt is markedly more

expensive, and thus unsuitable for use in large-scale experiments. Ampicillin trihydrate is soluble at the required concentrations and its

structural similarity to amoxicillin makes it a good substitute for our purposes.

108

μg/mL for tested -lactams in the wild-type background), and even in the absence of DsbA the recorded

MIC values ranged from 500 to 1000 g/mL. Since MIC values of such magnitude are of little relevance

to antibiotic treatment in clinical settings, this enzyme was not investigated further.

For the remaining three class D enzymes, the wild-type MIC values were restored for representative

antibiotics by re-insertion of the dsbA gene at the attTn7 site of the E. coli MC1000 dsbA chromosome

(Figure 3.2).

Figure 3.2 Complementation of dsbA restores the β-lactam MIC values for E. coli MC1000 expressing class D β-

lactamases. Re-insertion of dsbA at the attTn7 site of the E. coli MC1000 chromosome restores the β-lactam MIC values for

E. coli MC1000 dsbA harbouring pDM1-blaOXA-4 (cefuroxime MIC), pDM1-blaOXA-10 (aztreonam MIC), pDM1-blaOXA-198

(imipenem MIC). Graphs show MIC values (µg/mL) and represent two independent experiments.

109

3.2.2 Deletion of dsbA does not affect the integrity of the cell envelope in E. coli MC1000

The gentamicin, pDM1 and pDM1-blaL2-1 controls described above (Figure 3.1), showed that the effects

on the recorded MIC values were specific to the interaction of the -lactamases with DsbA. In addition

to these tests, and since DsbA assists the folding of hundreds of cell envelope proteins, a series of

experiments were performed to assess whether altered outer-membrane permeability or cell-envelope

integrity of the dsbA mutant strain could confound our results.215,251

The permeability of the outer membrane of the dsbA mutant was measured using the hydrophobic

fluorescent dye 1-N-phenylnaphthylamine (NPN), and was found to be no different to that of the

parental strain (Figure 3.3 A).338,361 Additionally, a vancomycin sensitivity assay was used to confirm

this result. Vancomycin is a bactericidal antibiotic targeting cell wall synthesis. Outer membrane porins

of Gram-negative bacteria are too small to allow the passage of large glycopeptides, such as

vancomycin, across the membrane and, thus sensitivity to this antibiotic is only observed in Gram-

positive bacteria.370 Any perturbation to the integrity of the outer membrane in a dsbA mutant, would

thus enable the passage of vancomycin to the periplasmic space and result in major decreases in MIC

values. With the MIC values of MC1000 dsbA being comparable to those of the parental strain, it was

concluded that DsbA absence does not result in any outer membrane defects (Figure 3.3 B), consistent

with the data reported by Denoncin et al.370

110

Figure 3.3 Deletion of dsbA has no effect on outer membrane permeability in E. coli MC1000. (A) The bacterial outer

membrane acts as a selective permeability barrier to hydrophobic molecules. Deletion of dsbA has no effect on the outer

membrane integrity of E. coli MC1000, as the hydrophobic fluorescent dye NPN crosses the outer membrane of E. coli

MC1000 and its dsbA mutant to the same extent. Conversely, exposure to the outer-membrane-permeabilizing antibiotic

colistin results in a significant increase in NPN uptake. Graph shows means ± SD, significance is indicated by *** = p < 0.001,

ns = non-significant. Significance was determined using one-way ANOVA with Bonferroni’s multiple comparison test; n=3;

6 degrees of freedom; F value=39.22; p=0.0007 (significance), p=0.99 (non-significance). (B) No major differences are

observed in the MIC values for the aminoglycoside antibiotic vancomycin confirming that the outer membrane of E. coli

MC1000 dsbA is not compromised, and hence blocks the entry of the antibiotic. For both experiments n=3; NPN assay data is

courtesy of Dr R. Christopher D. Furniss (APPENDIX II).338

The integrity of the entire cell envelope of the dsbA mutant was also assessed using the fluorescent dye

propidium iodide (PI). PI is a cationic hydrophilic dye that fluoresces upon intercalation with nucleic

acids. Under normal conditions PI freely crosses the outer membrane but is unable to cross the inner

membrane.357 The cell envelope integrity of the dsbA mutant was found to be no different to that of the

parental strain (Figure 3.4 A). Further, a chromogenic -galactosidase assay based on the substrate

CPRG was used to confirm these results.359 CPRG is excluded from the cytoplasm by the cell envelope,

and therefore its hydrolysis by the cytosolic -galactosidase is prevented. If both the inner and outer

membranes are compromised, release of -galactosidase results in CPRG breakdown and the

appearance of red colour. The red colouration of the dsbA mutant colonies was comparable to those of

the parent strain showing that the cell envelope is not compromised in the mutant strain (Figure 3.4 B).

111

Figure 3.4 Deletion of dsbA does not result in damage to the bacterial inner membrane cell envelope. (A) No difference

in basal PI uptake is seen between E. coli MC1000 and its dsbA mutant when both strains express superfolder GFP (sfGFP).

Fluorescence was used to distinguish live from dead cells. Addition of the inner-membrane-permeabilizing antimicrobial

peptide cecropin A371 to E. coli MC1000 induces robust inner-membrane permeabilization in the sfGFP-positive population

indicating compromised inner membrane. Graph shows means ± SD, significance is indicated by *** = p < 0.001, ns = non-

significant. Significance determined using one-way ANOVA with Bonferroni’s multiple comparison test; n=3; 6 degrees of

freedom; F value=61.84; p=0.0002 (significance), p=0.99 (non-significance). (B) No difference in the red colouration of E.

coli MC1000 and its dsbA mutant colonies is seen, suggesting that CPRG is excluded from the colonies to the same extent.

This confirms that the integrity of the cell envelope is not compromised in the mutant strain. Images of plates were converted

to the CMYK colour space in Adobe Photoshop CS4. Colonies were selected using the magic wand tool with consistent

tolerance and edge refinement, and their magenta levels were compared. Graph shows means ± SD, significance is indicated

by ns = non-significant. Significance determined using unpaired T-test with Welch’s correction; n=3; 4 degrees of freedom; t-

value=0.1136 p=0.9150 (non-significance). PI uptake data is courtesy of Dr R. Christopher D. Furniss (APPENDIX II).338

3.2.3 Deletion of dsbA does not affect the viability of E. coli MC1000

Bacterial growth of the E. coli MC1000 strains was assessed under standard conditions to determine

whether DsbA absence causes any significant growth defects that could potentially confound our

results. No differences in growth were observed for the dsbA mutant strain, when compared to its wild-

type counterpart (Figure 3.5).

112

Figure 3.5 Deletion of dsbA does not have drastic effects on the growth of E. coli MC1000. Growth curves of E. coli

MC1000 and its dsbA mutant show that bacterial growth remains largely unaffected by the absence of DsbA. SD is marked by

the light blue shaded area in each graph. Significance determined using unpaired T-test with Welch’s correction; n=3; 4 degrees

of freedom; t-value=3.049; p=0.0381 (significance). Courtesy of Dr R Christopher D Furniss (APPENDIX II).338

3.2.4 Class D -lactamases misfold in absence of DsbA

To understand the underlying mechanism resulting in the decreased MIC values observed for the dsbA

mutant strains, -lactamase protein levels were assessed by immunoblotting. When expressed in the

dsbA mutant all of the tested class D enzyme levels remained unchanged (Figure 3.6 A), in a similar

fashion to the amount of the control enzyme L2-1, which also remained unaffected when DsbA was

absent (Figure 3.6 B).

113

Figure 3.6 Class D -lactamase enzyme levels remain unaffected by the absence of DsbA. (A) Protein levels for the

cysteine-containing β-lactamases OXA-4, OXA-10, and OXA-198 are not affected when expressed in E. coli MC1000 dsbA.

OXA-4 is detected as two bands at ~ 28 kDa. (B) The amount of the control enzyme L2-1, which contains three cysteines but

no disulfide bonds, remains unaffected in the absence of DsbA.137,189 Protein levels of StrepII-tagged β-lactamases were

assessed using a Strep-Tactin-AP conjugate (OXA-10 and OXA-198) or a Strep-Tactin-HRP conjugate (OXA-4).

Representative blots from three independent experiments are shown. Molecular weight markers (M) are on the left, DnaK was

used as a loading control and solid black lines indicate where the membrane was cut.

To assess the activity of the tested -lactamase enzymes when they lacked their disulfide bonds,

nitrocefin hydrolysis assays were performed. Nitrocefin is a chromogenic cephalosporin with a -

lactam ring that can be easily hydrolysed. The opening of this key pharmacophore leads to a detectable

bathochromic shift from =390 nM to =486 nM.

Direct comparison of -lactamase lysates of E. coli MC1000 and its dsbA mutant showed that the

hydrolytic activity of these β-lactamases was significantly decreased in absence of DsbA (Table 5),

suggesting a folding defect that leads to loss of function. E. coli MC1000 carrying the pDM1 empty

vector showed minimal nitrocefin hydrolysis, consistent with the lack of an enzyme that can perform

-lactam ring breakdown. Both E. coli MC1000 and its dsbA mutant expressing the disulfide-free L2-

1 enzyme showed high levels of nitrocefin hydrolysis, but the concentration of hydrolysed nitrocefin

did not vary (Table 5). These results are consistent with the MIC assays (Figure 3.1). Notably, the

hydrolytic activity of -lactamases expressed in the dsbA deletion strain remained higher than that of

the empty vector. Thus, in the absence of DsbA, the -lactamases appear to be partially misfolded, and

as a result not as effective as when they have undergone oxidative folding in the parental strain, but not

entirely inactive; the fact that their protein levels do not decrease in the absence of DsbA is likely crucial

to basal activity maintenance (Figure 3.6). Together the MIC and functional data show that for class D

114

Table 5 The hydrolytic activities of tested β-lactamase enzymes are significantly decreased in the absence of DsbA. The

hydrolysis of the chromogenic β-lactam nitrocefin by cysteine-containing class D β-lactamases is impaired when these

enzymes are expressed in E. coli MC1000 dsbA. The hydrolytic activities of strains harbouring the empty vector or expressing

the control enzyme L2-1, containing three cysteines but no disulfide bond, show no dependence on DsbA. The “Enzyme

stability” column informs on the stability of each enzyme when lacking its disulfide bond; this was inferred from the

immunoblotting experiments presented in Figure 3.6. The “Nitrocefin hydrolysis” column shows the amount of nitrocefin

hydrolysed per 1 mg of bacterial cell pellet in 15 minutes. n=3, table shows means ±SD, significance is indicated by * = p <

0.05, ns = non-significant. Significance was determined using unpaired T-test with Welch’s correction; n=3; 3.417 degrees of

freedom, t-value=0.3927, p=0.7178 (non-significance) (for pDM1 strains); 2.933 degrees of freedom, t-value=0.3296,

p=0.7639 (non-significance) (for pDM1-blaL2-1 strains); 2.011 degrees of freedom, t-value=6.825, p=0.0205 (significance) (for

pDM1-blaOXA-4 strains); 2.005 degrees of freedom, t-value=6.811, p=0.0208 (significance) (for pDM1-blaOXA-10 strains); 2.025

degrees of freedom, t-value=5.629, p=0.0293 (significance) (for pDM1-blaOXA-198 strains).

Strain (MC1000) Enzyme stability Nitrocefin hydrolysis†

pDM1

dsbA pDM1 -

3.57±4.40

4.51±3.15ns

pDM1-blaL2-1

dsbA pDM1-blaL2-1 not degrading

56.17±4.90

57.83±3.73ns

pDM1-blaOXA-4

dsbA pDM1-blaOXA-4 not degrading

96.93±20.22

17.16±1.05*

pDM1-blaOXA-10

dsbA pDM1-blaOXA-10 not degrading

1420±3.11

1059.78±91.67*

pDM1-blaOXA-198

dsbA pDM1-blaOXA-198 not degrading

790.75±137.07

343.90±10.78* †nM.mg-1 pellet.15 min-1

cysteine-containing AMR determinants, absence of DsbA leads to reduced levels of enzymatic activity

resulting in the inability of these enzymes to confer resistance.

3.2.5 DsbA is a tractable target for class D -lactamases

Given the role of DsbA in folding of many virulence factors, the inhibition of the DSB system has been

proposed as a promising anti-virulence approach and some efforts have been made to develop inhibitors

for DsbA, its redox partner DsbB or both (see also section 1.6.5).216,236,306,322,325,327,329,331 These studies

have made the first steps towards the production of chemical compounds that inhibit the function of the

DSB proteins and provided a laboratory tool to test our DsbA inhibition strategy against AMR.

4,5-dichloro-2-(2-chlorobenzyl)pyridazin-3-one, termed “Compound 12” in Landeta et al., is a potent

laboratory inhibitor of E. coli DsbB and its analogues from closely related organisms.325 It was

developed through further optimization of lead compounds discovered in an adapted, disulfide-sensitive

115

-galactosidase assay in an effort to target DsbB. In particular, the reaction of the chromogenic substrate

X-Gal (5-bromo-4-chloro-3-indolyl--d-galactopyranoside) with the native form of -galactosidase

results in the formation of blue pigment called 5,5’-dibromo-4,4’-dichloro-indigo. In this case, -

galactosidase-MalF fusions were used to ensure enzyme translocation into the periplasm. Since -

galactosidase contains cysteines, DsbA acts on it in the cell envelope and the introduction of non-native

disulfide bonds leads to enzyme deactivation and loss of the blue pigment. Compounds capable of

inhibiting the DSB system by blocking the function of DsbB, were identified through the re-appearance

of blue pigmentation to the otherwise white colonies.325

Although the reoxidation of DsbA is primarily carried out by DsbB, in the absence of DsbB function

this process can also occur via small-molecule oxidants, like oxygen and cystine. The presence of

cystine in most laboratory media, including Mueller-Hinton (the gold standard for MIC measurments),

interferes with chemical inhibition of the DSB system, since DsbB-independent DsbA re-oxidation can

occur. For this reason, defined cysteine-free M63 media was used for any assays involving the chemical

inhibitor of DsbB. Furthermore, the bacterial load was adjusted for E. coli MC1000 so that the same

MIC values as the ones recorded in MH media were achieved in M63 media. Using these adjustments,

the effect of chemical inhibition of the DSB system on the activity of class D OXA enzymes was probed.

Before conducting MIC experiments, the ability of Compound 12 to chemically inhibit the function of

the DSB system was established. First, the motility of E. coli MC1000 in the presence of this compound

was tested, as impairment of DSB function is known to prevent the formation of the flagellar P-ring

component FlgI rendering cells immotile. Chemical inhibition of DsbB function, indeed impaired

bacterial motility, similar to a dsbA deletion (Figure 3.7 A, B).238,372 Further, the redox state of DsbA in

the presence of the compound was also assessed to probe whether it was being re-oxidized by DsbB.

This is a necessary step that occurs after each round of oxidative protein folding and allows DsbA to

remain active. Under normal growth conditions, DsbA was in its active oxidized form in the bacterial

periplasm (i.e. Cys30 and Cys33 formed a disulfide bond), showing that it was efficiently regenerated by

DsbB (Figure 3.7 C).360 By contrast, addition of the inhibitor to growing E. coli MC1000 cells resulted

in accumulation of inactive reduced DsbA and confirmed that DsbB function was impeded.

116

Figure 3.7 Chemical inhibition of the DSB system impedes DsbA re-oxidation and flagellar motility in E. coli MC1000.

(A) A functional DSB system is necessary for flagellar motility in E. coli. In the absence of DsbA, or upon addition of a

chemical inhibitor of the DSB system, the motility of E. coli MC1000 is significantly impeded. Representative images of

motility plates are shown. (B) Quantification of the growth halo diameters in the motility assays shown in panel. n=3, graph

shows means ±SD, significance is indicated by **** = p < 0.0001. Significance was determined using one-way ANOVA with

Bonferroni’s multiple comparison test; n=3; 6 degrees of freedom; F value=1878; p=0.000000002 (significance). (C) Addition

of the reducing agent DTT to E. coli MC1000 bacterial lysates allows the detection of DsbA in its reduced form (DsbAred)

during immunoblotting; this redox state of the protein, when labelled with the cysteine-reactive compound AMS, shows a 1

kDa size difference (lane 2) compared to oxidized DsbA as found in AMS-labelled but not reduced lysates of E. coli MC1000

(lane 3). Addition of a small-molecule inhibitor of DsbB to growing E. coli MC1000 cells also results in accumulation of

reduced DsbA (lane 4). E. coli MC1000 dsbA was used as a negative control for DsbA detection (lane 1). A representative blot

from two independent experiments is shown; DsbA was visualized using an anti-DsbA primary antibody and an AP-conjugated

secondary antibody. Molecular weight markers (M) are shown on the left. Courtesy of Dr R Christopher D Furniss

(APPENDIX II).338

Having confirmed the efficacy of the DsbB inhibitor, chemical inhibition of the DSB system was tested

to determine whether it could be a suitable strategy to impair the function of class D OXA-type -

lactamases. The addition of the inhibitor compound during MIC testing phenocopied the effects of a

dsbA deletion (Figure 3.8, Supplementary Table 3). The fold changes observed for OXA-4 and -198

were comparable to those recorded for the E. coli MC1000 dsbA mutant, while the OXA-10 fold

changes were of greater magnitude (Figure 3.1). The MIC values of for the aminoglycoside antibiotic

gentamicin, the pDM1 empty vector and the disulfide-free L2-1 -lactamase controls remained

117

unchanged upon the addition of the DsbB inhibitor to the growth medium. These results were not due

to altered cell growth, as addition of the compound did not affect the bacterial growth (Figure 3.9).

Moreover, the DSB inhibitor has previously been tested in the Mavridou lab to ensure that it specifically

inhibits the DSB system, and no off-target effects were observed.338

Figure 3.8 Chemical inhibition of the DSB system phenocopies the β-lactam MIC changes observed using E. coli

MC1000 dsbA mutant. Chemical inhibition of the DSB system reduces the MIC values of representative -lactam antibiotics

for E. coli MC1000 expressing disulfide-bond-containing β-lactamases in a similar manner to the deletion of dsbA (Figure

3.1). Graphs show MIC fold changes (i.e. MC1000 MIC (µg/mL) / MC1000 + DSB system inhibitor MIC (µg/mL)) for β-

lactamase-expressing E. coli MC1000 with and without addition of a DSB system inhibitor to the culture medium and show

three independent experiments. Black dotted lines indicate an MIC fold change of 2. No changes in MIC values are observed

for the aminoglycoside antibiotic control, gentamicin (white bars, MIC fold changes: < 2). In addition, no changes in MIC

values are observed for strains harbouring the empty vector control (pDM1).

118

Figure 3.9 Chemical inhibition of the DSB system has no effect on the growth of E. coli MC1000. Growth curves for E.

coli MC1000 with and without chemical inhibition of the DSB show that bacterial growth remains unaffected by the inhibitor

compound on the supplemented M63 media. DMSO was used a carrier. n=3, curves show SD as light blue shaded area.

3.2.5.1 Sensitisation of multidrug-resistant P. aeruginosa clinical isolates to existing

antibiotics can be achieved by compromising the DSB system

Following the successful chemical inhibition of the DSB system in E. coli MC1000 strains (Figure 3.8),

the same inhibitor was used on clinical P. aeruginosa isolates. Resistance to β-lactam antibiotics in P.

aeruginosa is partially caused by the interplay between resident β-lactamase(s) and the MexAB-OprM

efflux pump.373 All P. aeruginosa possess two resident β-lactamases, the inducible class C enzyme

AmpC and the constitutively expressed class D enzyme OXA-50 (Figure 3.11). However, compared to

-lactamases like the OXA family enzymes, these enzymes do not represent a major resistance

determinant in clinical P. aeruginosa and, in combination with efflux pumps, play a relatively minor

role in β-lactam resistance.374

Clinical isolates expressing a single OXA-family -lactamase were selected. Their resistance profiles

were determined for both -lactam and non--lactam antibiotics using MH agar and E-test strips, with

the exception of colistin MICs for which broth microdilution was used (Supplementary Table 4). All

isolates were multi-drug resistant. For these strains defined cysteine-free MOPS media was used and

bacterial loads were adjusted to achieve the same MIC values for representative -lactam antibiotics,

119

as recorded on MH media. Only two of the P. aeruginosa isolates (expressing OXA-198 and OXA-28,

a member of the OXA-10 phylogenetic family) could be taken forward for MIC testing using the DSB

system inhibitor (Table 6), as the defined nature of the MOPS media strongly impacted the bacterial

viability of the remaining clinical isolates. For both isolates tested, the addition of Compound 12 to the

growth media did not result in reduction of the recorded MIC values for any of the -lactam antibiotics

tested.

These results reflect the challenges presented when attempting to target the DSB system in some

pathogenic bacterial species. The fact that Compound 12, which was developed to inhibit the E. coli

DsbB protein, was not effective against P. aeruginosa either due to a lower P. aeruginosa membrane

permeability in comparison to E. coli, the low sequence similarity between PaDsbA and EcDsbA, as

well as the polymorphisms that are quite common in this system, (see section 1.6.3).215 In particular,

the P. aeruginosa genome carries at least two functionally redundant DsbB analogues.375 As Compound

12, has been shown to be effective against DsbB1, but less so against DsbB2 of P. aeruginosa PA14,

the DsbB protein redundancy along with the initial high levels of resistance of the clinical isolates tested

mean that any effects of DsbB inhibition were likely masked.325

In the absence of a suitable chemical inhibitor, to investigate the consequences of compromising the

DSB system for the resistance of P. aeruginosa clinical strains to -lactam drugs, dsbA1, the principal

pseudomonal dsbA gene, was deleted from the P. aeruginosa PA43417 and PAe191 clinical isolates,

expressing OXA-198 and OXA-19 (a member of the OXA-10 phylogenetic family and the most

disseminated OXA enzyme in clinical strains), respectively (Figure 3.10).351

Table 6 Chemical inhibition of the DSB system via DsbB shows no effects on multidrug-resistant P. aeruginosa clinical

isolates. Addition of a small-molecule inhibitor of DsbB does not result in MIC drop for the two P. aeruginosa strains tested.

MIC values determined using MH agar in accordance with the EUCAST guidelines are comparable to the values obtained

using defined cysteine-free media (MOPS agar, recorded below). The use of growth media lacking small-molecule oxidants

is required for the DSB system inhibitor to be effective. MIC values (g/ml) are representative of two independent experiments.

DMSO was used as a vehicle control; DsbB inhibitor 4,5-dichloro-2-(2-chlorobenzyl)pyridazin-3-one was used at final

concentration of 50 M. The following abbreviations are used: XM, cefuroxime; IP, imipenem; AT, aztreonam.

Strain Additive XM IP AT

P. aeruginosa EDI

(blaOXA-28)

DMSO

DMSO + inhibitor - -

16

16

P. aeruginosa PA43417

(blaOXA-198)

DMSO

DMSO + inhibitor

>256

>256

>32

>32

6

6

120

Figure 3.10 Absence of DsbA1, the principal pseudomonal DsbA analogue, sensitizes multidrug-resistant clinical P.

aeruginosa isolates to first-line and last-resort -lactam antibiotics. (A) Removal of DsbA1 from P. aeruginosa PA43417

expressing OXA-198 sensitises the strain to the first-line antibiotic piperacillin (median MIC of 12 g/mL). (B) Removal of

DsbA1 from P. aeruginosa PAe191 expressing OXA-19 sensitises the strain to the aztreonam (median MIC of 12 g/mL).

The graphs show MIC values (g/ml) from 3 independent experiments, the red dotted lines indicate the EUCAST clinical

breakpoints (16 g/mL for piperacillin and aztreonam, 8 g/mL for ceftazidime and 4 g/mL for imipenem).

Deletion of dsbA1 in P. aeruginosa 41437 led to sensitization of the isolate to the first-line β-lactam

piperacillin (Figure 3.10 A), while its absence in P. aeruginosa PAe191 resulted in significant drops in

the MIC values for ceftazidime (16-fold drop) and imipenem (5.33-fold drop) and sensitisation for

aztreonam (Figure 3.10 B).These results suggest that targeting disulfide bond formation could be useful

for the sensitization of many more clinically important Gram-negative species to existing -lactam

compounds.

3.2.5.2 In vivo clearance experiments in the Galleria mellonella infection model

To test this strategy in an infection context we performed in vivo clearance assays using the wax moth

model G. mellonella (Figure 3.11). Larvae were infected with the P. aeruginosa PA43417 clinical

isolate producing OXA-198 and its dsbA1 mutant, and infections were treated once with piperacillin at

final concentrations below the EUCAST breakpoint. Neither deletion of dsbA1 nor treatment with

piperacillin were sufficient to reliably clear the infection when applied alone. In some cases, deletion

of dsbA1 led to a significant decrease in the recovered bacterial load due to the fact that absence of the

121

principal DsbA protein affects the virulence of the pathogen.333 Nonetheless, in other cases, the infection

with the dsbA1 mutant was only ~40%-70% cleared. On the other hand, treatment of the dsbA1 mutant

with piperacillin resulted in a drastic (> 99% on average) reduction in bacterial load in the infected

larvae in agreement with the fact that in the absence of DsbA the ability of OXA-198 to hydrolyse β-

lactams is impaired (Figure 3.1). As OXA-198 is a broad-spectrum β-lactamase that cannot be

neutralized by classical β-lactamase inhibitors (Supplementary Table 1) and piperacillin is a first-line

antibiotic, these results further highlight the promise of our approach for future clinical applications.

Figure 3.11 Absence of DsbA1 from a P. aeruginosa clinical isolate expressing OXA-198 allows it to be cleared from

infected G. mellonella larvae by piperacillin. Neither deletion of dsbA1, nor treatment with Piperacillin (at a concentration

of 12 g/mL) is sufficient to reliably clear P. aeruginosa PA4317 from infected G. mellonella larvae. However, the

combination of both results in an average reduction in bacterial load that is greater than 99%. The graph shows the average

number of colony-forming units (CFU) recovered from infected larvae for each condition relative to the CFU recovered for

the untreated P. aeruginosa PA4317 strain. n = 8 groups infected on eight different days; each group contains five G. mellonella

larvae per condition except for one group which contains three G. mellonella larvae per condition. Graph shows means ±SD,

significance is indicated by ns = p > 0.05 * = p < 0.05, ** = p < 0.01, *** = p <0.001. Significance was determined using

Kruskal-Wallis test with Dunn’s multiple comparisons test; n=8; Kruskal-Wallis H=25.24, 3 degrees of freedom; p=<0.0001.

For multiple comparisons, p=0.0029 (P. aeruginosa versus P. aeruginosa dsbA1), p=<0.0001 (P. aeruginosa versus P.

aeruginosa dsbA1 treated with piperacillin), p=0.0369 (P. aeruginosa treated with piperacillin versus P. aeruginosa dsbA1

treated with piperacillin). Data courtesy of Dr Ronan R. McCarthy and Evgenia Maslova (Brunel University, APPENDIX

II).338

122

3.3 DISCUSSION

Mobile resistance genes create a serious problem for the use of clinically available therapeutics, due to

their potential for spreading amongst bacterial pathogens. OXA-type -lactamases, which were initially

isolated from Acinetobacter species represent an example of the threat that mobile resistance

determinants pose to clinical therapy. Their prevalence in P. aeruginosa isolates resulted in the

emergence of clinical strains with very high levels of resistance, such as the OXA-19-expressing P.

aeruginosa PAe191 isolate that was tested in this study.67 To explore the potential of targeting the DSB

system as a strategy for mitigating antibiotic resistance, the activity of representative class D -

lactamases from P. aeruginosa was tested in an E. coli MC1000 strain and its isogenic dsbA mutant.

Clinically meaningful decreases in their ability to confer resistance in the absence of DsbA was

observed in three out of four enzymes, OXA-4, OXA-10 and OXA-198 (Figure 3.1)xvii. While the

absence of the disulfide bond did not notably decrease the protein stability of the tested -lactamases at

physiological conditions (Figure 3.6), their hydrolytic efficiency was found to be impaired (Table 5).

This is in agreement with the computer simulations carried out by Simakov et al. who proposed that the

disulfide bond is responsible for the correct conformation of one of the enzyme’s key catalytic

components, the omega loop.368

A potent chemical inhibitor of DSB activity, Compound 12, which covalently attaches to one of the

essential cysteines of DsbB blocking its function, was tested on model E. coli MC1000 strains

expressing OXA-4, -10, and -198.325 In all cases the trends observed for the recorded MIC values

(Figure 3.8) were in agreement with the results obtained from the dsbA mutant strain (Figure 3.1)

suggesting that the DSB system could be targeted chemically, thus offering a potential route for

abrogating antimicrobial resistance. While direct chemical inhibition of the DSB system was also

attempted on several P. aeruginosa clinical isolates expressing disulfide-bond-containing OXA-family

enzymes, no changes in their MIC values were observed. This was likely due to the presence of two

functionally redundant DsbB analogues in P. aeruginosa; Compound 12 is an efficient inhibitor of only

one of these two proteins.39 In combination with the high levels of resistance of these strains, any

xvii Absence of the native disulfide bond of the OXA-18 -lactamase also led to decreased MIC values, nonetheless the basal resistance of this

enzyme, even in the absence of its disulfide, was at levels that were not clinically meaningful (MIC values of 500-1000 g/mL for all tested

-lactams), and thus was not pursued further.

123

Compound-12-mediated effects were likely masked. Notably, however, deletion of dsbA1 in P.

aeruginosa isolates expressing OXA-198 and OXA-19 led to sensitization to first- and last-line β-

lactam antibiotics (Figure 3.11). As such, targeting of DsbA1 in P. aeruginosa is a promising strategy

for breaking class D -lactamase-mediated resistance.

Overall, these observations provide evidence that inhibition of DsbA, a non-essential cell envelope

protein which is unique to bacteria, could be useful for the sensitization of clinically important P.

aeruginosa strains to existing antibiotics. In addition, previous work by Dr R. Christopher D. Furniss

focusing on class A and B plasmid-encoded -lactamases from pathogenic species such as E. coli,

Citrobacter freundii, K. pneumoniae or E. cloacae, also showed that targeting the DSB system impairs

-lactamase function in other highly-resistant Gram-negative organisms. To our knowledge, this is the

first report of a strategy that can target -lactamase enzymes from three different Ambler classes.

Moreover, it is noteworthy that with 25% of β-lactamases found in bacterial pathogens and organisms

capable of causing opportunistic infections containing two or more cysteine residues,338 many more

clinically relevant β-lactamases are likely to depend on DsbA. This opens a new avenue to reverse -

lactamase-mediated resistance in Gram-negative organisms, through the development of DsbA

inhibitors acting as broad-acting resistance breakers.338

124

4 THE IMPORTANCE OF DISULFIDE BOND FORMATION FOR

THE FUNCTION OF CHROMOSOMALLY-RESIDENT -

LACTAMASE ENZYMES

4.1 INTRODUCTION

Chromosomally-encoded resistance genes evolve naturally in communities of environmental bacteria,

as their presence offers fitness advantages against competing microorganisms.22 High levels of genetic

variation can lead to intrinsic multi-drug resistance through a range of mechanisms, including disparities

in metabolic pathways, divergent membrane permeabilities, and expression of efflux pumps or -

lactamase enzymes. In addition to directly causing infections, intrinsically resistant environmental

bacteria can act as reservoirs for the dissemination of novel antimicrobial resistance genes. This appears

to be the case with broadly disseminated -lactamase or mcr colistin resistance genes, which are

currently spreading with an alarming speed in clinical strains.376

S. maltophilia, one of the most intrinsically resistant bacterial species, is an opportunistic pathogen

primarily associated with serious nosocomial respiratory infections. Although the clinical prevalence

of this organism is smaller compared to its closely related P. aeruginosa, S. maltophilia infections are

associated with extremely poor clinical prognosis stemming from the activity of multiple efflux pumps

and the presence of an extensive array of other resistance mechanisms.79,137 As a result, S. maltophilia

infections are resistant to most currently available antibiotics, including -lactams, macrolides,

fluoroquinolones, aminoglycosides, chloramphenicol, tetracyclines and colistin.79 S. maltophilia

resistance to -lactam antibiotics is predominantly driven by the expression of two chromosomally-

encoded -lactamases, the disulfide-free class A L2-1 enzyme and the disulfide-bond-bearing class B3

metallo--lactamase L1-1.79,137 Interestingly, the hydrolytic spectra of these two enzymes appear to

have co-evolved to confer comprehensive protection against all available -lactam classes. The ESBL

L2-1 can efficiently degrade penicillins and monobactams, whilst the carbapenemase L1-1 breaks down

penicillins and cephalosporins.79,137,338 The current clinically used therapy relies on sensitivity of the

bacterium to the third generation cephalosporin Ceftazidime, arising from low L1-1 expression levels,

or on the combination of Trimethoprim and Sulfamethoxazole.137,377–379 However, both therapy options

125

suffer from mutational and horizontally acquired resistance.137,377,378 Although extreme, the traits of S.

maltophilia infections are often recapitulated in infections caused by other environmental bacteria that

act as opportunistic human pathogens, for example saprophytic species like Burkholderia and

Pseudomonas, which become near impossible to treat.79,373,380,381

Previous work in the Mavridou lab has shown that mobile class A and D β-lactamases rely on DsbA-

mediated disulfide bond formation for their stability and function (Chapter 3 and APPENDIX II).338

Additionally, two important chromosomally resident enzymes were also observed to be DsbA

dependent, the class A enzyme SME-1 from Serratia marcescens and the aforementioned class B3

metallo-β-lactamase L1-1 from S. maltophilia.79,338,382 The DsbA dependence of these two

representative chromosomal enzymes formed the basis for the work carried out in this chapter aiming

to investigate the potential of targeting the DSB system of inherently resistant organisms as a strategy

to sensitize them against existing β-lactam compounds.

Eight phylogenetically distinct and chromosomally encoded β-lactamase enzymes from P. aeruginosa

(BEL-1, CARB-2, AIM-1, OXA-50), Pseudomonas otitidis (POM-1), Burkholderia spp (BPS-1m),

Franciscella tularensis (FTU-1), and Serratia spp (SMB-1) were selected for this work. These proteins

cover the full range of hydrolytic spectra of -lactamase enzymes, while their expression, which causes

considerable treatment complications, originates from diverse chromosomal settings (Supplementary

Table 1).109,338,353,383–386 For example, Burkholderia pseudomallei is the causative agent of

melioidosisxviii, a disease with limited treatment options due to intrinsic multi-drug resistance of the

bacterium to penicillin, first- and second- generation cephalosporins, aminoglycosides, macrolides, and

colistin.380 The mutation of its resident -lactamase BPS-1 gene into the clinically observed BPS-1m

results in resistance to ceftazidime which means the loss of the choice treatment option for this

organism.213,380,385

In addition to testing enzymes from bacterial species with high levels of intrinsic resistance, this study

also expands on class B metallo--lactamases which were underrepresented in previous investigations

(Chapter 3 and APPENDIX II); the requirement for oxidative protein folding for three representative

class B enzymes, AIM-1, POM-1, and SMB-1, was investigated here.338 These -lactamases are of

xviii Melioidosis is a severe disease endemic to Southeast Asia and Northern Australia.

126

increasing clinical concern due to the fact that they cannot be neutralised by currently available classical

inhibitor compounds, since their hydrolytic mechanism is discrete compared to serine -lactamases

(Section 1.3.1.4).67,75,76 To date, most class B enzymes have been found on the chromosomes of

environmental bacteria that are usually responsible for opportunistic infections, such as the

aforementioned S. maltophilia, P. otitidis, or S. marcescens. However, increasing evidence on the

mobility of these genes is recently coming to light.76,79,109,387,388 For example the carbapenemase SMB-

1 has been isolated upstream of an ISCR1 insertion sequence in the genome of S. marcescens, a rare

opportunistic pathogen which causes, among others, urinary and respiratory infections, septicaemia and

meningitis. Sequence comparison showed 75% sequence identity of SMB-1 with a -lactamase from

an uncultured soil bacterium, AMO-1, while its flanking region was 84% identical to the gene upstream

of AMO-1.388 Given that this ISCR1 mobility element in S. marcescens has previously been shown to

be associated with a class 1 integron carrying kanamycin and chloramphenicol resistance genes, it is

likely that SMB-1 has originated and was mobilised from a similar environmental reservoir to the one

AMO-1 was found in.386 More generally, ISCR elements have been implicated in the mobilisation of

other class B3 -lactamases outside S. marcescens, including SPM-1, NDM-1, or AIM-1.109,389,390

AIM-1, an emerging class B metallo-carbapenemase, is of particular importance in this study, given

both its unusually high number of disulfide bonds (all six of its cysteine residues are covalently linked

by disulfide bridges) and its location on the accessory part of the P. aeruginosa chromosome something

that increases its mobility potential.81 Similar chromosomal positions have been reported for the genes

of pseudomonal enzymes BEL-1 and CARB-2, which are associated with mobile genetic elements.353,391

Altogether the capacity of these -lactamases to spread to other strains and species, becomes

increasingly important in opportunistic pathogens and further exacerbates their already marked

resistance to antimicrobials in clinical settings.109,386,392 Further corroborating this, the genomes of

multidrug resistant S. maltophilia strains have also been observed to support great environmental

adaptability and have exhibited the ability to include and retain novel antimicrobial genes even after the

loss of selective pressure, a trait shared with Pseudomonas spp.79,80,393

127

4.2 RESULTS

4.2.1 The activity of cysteine-containing chromosomally encoded β-lactamase enzymes is

dependent on DsbA

In a similar manner to the work carried out with class D mobile -lactamase enzymes (Chapter 3), the

eight selected chromosomal β-lactamases were cloned and expressed in the E. coli MC1000 K-12 strain

and its isogenic dsbA mutant. MIC values against a panel of -lactam antibiotics, in accordance with

the known hydrolytic profiles of each enzyme, were recorded for the resulting strains. The

aminoglycoside antibiotic gentamicin, which cannot be inactivated by -lactamase enzymes, as well as

strains carrying the pDM1 empty vector or expressing the -lactamases L2-1 (from S. maltophilia) or

LUT-1 (from P. luteola) were included as negative controls. Both control enzymes contain two or more

cysteine residues but lack disulfide bonds. This is because they are transported pre-folded to the

periplasm via the Tat pathway, rather than by the Sec system. In the case of L2-1, Tat-dependent

transport has been experimentally confirmed189, whilst LUT-1 contains a predicted Tat signal peptide

(SignalP 5.0213 likelihood scores: Sec/SPI = 0.0572, Tat/SPI = 0.9312, Sec/SPII (lipoprotein) = 0.0087,

other = 0.0029).

It the absence of DsbA markedly reduced MIC values (>2-fold) were observed for at least one antibiotic

for all tested enzymes, except FTU-1, when compared to the parental strain (Figure 4.1, Supplementary

Table 2). Importantly, the sub-set of tested enzymes associated with mobile genetic elements were

dependent on DsbA, as were all three representatives of class B3 metallo-β-lactamases, POM-1, AIM-

1 and SMB-1. The latter is consistent with previous results where the class B3 L1-1 enzyme encoded

on the chromosome of S. maltophilia showed dependence on DsbA (APPENDIX II).338 Most of the

affected enzymes cannot be inhibited by classical -lactamase inhibitors, further expanding the group

of -lactamases that can be targeted by compromising the process of oxidative protein folding in the

cell envelope.338,394

128

Figure 4.1 Antimicrobial resistance mediated by chromosomally resident -lactamases depends on disulfide bond

formation. β-lactam minimum inhibitory concentration (MIC) values for E. coli MC1000 expressing a range of disulfide-

bond-containing β-lactamases (Ambler classes A, B and D) decrease in the absence of DsbA. For E. coli MC1000 expressing

FTU-1 and that of its dsbA mutant, the graph shows only a two-fold decrease between the ampicillin MIC value. This effect

corresponds to a reduction in MIC of 250 g/mL (Supplementary Table 2). No changes in MIC values are recorded for the

aminoglycoside antibiotic gentamicin (white bars, MIC fold changes: <2) confirming that absence of DsbA does not

compromise the general ability of MC1000 dsbA strain to resist antibiotic stress. No changes in MIC values are observed for

strains harbouring the empty vector control (pDM1) or those expressing the disulfide free β-lactamase controls L2-1 and LUT-

1. Graphs show MIC fold changes (MIC fold changes: > 2, fold change defined as MC1000 MIC (µg/mL) / MC1000 dsbA

129

(µg/mL),) for β-lactamase-expressing E. coli MC1000 and its dsbA mutant and show three independent experiments. Black

dotted lines indicate an MIC fold change of 2.

MIC values recorded for strains expressing the class A -lactamase FTU-1 from the human pathogen

F. tularensis only decreased by 2-fold in the absence of DsbA (Figure 4.1).384 Nonetheless, this effect

corresponded to a reduction in the Amoxicillin MIC value by 250 g/mL (Supplementary Table 2),

indicating that the function of this enzyme is affected by the disulfide bond loss even though substantial

hydrolytic activity is retained in the mutant strain. This -lactamase was not investigated further.

Finally, the MIC values for the aminoglycoside antibiotic gentamicin, E. coli strains harbouring the

pDM1 empty vector or strains expressing the disulfide-free -lactamases L2-1 or LUT-1 remained

unaffected (Figure 4.1, Supplementary Table 2).These results, along with the cell envelope integrity

checks presented in section 3.2.2 for this background, confirm that the observed MIC effects were

specific to the tested resistance proteins and their interaction with DsbA and not a result of a general

inability of the dsbA mutant to resist antibiotic stress. This was further supported by the fact that

complementation of dsbA restores the recorded MIC values to wild-type levels for all of the affected

enzymes (Figure 4.2).

130

Figure 4.2 Complementation of dsbA restores the β-lactam MIC values for E. coli MC1000 expressing β-lactamases.

Re-insertion of dsbA at the attTn7 site of the E. coli MC1000 chromosome restores the β-lactam MIC values for E. coli

MC1000 dsbA harbouring pDM1-blaBEL-1 (ceftazidime MIC), pDM1-blaBPS-1m (ceftazidime MIC), pDM1-blaCARB-2

(cefuroxime MIC), pDM1-blaFTU-1 (ampicillin MIC), pDM1-blaAIM-1 (ceftazidime MIC), pDM1-blaPOM-1 (imipenem MIC),

pDM1-blaSMB-1 (ceftazidime MIC) and pDM1-blaOXA-50 (amoxicillin MIC). Graphs show MIC values (µg/mL) from two

independent experiments.

4.2.2 Chromosomally encoded β-lactamase enzymes degrade or misfold in the absence of

DsbA

To gain insight into how impairment of disulfide bond formation impacts the production or activity of

the seven enzymes further tested in this chapter, we assessed their protein levels by immunoblotting in

the presence and absence of DsbA. For five of the tested β-lactamases (AIM-1, BEL-1, OXA-50,

CARB-2 and SMB-1) deletion of dsbA resulted in drastically reduced protein levels demonstrating that

without their disulfide bonds these proteins are unstable and ultimately degrade (Figure 4.3 A). In the

131

case of BPS-1m and POM-1, enzyme levels remained unchanged in the absence of dsbA (Figure 4.3

B).

In all tested cases, β-lactamase enzymes were also significantly less able to hydrolyse the chromogenic

β-lactam substrate nitrocefin (Table 7), consistent with the reduced MIC values recorded in the absence

of DsbA (Figure 4.1). For BPS-1m and POM-1, in particular, this shows that the absence of their

disulfide bond leads to a folding defect resulting in loss of function, despite the fact that enzyme

degradation was not observed (Figure 1.16). The protein levels of the control enzyme L2-1 and its

hydrolytic activity remained unaffected by the dsbA deletion (Figure 4.3, Table 7). Further, the strain

carrying the empty vector control, pDM1, was also unaffected by the absence of DsbA (Table 7).

Figure 4.3 The majority of tested β-lactamase enzymes become unstable in the absence of DsbA. (A) Protein levels were

drastically reduced for the cysteine-containing β-lactamases AIM-1, BEL-1, OXA-50, CARB-2, and SMB-1 when expressed

in E. coli MC1000 dsbA. The amount of the control enzyme L2-1, which contains three cysteines but no disulfide bonds

remains unaffected. (B) Protein levels of BPS-1m, and POM-1 β-lactamases are largely unaffected by the absence of DsbA.

Protein levels of StrepII-tagged β-lactamases were assessed using a Strep-Tactin-AP conjugate; representative blots from three

independent experiments are shown. Molecular weight markers (M) are on the left, DnaK was used as a loading control and

solid black lines indicate where the membrane was cut.

132

Table 7 The hydrolytic activities of tested β-lactamase enzymes are significantly decreased in the absence of DsbA. The

hydrolysis of the chromogenic β-lactam nitrocefin by cysteine-containing β-lactamases is impaired when these enzymes are

expressed in E. coli MC1000 dsbA. The hydrolytic activities of strains harbouring the empty vector or expressing the control

enzyme L2-1, containing three cysteines but no disulfide bonds, show no dependence on DsbA. The “Enzyme stability”

column informs on the stability of each enzyme when it is lacking its disulfide bond(s); this was inferred from the

immunoblotting experiments presented in Figure 4.3. The “Nitrocefin hydrolysis” column shows the amount of nitrocefin

hydrolysed per mg of bacterial cell pellet in 15 minutes. n=3, table shows means ±SD, significance is indicated by * = p <

0.05, ns = non-significant. Significance was determined using unpaired T-test with Welch’s correction; n=3; 3.417 degrees of

freedom, t-value=0.3927, p=0.7178 (non-significance) (for pDM1 strains); 2.933 degrees of freedom, t-value=0.3296,

p=0.7639 (non-significance) (for pDM1-blaL2-1 strains); 2.021 degrees of freedom, t-value=7.549, p=0.0166 (significance) (for

pDM1-blaBEL-1 strains); 2.146 degrees of freedom, t-value=9.153, p=0.0093 (significance) (for pDM1-blaCARB-1 strains); 2.320

degrees of freedom, t-value=5.668, p=0.0210 (significance) (for pDM1-blaAIM-1 strains); 2.345 degrees of freedom, t-

value=15.02, p=0.0022 (significance) (for pDM1-blaSMB-1 strains); 3.316 degrees of freedom, t-value=4.353, p=0.0182

(significance) (for pDM1-blaOXA-50 strains); 3.416 degrees of freedom, t-value=13.68, p=0.0004 (significance) (for pDM1-

blaBPS-1m strains); 3.998 degrees of freedom, t-value=4.100, p=0.0149 (significance) (for pDM1-blaPOM-1 strains).

Strain (MC1000) Enzyme stability Nitrocefin hydrolysis†

pDM1

dsbA pDM1 -

1.82 ± 3.20

0.96 ± 2.06ns

pDM1-blaL2-1

dsbA pDM1-blaL2-1 not degrading

82.66 ± 1.26

82.99 ± 0.77ns

pDM1-blaBEL-1

dsbA pDM1-blaBEL-1 degrading

92.66 ± 17.99

14.05 ± 1.31*

pDM1-blaCARB-2

dsbA pDM1-blaCARB-2 degrading

484.67 ± 72.09

96.79 ± 13.80*

pDM1-blaAIM-1

dsbA pDM1-blaAIM-1 degrading

109.45 ± 5.14

47.77 ± 18.13*

pDM1-blaSMB-1

dsbA pDM1-blaSMB-1 degrading

144.96 ± 14.69

12.11 ± 4.33*

pDM1-blaOXA-50

dsbA pDM1-blaOXA-50 degrading

37.30 ± 7.36

15.59 ± 4.51*

pDM1-blaBPS-1m

dsbA pDM1-blaBPS-1m not degrading

30.55 ± 1.14

13.99 ± 1.76*

pDM1-blaPOM-1

dsbA pDM1-blaPOM-1 not degrading

27.87 ± 1.70

22.23 ± 1.67* †nM.mg-1 pellet.15 min-1

4.2.3 Deletion of dsbA1 compromises the function of the intrinsic β-lactamase OXA-50 in

P. aeruginosa laboratory strains and clinical isolates

The resistance to some β-lactam antibiotics in P. aeruginosa isolates is a consequence of the interplay

between the MexAB-OprM efflux pump and the resident β-lactamase(s) AmpC and OXA-50.373 When

expressed in E. coli, OXA-50 conferred increased amoxicillin and cefuroxime MICs, which were

reduced in the absence of DsbA (Figure 4.1, Supplementary Table 2) due to enzyme degradation (Figure

133

4.3). Given the caveats with chemical inhibition of the pseudomonal DSB system activity (see section

3.2.5), effects of the lack of DsbA activity on β-lactam resistance conferred by the resident OXA-50

enzyme were assessed by deleting of the principal dsbA gene, dsbA1, from three P. aeruginosa strains

(the clinical isolate PA14 and two PAO1 strains from different sources; Table 2).355,395 MIC values were

recorded for the ureidopenicillin piperacillin, alone and in combination with the β-lactamase inhibitor

tazobactam; piperacillin-tazobactam formulations are clinically used for the treatment of P. aeruginosa

infections.396 In all strains the deletion of dsbA1 consistently decreased the piperacillin/piperacillin-

tazobactam MIC by two-fold (Figure 4.4), suggesting that the effect of dsbA deletion on antibiotic

tolerance can be observed in P. aeruginosa, even when the function of a relatively “minor” resistance

determinant is impaired. The results for both piperacillin and piperacillin-tazobactam treatments are the

same; OXA-50 cannot be inhibited by tazobactam (Supplementary Table 1).

Figure 4.4 Absence of DsbA1, the principal pseudomonal DsbA analogue, from P. aeruginosa laboratory strains and

clinical isolates expressing OXA-50 results in a two-fold decrease in their β-lactam MIC values. Results are shown for

the clinical P. aeruginosa PA14 strain, as well as the laboratory P. aeruginosa PAO1 LA and P. aeruginosa PAO1 LD strains

and their respective dsbA1 mutants. Piperacillin and piperacillin / tazobactam MIC values were determined for all strains.

Graphs show MIC values (g/mL) and show three independent experiments, the red dotted lines indicate the EUCAST clinical

breakpoints (16 g/mL).

134

4.2.4 Deletion of dsbA1 results in sensitization of P. aeruginosa clinical isolates to existing

β-lactam antibiotics

Given the promising results above, deletion of dsbA1 in P. aeruginosa was carried out for two

multidrug-resistant clinical isolates expressing the class B3 metallo-β-lactamase AIM-1. In contrast to

OXA-50, this β-lactamase confers high-level resistance to piperacillin-tazobactam as well as to the third

generation cephalosporin ceftazidime, an anti-pseudomonal β-lactam used for the treatment of critically

ill patients (Supplementary Table 4).109 MICs for piperacillin, piperacillin-tazobactam and ceftazidime

were determined for the two P. aeruginosa isolates (strains G4R7 and G6R7; Supplementary Table 4)

and their dsbA1 mutants (Figure 4.5).

For both strains the deletion of dsbA1 resulted in decrease MIC values for all tested antibiotics (Figure

4.5). The G4R7 strain was less affected by the loss of DsbA1, which resulted in a substantial 4-fold

decrease in piperacillin MIC values but did not sensitise the isolate to the first-line antibiotic.

Nonetheless, sensitisation to the third-generation cephalosporin ceftazidime was observed. The deletion

of DsbA1 was less tolerated by the G6R7 strain, where sensitisation to all three tested treatments was

observed (Figure 4.5). Consistent with the fact that currently available β-lactamase inhibitors do not

counteract metallo-β-lactamases, the MICs of both strains remained identical in the presence of

tazobactam in the media.62

135

Figure 4.5 Absence of DsbA1, the principal pseudomonal DsbA analogue, sensitises P. aeruginosa clinical isolates

expressing AIM-1 to penicillins and cephalosporins. Deletion of dsbA1 in the AIM-1 expressing P. aeruginosa G4R7

clinical isolate sensitizes this strain to ceftazidime (ceftazidime MIC of 8 g/mL) and results in reduction of the piperacillin

MIC value by over 192 g/mL. Deletion of dsbA1 in the AIM-1 expressing P. aeruginosa G6R7 clinical isolate sensitizes this

strain to piperacillin (piperacillin MIC of 8 g/mL) and ceftazidime (ceftazidime MIC of 6 g/mL). For both AIM-1 expressing

strains the piperacillin MIC values remain unchanged by the addition of tazobactam to the growth media, since AIM-1 is not

inhibited by tazobactam (Supplementary Table 1). Graphs show MIC values (g/mL) and show three independent experiments,

the red dotted lines indicate the EUCAST clinical breakpoints (16 g/mL for piperacillin and 8 g/mL for ceftazidime).

136

4.2.5 Deletion of dsbA1 and dsbL1 results in increased sensitivity of a S. maltophilia clinical

isolate to ceftazidime

To investigate whether the strategy of abrogating oxidative protein folding, which showed promise in

decreasing the resistance of P. aeruginosa clinical strains, could be used for other resistant bacterial

pathogens, S. maltophilia was tested. This organism, often found to colonise the lungs of cystic fibrosis

patients along with P. aeruginosa, was selected because of its high levels of resistance and because of

the fact that its resident L1-1 carbapenemase is dependent on DsbA; for the disulfide-harbouring L1-1

enzyme significant drops in MIC values were observed when expressed in E. coli K-12 lacking dsbA

(Supplementary Table 2).79,338

Two adjacent DSB oxidase genes, dsbA1 and dsbL1, were deleted from the multidrug-resistant S.

maltophilia GUE clinical isolate expressing both L2-1 and L1-1 -lactamases (Supplementary Table

4). Ceftazidime MIC values were recorded as previous work has attributed the high aztreonam MIC

values of this organism exclusively to the activity of the disulfide-free L2-1 -lactamase.137,338,378,379

The deletion of both dsbA1 and dsbL1 resulted in robust drops (8-fold) in the ceftazidime MIC value

for the clinical strain (Figure 4.6). No ceftazidime breakpoint is available for S. maltophilia, due to the

inherently high levels of resistance of this bacterium to cephalosporins afforded by L1-1. Nonetheless,

the recorded MICs for the dsbA dsbL mutant strain were 1-2 g/mL, values that are 4-8-fold lower than

the ceftazidime EUCAST clinical breakpoint for P. aeruginosa (8 g/mL), a closely related organism.

137

Figure 4.6 Absence of DsbA1 and of its analogue DsbL1 significantly decreases the MIC of the S. maltophilia GUE

clinical isolate, expressing L2-1 and L1-1, to ceftazidime. No ceftazidime breakpoint is available for S. maltophilia due to

the inherently high levels of resistance of this bacterium to cephalosporins. For comparison purposes, the black dotted line

represents the ceftazidime EUCAST clinical breakpoint for P. aeruginosa (MIC of 8 g/mL), a closely related organism.

Graph shows MIC values (g/mL) and from three independent experiments.

138

4.3 DISCUSSION

A panel of chromosomally-resident -lactamases exhibiting a range of hydrolytic spectra and driving

intrinsic resistance in their respective host organisms were investigated for their dependence on the

DSB system. Understanding the importance of the native disulfide bonds of these enzymes for their

stability and function allows us to assess the potential of targeting the DSB system as a strategy to

overcome species-specific resistance and treat infections with high-case fatality rates.

Eight -lactamase genes, encoding the BEL-1, BPS-1m, CARB-2, FTU-1, AIM-1, POM-1, SMB-1 and

OXA-50 enzymes, were cloned into an IPTG-inducible vector and expressed in the E. coli MC1000

K12 strain and its isogenic dsbA mutant. Their ability to confer resistance to -lactam antibiotics in the

absence of DsbA was drastically decreased for the majority of the enzymes tested (Figure 4.1). These

included -lactamases associated with mobile genetic elements (BEL-1, CARB-2, AIM-1, and SMB-

1), as well as metallo-β-lactamases found in environmental bacteria (POM-1, AIM-1, SMB-1, and L1-

1338) that cannot be neutralized via classical -lactamase inhibitors (Figure 4.1).109,394 Notably, the

deletion of dsbA, led to substantial ceftazidime MIC drops for the clinically observed extended-

spectrum enzyme BPS-1m which is responsible for the failure of one of the few viable treatments for

melioidosis (Figure 4.1).397,398 Further, the absence of their native disulfide bonds decreased the protein

stability of five out of the seven affected enzymes (Figure 4.3), and impaired the hydrolytic efficiency

of all tested -lactamases (Figure 4.3, Table 7). This is in line with the effects observed previously in

the Mavridou lab338 as well as our understanding of the role of disulfide bonds in the cell envelope.

In addition to BPS-1m, substantial MIC drops were also observed for the class B3 metallo-β-lactamase

AIM-1 (Figure 4.1), which is of clinical concern.76 Moreover, deletion of the primary dsbA gene from

the chromosomes of two multidrug-resistant P. aeruginosa clinical strains expressing this enzyme,

resulted in sensitisation to piperacillin and ceftazidime (Figure 4.5), which are drugs that are routinely

used to treat pseudomonal infections. Further promise of this approach can be seen in the results using

an almost pan-resistant clinical isolate of S. maltophilia expressing the resident L2-1 and L1-1 enzymes.

In this case, robust 8-fold drops in the ceftazidime MIC values were recorded when the genes encoding

the DSB oxidases DsbA and DsbL were deleted (Figure 4.6). This brought down the ceftazidime MIC

of the mutant strain at values that are 4-8-fold lower than the EUCAST clinical breakpoint of P.

aeruginosa, a closely related bacterium to S. maltophilia. Since the DSB system of S. maltophilia is

largely uncharacterized, and the dsbA and dsbL oxidase genes are located next to each other on its

139

chromosome, it was more practical to delete both oxidase genes simultaneously for this first test of the

role of DSB proteins in β-lactamase-mediated resistance; performing gene deletions in clinical isolates

of S. maltophilia is particularly challenging. Further experiments using single-gene deletions would be

required to investigate whether the activity of only one or both oxidases are necessary for the folding

of L1-1.

Together the results presented in this chapter show that the use of future DsbA inhibitors as antibiotic

adjuvants has the potential to help overcome intrinsic resistance of several bacterial species to existing

-lactam antibiotics.

140

5 THE IMPORTANCE OF DISULFIDE BOND FORMATION FOR

THE FUNCTION OF RESISTANCE-NODULATION-DIVISION

EFFLUX PUMPS

5.1 INTRODUCTION

Efflux pumps are crucial for the upkeep of cellular homeostasis in Gram-negative bacteria. They protect

the cytoplasm of the cell by removing a range of toxic substances, including solvents, metabolic by-

products or antimicrobial agents, from the periplasm.32,49,399 In the case of some pathogens, effluxed

substrates also play an active role in the development of biofilms, extracellular structures that exacerbate

antibiotic resistance by limiting antibiotic access and decreasing the achievable intracellular

concentrations of useful drugs.37,400 The majority of characterised antibiotic efflux pump protein

families are localised in the cytoplasmic membrane, where they counteract the spontaneous diffusion

of deleterious compounds into the cytoplasm (Figure 1.1).399 There are a few notable exceptions, with

the most important one being Resistance-Nodulation-Division (RND) efflux pumps, complex molecular

machineries that span the entire cell envelope with their tripartite structure.399,401 In addition to the

cytoplasmic pump, these macromolecular assemblies also have an outer membrane channel and a

periplasmic adaptor protein.402,403 Bacteria often encode more than one RND pump, and partial

redundancy between different pumps or their components, allows for extensive protection of the cell

against cytoplasmically-active antibiotics, which are efficiently expelled from the periplasm to the

extracellular millieu.404

The best characterised RND efflux pump is the E. coli AcrAB-TolC, the assembled structure of which

has just recently been observed in situ.402,403 AcrAB-TolC is composed of the periplasmic component

AcrA, the cytoplasmic membrane protein AcrB, and an outer-membrane channel, TolC (Figure

5.1).402,403,405 All three proteins are essential for the assembly of a functional pump with activity against

substrates like tetracycline, erythromycin, nalidixic acid or chloramphenicol.38,52,402,403,405 Pump

assembly is mediated by AcrA interactions with AcrB, TolC and a small conserved protein AcrZ.47,406,407

Substrate efflux begins by its binding on a loose protomer of the AcrB trimer, which is embedded in

the inner membrane.47,406 Subsequent binding on the tight protomer results in a conformational change

that allows substrate binding on the last of the three AcrB protomers, the open protomer. This in turn

141

results in expulsion of the substrate into the AcrA pump tunnel.406,408 Ultimately, the substrate exits the

cell through the TolC outer-membrane channel, a large trimeric -barrel structure composed of 12

transmembrane -sheets and 12 periplasmic -helices as well as an periplasmic domain.408

Figure 5.1 Structure of the E. coli AcrAB-TolC efflux pump. The RND pump spans across the whole length of the periplasm

and expels a wide range of substrates, including macrolides, chloramphenicol or nalidixic acid, to the extracellular

environment. It is composed of the outer-membrane-embedded TolC trimer, the inner-membrane-embedded AcrB trimer and

the periplasmic hexamer AcrA. PDB code: 5V5S.66,403

142

Expression of the AcrAB-TolC pump in E. coli is dependent on many factors and is, more importantly,

highly responsive to environmental parameters such as the presence of bile, fatty acids, antibiotics or

cationic peptides.409 Genetically, at the local level, acrAB falls under the control of two local repressors,

AcrS/EnvR and AcrR, the genes of which are both located upstream of the acrAB operon.50,51,409 Li et

al. have shown that a variety of antimicrobial compounds are responsible for repression of AcrR, leading

to pump expression.32,51 At the global level, pump expression control is also driven from the multiple

antibiotic resistance operon, Mar, and its transcriptional activator MarA.50,51,409 In addition, a high

degree of sequence homology between MarA, SoxS and Rob (activators implicated in antibiotic and

superoxide resistance) also allows SoxS/Rob-associated pump expression.410 Under normal growth

conditions, MarA, and as a result efflux pump expression, is repressed by MarR.410 Overexpression of

SoxS or Rob, as well as repression or mutation of MarR have been commonly observed in resistant E.

coli isolates with increased AcrAB-TolC expression.409,411,412 Interestingly, these genome mutations

often appear together, demonstrating how complex regulation of efflux pump expression is. Despite

their complex nature, the regulatory control networks characterised in E. coli are broadly conserved in

other species expressing analogues of AcrAB-TolC.409

Although RND pumps are large macromolecular assemblies, they contain very few or sometimes no

cysteine residues in Enterobacteriaceae. As such, they do not contain any disulfide bonds that would

clearly link them to the function of the DSB system.406 However, given the role of DsbA in the folding

of more than 300 extra-cytoplasmic proteins and in the maintenance of cell envelope homeostasis, it is

possible that changes in periplasmic proteostasis occurring in its absence could indirectly influence

efflux pump function.215,242 An examination of the literature on this topic revealed very little. An

unbiased genetic screen by Weatherspoon-Griffin et al., using the Keio collection and an effluxed

antimicrobial peptide called protamine, showed a significant drop in colony forming units (CFU) and

in the survival of E. coli in a disulfide bond formation impaired background (dsbA deletion).179 The

effects recorded in this study were of comparatively lesser impact than those caused by the deletion of

the essential pump components, AcrA and TolC. Further, results from this work also indicated that the

potential link between DsbA and efflux activity relates to the periplasmic chaperone/protease DegP,

although a clear justification for this connection was not given.179 In addition, a study by Hayashi et al.

reported that antibiotic resistance in B. cepacia, an opportunistic Gram-negative pathogen commonly

found in chronic lung infections, is dependent on its DSB system.413,414 A dsbA deletion in B. cepacia

lowered its resistance levels to several antibiotics that are likely efflux substrates for efflux pumps

homologous to those found in Alcaligenes eutrophus or B. pseudomallei.413,415–417 Together, these

143

observations suggest that, despite the absence of disulfide bonds in efflux pumps, there might be a link

between RND pump-mediated resistance and the DSB system.

To investigate the importance of DSB activity for the function of RND efflux pumps, AcrAB-TolC was

selected as a model system. The reason for this choice is that this particular efflux pump has been

extensively characterised and is known to be homologous to RND pumps of other highly resistant

organisms, such as P. aeruginosa (MexAB-OprM) and A. baumannii (AdeABC).48,418,419 MIC values

were measured for a panel of AcrAB-TolC antibiotic substrates, erythromycin, chloramphenicol, and

nalidixic acid, in the wild-type E. coli MG1655 strain and its isogenic dsbA mutant.21,52 Immunoblotting

analysis performed on key pump components and the periplasmic protease/chaperone DegP, which is

known to be important for AcrA proteostasis, to further tease apart the mechanism of the effects

observed.420,421

144

5.2 RESULTS

5.2.1 Deletion of dsbA in E. coli MG1655 does not affect the outer or the inner membrane

permeability

Assessment of the effects of the disruption of periplasmic proteostasis caused by the absence of DsbA

on efflux pump function was carried out in the efflux-active E. coli MG1655 K-12 strain and its isogenic

dsbA mutant. As with the E. coli MC1000 strains (Chapters 3 and 4), the permeability of the outer

membrane of the dsbA mutant was examined using the fluorescent dye 1-N-phenylnaphthylamine

(NPN) and the obtained results were confirmed using a vancomycin MIC assay.361,370 Both experiments

showed that the outer-membrane permeability of the dsbA mutant is no different from that of the

parental strain (Figure 5.2).

Figure 5.2 Deletion of dsbA has no effect on the membrane permeability or on the outer membrane integrity of E. coli

MG1655. (A) The hydrophobic fluorescent dye NPN crosses the outer membrane of E. coli MG1655 and its dsbA mutant to

the same extent. Conversely, exposure to the outer-membrane-permeabilizing antibiotic colistin results in a significant increase

in NPN uptake Graph shows means ± SD, significance is indicated by *** = p < 0.001, ns = non-significant. Significance was

determined using a one-way ANOVA with Bonferroni’s multiple comparison test; n=3; 6 degrees of freedom; F value=261.4;

p=0.00000055 (significance), p=0.0639 (non-significance). (B) No major differences are observed in the MIC values for the

aminoglycoside antibiotic vancomycin confirming that the outer membrane of E. coli MG1655 dsbA is not compromised, and

hence blocks the entry of the antibiotic. Graph shows n=3. NPN assay data is courtesy of Dr R. Christopher D. Furniss

(APPENDIX II).338

145

The integrity of the entire cell envelope of the dsbA mutant was also tested using the fluorescent dye

propidium iodide (PI) as well as a chromogenic CPRG assay.357,359 In both cases, there were no

significant differences between the cell envelope integrity of the wild-type strain and its dsbA mutant

(Figure 5.3).

Figure 5.3 Deletion of dsbA does not result in damage to the bacterial cell envelope. (A) No difference in basal PI uptake

is seen between E. coli MG1655 and its dsbA mutant. Both strains express sfGFP, and fluorescence was used to distinguish

live from dead cells. Addition of the inner-membrane-permeabilizing antimicrobial peptide cecropin A371 to E. coli MG1655

induces robust inner membrane permeabilization in the sfGFP-positive population indicating that the inner membrane becomes

compromised. Graph shows means ±SD, significance is indicated by **** = p < 0.0001, ns = non-significant. Significance

was determined using a one-way ANOVA with Bonferroni’s multiple comparison test; n=3; 6 degrees of freedom; F

value=77.49; p=0.0001 (significance), p=0.9999 (non-significance). (B) No difference in the red colouration of E. coli

MG1655 and its dsbA mutant colonies is seen, suggesting that CPRG is excluded from the colonies to the same extent. This

confirms that the integrity of the cell envelope is not compromised in the dsbA mutant strain. Images of plates were converted

to the CMYK colour space in Adobe Photoshop CS4. Colonies were selected using the magic wand tool with consistent

tolerance and edge refinement, and their magenta levels were compared. Graph shows means ± SD, significance is indicated

by ns = non-significant. Significance determined using an unpaired T-test with Welch’s correction; n=3; 4 degrees of freedom;

t-value=0.02647 p=0.9801 (non-significance). PI uptake data is courtesy of Dr R. Christopher D. Furniss (APPENDIX II).338

146

5.2.2 Deletion of dsbA in E. coli MG1655 causes only minor decreases in bacterial viability

Bacterial growth of the two E. coli MG1655 strains were assessed to ensure that the dsbA deletion did

not cause any major growth defects that could confound subsequent MIC assay results. Although

significant, only a small decrease in the growth rate of the dsbA mutant was observed compared to the

parental strain (Figure 5.4). Based on the cell-envelope integrity data presented in section 5.2.1 this

minor difference in growth rate is unlikely to affect the assays presented below.

Figure 5.4 Deletion of dsbA causes a small defect in the growth of E. coli MG1655. Growth curves of E. coli MG1655 and

its isogenic dsbA mutant show that bacterial growth rate and final OD600 achieved decrease slightly in absence of dsbA.

Significance determined using unpaired T-test with Welch’s correction; n=3; 4 degrees of freedom; t-value=5.716; p=0.0046

(significance).

147

5.2.3 RND efflux pump function is compromised in the absence of DsbA

Erythromycin, chloramphenicol and nalidixic acid are known AcrAB-TolC substrates.21,38,52 MIC

values for all three antibiotics were determined for the E. coli MG1655, its dsbA mutant, as well as

mutant in acrA which is an essential AcrAB-TolC pump component. The aminoglycoside antibiotic

gentamicin is not an AcrAB-TolC substrate and was used as control to assess the general ability of the

mutant strains to tolerate antibiotic stress. Decreases in MIC values were observed for both mutant

strains, with the acrA mutant being much more affected than the dsbA one; the MIC values measured

for the dsbA mutant were visibly lower than those of the parental strain, but higher than the ones

recorded for the acrA mutant. (Figure 5.5). Although less substantial than the MIC drops for E. coli

MG1655 acrA, the decreases in MIC value for the dsbA mutant were robust and reproducible, and

importantly were not observed for the non-substrate gentamycin. These effects are also in agreement

with previous studies reporting that deletion of dsbA increases the sensitivity of E. coli to dyes like

acridine orange and pyronin Y which are known substrates of AcrAB-TolC.237

The observed phenotype could be fully reversed by complementation of dsbA into the Tn7 site of the

E. coli MG1655 chromosome (Figure 5.6), further indicating a link between DsbA-mediated protein

homeostasis and efflux pump function.

Figure 5.5 Antimicrobial resistance mediated by a tripartite efflux pump, AcrAB-TolC, of E. coli MG1655 is affected

in absence of DsbA. Deletion of dsbA decreases the erythromycin, chloramphenicol and nalidixic acid MIC values for E. coli

MG1655, but no effects are detected for the non-substrate antibiotic gentamicin. The essential pump component AcrA serves

as a positive control. Graphs show MIC values (µg/mL) and shows three independent experiments.

148

Figure 5.6 Complementation of dsbA restores efflux-pump substrate MIC values for E. coli MG1655. Re-insertion of

dsbA at the attTn7 site of the E. coli MG1655 chromosome restores erythromycin, chloramphenicol and nalidixic acid MIC

values for MG1655 dsbA. Graphs show MIC values (µg/mL) and shows two independent experiments.

5.2.4 Compromised function of RND efflux pumps is due to altered periplasmic

proteostasis

To further elucidate the role of DsbA activity in efflux pump function we performed immunoblotting.

Since RND efflux pump proteins do not contain any disulfide bonds, the decreases in MIC values for

pump substrates in the absence of dsbA (Figure 5.5) are likely mediated by additional cell-envelope

components. As indicated by Weatherspoon-Griffin et al. the protease DegP, a previously identified

DsbA substrate, was a promising candidate. DegP degrades a range of misfolded extra-cytoplasmic

proteins including, but not limited to, subunits of higher order protein complexes and proteins lacking

their native disulfide bonds.179,422,423 In a dsbA mutant the substrate burden on DegP would likely be

dramatically increased. Additionally, DegP itself would not function optimally due to absence of its

disulfide bond.424 Consequently, protein turn over in the cell envelope would be less efficient. AcrA, an

essential component of AcrAB-Tolc RND efflux pump, is cleared by DegP when it becomes

misfolded.420,421 Thus, reduction in DegP efficiency in a dsbA mutant could result in accumulation of

non-functional AcrA in the periplasm and interfere with pump function.424

149

Immunoblotting of E. coli MG1655 and its mutants showed DegP degradation leading to the reduction

in the pool of active enzyme at any given time (Figure 5.7 A). In both, dsbA and degP backgrounds,

accumulation of AcrA was observed (Figure 5.7 B) showing that in both strains AcrA was not cleared

efficiently. No accumulation was detected for the control outer-membrane protein TolC (Figure 5.7 C),

which is not a DegP substrate.425 Thus, in the absence of DsbA, inefficient DegP-mediated periplasmic

proteostasis impacts RND efflux pump function through accumulation of non-functional AcrA. These

results are consistent with the study by Weatherspoon-Griffin et al., who did not observe any additive

effects in a dsbA/degP mutant and suggested that these two proteins are part of the same pathway.179

Figure 5.7 RND efflux pump function is impaired in the absence of DsbA due to accumulation of unfolded AcrA

resulting from insufficient DegP activity. (A) In the absence of DsbA the pool of active DegP is reduced. In E. coli MG1655

(lane 1), DegP is detected as a single band, corresponding to the intact active enzyme. In E. coli MG1655 dsbA (lane 2), an

additional lower molecular weight band of equal intensity is present, indicating that DegP is degraded in the absence of its

disulfide bond.423,424 DegP protein levels were assessed using an anti-DegP primary antibody and an HRP-conjugated

secondary antibody. E. coli MG1655 degP was used as a negative control for DegP detection (lane 3); the red arrow indicates

the position of intact DegP. (B) The RND pump component AcrA accumulates to the same extent in the E. coli MG1655 dsbA

and degP strains, indicating that in both strains protein clearance is affected. AcrA protein levels were assessed using an anti-

AcrA primary antibody and an HRP-conjugated secondary antibody. E. coli MG1655 acrA was used as a negative control for

AcrA detection; the red arrow indicates the position of the AcrA band. (C) TolC, the outer-membrane channel of the AcrAB-

TolC pump, does not accumulate in a dsbA or a degP mutant. TolC is not a DegP substrate, hence similar TolC protein levels

are detected in E. coli MG1655 (lane 1) and its dsbA (lane 2) and degP (lane 3) mutants.425 TolC protein levels were assessed

using an anti-TolC primary antibody and an HRP-conjugated secondary antibody. E. coli MG1655 tolC was used as a negative

control for TolC detection (lane 4); the red arrow indicates the position of the bands originating from TolC. For all panels, a

representative blot from three independent experiments is shown. Molecular weight markers (M) are on the left, DnaK was

used as a loading control and solid black lines indicate where the membrane was cut.

150

5.2.5 DsbA as a tractable target for RND efflux pumps

The intrinsic resistance mediated by RND efflux pumps is regulated by the MarA transcriptional

activator, which, in turn, is under the control of a the MarR repressor. Mutations in marR that derepress

MarA and that cause constitutive expression of AcrAB are commonly observed in clinical isolates with

increased efflux.405,410,426,427 An E. coli MG1655 marR mutant with de-regulated efflux pump expression

was used to investigate whether the negative effects observed in the efflux pump activity of E. coli dsbA

strain could also be recapitulated in cases of increased efflux pump expression.

While MIC drops were recorded for all efflux substrates tested on E. coli MG1655 dsbA (Figure 5.5),

chloramphenicol was selected to further test the ability of the de-regulated E. coli MG1655 strain to

resist antibiotic stress. The reason for this choice was that a EUCAST clinical breakpoint is available

for this antibiotic when used against Gram-negative bacteria (E. coli strains with MIC of 8 μg/mL or

below are classified as sensitive).428 Notably sensitisation to chloramphenicol was already observed for

E. coli MG1655 upon deletion of dsbA (Figure 5.5, MIC value = 6, sensitive compared to MIC value =

12, resistant),428 indicating that even the indirect effects resulting from compromising disulfide bond

formation are potentially clinically important. This sensitisation trend also held true when

chloramphenicol MIC values were compared between the de-repressed E. coli MG1655 marR strain

and its dsbA mutant (Figure 5.8).

Figure 5.8 Deletion of dsbA sensitizes the efflux-active E. coli MG1655 strain to chloramphenicol. Sensitization is also

observed for the dsbA mutant of the deregulated E. coli MG1655 marR strain (chloramphenicol MIC of 6 g/mL). The data

presented in the shaded blue and light blue bars were also used to generate Figure 5.7. The graph shows MIC values (g/ml)

from two independent experiments.

151

The already high level of efflux activity in the parental E. coli MG1655 meant that deletion of marR

did not result in a change in the recorded chloramphenicol MIC value compared to the parental strain

(Figure 5.8). Nonetheless, immunoblotting analysis showed that the expression of the AcrA component

of the pump was increased in the absence of MarR in comparison to the parental strain (Figure 5.9),

further validating that abrogating DsbA activity robustly compromises efflux pump function.

Figure 5.9 Deletion of marR results in increased expression of the AcrAB pump. MarR deletion increases the amount of

AcrA (lane 2) compared to the parental strain (lane 1). Expression of the AcrAB pump was assessed using an anti-AcrA

primary antibody and an HRP-conjugated secondary antibody. E. coli MG1655 acrA was used as a negative control for AcrA

detection (lane 3); the red arrow indicates the position of the AcrA band. A representative blot from two independent

experiments is shown; molecular weight markers (M) are shown on the left and DnaK was used as a loading control.

152

5.3 DISCUSSION

RND efflux pumps, which effectively decrease cell envelope concentrations of useful antibiotic

compounds, are an important contributor to the antimicrobial resistance profiles of many Gram-negative

bacterial pathogens. The dependence of the function of E. coli AcrAB-TolC-like efflux pumps on DsbA

was tested using E. coli MG1655 and its dsbA, degP, and acrA mutants. In agreement with our initial

hypothesis, modest but robust decreases in MIC values were observed for all three efflux pump

substrates tested (Figure 5.5). In addition, loss of dsbA sensitised E. coli MG1655 to chloramphenicol,

both in the wild-type background and in the absence of the marR, a mutation that was used to simulate

an efflux-pump-overexpression phenotype commonly observed in clinical strains (Figure 5.8). The

defects observed in the absence of DsbA were driven by the accumulation of non-functional AcrA in a

DegP-dependent manner.

By extension, other efflux pumps containing AcrA-like components are also likely to depend on DegP

for their homeostasis in the periplasm, and their function could be compromised by DsbA or DegP

inhibition. More importantly though, the results presented in this chapter demonstrate that cell envelope

protein homeostasis is crucial for the optimal function of RND efflux pumps and indicate that

periplasmic proteostasis pathways could have significant yet untapped potential for the development of

novel antibacterial strategies against these resistant determinants. The development of broad-acting

clinically applicable efflux pump inhibitors has been challenging to date.32,37 As an alternative,

inhibition of DsbA, DegP or of other cell envelope proteostasis pathways (chaperones and proteases)

could offer a new avenue for the development of efficient strategies to compromise the function of these

elusive macromolecular apparatuses.

153

6 THE IMPORTANCE OF DISULFIDE BOND FORMATION FOR

THE EXPANSION OF THE HYDROLYTIC SPECTRUM OF -

LACTAMASE ENZYMES

6.1 INTRODUCTION

Mutation-driven evolution of narrow-spectrum -lactamases into extended-spectrum and carbapenem-

hydrolysing enzymes is a major contributor to the increasing emergence of resistance to -lactam

compounds, the most prescribed antibiotics worldwide. Numerous studies have shown that vertical

evolution, driven by incremental changes in the primary sequence of narrow-spectrum -lactamases,

can lead to the rapid emergence of enzymes with higher hydrolytic capabilities. Increases in protein

expression levels aside, resistance to -lactam antibiotics has been shown to expand mostly through

two types of amino-acid changes, ‘new-function’ and ‘non-functional’ point substitutions and sequence

modifications.181 This, in addition to the extensive mobility of these promiscuous resistance

determinants, poses critical challenges to modern infection treatment strategies.

Stability-function trade-offs determine the evolutionary success of mutational evolution.181 Key

catalytic residues are evidently retained to prevent detrimental loss of enzymatic activity.181–183,185 New-

function substitutions drive structural changes in the generally unstructured parts of the sequence

surrounding the active site of the enzyme, often causing protein instability, due to the already inherent

strain in the active-site area.181–183,185,429 By contrast, ‘non-functional’ substitutions often act as

compensatory changes that allow protein stabilisation.181,183,430 Both types of primary sequence changes

frequently occur, albeit one at a time, and their functional effects are, overall, additive.181,431

The narrow-spectrum class A -lactamase TEM-1 only confers resistance to first- and second-

generation penicillins. However, the Arg162Ser substitution, which transforms TEM-1 to TEM-12, leads

to resistance to third-generation cephalosporins, like ceftazidime, while the addition of the stability-

inducing substitution Glu104Lys (TEM-26) further expands the spectrum of this enzyme to

monobactams, such as aztreonam.432 Interestingly, mutations resulting in the Glu104Lys on its own, as

observed in TEM-17, do not result in efficient hydrolysis of either ceftazidime or aztreonam.432 It should

154

be noted that expansion of the hydrolytic activity of -lactamases through new-function mutations often

concurrently results in increased affinity for -lactamase inhibitors.133,140,433–436 Increased sensitivity to

sulbactam, clavulanic acid, or tazobactam has been observed in the case of both TEM-12 and TEM-

26.432 Unfortunately, occurrence of gene mutations that allow enzymes to escape mutation-driven

sensitivity to classical inhibitors have also been reported for several -lactamases, including AmpC, a

resident enzyme of P. aeruginosa, or KPC-3 and OXA-48 from K. pneumoniae.133,140,437 In particular,

Fröhlich et al. showed that gradual laboratory exposure of the class D carbapenemase OXA-48 to

ceftazidime-avibactam combinations resulted in the development of ceftazidime (Pro68Ala) as well as

ceftazidime-avibactam (Pro68Ala/Tyr211Ser) resistance through increased flexibility of the active site.437

A 1987 study by Schultz et al. showed that the activity of TEM-1 did not depend on its native disulfide

bond at low temperatures, although substrate hydrolysis was impacted at 42°C.438 A similar evaluation

by Dr R. Christopher D. Furniss (postdoctoral research associate, Mavridou lab) on SHV-27 showed

that its hydrolytic activity against a second-generation cephalosporin, cefuroxime, is unaffected at 37°C,

but decreases under temperature stress (42°C) (APPENDIX II).338 SHV-27 has a single amino acid

substitution (Asp156Gly) with respect to the narrow-spectrum enzyme SHV-1.78,439 By contrast the

activity of a range of both mobile and chromosomal -lactamases with broader hydrolytic spectra

clearly depends on the disulfide-forming protein DsbA even at physiological conditions (Chapters 3, 4,

and APPENDIX II).338 More importantly, strains expressing enzymes capable of hydrolysing more

complex -lactams, such as ceftazidime, imipenem or aztreonam, exhibit larger decreases in their MIC

values when disulfide bond formation is abrogated through the deletion of dsbA, compared to strains

expressing narrower spectrum enzymes (Chapters 3, 4, and APPENDIX II). These observations suggest

that disulfide bond formation is important for the enzymatic function of β-lactamases with broader

hydrolytic activities, i.e. enzymes that present the greatest clinical challenge. This was further

confirmed by the work of Dr R. Christopher D. Furniss on class A -lactamases from the GES

phylogenetic family, for which he showed that expansion of the hydrolytic spectrum of these enzymes

increased their dependence on their native disulfide bonds (APPENDIX II).338

We hypothesized that in narrow-spectrum enzymes non-covalent interactions are sufficient to maintain

the integrity of their native structures, and thus to preserve their hydrolytic activities. Further, we posited

that disulfide bonds become crucial for function when the structures of -lactamases are challenged by

de-stabilising new-function substitutions occurring during expansion of their hydrolytic spectra through

mutation-driven evolution. Given that ‘new-function’ variations generally have de-stabilising effects

155

on evolving enzymes, the presence of a covalent bond (for example a disulfide bridge) could act as a

key stabilising element holding the active site together.440 This is supported by computational studies

by Szarecka et al. and Tokuriki et al., which showed that the disulfide bond stabilizes the structure of

OXA -lactamases during their evolution into enzymes with new hydrolytic functions.181,366 In this

chapter, narrow-spectrum -lactamases SHV-1 and TEM-1 were used to test these hypotheses on the

role of oxidative protein folding in the evolution of enzymes of broad hydrolytic spectra.

Laboratory evolution of TEM-1 and SHV-1 was to be performed by the gradual exposure of E. coli

strains expressing each enzyme to a -lactam compound that could not be efficiently hydrolysed by the

starting -lactamase protein. This process has been documented to lead to the emergence of extended-

spectrum enzymes for several -lactamase families, including TEM and SHV. In the case of SHV-1, a

narrow-spectrum -lactamase first identified in K. pneumoniae only capable of hydrolysing penicillin

and ampicillin,441 point mutants SHV-6 (Asp179Ala) and SHV-2 (Gly238Ser) hydrolyse extended-

spectrum cephalosporins, such as ceftazidime and cefotaxime.77,441,442 An Asp179Asn substitution in the

same site as in SHV-6, expands the hydrolytic activity even more, and allows hydrolysis of

monobactams.443 An overview of key amino acid variations reported to enable the expansion of the

hydrolytic spectra of TEM-1 and SHV-1 is shown in Table 8.

Table 8 An overview of commonly occurring amino acid substitutions in TEM-1 and SHV-1 -lactamases that mediate

expansion of their hydrolytic activity.78,432,436,444–448 The “Spectrum” column refers to the hydrolytic spectrum of each variant

enzyme; narrow-spectrum β-lactamases (NS), extended spectrum β-lactamases (ESBL) or carbapenemases.

Enzyme Spectrum Substitution

TEM-1 NS -

TEM-12 ESBL Arg162Ser

TEM-15 ESBL Gly236Ser/Glu104Lys

TEM-17 ESBL Glu104Lys

TEM-19 ESBL Gly236Ser

TEM-26 ESBL Arg162Ser/Glu104Lys

SHV-1 NS -

SHV-2 ESBL Gly238Ser

SHV-5 ESBL Gly238Ser /Glu256Lys

SHV-38 carbapenemase Ala157Val

SHV-31 ESBL Glu256Lys

SHV-144 ESBL Ala157Val/Leu33Gln

156

To determine whether the emergence of broader spectrum enzymes depends on the presence of their

native disulfide bonds, and by extension on the activity of the DSB oxidase DsbA, generation of some

of these naturally occurring mutants of cysteine-containing -lactamases by way of experimental

evolution was carried out. Application of antibiotic pressure and increase in the hydrolytic activity of

each enzyme through the occurrence of de-stabilising new-function mutations, was expected to render

native disulfide bonds increasingly more important for enzyme stability and function. One can imagine

that as the active site “loosens up” in order to efficiently accommodate a wider range of antibiotic

substrates,185,436,443 the flexibility in this protein region increases and the presence of the disulfide bond

could be important for stabilising the overall enzyme fold and preserving its hydrolytic activity.

157

6.2 EXPERIMENTAL DESIGN

6.2.1 Deletion of dsbA does not severely impact the resistance to -lactams conferred by

the narrow-spectrum -lactamases SHV-1 and TEM-1

The genes for the wild-type TEM-1 and SHV-1 enzymes, as well as their single-cysteine variants

(Cys86Ala TEM-1 and Cys54Ala SHV-1) were codon optimised and cloned into the pDM2 vector, where

constitutive protein expression is driven by a strong synthetic Biofab promoter. The -lactamases were

expressed in the E. coli MC1000 K-12 strain and its isogenic dsbA mutant, and their MIC values were

recorded for a panel of -lactam antibiotics. The aminoglycoside antibiotic gentamicin, which is not

hydrolysed by -lactamase enzymes, and strains carrying the pDM2 empty vector were also included,

as negative controls.

At 37°C the absence of DsbA did not result in a substantial decrease in the -lactam MIC values for

either of the tested enzymes (Figure 6.1 A, MIC fold changes: < 2). At 42°C, MIC values for SHV-1

remained unaffected by the deletion of dsbA, but the ceftazidime MIC for E. coli MC1000 dsbA

harbouring pDM2-blaTEM-1 was 3-fold lower than that of the parental strain (Figure 6.1 B). As expected,

no decreases in MIC values were recorded for the aminoglycoside antibiotic gentamicin, nor strains

carrying the pDM2 empty vector at either tested temperature (Figure 6.1). These observations are

consistent with the results of Schultz et al. (TEM-1) and Furniss et al. (SHV-27, APPENDIX II); the

disulfide bonds do not seem to play a major role for protein function at 37°C, but at higher temperatures,

at which protein stability is challenged, their presence becomes more important for substrate

hydrolysis.338,438

158

Figure 6.1 Antimicrobial activity of narrow-spectrum enzymes, SHV-1 and TEM-1 does not dependent on DsbA. (A)

At 37°C deletion of dsbA does not affect the MIC values for E. coli MC1000 strains harbouring pDM2-blaSHV-1 or pDM2-

blaTEM-1. (B) At 42°C deletion of dsbA does not affect the MIC values for E. coli MC1000 strains harbouring pDM2-blaSHV-1.

Ceftazidime MIC values for pDM2-blaTEM-1 show a notable 3-fold reduction in the absence of their native disulfide bond. No

changes in MIC values are observed for the aminoglycoside antibiotic gentamicin (white bars, MIC fold changes: < 2)

confirming that absence of DsbA does not compromise the general ability of this strain to resist antibiotic stress. Further, no

changes in MIC values are observed for strains harbouring the empty vector control (pDM2). Graphs show MIC fold changes

(MIC fold changes: > 2, fold change defined as MC1000 MIC (µg/mL) / MC1000 dsbA (µg/mL),) for β-lactamase-expressing

E. coli MC1000 and its dsbA mutant. Three independent experiments are shown for MIC values at 37°C, and one independent

experiment is shown for MIC values at 42°C. Black dotted lines indicate an MIC fold change of 2.

Ceftazidime is the least complex compound of the tested -lactam antibiotics, against which both TEM-

1 and SHV-1 enzymes show very little hydrolytic activity (Supplementary Table 6). As such, it was

selected as the most appropriate compound to be used during the experimental evolution step.

6.2.2 Setup of the experimental evolution experiment

Laboratory evolution of -lactamase enzymes had been reported extensively in the existing literature.

This has been carried out using broth-based or plate-based methods. Recently, Fröhlich et al. evolved

the hydrolytic spectrum of a Class D enzyme, OXA-48, by gradual exposure to ceftazidime or

ceftazidime-avibactam. Strains expressing the wild-type carbapenemase OXA-48 are susceptible to

159

both -lactam treatments, nonetheless, resistance development has been observed in clinical strains of

Enterobacteriaceae.437,449 In the work of Fröhlich et al., overnight cultures of strains expressing OXA-

48 were grown in MH broth, spun down and resuspended in a set volume of fresh MH media. 100 L

of that bacterial suspension was plated on MH agar containing increasing concentrations of ceftazidime

or ceftazidime-avibactam (up to 64x MIC).437 After overnight incubation, colonies from the plate with

the highest concentration of ceftazidime were recovered, grown in fresh MH broth and used as a starter

culture for a second round of passaging.437 The emergence of resistant colonies stopped when the

evolving strains reached a ceftazidime MIC of 32 g/mL; at this stage isolated colonies were tested for

the ceftazidime susceptibility.437

An experimental evolution experiment for TEM-1 and SHV-1 was set up using the above method.

However, plating of the bacterial suspensions of E. coli MC1000 strains expressing TEM-1 and SHV-

1 without performing a normalisation step, as was done in the study by Fröhlich et al., resulted in

extensive bacterial growth at all concentrations of ceftazidime used (Table 9).437 Given that the load

used by Fröhlich et al. was optimised for the experimental evolution of a carbapenemase enzyme, and

resulted in no growth at above 32 g/mL of ceftazidime, further inoculum optimisation for strains

expressing narrow-spectrum enzymes was needed.437 Nonetheless, it should be noted that, despite the

high bacterial loads on our selection plates, trends showing a reduced ability for resistance evolution

had started to emerge for SHV-1, especially at high ceftazidime concentrations. More specifically, the

ability to evolve resistance against ceftazidime is lower in the absence of DsbA (50-fold decrease in the

number of resistant colonies at 256 g/mL of ceftazidime) or for the single-cysteine SHV-1 variant,

which can also not form a disulfide bond (25-fold decrease in the number of resistant colonies at 256

g/mL of ceftazidime) (Table 9).

Parental E. coli MC1000 background expressing either wild-type TEM-1 or wild-type SHV-1 were

used to determine the most suitable bacterial load for each set of bacterial strains. The obtained OD600

of bacterial suspensions after an overnight growth were standardised and dilutions were performed.

Subsequently, test inocula were plated on ceftazidime-containing plates of increasing concentrations

and incubated at 37°C for 18 hours (Table 10). Enumeration of the recovered colonies on each plate

allowed the identification of the optimal inoculum to be used for the subsequent experimental evolution

step.

160

Table 9 E. coli MC1000 strains constitutively expressing SHV-1 and TEM-1 -lactamases as well as the single-cysteine

variants of these enzymes were plated on ceftazidime containing plates (0.5-64x MIC). High bacterial loads were

recovered for all concentration of ceftazidime for all strains. “L” strands for “bacterial lawns and “U” stands for “uncountable”

i.e. too many to count.

In contrast to the approach employed by Fröhlich et al., standardisation of the OD600 of the overnight

cultures and selection of an optimal bacterial load for each set of strains ensures the reproducibility of

our results and allows tailoring of the experimental setup to the hydrolytic profile of each enzyme.437 A

schematic presenting our optimised method for performing experimental evolution on strains

expressing TEM-1 and SHV-1 enzymes using ceftazidime pressure is shown in Figure 6.2.

Strain (MC1000) 1 2 4 8 16 32 64 128 256

pDM2-blaSHV-1 L L L L L L U U Est. 300

dsbA pDM2-blaSHV-1 L L L L L L U Est. 200 6

pDM2-blaSHV-1 C54A L L L L L L U Est. 180 12

- 0.125 0.25 0.5 1 2 4 8 16

pDM2-blaTEM-1 - L L L L L L L L

dsbA pDM2-blaTEM-1 - L L L L L L L L

pDM2-blaTEM-C86A - L L L L L L L L

161

Table 10 Bacterial suspensions resulting from overnight growth of E. coli MC1000 strains constitutively expressing

wild-type SHV-1 and TEM-1 -lactamases. Wild-type MC1000 strains harbouring the constitutively expressed pDM2

plasmid encoding the SHV-1 and TEM-1 -lactamases were standardised to OD600 0.5, 1.0 or 2.0 and 10-100L of the neat

suspensions or of further dilutions were plated on ceftazidime-containing plates at 0.5-64x MIC. Enumeration of colonies on

each plate allowed the selection of the optimal bacterial load for the experimental evolution step; bacterial loads resulting in

isolated colonies on plates containing ceftazidime at lower concentrations per plate were counted and loads for the evolution

were selected such that the highest concentration plates were less than 4x MIC and single colonies could be isolated to allow

enumeration and characterisation. “L” strands for “bacterial lawns and “U” stands for “uncountable” i.e. too many to count.

MC1000 pDM2-blaSHV-1 Ceftazidime (g/mL)

OD600 Dilution Load

(L) 1 2 4 8 16 32 64 128 256

2.0

NA

100

L

L L U U 19 1 - - -

NA

10 L L U 107 - - - - -

1:100

100 L L 125 12 - - - - -

1:200

100 L L 57 4 - - - - -

1:400 100 L L 45 1 - - - - -

MC1000 pDM2-blaTEM-1 Ceftazidime (g/mL)

OD600 Dilution Load

(L) 0.125 0.25 0.5 1 2 4 8 16

2.0

NA 100 L L L L L L L L

NA 10 L L L L L L L L

1:100 100 L L L L L L L L

1:200 100 L L L L L L L L

1:400 100 L L L L L L L L

1.0

NA 100 L L L L L L L -

NA 10 L L L L L L L -

1:100 100 L L L L L L L -

1:200 100 L L L L L L L -

1:400 100 L L L L L L L -

0.5

NA 100 U U U 4 1 1 - -

NA 10 U U U 1 - - - -

1:100 100 Est.

350 U

Est.

500 - - - - -

1:200 100 U U 250 - - - - -

1:400 100 U Est.

200 1 - - - - -

162

Figure 6.2 Schematic of the experimental evolution method to be used for strains expressing TEM-1 and SHV-1 -

lactamases. E. coli strains harbouring the pDM2-blaSHV-1 and pDM2-blaSHV-1 C54A plasmids are grown in tetracycline-

supplemented MH broth at 37°C overnight. Cultures are spun down at 4000x g, resuspended in fresh MH broth, and

standardised to OD600 2.0 for SHV-1- and OD600 0.5 for TEM-1-producing strains. Standardized cultures are diluted 1:100

(SHV-1) or 1:200 (TEM-1) and 100L is plated on ceftazidime-containing MH agar (0.5-64x MIC). Following an 18-hour

incubation period, colonies from the plate containing the highest concentration of ceftazidime are collected, pooled in a new

overnight culture (MH broth supplemented with tetracycline) and re-exposed to a new ceftazidime gradient the next day. MIC

values of representative colonies are determined using E-test strips. Method adapted from Fröhlich et al.437

163

6.3 SHV-1 PILOT STUDY RESULTS

6.3.1 Absence of DsbA decreases the potential for evolution of antibiotic resistance to

ceftazidime upon exposure to increasing antibiotic concentrations

E. coli MC1000 strains constitutively expressing the narrow-spectrum enzyme SHV-1 were subjected

to increasing concentrations of ceftazidime, ranging from 0.5x to 64x of the initial MIC values (4 g/mL

for E. coli MC1000 pDM2-blaSHV-1, and 2 g/mL for E. coli MC1000 dsbA pDM2-blaSHV-1 and MC1000

pDM2-blaSHV-1 C54A). The number of recovered colonies per plate was enumerated, and colonies from

the plate with the highest ceftazidime concentration were pooled for the next round of passaging. In

total three passaging steps were carried out (I-III, Table 11), before colonies with increased ability to

withstand ceftazidime stress stopped emerging. Evolution of resistance to ceftazidime was observed in

all three tested backgrounds, nonetheless varying efficiency in resistance evolution was recorded for

each strain (Table 11).

E. coli MC1000 expressing wild-type SHV-1 produced the greatest number of colonies on increasing

concentrations of ceftazidime throughout rounds I and II of passaging (Table 11). By contrast, resistance

evolution in the absence of DsbA was less efficient, and although resistant colonies emerged at the same

ceftazidime concentration as for the E. coli MC1000 strain, their numbers were greatly decreased.

Surprisingly, the strain expressing the single-cysteine SHV-1 mutant showed a similar behaviour to the

E. coli MC1000 dsbA background for round I of passaging but seemed to behave like the parental strain

during round II. Despite inter-strain differences, these results indicate that oxidative protein folding is

important for the ability of strains expressing narrow-spectrum -lactam to evolve resistance to more

complex -lactam antibiotics.

164

Table 11 E. coli MC1000 strains expressing SHV-1 and its single-cysteine variant, develop resistance upon exposure to

increasing ceftazidime concentrations. Three rounds of passaging were performed (I-III) in each of which 0.5-64x MIC

concentrations of ceftazidime were used. Background MIC values for the strains were as follows: E. coli MC1000 pDM2-

blaSHV-1, MIC of 4 g/mL; E. coli MC1000 dsbA pDM2-blaSHV-1 and E. coli MC1000 pDM2-blaSHV-1 C54A, MIC of 2 g/mL.

The following abbreviations are used: “L” strands for “bacterial lawns and “U” stands for “uncountable” i.e. too many to

count. The star ‘*’ denotes poor bacterial growth.

Passage Strain (MC1000) 1 2 4 8 16 32 64 128 256

I

pDM2-blaSHV-1 NA U U 172 - - - - -

dsbA pDM2-blaSHV-1 L U 13 21 - - - - -

pDM2-blaSHV-1 C54A L U 8 15 1 - - - -

8 16 32 64 128 256 512 1024 2048

II

pDM2-blaSHV-1 L U Est. 750 - - - - - -

dsbA pDM2-blaSHV-1 L U 19 - - - - - -

pDM2-blaSHV-1 C54A NA U Est.950 3 - - - - -

32 64 128 256 512 1024 2048 4096 8192

III

pDM2-blaSHV-1 U 18* - - - - - - -

dsbA pDM2-blaSHV-1 U 10* - - - - - - -

pDM2-blaSHV-1 C54A U 4* - - - - - - -

6.3.2 Characterisation of evolved SHV-1 expressing strains

Three evolved colonies per strain background generated during passage II, were isolated from the plate

containing 32g/mL ceftazidime, and their MICs values were determined for a panel of -lactam

antibiotics (henceforth referred to as “Evolved” strains). In addition, their -lactamase-expressing

plasmids were isolated and transformed into the original E. coli MC1000 or E. coli MC1000 dsbA

backgrounds; MIC values for these strains were also determined for the same antibiotics (henceforth

referred to as “Original” strains). For all nine characterised evolved isolates, increased MIC values for

complex -lactam antibiotics were recorded.

Evolution of E. coli MC1000 pDM2-blaSHV-1 resulted in two to four-fold increase in ceftazidime MIC

values, while only modest changes in the MIC values for other antibiotics, such as cefuroxime or

aztreonam (Figure 6.3, Supplementary Table 7) were recorded. Importantly, MIC values were

comparable for both the evolved and the original strains, suggesting that increased resistance emerged

165

from mutational changes occurring on the resistance plasmid, and likely either on the -lactamase

sequence or its promoter region.

Figure 6.3 Determination of -lactam MIC values (µg/mL) of three isolates of E. coli MC1000 pDM2-blaSHV-1 obtained

during passage II from plates containing 32 g/mL of ceftazidime. Graphs labelled “Evolved” show the MIC values of the

three isolates (#1, #2, #3) directly isolated from the experimental evolution experiment. Graphs labelled “Original” graphs

show the MIC values of the un-evolved E. coli MC1000 harbouring the plasmids obtained from the three evolved isolates.

“Background MICs” are the MIC values of un-evolved E. coli MC1000 pDM2-blaSHV-1. No changes to the MIC values were

observed for the aminoglycoside, gentamicin (white bars), which served as a negative control. Three independent experiments

are shown for the background MIC values, and one independent experiment is shown for Evolved and Original strains.

In the absence of DsbA, emerging resistant colonies had increased MIC values for cefuroxime (6-8-

fold), ceftazidime (8-16-fold), and aztreonam (4-8-fold) (Figure 6.4, Supplementary Table 7). However,

these high MIC values were not recapitulated when the evolved plasmids were transformed into the

original dsbA mutant background (Figure 6.4). Moreover, transformation of the evolved plasmids into

the parental strain, which can introduce the native disulfide bond into the SHV enzyme, also failed to

reproduce the high MICs recorded for the evolved dsbA mutants. Together, these results suggest that

resistance in this background is not only less likely to evolve (Table 11) but is also mostly arising from

changes to the strain background rather than modifications to the enzyme sequence.

166

Figure 6.4 Determination of -lactam MIC values (µg/mL) of three isolates of E. coli MC1000 dsbA pDM2-blaSHV-1

obtained during passage II from plates containing 32 g/mL of ceftazidime. Graphs labelled “Evolved” show the MIC

values of the three isolates (#1, #2, #3) directly isolated from the experimental evolution experiment. Graphs labelled

“Original” graphs show the MIC values of the un-evolved E. coli MC1000 or its dsbA mutant harbouring the plasmids obtained

from the three evolved isolates. “Background MICs” are the MIC values of un-evolved E. coli MC1000 dsbA pDM2-blaSHV-

1. No changes to the MIC values were observed for the aminoglycoside, gentamicin (white bars), which served as a negative

control. Three independent experiments are shown for the background MIC values, and one independent experiment is shown

for Evolved and Original strains.

Evolution of the single-cysteine SHV-1 mutant led to intermediate phenotypes. The evolved strains

showed increased MIC values for cefuroxime (2-3-fold) and ceftazidime (8-12-fold), but values for

aztreonam remain unchanged (Figure 6.5, Supplementary Table 7). Transformation of the evolved

pDM2-blaSHV-1 C54A plasmid into the original E. coli MC1000 background, conferred lower levels of

ceftazidime resistance for isolates #1 and #2, suggesting that like with the dsbA mutant strains, the

increased resistance could be arising from changes to the strain, albeit at lesser extents.

167

Figure 6.5 Determination of -lactam MIC values (µg/mL) of three isolates of E. coli MC1000 pDM2-blaSHV-1 C54A

obtained during passage II from plates containing 32 g/mL of ceftazidime. Graphs labelled “Evolved” show the MIC

values of the three isolates (#1, #2, #3) directly isolated from the experimental evolution experiment. Graphs labelled

“Original” graphs show the MIC values of the un-evolved E. coli MC1000 harbouring the plasmids obtained from the three

evolved isolates. “Background MICs” are the MIC values of un-evolved E. coli MC1000 pDM2-blaSHV-1 C54A. No changes to

the MIC values were observed for the aminoglycoside, gentamicin (white bars), which served as a negative control. Three

independent experiments are shown for the background MIC values, and one independent experiment is shown for Evolved

and Original strains.

In all three strain backgrounds, no changes to carbapenem susceptibility were recorded (Supplementary

Table 7). Further, no changes in the MIC values of the gentamicin control were recorded (Figure 6.3,

Figure 6.4, Figure 6.5), indicating that any changes to the strains during evolution are not affecting the

general ability of bacteria to withstand antibiotic stress. It should be noted that due to time limitations,

the MIC values reported here represent a single MIC experiment, and biological replicates would be

needed to fully confirm the reliability of the presented data.

6.3.3 Increase of the hydrolytic spectrum of SHV-1 does not affect bacterial fitness

Evolution of antibiotic resistance is often linked to fitness costs.450 For this reason bacterial growth of

the evolved, as well as the original E. coli MC1000 strains harbouring the evolved plasmids, was

168

assessed under standard growth conditions to determine whether the experimental evolution process

resulted in the generation of strains with significant growth defects. No differences in growth were

observed for any of the evolved strains, when compared to their original background counterparts

carrying evolved plasmids that were generated during evolution (Figure 6.6).

Figure 6.6 The experimental evolution process does not affect the fitness of any of the evolved strains. Growth rates of

all three evolved strain backgrounds under standard growth conditions were compared to the growth of the original strain

backgrounds harbouring the plasmids obtained from the evolved strains. SD is marked by the light blue and light grey shaded

area in each graph. Significance determined using unpaired T-test with Welch’s correction; n=3; 4 degrees of freedom. For

multiple comparisons: t-value=2.118, p=0.1016 (non-significance, Original MC1000 evolved pDM2-blaSHV-1 #3 versus

Evolved MC1000 pDM2-blaSHV-1 #3 ); t-value=1.783, p=0.149 (non-significance, Original MC1000 evolved dsbA pDM2-

blaSHV-1 #3 versus Evolved MC1000 dsbA pDM2-blaSHV-1 #3); t-value=0.03789, p=0.9716 (non-significance, Original

MC1000 dsbA evolved dsbA pDM2-blaSHV-1 #3 versus Evolved MC1000 dsbA pDM2-blaSHV-1 #3); t-value=1.360, p=0.2455

(non-significance, Original MC1000 evolved pDM2-blaSHV-1 C54A #3 versus Evolved MC1000 pDM2-blaSHV-1 C54A #3).

169

6.4 DISCUSSION

Native disulfide bonds play a key role in the stability, folding and function of many extra-cytoplasmic

proteins in Gram-negative bacteria.236 Further, the importance of the DSB system for the activity and

stability of class A, B, and D -lactamases has been investigated in the previous chapters (chapter 3, 4

and APPENDIX II). These investigations showed an increased dependence of broad-spectrum -

lactamase enzymes on the DSB oxidase DsbA, compared to narrow-spectrum ones. We hypothesized

that this was due to the fact that the looser active site of broader-spectrum enzymes, necessary to allow

the binding of more complex -lactams, might lead to increased protein flexibility and thus greater

dependence on the stability afforded by the presence of a covalent linkage, like a disulfide bond.185,436,443

Amino acid substitutions in the primary sequences of -lactamase enzymes are one of the key factors

behind the emergence of resistance evolution in this family of proteins. Point changes in numerous

narrow-spectrum hydrolases have been observed to give rise to extended-spectrum and carbapenemase

enzymes that are of great concern to current antibiotic therapy. New-function gene mutations driving

the expansion of the hydrolytic spectrum of the enzymes, are commonly supported by secondary non-

functional mutations, which result in increased protein stability. Extending this to disulfide bonds,

which are known to generally increase protein stability, this chapter focused on testing whether the

process of oxidative protein folding is key for the evolution of broad-spectrum cysteine-containing -

lactamases.451

A pilot experimental evolution study of the narrow-spectrum SHV-1 -lactamase from K. pneumoniae

was performed in presence and absence of its native disulfide bond, using an adaptation of the method

developed by Fröhlich et al.437 Strains expressing SHV-1 were exposed to gradually increasing

concentrations of the complex -lactam ceftazidime, which they could not hydrolyse prior to this

exposure. By abrogating the capacity for disulfide bond formation via deletion of dsbA or removal of

one of the two conserved cysteine residues of the enzyme, 8- to 40-fold decrease in the number of

breakthrough colonies was observed at ceftazidime concentrations higher than the initial strain MIC,

compared to the parental strain.

-lactam MIC values were recorded for three evolved colonies per strain background and in all cases

increase in resistance to complex -lactams was confirmed. However, further MIC testing showed that

170

in the absence of the enzyme’s native disulfide bond this increased capacity to survive -lactam

antibiotic pressure may be arising from mutations on the strain background rather than the -lactamase

gene. This is particularly notable in the dsbA mutant strain, where evolved strains seem to have the

highest -lactam MIC values, but transformation of the evolved plasmids into the original strain

backgrounds does not recapitulate these MICs. This suggests that evolutionary pressure on disulfide-

free SHV-1 enzymes, not only leads to the emergence of drastically fewer resistant colonies, but also

that resistance does not originate from the evolution of the hydrolytic enzyme.

Despite these promising results, further work is required to fully elucidate the role of the DSB system

in -lactamase evolution. Briefly, this includes 1) DNA sequencing of the evolved plasmids for at least

10% of the resistant colonies obtained; 2) characterisation of the resistant colonies used for sequencing

in a similar manner as performed in this chapter (-lactam MIC value determination in the evolved and

original strain backgrounds); 3) repeating the experimental evolution step on strains expressing SHV-1

using ceftazidime, but also another -lactam antibiotic, like cefuroxime or aztreonam; 4) testing of the

experimental evolution process on the second narrow-spectrum enzyme, TEM-1.

171

7 DISCUSSION AND FUTURE WORK

The increasing emergence of resistant bacterial strains represents a silent pandemic that results in

hundreds of thousands of deaths every year.452 Despite extensive efforts over the past decades, few

novel therapeutics are in development and even fewer are successfully reaching the clinics. Generation

of new antibiotic compounds is particularly challenging for many Gram-negative bacteria, such as E.

coli, P. aeruginosa, K. pneumoniae and A. baumannii, which are currently classified as microorganisms

of critical importance.452 Therefore, in a race to extend the lifetime of already approved drugs, new

strategies aiming to break antimicrobial resistance are of growing interest. In this work the DSB system,

a conserved oxidative protein folding pathway in the cell envelope of Gram-negative bacterial species,

is characterised for its role in intrinsic and acquired resistance caused by -lactamase enzymes and

multi-drug efflux pumps, in order to establish whether it could be a novel target against antimicrobial

resistance.

-lactamase enzymes are expressed from a variety of genetic locations and are major contributors to

multidrug-resistant phenotypes due to their ability to hydrolyse different classes of -lactam antibiotics,

from penicillins to carbapenems and monobactams. As such, new approaches that would allow us to

inhibit their hydrolytic activity have great potential for targeting antibiotic resistance. In this work a

total of 13 phylogenetically distinct -lactamase enzymes from different Ambler classes and organisms

(B. pseudomallei, F. tularensis, P. aeruginosa, P. otitidis, S. marcescens and S. maltophilia) were

characterised in the presence and absence of DsbA to assess their functional dependence on this key

oxidative folding catalyst. For most of these enzymes, decrease in protein stability and/or hydrolytic

activity resulted in decreased MIC values against multiple -lactam antibiotics, both in model E. coli

strains and in multidrug-resistant P. aeruginosa and S. maltophilia clinical isolates. Notably the greatest

effects were observed for enzymes with broader hydrolytic activities, whilst enzymes of narrower

spectra, such as CARB-2 or FTU-1, were less dependent on their disulfide bonds (Figure 4.1). In vivo

clearance assays using the wax moth model confirmed that in infections of G. mellonella larvae with a

P. aeruginosa isolate expressing the carbapenemase OXA-198, deletion of dsbA in combination with

piperacillin treatment robustly decreased the bacterial load of the infection by more than 99% (Figure

3.11). This result demonstrates the promise of this approach as a strategy to break antimicrobial

resistance. Time constraints prevented further investigation of the effects of abrogating the function of

the DSB system in vivo. Future work should include the use of the same G. mellonella infection model

172

to investigate whether these observations can be recapitulated in other clinical isolates, for example P.

aeruginosa strains expressing AIM-1 or OXA-19, which were sensitised to -lactam antibiotics when

dsbA was deleted, or inherently resistant organisms, like S. maltophilia, where deletion of DSB genes

resulted in decrease of its ability to tolerate cephalosporin antibiotics (Figure 4.5, Figure 4.6).

Ultimately, a true trial of this strategy would require testing it in a relevant murine infection model,

such as the sub-cutaneous abscess model.453 Implementation of this model is straightforward, enables

highly reproducible chronic infections, and has been shown to be compatible with most organisms of

interest, like P. aeruginosa. The model involves an injection of bacteria underneath the thin skeletal

muscle at the right dorsum on the back of shaved mice. This results in bacterial growth at high densities

within a raised lump around the inoculation point, and subsequent formation of an abscess above the

lump. Bacteria can persist in the abscess for days and resultant infections can be treated by injection of

antibiotics directly into the inflamed tissue. Reduction of bacterial load results in decrease of the size

of the abscess allowing visual monitoring of the treatment; additionally, infection progress can also be

assessed by CFU counts per abscess after sacrificing the animals. To design such in vivo experiments,

especially ones involving vertebrate infection models, it is also important to consider that dsbA gene

deletions are not an ideal strategy for abrogating DSB function; loss of oxidative protein folding can

often cause misfolding of virulence factors and affect the ability of pathogens to establish an

infection.216,245 This means that the use of gene deletions could introduce confounding factors that would

prevent the investigation of the role of DsbA in antibiotic resistance. In the absence of an appropriate

DSB system chemical inhibitor for in vivo experiments, a solution to this problem would come from

using inducible dsbA knockdowns, which could be induced at the point of antibiotic treatment. The

dsbA knockdowns would be based on inducible Mobile-CRISPRi gene silencing, which has been shown

to be effective in both P. aeruginosa and K. pneumoniae infections.454 This way the experimental setup

would overall be simulating, as closely as possible, the treatment of human resistant infections.

The importance of the activity of the DSB system for the folding and function of -lactamase enzymes

demonstrated previously in the Mavridou lab, and further proven in this work, raised the question of

whether other AMR determinants, located in the bacterial cell envelope, might depend on this oxidative

folding pathway (Chapters 3, 4, and APPENDIX II).338 A prominent periplasmic mechanism of

resistance in Gram-negative species is the expulsion of toxic compounds from the cell envelope through

the activity of RND efflux pumps. MIC value determination for antibiotics that are pump substrates

showed that the function of the E. coli AcrAB-TolC RND pump is compromised when dsbA is deleted

173

(Figure 5.5). Despite its considerable size, the AcrAB-TolC assembly does not contain any disulfide

bonds. As such, it is not directly acted on by DsbA, and the observed effects had to be the result of

additional protein interactions. The chaperone/protease DegP was identified as the link between efflux

pump efficiency and oxidative folding (Figure 5.7). DegP is a substrate of DsbA and plays a key role

in periplasmic proteostasis under temperature or reductive stress conditions. Absence of its native

disulfide results in its decreased stability and leads to suboptimal clearing of its substrates, including

AcrA, especially in the presence of reductive stress caused by the absence of DsbA that results in overall

increased DegP substrate load.420 Immunoblotting experiments using a labile AcrA counterpart, AcrA

Leu222Gln, in the presence and absence of DsbA/DegP would provide independent confirmation of the

inability of DegP to deal with its substrate load when oxidative protein folding is abrogated, and

strengthen the link between DsbA function and efflux pump activity.420 In addition to the fact that efflux

activity is compromised due to insufficient periplasmic proteostasis when DsbA is absent, the

possibility that DsbA might act as a chaperone for efflux pump components needs to also be considered.

Disulfide-independent DsbA chaperone activity has been shown in the literature, although experimental

validation is scarce.455 To this end, a DsbA variant where the active site cysteines would be replaced by

alanine or serine residues could be used to test whether any efflux pump components are chaperoned

by DsbA. This would be assessed by immunoblotting of each efflux pump component in the absence

of DsbA or in the presence of cysteine-free DsbA. As AcrA accumulates when dsbA is deleted, here we

would be looking for decrease of AcrB or TolC protein levels in the absence of DsbA and reversal of

this phenotype in the presence of cysteine-free DsbA.

Deletion of dsbA was found to sensitise an E. coli strain to chloramphenicol (Figure 5.8). This indicates

that indirectly compromising resistant determinants via affecting periplasmic proteostasis might be a

strategy worth considering, in particular, for resistance targets such as efflux pumps, for which years of

research have generated very few successful inhibitors.156 As such, it would generally be worth to

investigate the role of other cell envelope folding catalysts (for example chaperones or proteases like

SurA, Skp, FkpA, YfgM, PpiA/D, Spy, HdeA/B, DegQ) in the context of antibiotic resistance.214

Surprisingly, despite the proteinaceous nature of many resistance determinants that are localized in the

cell envelope, the role of these folding catalysts in safeguarding their integrity remains largely

unexplored. This is likely due to the fact that most studies on the function of these proteins have been

carried out in susceptible model strains, and more specifically the E. coli K-12. Therefore, it would be

crucial that any future studies on the role of these proteostasis systems in mechanisms of resistance are

carried out on strains that have either been engineered to express specific resistance proteins, as we did

here, or are resistant clinical isolates.

174

In addition to new interventions that would abrogate the function of intrinsic or mobile resistance

mechanisms, it is also important to find ways to delay or prevent antibiotic resistance evolution. -

lactamase enzymes encompass almost 4,000 discrete proteins of diverse hydrolytic capabilities.456 Some

-lactamase families have several hundreds of members and are thus ideal to test such anti-evolution

approaches. In a pilot study using the narrow-spectrum -lactamase SHV-1, it was shown that absence

of the native disulfide bond of the enzyme decreased the efficiency of mutational acquisition of

resistance by 88-98% in response to the applied antibiotic pressure (Table 11). This indicates that the

DSB system is not only a promising target for the abrogation of resistance, but its inhibition would also

block the evolution of narrow-spectrum enzymes to enzymes that can hydrolyse complex -lactams,

which are drugs of clinical importance.

The experiments presented in Chapter 6 of this thesis are only preliminary and numerous additional

tests need to be carried out to confirm the importance of disulfide bond formation for the evolution of

-lactamase-mediated resistance. Namely the following steps will be taken:

1) The experiments on SHV-1 will be repeated using antibiotic pressure applied by ceftazidime;

additionally, other complex -lactams like aztreonam will be tested. In addition to a dsbA mutant and a

strain expressing a single-cysteine variant of SHV-1, a strain producing a double-cysteine -lactamase

variant will also be included. These experiments will confirm the reproducibility of our original

observations on the effects of the absence of disulfide bond formation for the evolution of -lactamase

resistance.

2) For at least 10% of the evolved resistant colonies emerging for each of the tested strains, next

generation whole-plasmid sequencing will be performed using the Illumina MiSeq platform in order to

identify mutations, both in the -lactamase DNA sequence and in the rest of the plasmid, that are

responsible for evolution of resistance. This will be complemented by determination of -lactam MIC

values of evolved strains and of original strains harbouring the evolved plasmids. In cases of

discrepancy between the two sets of MIC values, something expected to happen mostly in strains

expressing enzymes without their native disulfide which are struggling to expand their hydrolytic

spectra, whole-genome sequencing of the evolved strains will be carried out. This will provide

information on the origins of the resistance in strains expressing disulfide-free -lactamases.

175

3) Specific amino acid substitutions that expand the hydrolytic spectrum of SHV-1 have been both

generated in the lab and identified in clinical samples.439,444,446,447 Representative broad-spectrum SHV-

1 point mutants (Gly255Ser, Ala157Val, Glu256Lys, Glu255Ser/Gln256Lys, Ala157Val/Leu33Gln) will be

constructed in the same expression system and their hydrolytic activity will be determined using MIC

assays for a panel of different classes of β-lactam drugs, in the presence and absence of DsbA. This

approach will identify the point during the reported evolution of this enzyme at which disulfide bonds

might become essential for function and thus for evolution of resistance.

4) All of the above experiments will also be performed for strains expressing the narrow-spectrum

enzyme TEM-1. In this case, the following point mutants will be used to increase its hydrolytic activity:

Glu104Lys, Gly236Ser, Arg162Ser, Gly236Ser/Glu104Lys and Arg162Ser/Glu104Ser. Like with SHV-1, these

rationally mutated enzymes with extended hydrolytic spectra are expected to have increased

dependence on their the native disulfide bond, as seen with other enzymes tested in this thesis or

previously (APPENDIX II).338

In summary, this work characterises further the role of the DSB system, and especially of the principal

oxidase DsbA, in antimicrobial resistance with the aim to assess its potential uses as a novel target for

the simultaneous inhibition of several periplasmic resistance determinants. Further, it hints at the

importance of disulfide bonds for the evolution of narrow-spectrum -lactamases to enzymes that can

hydrolyse complex -lactams, including antibiotics of last resort. Given the already known contribution

of the DSB system in bacterial virulence, targeting this non-essential bacterial pathway could have

multiple beneficial outcomes in decreasing bacterial pathogenicity, breaking antibiotic resistance and

blocking resistance evolution.216,236

176

8 REFERENCES

1. Susser, M. & Susser, E. Choosing a future for epidemiology: I. Eras and paradigms. Am. J.

Public Health 86, 668–673 (1996).

2. Nutton, V. The seeds of disease: An explanation of contagion and infection from the Greeks to

the Renaissance. Med. Hist. 27, 1–34 (1983).

3. Committee National Research Council (US). Science, Medicine and Animals. CWL Publishing

Enterprises, Inc., Madison (National Academies Press, 2004).

4. Byrne, J. P. Encyclopedia of the Black Death. (2012).

5. Kong, K.-F., Schneper, L. & Mathee, K. Beta-lactam Antibiotics: From Antibiosis to Resistance

and Bacteriology. Apmis 118, 1–36 (2010).

6. Young, K. D. The Selective Value of Bacterial Shape. Microbiol. Mol. Biol. Rev. 70, 660–703

(2006).

7. Loutet, S. A. & Valvano, M. A. Extreme antimicrobial peptide and polymyxin B resistance in

the genus Burkholderia. Front. Microbiol. 1, 1–8 (2011).

8. Yang, D. C., Blair, K. M. & Salama, N. R. Staying in Shape: the Impact of Cell Shape on

Bacterial Survival in Diverse Environments. Microbiology Mol. Biol. Rev. 80, 187–203 (2016).

9. Glauert, A. M. & Thornley, M. J. The topography of the bacterial cell wall. Annu. Rev.

Microbiol. 23, 159–198 (1969).

10. Raetz, C. R. H. & Whitfield, C. Lipopolysaccharide endotoxins. Annu. Rev. Biochem. 71, 635–

700 (2002).

11. Vollmer, W., Blanot, D. & De Pedro, M. A. Peptidoglycan structure and architecture. FEMS

Microbiol. Rev. 32, 149–167 (2008).

12. Silhavy, T. J., Kahne, D. & Walker, S. The Bacterial Cell Envelope. Cold Spring Harb Perspect

Biol 2, 1–16 (2010).

13. Aminov, R. I., Otto, M. & Sommer, A. A brief history of the antibiotic era: lessons learned and

challenges for the future. Front. Microbiol. 1, 1–7 (2010).

14. Bassett, E. J., Keith, M. S., Armelagos, G. J., et al. Tetracycline-Labeled Human Bone from

Ancient Sudanese Nubia (A.D. 35). Science. 209, 94–95 (1980).

15. Emmerich, R. & Low, O. Bakteriotytische Enzyme als Ursaehe der erworbenen Immunitat und

die Heilung yon Infeetionskrankheiten durch dieselben. Zeitschrift für Hyg. und Infekt. 31, 1–65

(1899).

16. Ehrlich, P. & Hata, S. Experimentelle Chemotherapie der Spirilosen. (Julius Springer, 1910).

177

17. Domagk, G. Ein Beitrag zur Chemotherapie der bakteriellen Infektionen. Dtsch. Med.

Wochenschr 61, (1935).

18. Davies, J. & Davies, D. Origins and Evolution of Antibiotic Resistance. Microbiology Mol. Biol.

Rev. 74, 417–433 (2010).

19. Harbarth, S., Kahlmeter, G., Kluytmans, J., et al. Global priority list of antibiotic-resistant

bacteria to guide research, discovery, and development of new antibiotics. (2017).

20. Tacconelli, E., Carrara, E., Savoldi, A., et al. Discovery, research, and development of new

antibiotics: The WHO priority list of antibiotic-resistant bacteria and tuberculosis. Lancet Infect.

Dis. 18, 318–327 (2018).

21. Olivares, J., Bernardini, A., Garcia-Leon, G., et al. The intrinsic resistome of bacterial

pathogens. Front. Microbiol. 4, 1–15 (2013).

22. Munita, J. M. & Arias, C. A. Mechanisms of Antibiotics Resistance. Microbiol Spectr 4, 1–37

(2016).

23. Bengtsson-Palme, J., Kristiansson, E. & Larsson, D. G. J. Environmental factors influencing the

development and spread of antibiotic resistance. FEMS Microbiol. Rev. 42, 68–80 (2018).

24. Lister, P. D., Wolter, D. J. & Hanson, N. D. Antibacterial-resistant Pseudomonas aeruginosa:

Clinical impact and complex regulation of chromosomally encoded resistance mechanisms.

Clin. Microbiol. Rev. 22, 582–610 (2009).

25. Gellatly, S. L. & Hancock, R. E. W. Pseudomonas aeruginosa: New insights into pathogenesis

and host defenses. Pathog. Dis. 67, 159–173 (2013).

26. Nikaido, H. Outer Membrane Barrier as a Mechanism of Antimicrobial Resistance. Antimicrob.

Agents Chemother. 33, 1831–1836 (1989).

27. Yoshimura, F. & Nikaido, H. Diffusion of -Lactam Antibiotics Through the Porin Channels of

Escherichia coli K-12. Antimicrob. Agents Chemother. 27, 84–92 (1985).

28. Strateva, T. & Yordanov, D. Pseudomonas aeruginosa - A phenomenon of bacterial resistance.

J. Med. Microbiol. 58, 1133–1148 (2009).

29. Pechdre, J.-C. & Kohler, T. Patterns and modes of -lactam resistance in Pseudomonas

aeruginosa. Clin. Microbiol. Infect. 5, 515–518 (1999).

30. Whitchurch, C. B., Tolker-Nielsen, T., Ragas, P. C., et al. Extracellular DNA Required for

Bacterial Biofilm Formation. Science (80-. ). 295, 1487 (2002).

31. Costerton, J. W., Stewart, P. S. & Greenberg, E. P. Bacterial Biofilms : A Common Cause of

Persistent Infections. Science (80-. ). 284, 1318–1323 (1999).

32. Sun, J., Deng, Z. & Yan, A. Bacterial multidrug efflux pumps: Mechanisms, physiology and

pharmacological exploitations. Biochem. Biophys. Res. Commun. 453, 254–267 (2014).

178

33. Aeschlimann, J. R. The role of multidrug efflux pumps in the antibiotic resistance of

Pseudomonas aeruginosa and other Gram-negative bacteria. Pharmacotherapy 23, 916–924

(2003).

34. Bazzini, S., Udine, C., Sass, A., et al. Deciphering the role of RND efflux transporters in

Burkholderia cenocepacia. PLoS One 6, 1–15 (2011).

35. Yonehara, R., Yamashita, E. & Nakagawa, A. Crystal structures of OprN and OprJ, outer

membrane factors of multidrug tripartite efflux pumps of Pseudomonas aeruginosa. Proteins

Struct. Funct. Bioinforma. 84, 759–769 (2016).

36. Aínsa, J. A., Blokpoel, M. C. J., Otal, I., et al. Molecular cloning and characterization of Tap, a

putative multidrug efflux pump present in Mycobacterium fortuitum and Mycobacterium

tuberculosis. J. Bacteriol. 180, 5836–5843 (1998).

37. Laws, M., Shaaban, A. & Rahman, K. M. Antibiotic resistance breakers: Current approaches

and future directions. FEMS Microbiol. Rev. 43, 490–516 (2019).

38. Du, D., Wang-Kan, X., Neuberger, A., et al. Multidrug efflux pumps: structure, function and

regulation. Nat. Rev. Microbiol. 16, 523–539 (2018).

39. Poole, K. Bacterial stress responses as determinants of antimicrobial resistance. J. Antimicrob.

Chemother. 67, 2069–2089 (2012).

40. Dawson, R. J. P. & Locher, K. P. Structure of a bacterial multidrug ABC transporter. Nature

443, 180–185 (2006).

41. Bogomolnaya, L. M., Andrews, K. D., Talamantes, M., et al. The ABC-type efflux pump

MacAB protects Salmonella enterica serovar typhimurium from oxidative stress. MBio 4, 1–11

(2013).

42. Fitzpatrick, A. W. P., Llabrés, S., Neuberger, A., et al. Structure of the MacAB-TolC ABC-type

tripartite multidrug efflux pump. Nat. Microbiol. 2, 1–20 (2017).

43. ZhHeng, J., ZHao, Y., Liu, M., et al. Substrate-bound structure of the E. coli multidrug resistance

transporter MdfA. Cell Res. 25, 1060–1073 (2015).

44. Horiyama, T. & Nishino, K. AcrB, AcrD, and MdtABC multidrug efflux systems are involved

in enterobactin export in Escherichia coli. PLoS One 9, 1–7 (2014).

45. He, X., Szewczyk, P., Karyakin, A., et al. Structure of a cation-bound multidrug and toxic

compound extrusion transporter. Nature 467, 991–994 (2010).

46. Chen, Y. J., Pornillos, O., Lieu, S., et al. X-ray structure of EmrE supports dual topology model.

Proc. Natl. Acad. Sci. U. S. A. 104, 18999–19004 (2007).

47. Du, D., van Veen, H. W., Murakami, S., et al. Structure, mechanism and cooperation of bacterial

multidrug transporters. Curr. Opin. Struct. Biol. 33, 76–91 (2015).

179

48. Daury, L., Orange, F., Taveau, J. C., et al. Tripartite assembly of RND multidrug efflux pumps.

Nat. Commun. 7, 1–8 (2016).

49. Elkins, C. a & Nikaido, H. Substrate Specificity of the RND-Type Multidrug Ef ux Pumps AcrB

and AcrD of Escherichia coli is Determined Predominantly by Two Large Periplasmic Loops.

J. Bacteriol. 184, 6490–6498 (2002).

50. Ma, D., ALberti, M., Lynch, C., et al. The local repressor AcrR plays a modulating role in the

regulation of acrAB genes of Escherichia coli by global stress signals. Mol. Microbiol. 101–112

(1996).

51. Li, M., Gu, R., Su, C., et al. Crystal Structure of the Transcriptional Regulator AcrR from

Escherichia coli. J. Mol. Biol. 374, 591–603 (2007).

52. Chollet, R., Chevalier, J., Bryskier, A., et al. The AcrAB-TolC pump is involved in macrolide

resistance but not in telithromycin efflux in Enterobacter aerogenes and Escherichia coli.

Antimicrob. Agents Chemother. 48, 3621–3624 (2004).

53. Wilke, M. S., Heller, M., Creagh, A. L., et al. The crystal structure of MexR from Pseudomonas

aeruginosa in complex with its antirepressor ArmR. Proc. Natl. Acad. Sci. U.S.A. 105, 14832–

15837 (2008).

54. Ni, R. T., Onishi, M., Mizusawa, M., et al. The role of RND-type efflux pumps in multidrug-

resistant mutants of Klebsiella pneumoniae. Sci. Rep. 10, (2020).

55. Biswas, S., Brunel, J.-M., Dubus, J.-C., et al. Colistin : an update on the antibiotic of the 21st

century. Expert Rev. Anti. Infect. Ther. 10, 917–934 (2012).

56. Rozalski, A., Sidorczyk, Z. & Kotełko, K. Potential Virulence Factors of Proteus Bacilli.

Microbilogy Mol. Biol. Rev. 61, 65–89 (1997).

57. Olaitan, A. O., Morand, S. & Rolain, J.-M. Mechanisms of polymyxin resistance: acquired and

intrinsic resistance in bacteria. Front. Microbiol. 5, 1–18 (2014).

58. Gunn, J. S., Ryan, S. S., Velkinburgh, J. C. V. A. N., et al. Genetic and Functional Analysis of

a PmrA-PmrB-Regulated Locus Necessary for Lipopolysaccharide Modification, Antimicrobial

Peptide Resistance, and Oral Virulence of Salmonella enterica Serovar Typhimurium. Infect.

Immun. 68, 6139–6146 (2000).

59. Gunn, J. S. & Miller, S. I. PhoP-PhoQ Activates Transcription of pmrAB, Encoding a Two-

Component Regulatory System Involved in Salmonella typhimurium Antimicrobial Peptide

Resistance. J. Bacteriol. 178, 6857–6864 (1996).

60. Abraham, N. & Kwon, D. H. A single amino acid substitution in PmrB is associated with

polymyxin B resistance in clinical isolate of Pseudomonas aeruginosa. FEMS Microbiol. Lett.

298, 249–254 (2009).

180

61. Barrow, K. & Kwon, D. H. Alterations in Two-Component Regulatory Systems of phoPQ and

pmrAB Are Associated with Polymyxin B Resistance in Clinical Isolates of Pseudomonas

aeruginosa. Antimicrob. Agents Chemother. 53, 5150–5154 (2009).

62. Tooke, C. L., Hinchliffe, P., Bragginton, E. C., et al. β-Lactamases and β-Lactamase Inhibitors

in the 21st Century. J. Mol. Biol. 431, 3472–3500 (2019).

63. Macheboeuf, P., Contreras-Martel, C., Job, V., et al. Penicillin binding proteins: Key players in

bacterial cell cycle and drug resistance processes. FEMS Microbiol. Rev. 30, 673–691 (2006).

64. Van Heijenoort, J. Recent advances in the formation of the bacterial peptidoglycan monomer

unit. Nat. Prod. Rep. 18, 503–519 (2001).

65. Filipe, S. R. & Tomasz, A. Inhibition of the expression of penicillin resistance in Streptococcus

pneumoniae by inactivation of cell wall muropeptide branching genes. Proc. Natl. Acad. Sci.

U.S.A. 97, 4891–4896 (2000).

66. Schrödinger, L. The PyMOL Molecular Graphics System v.2.1.0.

67. Bozcal, E. & Dagdeviren, M. Toxicity of β-Lactam Antibiotics: Pathophysiology, Molecular

Biology and Possible Recovery Strategies. Poisoning - From Specif. Toxic Agents to Nov. Rapid

Simpl. Tech. Anal. 87–105 (2017) doi:10.5772/intechopen.70199.

68. Tipper, D. J. & Strominger, J. L. Mechanism of action of penicillins: A proposal based on their

structural similarity to acyl-D-alanyl-D-alanine. Proc. Natl. Acad. Sci. U.S.A. 54, 1133–1140

(1965).

69. Tipper, D. J. & Strominger, J. L. Biosynthesis of the Peptidoglycan of Bacterial Cell Walls. J.

Biol. Chem. 243, 3169–3179 (1968).

70. Wise, E. M. & Park, J. T. Penicillin: its basic site of action as an inhibitor of a peptide cross-

linking reaction in cell wall mucopeptide synthesis. Proc. Natl. Acad. Sci. U.S.A. 54, 75–81

(1965).

71. Patrick, G. L. Antibiotic Agents. in Introduction to Medicinal Chemistry 413–467 (2013).

72. Zeng, X. & Lin, J. Beta-lactamase induction and cell wall metabolism in Gram-negative

bacteria. Front. Microbiol. 4, 1–9 (2013).

73. Abraham, E. P. & Chain, E. An Enzyme from Bacteria able to Destroy Penicillin. Nature 146,

837–837 (1940).

74. Ambler, R. P. The Structure of -lactamases. Phil TransR Soc L. 289, 321–331 (1980).

75. Hong, D. J., Bae, I. K., Jang, I. H., et al. Epidemiology and characteristics of metallo--

lactamase-producing Pseudomonas aeruginosa. Infect. Chemother. 47, 81–97 (2015).

76. Selleck, C., Larrabee, J. A., Harmer, J., et al. AIM-1: An Antibiotic-Degrading Metallohydrolase

That Displays Mechanistic Flexibility. Chem. A Eur. J. 22, 17704–17714 (2016).

181

77. Barthelemy, M., Peduzzi, J., Yaghlane, H. Ben, et al. Single amino acid substitution between

SHV-1 -lactamase and cefotaxime-hydrolyzing SHV-2 enzyme. FEBS Lett. 231, 217–220

(1988).

78. Chang, F.-Y., Siu, L. K., Fung, C.-P., et al. Diversity of SHV and TEM -Lactamases in

Klebsiella pneumoniae : Gene Evolution in Northern Taiwan and Two Novel -Lactamases,

SHV-25 and SHV-26. Antimicrob. Agents Chemother. 45, 2407–2413 (2001).

79. Brooke, J. S. Stenotrophomonas maltophilia: An emerging global opportunistic pathogen. Clin.

Microbiol. Rev. 25, 2–41 (2012).

80. Berg, G., Roskot, N. & Smalla, K. Genotypic and phenotypic relationships between clinical and

environmental isolates of Stenotrophomonas maltophilia. J. Clin. Microbiol. 37, 3594–3600

(1999).

81. Mathee, K., Narasimhan, G., Valdes, C., et al. Dynamics of Pseudomonas aeruginosa genome

evolution. Proc. Natl. Acad. Sci. U.S.A. 105, 3100–3105 (2008).

82. Ogawara, H. Possible drugs for the treatment of bacterial infections in the future: anti-virulence

drugs. J. Antibiot. (Tokyo). (2020) doi:10.1038/s41429-020-0344-z.

83. Thomas, C. M., Nielsen, K. M. & N, S. P. Mechanisms of, and barriers to, horizontal gene

transfer between bacteria. Nat. Rev. Microbiol. 3, 711–721 (2005).

84. Ravi, J., Rivera, M. C. & Lake, J. A. Horizontal gene transfer among genomes: The complexity

hypothesis. Proc. Natl. Acad. Sci. U.S.A. 96, 3801–3806 (1999).

85. Sparling, P. F. Genetic Transformation of Neisseria gonorrhoeae Streptomycin Resistance. J.

Bacteriol. 92, 1364–1371 (1966).

86. Lorenz, M. G. & Wackernagel, W. Bacterial Gene Transfer by Natural Genetic Transformation

in the Environment. Microbiol. Rev. 58, 563–602 (1994).

87. Paget, E. & Simonet, P. On the track of natural transformation in soil. FEMS Microbiol. Ecol.

15, 109–118 (1994).

88. Lorenz, M. G., Gerjets, D. & Wackernagel, W. Release of transforming plasmid and

chromosomal DNS from two cultured soil bacteria. Arch Micrbiol 4, 319–326 (1991).

89. Ueda, S. & Hara, T. Studies on nucleic acid production and application. I. Production of

extracellular DNA by Pseudomonas sp. KYU-1 [industrial fermentations]. J. Appl. Biochem. 3,

1–10 (181AD).

90. Rozenberg-Arska, M., Salters, E. C., Van Strijp, J. A., et al. Degradation of Escherichia coli

Chromosomal and Plasmid DNA in Serum. J. Gen. Microbiol. 130, 217–222 (1983).

91. Cehovin, A., Simpson, P. J., Mcdowell, M. A., et al. Specific DNA recognition mediated by a

type IV pilin. Proc. Natl. Acad. Sci. U.S.A. 110, 3065–3070 (2013).

182

92. Chen, I. & Dubnau, D. DNA Uptake During Bacterial Transformation. Nat. Rev. Microbiol. 2,

1–9 (2004).

93. Griffiths, A. J., Miller, J. H., Suzuki, D. T., et al. Transduction. in An Introduction to Genetic

Analysis (2000).

94. Torres-Barceló, C. The disparate effects of bacteriophages on antibiotic-resistant bacteria.

Emerg. Microbes Infect. 7, 1–12 (2018).

95. Ikeda, H. & Tomizawa, J. Transducing Fragments in Generalized Transduction by Phage PI. J.

Mol. Biol. 14, 85–109 (1965).

96. Zinder, N. D. & Lederberg, J. Genetic Exchange in Salmonella. J. Bacteriol. 64, 679–699

(1952).

97. Morse, M. L., Lederberg, E. M. & Lederberg, J. Transduction in Escherichia coli K-12. Genetics

41, 142–156 (1955).

98. Lang, A. S., Zhaxybayeva, O. & Beatty, J. T. Gene transfer agents: phage-like elements of

genetic exchange. Nat. Rev. Microbiol. 10, 472–482 (2013).

99. Haaber, J., Leisner, J. J., Cohn, M. T., et al. Bacterial viruses enable their host to acquire

antibiotic resistance genes from neighbouring cells. Nat. Commun. 7, 1–8 (2016).

100. Waldor, M. K. & Mekalanos, J. J. Lysogenic Conversion by a Filamentous Phage Encoding

Cholera Toxin. Science (80-. ). 272, 1910–1914 (1996).

101. Graf, F. E., Palm, M., Warringer, J., et al. Inhibiting conjugation as a tool in the fight against

antibiotic resistance. Drug Dev. Res. 80, 19–23 (2019).

102. Manson, J. M., Hancock, L. E. & Gilmore, M. S. Mechanism of chromosomal transfer of

Enterococcus faecalis pathogenicity island, capsule, antimicrobial resistance, and other traits.

Proc. Natl. Acad. Sci. U.S.A. 107, 12269–12274 (2010).

103. Domingues, S., Silva, G. J. & Nielsen, K. M. Integrons - Vehicles and pathways for horizontal

dissemination in bacteria. Mob. Genet. Elem. 2 5, 211–223 (2012).

104. Domingues, S., Nielsen, K. M. & da Silva, G. J. Various pathways leading to the acquisition of

antibiotic resistance by natural transformation. Mob. Genet. Elem. 2 6, 257–260 (2012).

105. Stokes, H. W. & Gillings, M. R. Gene flow, mobile genetic elements and the recruitment of

antibiotic resistance genes into Gram-negative pathogens. FEMS Microbiol. Rev. 35, 790–819

(2011).

106. Mahillon, J. & Chandler, M. Insertion Sequences. Microbiology Mol. Biol. Rev. 62, 725–774

(1998).

107. Siguier, P., Gourbeyre, E. & Chandler, M. Bacterial insertion sequences: their genomic impact

and diversity. FEMS Microbiol. Rev. 38, 865–891 (2014).

183

108. Leiros, H. K. S., Borr, P. S., Brandsdal, B. O., et al. Crystal structure of the mobile metallo-β-

lactamase AIM-1 from Pseudomonas aeruginosa: Insights into antibiotic binding and the role

of Gln157. Antimicrob. Agents Chemother. 56, 4341–4353 (2012).

109. Yong, D., Toleman, M. A., Bell, J., et al. Genetic and biochemical characterization of an

acquired subgroup B3 metallo-β-lactamase gene, blaAIM-1, and its unique genetic context in

Pseudomonas aeruginosa from Australia. Antimicrob. Agents Chemother. 56, 6154–6159

(2012).

110. Sherratt, D. J. Plasmids Review. Cell 3, 189–195 (1974).

111. Smillie, C., Garcilla, M. P., Francia, M. V., et al. Mobility of Plasmids. Microbilogy Mol. Biol.

Rev. 74, 434–452 (2010).

112. Couturier, M., Bex, F., Bergquist, P. L., et al. Identification and Classification of Bacterial

Plasmids. Microbiol. Rev. 52, 375–395 (1988).

113. Johnson, T. J. & Nolan, L. K. Pathogenomics of the Virulence Plasmids of Escherichia coli.

Microbiol. Mol. Biol. Rev. 73, 750–774 (2009).

114. Heinemann, J. A. & Traavik, T. Problems in monitoring horizontal gene transfer in field trials

of transgenic plants. Nat. Biotechnol. 22, 1105–1110 (2004).

115. Dempsey, L. A. & Dubnau, D. A. Identification of Plasmid and Bacillus subtilis Chromosomal

Recombination Sites Used for pE194 Integration. J. Bacteriol. 171, 2856–2865 (1989).

116. Majewski, J. & Cohan, F. M. DNA Sequence Similarity Requirements for Interspecific

Recombination in Bacillus. Genetics 153, 1525–1533 (1999).

117. Guerin, É., Cambray, G., Sanchez-Alberola, N., et al. The SOS Response Controls Integron

Recombination. Science (80-. ). 324, 1034 (2009).

118. Gillings, M. R. Integrons : Past , Present , and Future Structure of Integrons. Microbiology Mol.

Biol. Rev. 78, 257–277 (2014).

119. Woodford, N. & Ellington, M. J. The emergence of antibiotic resistance by mutation. Clin.

Microbiol. Infect. 13, 5–18 (2006).

120. Sensi, P. SESSION I History of the Development of Rifampin. Rev. Infect. Dis. 5, 5402–5406

(1983).

121. Floss, H. G. & Yu, T. Rifamycin - Mode of Action, Resistance, and Biosynthesis. Chem. Rev.

106, 621–632 (2005).

122. Heep, M., Beck, D., Bayerdo, E., et al. Rifampin and Rifabutin Resistance Mechanism in

Helicobacter pylori. Antimicrob. Agents Chemother. 43, 1497–1499 (1999).

123. Wehrli, W. Rifampin: Mechanisms of Action and Resistance. Rev. Infect. Dis. 5, 407–411

(1983).

184

124. Ramaswamy, S. & Musser, J. M. Molecular genetic basis of antimicrobial agent resistance in

Mycobacterium tuberculosis: 1998 update. Tuber. Lung Dis. 79, 3–29 (1998).

125. Bradford, P. A. Extended-Spectrum -Lactamases in the 21st Century: Characterization,

Epidemiology, and Detection of This Important Resistance Threat. Clin. Microbiol. Rev. 14,

933–951 (2001).

126. El’Garch, F., Jeannot, K., Hocquet, D., et al. Cumulative effects of several nonenzymatic

mechanisms on the resistance of Pseudomonas aeruginosa to aminoglycosides. Antimicrob.

Agents Chemother. 51, 1016–1021 (2007).

127. Buckner, M. M., Cisa, M. L., Meek, R. W., et al. HIV Drugs Inhibit Transfer of Plasmids

Carrying Extended-Spectrum -Lactamase and Carbapenemase Genes. MBio 11, 1–18 (2020).

128. Reading, C. & Cole, M. Clavulanic acid: a beta lactamase inhibiting beta lactam from

Streptomyces clavuligerus. Antimicrob. Agents Chemother. 11, 852–857 (1977).

129. Reading, C., Farmer, T. & Cole, M. The β-lactamase stability of amoxycillin with the β-

lactamase inhibitor, clavulanic acid. J. Antimicrob. Chemother. 11, 27–32 (1983).

130. Leflon-Guibout, V., Speldooren, V., Heym, B., et al. Epidemiological survey of amoxicillin-

clavulanate resistance and corresponding molecular mechanisms in Escherichia coli isolates in

France: New genetic features of bla(TEM) genes. Antimicrob. Agents Chemother. 44, 2709–

2714 (2000).

131. Mohanty, S., Singhal, R., Sood, S., et al. Comparative in vitro activity of beta-lactam/beta-

lactamase inhibitor combinations against gram negative bacteria. Indian J. Med. Res. 122, 425–

428 (2005).

132. Penwell, W. F., Shapiro, A. B., Giacobbe, R. A., et al. Molecular mechanisms of sulbactam

antibacterial activity and resistance determinants in Acinetobacter baumannii. Antimicrob.

Agents Chemother. 59, 1680–1689 (2015).

133. Russ, D., Glaser, F., Tamar, E. S., et al. Escape mutations circumvent a tradeoff between

resistance to a beta-lactam and a beta-lactamase inhibitor. Nat. Commun. 11, 1–9 (2020).

134. Ehmann, D. E., Jahić, H., Ross, P. L., et al. Avibactam is a covalent, reversible, non-β-lactam

β-lactamase inhibitor. Proc. Natl. Acad. Sci. U.S.A. 109, 11663–11668 (2012).

135. Zhanel, G. G., Lawson, C. D., Adam, H., et al. Ceftazidime-avibactam: A novel

cephalosporin/β-lactamase inhibitor combination. Drugs 73, 159–177 (2013).

136. Tuon, F. F., Rocha, J. L. & Formigoni-Pinto, M. R. Pharmacological aspects and spectrum of

action of ceftazidime–avibactam: a systematic review. Infection 46, 165–181 (2018).

137. Calvopiña, K., Hinchliffe, P., Brem, J., et al. Structural/mechanistic insights into the efficacy of

nonclassical β-lactamase inhibitors against extensively drug resistant Stenotrophomonas

185

maltophilia clinical isolates. Mol. Microbiol. 106, 492–504 (2017).

138. Petty, L. A., Henig, O., Patel, T. S., et al. Overview of meropenem-vaborbactam and newer

antimicrobial agents for the treatment of carbapenem-resistant Enterobacteriaceae. Infect. Drug

Resist. 11, 1461–1472 (2018).

139. US Food and Drug Administration. VABOMERE (meropenem and vaborbactam) for injection.

US FDA 1–24 (2017).

140. Shields, R. K., Chen, L., Cheng, S., et al. Emergence of Ceftazidime-Avibactam Mutations

during Treatment of Carbapenem-Resistant Klebsiella pneumoniae Infections. Antimicrob

Agents Chemother. 61, 1–11 (2017).

141. Krause, K. M., Serio, A. W., Kane, T. R., et al. Aminoglycosides: An overview. Cold Spring

Harb. Perspect. Med. 6, 1–18 (2016).

142. Kotra, L. P., Haddad, J. & Mobashery, S. Aminoglycosides: Perspectives on mechanisms of

action and resistance and strategies to counter resistance. Antimicrob. Agents Chemother. 44,

3249–3256 (2000).

143. Ramirez, M. S. & Tolmasky, M. E. Aminoglycoside Modifying Enzymes. Drug Resist Updat

13, 151–171 (2010).

144. Llano-Sotelo, B., Azucena, E. F., Kotra, L. P., et al. Aminoglycosides modified by resistance

enzymes display diminished binding to the bacterial ribosomal aminoacyl-tRNA site. Chem.

Biol. 9, 455–463 (2002).

145. Livermore, D. M. Of Pseudomonas, porins, pumps and carbapenems. J. Antimicrob. Chemother.

47, 247–250 (2001).

146. Landman, D., Georgescu, C., Martin, D. A., et al. Polymyxins revisited. Clin. Microbiol. Rev.

21, 449–465 (2008).

147. Lin, L., Nonejuie, P., Munguia, J., et al. Azithromycin Synergizes with Cationic Antimicrobial

Peptides to Exert Bactericidal and Therapeutic Activity Against Highly Multidrug-Resistant

Gram-Negative Bacterial Pathogens. EBioMedicine 2, 690–698 (2015).

148. Vaara, M. Covalent structure of two novel neutrophile leucocyte-derived proteins of porcine and

human origin. Microbiol. Rev. 56, 395–411 (1992).

149. Falagas, M. E. & Kasiakou, S. K. Toxicity of polymyxins: A systematic review of the evidence

from old and recent studies. Crit. Care 10, (2006).

150. Viljanen, P. & Vaara, M. Susceptibility of gram-negative bacteria to polymyxin B nonapeptide.

Antimicrob. Agents Chemother. 25, 701–705 (1984).

151. Ofek, I., Cohen, S., Rahmani, R., et al. Antibacterial synergism of polymyxin B nonapeptide

and hydrophobic antibiotics in experimental Gram-negative infections in mice. Antimicrob.

186

Agents Chemother. 38, 374–377 (1994).

152. Giacometti, A., Cirioni, O., Kamysz, W., et al. In vitro activity and killing effect of temporin A

on nosocomial isolates of Enterococcus faecalis and interactions with clinically used antibiotics.

J. Antimicrob. Chemother. 55, 272–274 (2005).

153. Simmaco, M., Mignogna, G., Canofeni, S., et al. Temporins, antimicrobial peptides from the

European red frog Rana temporaria. Eur. J. Biochem. 242, 788–792 (1996).

154. Li, C., Lewis, M. R., Gilbert, A. B., et al. Antimicrobial activities of amine- and guanidine-

functionalized cholic acid derivatives. Antimicrob. Agents Chemother. 43, 1347–1349 (1999).

155. Alakomi, H. L., Paananen, A., Suihko, M. L., et al. Weakening effect of cell permeabilizers on

Gram-negative bacteria causing biodeterioration. Appl. Environ. Microbiol. 72, 4695–4703

(2006).

156. Sharma, A., Gupta, V. K. & Pathania, R. Efflux pump inhibitors for bacterial pathogens: From

bench to bedside. Indian J. Med. Res. 149, 129–145 (2019).

157. Radchenko, M., Symersky, J., Nie, R., et al. Structural basis for the blockade of MATE

multidrug efflux pumps. Nat. Commun. 6, 1–11 (2015).

158. Gupta, S., Cohen, K. A., Winglee, K., et al. Efflux inhibition with verapamil potentiates

bedaquiline in Mycobacterium tuberculosis. Antimicrob. Agents Chemother. 58, 574–576

(2014).

159. Singh, M., Ramdas, J., Sristava, K., et al. Effect of efflux pump inhibitors on the susceptibility

of Mycobacterium tuberculosis to isoniazid. Indian J. Med. Res. 133, 535–540 (2011).

160. Sekyere, J. O. & Amoako, D. G. Carbonyl cyanide m-chlorophenylhydrazine (CCCP) reverses

resistance to colistin, but not to carbapenems and tigecycline in multidrug-resistant

Enterobacteriaceae. Front. Microbiol. 8, 1–9 (2017).

161. Lomovskaya, O., Warren, M. S., Lee, A., et al. Evaluation of antibacterial effect of Vernonia

anthelmintica seed extract and its synergistic effect with antibiotics on resistant. Antimicrob

Agents Chemother. 45, 105–116 (2001).

162. Bhattacharyya, T., Sharma, A., Akhter, J., et al. The small molecule IITR08027 restores the

antibacterial activity of fluoroquinolones against multidrug-resistant Acinetobacter baumannii

by efflux inhibition. Int. J. Antimicrob. Agents 50, 219–226 (2017).

163. Fenosa, A., Fusté, E., Ruiz, L., et al. Role of TolC in Klebsiella oxytoca resistance to antibiotics.

J. Antimicrob. Chemother. 63, 668–674 (2009).

164. Jeon, B. & Zhang, Q. Sensitization of Campylobacter jejuni to fluoroquinolone and macrolide

antibiotics by antisense inhibition of the CmeABC multidrug efflux transporter. J. Antimicrob.

Chemother. 63, 946–948 (2009).

187

165. Rozwandowicz, M., Brouwer, M. S. M., Fischer, J., et al. Plasmids carrying antimicrobial

resistance genes in Enterobacteriaceae. J. Antimicrob. Chemother. 73, 1121–1137 (2018).

166. Rogers, B. A., Sidjabat, H. E., Paterson, D. L., et al. Prolonged carriage of resistant E. coli by

returned travellers: Clonality, risk factors and bacterial characteristics. Eur. J. Clin. Microbiol.

Infect. Dis. 31, 2413–2420 (2012).

167. Kamruzzaman, M., Shoma, S., Thomas, C. M., et al. Plasmid interference for curing antibiotic

resistance plasmids in vivo. PLoS One 12, 1–20 (2017).

168. Buckner, M. M. C., Ciusa, M. L. & Piddock, L. J. V. Strategies to combat antimicrobial

resistance: Anti-plasmid and plasmid curing. FEMS Microbiol. Rev. 42, 781–804 (2018).

169. Garcillán-Barcia, M. P., Jurado, P., González-Pérez, B., et al. Conjugative transfer can be

inhibited by blocking relaxase activity within recipient cells with intrabodies. Mol. Microbiol.

63, 404–416 (2007).

170. Ripoll-Rozada, J., Garcia-Cazoria, Y., Getino, M., et al. Type IV traffic ATPase TrwD as

molecular target to inhibit bacterial conjugation. Mol. Microbiol. 100, 912–921 (2016).

171. Sayer, J. R., Walldén, K., Pesnot, T., et al. 2- and 3-substituted imidazo[1,2-a]pyrazines as

inhibitors of bacterial type IV secretion. Bioorganic Med. Chem. 22, 6459–6470 (2014).

172. Shaffer, C. L., Good, J. A. D., Kumar, S., et al. Peptidomimetic small molecules disrupt type IV

secretion system activity in diverse bacterial pathogens. MBio 7, 1–10 (2016).

173. Anfinsen, C. B. Principles that Govern the Folding of Protein Chains. Science (80-. ). 181, 223–

230 (1973).

174. Krojer, T., Sawa, J., Schäfer, E., et al. Structural basis for the regulated protease and chaperone

function of DegP. Nature 453, 885–890 (2008).

175. Behrens, S., Maier, R., De Cock, H., et al. The SurA periplasmic PPIase lacking its parvulin

domains functions in vivo and has chaperone activity. EMBO J. 20, 285–294 (2001).

176. Shen, Q. T., Bai, X. C., Chang, L. F., et al. Bowl-shaped oligomeric structures on membranes

as DegP’s new functional forms in protein quality control. Proc. Natl. Acad. Sci. U. S. A. 106,

4858–4863 (2009).

177. Walton, T. A., Sandoval, C. M., Fowler, C. A., et al. The cavity-chaperone Skp protects its

substrate from aggregation but allows independent folding of substrate domains. Proc. Natl.

Acad. Sci. U. S. A. 106, 1772–1777 (2009).

178. Weski, J. & Ehrmann, M. Genetic Analysis of 15 Protein Folding Factors and Proteases of the

Escherichia coli Cell Envelope. J. Bacteriol. 194, 3225–3233 (2012).

179. Weatherspoon-griffin, N., Yang, D., Kong, W., et al. The CpxR / CpxA Two-component

Regulatory System Up-regulates the Multidrug Resistance Cascade to Facilitate Escherichia coli

188

Resistance to a Model Antimicrobial Peptide. J. Biol. Chem. bio 289, 32571–32582 (2014).

180. Sabate, R., De Groot, N. S. & Ventura, S. Protein folding and aggregation in bacteria. Cell. Mol.

Life Sci. 67, 2695–2715 (2010).

181. Tokuriki, N., Stricher, F., Serrano, L., et al. How protein stability and new functions trade off.

PLoS Comput. Biol. 4, 35–37 (2008).

182. Bartlett, G. J., Porter, C. T., Borkakoti, N., et al. Analysis of catalytic residues in enzyme active

sites. J. Mol. Biol. 324, 105–121 (2002).

183. Beadle, B. M. & Shoichet, B. K. Structural bases of stability-function tradeoffs in enzymes. J.

Mol. Biol. 321, 285–296 (2002).

184. Herzberg, O. & Moult, J. Analysis of the steric strain in the polypeptide backbone of protein

molecules. Proteins Struct. Funct. Bioinforma. 11, 223–229 (1991).

185. Thomas, V. L., McReynolds, A. C. & Shoichet, B. K. Structural Bases for Stability-Function

Tradeoffs in Antibiotic Resistance. J. Mol. Biol. 396, 47–59 (2010).

186. Frain, K. M., Robinson, C. & van Dijl, J. M. Transport of Folded Proteins by the Tat System.

Protein J. 38, 377–388 (2019).

187. Rapoport, T. A., Li, L. & Park, E. Structural and Mechanistic Insights into Protein Translocation.

Annu. Rev. Cell Dev. Biol. 33, 369–390 (2017).

188. Hiroyuki, M. & Koreaki, I. The Sec protein-translocation pathway. TRENDS Microbiol. 9, 494–

499 (2001).

189. Pradel, N., Delmas, J., Wu, L. F., et al. Sec- and Tat-dependent translocation of β-lactamases

across the Escherichia coli inner membrane. Antimicrob. Agents Chemother. 53, 242–248

(2009).

190. Ast, T., Cohen, G. & Schuldiner, M. A network of cytosolic factors targets SRP-independent

proteins to the endoplasmic reticulum. Cell 152, 1134–1145 (2013).

191. Josefsson, L. G. & Randall, L. L. Different exported proteins in E. coli show differences in the

temporal mode of processing in vivo. Cell 25, 151–157 (1981).

192. Schibich, D., Gloge, F., Pöhner, I., et al. Global profiling of SRP interaction with nascent

polypeptides. Nature 536, 219–223 (2016).

193. Egea, P. F., Shan, S. O., Napetschnig, J., et al. Substrate twinning activates the signal recognition

particle and its receptor. Nature 427, 215–221 (2004).

194. Dilks, K., Rose, R. W., Hartmann, E., et al. Prokaryotic utilization of the twin-arginine

translocation pathway: A genomic survey. J. Bacteriol. 185, 1478–1483 (2003).

195. Urbanus, M. L., Scotti, P. A., Fröderberg, L., et al. Sec-dependent membrane protein insertion:

Sequential interaction of nascent FtsQ with SecY and YidC. EMBO Rep. 2, 524–529 (2001).

189

196. Nielsen, H. & Engelbrecht, J. Identification of prokaryotic and eukaryotic signal peptides and

prediction of their cleavage sites Artificial neural networks have been used for many biological.

Protein Eng. 10, 1–6 (1997).

197. Hessa, T., Kim, H., Bihlmaier, K., et al. Recognition of transmembrane helices by the

endoplasmic reticulum translocon. Nature 433, 377–381 (2005).

198. Driessen, A. J. M. SecB, a molecular chaperone with two faces. Trends Microbiol. 9, 193–196

(2001).

199. Allen, W. J., Corey, R. A., Oatley, P., et al. Two-way communication between SecY and SecA

suggests a brownian ratchet mechanism for protein translocation. Elife 5, 1–23 (2016).

200. Bauer, B. W., Shemesh, T., Chen, Y., et al. A ‘push and slide’ mechanism allows sequence-

insensitive translocation of secretory proteins by the SecA ATPase. Cell 157, 1416–1429 (2014).

201. Park, E. & Rapoport, T. A. Preserving the membrane barrier for small molecules during bacterial

protein translocation. Nature 473, 239–242 (2011).

202. Petru, M., Wideman, J., Moore, K., et al. Evolution of mitochondrial TAT translocases

illustrates the loss of bacterial protein transport machines in mitochondria. BMC Biol. 16, 1–14

(2018).

203. Palmer, T. & Berks, B. C. The twin-arginine translocation (Tat) protein export pathway. Nat.

Rev. Microbiol. 10, 483–496 (2012).

204. Berks, B. C., Sargent, F. & Palmer, T. The Tat protein export pathway. Mol. Microbiol. 35, 260–

274 (2000).

205. Cline, K. Mechanistic aspects of folded protein transport by the twin arginine translocase (Tat).

J. Biol. Chem. 290, 16530–16538 (2015).

206. Patel, R., Smith, S. M. & Robinson, C. Protein transport by the bacterial Tat pathway. Biochim.

Biophys. Acta 1843, 1620–1628 (2014).

207. Gohlke, U., Pullan, L., McDevitt, C. A., et al. The TatA component of the twin-arginine protein

transport system forms channel complexes of variable diameter. Proc. Natl. Acad. Sci. U.S.A.

102, 10482–10486 (2005).

208. DeLisa, M. P., Tullman, D. & Georgiou, G. Folding quality control in the export of proteins by

the bacterial twin-arginine translocation pathway. Proc. Natl. Acad. Sci. U. S. A. 100, 6115–

6120 (2003).

209. Maurer, C., Panahandeh, S., Moser, M., et al. Impairment of twin-arginine-dependent export by

seemingly small alterations of substrate conformation. FEBS Lett. 583, 2849–2853 (2009).

210. Ize, B., Stanley, N. R., Buchanan, G., et al. Role of the Escherichia coli Tat pathway in outer

membrane integrity. Mol. Microbiol. 48, 1183–1193 (2003).

190

211. Krishnappa, L., Monteferrante, C. G. & van Dijl, J. M. Degradation of the twin-arginine

translocation substrate YwbN by extracytoplasmic proteases of Bacillus subtilis. Appl. Environ.

Microbiol. 78, 7801–7804 (2012).

212. McDonough, J. A., Hacker, K. E., Flores, A. R., et al. The twin-arginine translocation pathway

of Mycobacterium smegmatis is functional and required for the export of mycobacterial β-

lactamases. J. Bacteriol. 187, 7667–7679 (2005).

213. Almagro Armenteros, J. J., Tsirigos, K. D., Sønderby, C. K., et al. SignalP 5.0 improves signal

peptide predictions using deep neural networks. Nat. Biotechnol. 37, 420–423 (2019).

214. Goemans, C., Denoncin, K. & Collet, J. F. Folding mechanisms of periplasmic proteins.

Biochim. Biophys. Acta - Mol. Cell Res. 1843, 1517–1528 (2014).

215. Dutton, R. J., Boyd, D., Berkmen, M., et al. Bacterial species exhibit diversity in their

mechanisms and capacity for protein disulfide bond formation. Proc. Natl. Acad. Sci. U.S.A.

105, 11933–11938 (2008).

216. Heras, B., Shouldice, S. R., Totsika, M., et al. DSB proteins and bacterial pathogenicity. Nat.

Rev. Microbiol. 7, 215–225 (2009).

217. Bulleid, N. J. Disulfide Bond Formation in the Mammalian Endoplasmic Reticulum. Cold

Spring Harb. Perspect. Biol. 1–12 (2012).

218. Shouldice, S. R., Heras, B., Walden, P. M., et al. Structure and function of DsbA, a key bacterial

oxidative folding catalyst. Antioxidants Redox Signal. 14, 1729–1760 (2011).

219. Goldberger, R. F., Epstein, C. J. & Anfinsen, C. B. Acceleration of reactivation of reduced

bovine pancreatic ribonuclease by a microsomal system from rat liver. J. Biol. Chem. 238, 628–

635 (1963).

220. Frand, A. R., Cuozzo, J. W. & Kaiser, C. A. Pathways for protein disulphide bond formation.

Cell Biol. 10, 203–210 (2000).

221. Fuchs, S., De Lorenzo, F. & Anfinsen, C. B. Studies on the mechanism of the enzymic catalysis

of disulfide interchange in proteins. J. Biol. Chem. 242, 398–402 (1967).

222. Pollard, M. G., Travers, K. J. & Weissman, J. S. Ero1p : A Novel and Ubiquitous Protein with

an Essential Role in Oxidative Protein Folding in the Endoplasmic Reticulum. Mol. Cell 1, 171–

182 (1998).

223. Gross, E., Sevier, C. S., Vala, A., et al. A new FAD-binding fold and intersubunit disulfide

shuttle in the thiol oxidase Erv2p. Nat. Struct. Biol. 9, 61–67 (2002).

224. Tian, G., Xiang, S., Noiva, R., et al. The crystal structure of yeast protein disulfide isomerase

suggests cooperativity between its active sites. Cell 124, 61–73 (2006).

225. Frand, A. R. & Kaiser, C. A. Two pairs of conserved cysteines are required for the oxidative

191

activity of Ero1p in protein disulfide bond formation in the endoplasmic reticulum. Mol. Biol.

Cell 11, 2833–2843 (2000).

226. Schwaller, M., Wilkinson, B. & Gilbert, H. F. Reduction-reoxidation cycles contribute to

catalysis of disulfide isomerization by protein-disulfide isomerase. J. Biol. Chem. 278, 7154–

7159 (2003).

227. Wang, X., Wang, L., Wang, X., et al. Structural insights into the peroxidase activity and

inactivation of human peroxiredoxin 4. Biochem. J. 441, 113–118 (2012).

228. Zito, E., Melo, E. P., Yang, Y., et al. Oxidative Protein Folding by an Endoplasmic Reticulum-

Localized Peroxiredoxin. Mol. Cell 40, 787–797 (2010).

229. Chatzi, A., Manganas, P. & Tokatlidis, K. Oxidative folding in the mitochondrial intermembrane

space: A regulated process important for cell physiology and disease. Biochim. Biophys. Acta

1863, 1298–1306 (2016).

230. Allen, S., Balabanidou, V., Sideris, D. P., et al. Erv1 mediates the Mia40-dependent protein

import pathway and provides a functional link to the respiratory chain by shuttling electrons to

cytochrome c. J. Mol. Biol. 353, 937–944 (2005).

231. Banci, L., Bertini, I., Cefaro, C., et al. MIA40 is an oxidoreductase that catalyzes oxidative

protein folding in mitochondria. Nat. Struct. Mol. Biol. 16, 198–206 (2009).

232. Kawano, S., Yamano, K., Naoé, M., et al. Structural basis of yeast Tim40/Mia40 as an oxidative

translocator in the mitochondrial intermembrane space. Proc. Natl. Acad. Sci. U.S.A. 106,

14403–14407 (2009).

233. Sideris, D. P., Petrakis, N., Katrakili, N., et al. A novel intermembrane space-targeting signal

docks cysteines onto Mia40 during mitochondrial oxidative folding. J. Cell Biol. 187, 1007–

1022 (2009).

234. Banci, L., Bertini, I., Cefaro, C., et al. Molecular chaperone function of Mia40 triggers

consecutive induced folding steps of the substrate in mitochondrial protein import. Proc. Natl.

Acad. Sci. U.S.A. 107, 20190–20195 (2010).

235. Thorpe, C. & Coppock, D. L. Generating Disulfides in Multicellular Organisms : Emerging

Roles for a New Flavoprotein Family. J. Biol. Chem. 282, 13929–13933 (2007).

236. Landeta, C., Boyd, D. & Beckwith, J. Disulfide bond formation in prokaryotes. Nat. Microbiol.

3, 270–280 (2018).

237. Bardwell, J. C., McGovern, K. & Beckwith, J. Identification of a protein required for disulfide

bond formation in vivo. Cell 67, 581–589 (1991).

238. Dailey, F. E. & Berg, H. C. Mutants in disulfide bond formation that disrupt flagellar assembly

in Escherichia coli. Proc. Natl. Acad. Sci. U.S.A. 90, 1043–1047 (1993).

192

239. Missiakas, D., Schwager, F. & Raina, S. Identification and characterization of a new disulfide

isomerase-like protein (DsbD) in Escherichia coli. EMBO J. 14, 3415–3424 (1995).

240. Bardwell, J. C., Lee, J. O., Jander, G., et al. A pathway for disulfide bond formation in vivo.

Proc. Natl. Acad. Sci. U.S.A. 90, 1038–42 (1993).

241. Messens, J. & Collet, J.-F. Pathways of disulfide bond formation in Escherichia coli. Int. J.

Biochem. Cell Biol. 38, 1050–1062 (2006).

242. Martin, J. L., Bardwell, J. C. A. & Kuriyan, J. Crystal structure of the DsbA protein required for

disulphide bond formation in vivo. Nature 365, 464–468 (1993).

243. Ren, G., Stephan, D., Xu, Z., et al. Properties of the thioredoxin fold superfamily are modulated

by a single amino acid residue. J. Biol. Chem. 284, 10150–10159 (2009).

244. Guddat, L. W., Bardwell, J. C. A., Zander, T., et al. The uncharged surface features surrounding

the active site of Escherichia coli DsbA are conserved and are implicated in peptide binding.

Protein Sci. 6, 1148–1156 (1997).

245. Hatahet, F., Boyd, D. & Beckwith, J. Disulfide bond formation in prokaryotes: History, diversity

and design. Biochim. Biophys. Acta 1844, 1402–1414 (2014).

246. Paxman, J. J., Borg, N. A., Horne, J., et al. The structure of the bacterial oxidoreductase enzyme

DsbA in complex with a peptide reveals a basis for substrate specificity in the catalytic cycle of

DsbA enzymes. J. Biol. Chem. 284, 17835–17845 (2009).

247. Inaba, K., Murakami, S., Suzuki, M., et al. Crystal Structure of the DsbB-DsbA Complex

Reveals a Mechanism of Disulfide Bond Generation. Cell 127, 789–801 (2006).

248. Eisenberg, D., Schwarz, E., Komaromy, M., et al. Analysis of membrane and surface protein

sequences with the hydrophobic moment plot. J. Mol. Biol. 179, 125–142 (1984).

249. Guddat, L. W., Bardwell, J. C. A. & Martin, J. L. Crystal structures of reduced and oxidized

DsbA: Investigation of domain motion and thiolate stabilization. Structure 6, 757–767 (1998).

250. Guddat, L. W., Bardwell, J. C. A., Glockshuber, R., et al. Structural analysis of three His32

mutants of DsbA: Support for an electrostatic role of His32 in DsbA stability. Protein Sci. 6,

1893–1900 (1997).

251. Kadokura, H., Tian, H., Zander, T., et al. Snapshots of DsbA in Action: Detection of Proteins in

the Process of Oxidative Folding. Science (80-. ). 303, 534–537 (2004).

252. Charbonnier, J.-B., Belin, P., Moutiez, M., et al. On the role of the cis-proline residue in the

active site of DsbA. Protein Sci. 8, 96–105 (2008).

253. Chen, J., Song, J. L., Zhang, S., et al. Chaperone activity of DsbC. J. Biol. Chem. 274, 19601–

19605 (1999).

254. Couprie, J., Vinci, F., Dugave, C., et al. Investigation of the DsbA mechanism through the

193

synthesis and analysis of an irreversible enzyme-ligand complex. Biochemistry 39, 6732–6742

(2000).

255. Tinsley, C. R., Voulhoux, R., Beretti, J. L., et al. Three homologues, including two membrane-

bound proteins, of the disulfide oxidoreductase DsbA in Neisseria meningitidis: Effects on

bacterial growth and biogenesis of functional type IV pili. J. Biol. Chem. 279, 27078–27087

(2004).

256. Kortemme, T. & Creighton, T. E. Ionisation of cysteine residues at the termini of model α-helical

peptides. Relevance to unusual thiol pKa values in proteins of the thioredoxin family. J. Mol.

Biol. 253, 799–812 (1995).

257. Grauschopf, U., Winther, J. R., Korber, P., et al. Why is DsbA such an oxidizing disulfide

catalyst? Cell 83, 947–955 (1995).

258. Holmgren, A. Thioredoxin structure and mechanism: conformational changes on oxidation of

the active-site sulfhydryls to a disulfide. Structure 3, 239–243 (1995).

259. Jeng, M. F., Reymond, M. T., Tennant, L. L., et al. NMR characterization of a single-cysteine

mutant of Escherichia coli thioredoxin and a covalent thioredoxin-peptide complex. Eur. J.

Biochem. 257, 299–308 (1998).

260. Carvalho, A. T. P., Fernandes, P. A., Swart, M., et al. Role of the Variable Active Site Residues

in the Function of Thioredoxin Family Oxidoreductases. J. Comput. Chem. 30, 710–724 (2008).

261. Kadokura, H. & Beckwith, J. Detecting Folding Intermediates of a Protein as It Passes Through

the Bacterial Translocation Channel. Cell 138, 1164–1173 (2009).

262. Malojčić, G., Owen, R. L., Grimshaw, J. P. A., et al. Preparation and structure of the charge-

transfer intermediate of the transmembrane redox catalyst DsbB. FEBS Lett. 582, 3301–3307

(2008).

263. Zhou, Y., Cierpicki, T., Flores Jimenez, R. H., et al. NMR Solution Structure of the Integral

Membrane Enzyme DsbB: Functional Insights into DsbB-Catalyzed Disulfide Bond Formation.

Mol. Cell 31, 896–908 (2008).

264. Inaba, K., Murakami, S., Nakagawa, A., et al. Dynamic nature of disulphide bond formation

catalysts revealed by crystal structures of DsbB. EMBO J. 28, 779–791 (2009).

265. Bushweller, J. H. Protein Disulfide Exchange by the Intramembrane Enzymes DsbB, DsbD, and

CcdA. J. Mol. Biol. (in Press. (2020) doi:10.1016/j.jmb.2020.04.008.

266. Łasica, A. M. & Jagusztyn-Krynicka, E. K. The role of Dsb proteins of Gram-negative bacteria

in the process of pathogenesis. FEMS Microbiol. Rev. 31, 626–636 (2007).

267. Inaba, K., Takahashi, Y. H., Ito, K., et al. Critical role of a thiolate-quinone charge transfer

complex and its adduct form in de novo disulfide bond generation by DsbB. Proc. Natl. Acad.

194

Sci. U.S.A. 103, 287–292 (2006).

268. Jander, G., Martin, N. L. & Beckwith, J. Two cysteines in each periplasmic domain of the

membrane protein DsbB are required for its function in protein disulfide bond formation. EMBO

J. 13, 5121–5127 (1994).

269. Kobayashi, T. & Ito, K. Respiratory chain strongly oxidizes the CXXC motif of DsbB in the

Escherichia coli disulfide bond formation pathway. EMBO J. 18, 1192–1198 (1999).

270. Bader, M. W., Xie, T., Yu, C. A., et al. Disulfide bonds are generated by quinone reduction. J.

Biol. Chem. 275, 26082–26088 (2000).

271. Inaba, K. & Ito, K. Paradoxical redox properties of DsbB and DsbA in the protein disulfide-

introducing reaction cascade. EMBO J. 21, 2646–2654 (2002).

272. Inaba, K., Takahashi, Y. H. & Ito, K. Reactivities of quinone-free DsbB from Escherichia coli.

J. Biol. Chem. 280, 33035–33044 (2005).

273. Kadokura, H. & Beckwith, J. Four cysteines of the membrane protein DsbB act in concert to

oxidize its substrate DsbA. EMBO J. 21, 2354–2363 (2002).

274. McCarthy, A. A., Haebel, P. W., Törrönen, A., et al. Crystal structure of the protein disulfide

bond isomerase, DsbC, from Escherichia coli. Nat. Struct. Biol. 7, 196–199 (2000).

275. Ito, K. & Inaba, K. The disulfide bond formation (Dsb) system. Curr. Opin. Struct. Biol. 18,

450–458 (2008).

276. Katzen, F. & Beckwith, J. Role and location of the unusual redox-active cysteines in the

hydrophobic domain of the transmembrane electron transporter DsbD. Proc. Natl. Acad. Sci. U.

S. A. 100, 10471–10476 (2003).

277. Gleiter, S. & Bardwell, J. C. A. Disulfide bond isomerization in prokaryotes. Biochim. Biophys.

Acta 1783, 530–534 (2008).

278. Rietsch, A., Belin, D., Martin, N., et al. An in vivo pathway for disulfide bond isomerization in

Escherichia coli. Proc. Natl. Acad. Sci. U. S. A. 93, 13048–13053 (1996).

279. Rietsch, A., Bessette, P., Georgiou, G., et al. Reduction of the periplasmic disulfide bond

isomerase, DsbC, occurs by passage of electrons from cytoplasmic thioredoxin. J. Bacteriol.

179, 6602–6608 (1997).

280. Segatori, L., Murphy, L., Arredondo, S., et al. Conserved role of the linker α-helix of the

bacterial disulfide isomerase DsbC in the avoidance of misoxidation by DsbB. J. Biol. Chem.

281, 4911–4919 (2006).

281. Haebel, P. W., Goldstone, D., Katzen, F., et al. The disulfide bond isomerase DsbC is activated

by an immunoglobulin-fold thiol oxidoreductase: Crystal structure of the DsbC-DsbDα

complex. EMBO J. 21, 4774–4784 (2002).

195

282. Heras, B., Edeling, M. A., Schirra, H. J., et al. Crystal structures of the DsbG disulfide isomerase

reveal an unstable disulfide. Proc. Natl. Acad. Sci. U.S.A. 101, 8876–8881 (2004).

283. Liu, X. Q. & Wang, C. C. Disulfide-dependent folding and export of Escherichia coli DsbC. J.

Biol. Chem. 276, 1146–1151 (2001).

284. Shouldice, S. R., Cho, S. H., Boyd, D., et al. In vivo oxidative protein folding can be facilitated

by oxidation-reduction cycling. Mol. Microbiol. 75, 13–28 (2010).

285. Kpadeh, Z. Z., Jameson-Lee, M., Yeh, A. J., et al. Disulfide bond oxidoreductase DsbA2 of

Legionella pneumophila exhibits protein disulfide isomerase activity. J. Bacteriol. 195, 1825–

1833 (2013).

286. Bader, M. W., Hiniker, A., Regeimbal, J., et al. Turning a disulfide isomerase into an oxidase :

DsbC mutants that imitate DsbA. EMBO Kournal 20, 1555–1562 (2001).

287. Hiniker, A., Collet, J. F. & Bardwell, J. C. A. Copper stress causes an in vivo requirement for

the Escherichia coli disulfide isomerase DsbC. J. Biol. Chem. 280, 33785–33791 (2005).

288. Depuydt, M., Leonard, S. E., Vertommen, D., et al. A Periplasmic Reducing System Protects

Single Cysteine Residues from Oxidation. Science (80-. ). 326, 1109–1111 (2009).

289. Crooke, H. & Cole, J. The biogenesis of c-type cytochromes in Escherichia coli requires a

membrane-bound protein, DipZ, with a protein disulphide isomerase-like domain. Mol.

Microbiol. 15, 1139–1150 (1995).

290. Katzen, F. & Beckwith, J. Transmembrane electron transfer by the membrane protein DsbD

occurs via a disulfide bond cascade. Cell 103, 769–779 (2000).

291. Hiniker, A., Vertommen, D., Bardwell, J. C. A., et al. Evidence for conformational changes

within DsbD: Possible role for membrane-embedded proline residues. J. Bacteriol. 188, 7317–

7320 (2006).

292. Rozhkova, A., Stirnimann, C. U., Frei, P., et al. Structural basis and kinetics of inter- and

intramolecular disulfide exchange in the redox catalyst DsbD. EMBO J. 23, 1709–1719 (2004).

293. Cho, S. H., Parsonage, D., Thurston, C., et al. A new family of membrane electron transporters

and its substrates, including a new cell envelope peroxiredoxin, reveal a broadened reductive

capacity of the oxidative bacterial cell envelope. MBio 3, 1–11 (2012).

294. Porat, A., Cho, S. H. & Beckwith, J. The unusual transmembrane electron transporter DsbD and

its homologues: A bacterial family of disulfide reductases. Res. Microbiol. 155, 617–622 (2004).

295. Goulding, C. W., Sawaya, M. R., Parseghian, A., et al. Thiol-disulfide exchange in an

immunoglobulin-like fold: Structure of the N-terminal domain of DsbD. Biochemistry 41, 6920–

6927 (2002).

296. Goldstone, D., Haebel, P. W., Katzen, F., et al. DsbC activation by the N-terminal domain of

196

DsbD. Proc. Natl. Acad. Sci. U.S.A. 98, 9551–9556 (2001).

297. Mavridou, D. A. I., Saridakis, E., Kritsiligkou, P., et al. Oxidation state-dependent protein-

protein interactions in disulfide cascades. J. Biol. Chem. 286, 24943–24956 (2011).

298. Kim, J. H., Kim, S. J., Jeong, D. G., et al. Crystal structure of DsbDγ reveals the mechanism of

redox potential shift and substrate specificity. FEBS Lett. 543, 164–169 (2003).

299. Stirnimann, C. U., Rozhkova, A., Grauschopf, U., et al. High-resolution Structures of

Escherichia coli cDsbD in Different Redox States: A Combined Crystallographic, Biochemical

and Computational Study. J. Mol. Biol. 358, 829–845 (2006).

300. Stewart, E. J., Katzen, F. & Beckwith, J. Six conserved cysteines of the membrane protein DsbD

are required for the transfer of electrons from the cytoplasm to the periplasm of Escherichia coli.

EMBO J. 18, 5963–5971 (1999).

301. Chung, J., Chen, T. & Missiakas, D. Transfer of electrons across the cytoplasmic membrane by

DsbD, a membrane protein involved in thiol-disulphide exchange and protein folding in the

bacterial periplasm. Mol. Microbiol. 35, 1099–1109 (2000).

302. Zhou, Y. & Bushweller, J. H. Solution Structure and Elevator Mechanism of the Membrane

Electron Transporter CcdA. Physiol. Behav. 25, 163–169 (2018).

303. Drew, D. & Boudker, O. Shared Molecular Mechanisms of Membrane Transporters. Annu. Rev.

Biochem. 85, 543–572 (2016).

304. Cho, S. H. & Beckwith, J. Two snapshots of electron transport across the membrane: Insights

into the structure and function of DsbD. J. Biol. Chem. 284, 11416–11424 (2009).

305. Cho, S. H., Porat, A., Ye, J., et al. Redox-active cysteines of a membrane electron transporter

DsbD show dual compartment accessibility. EMBO J. 26, 3509–3520 (2007).

306. Heras, B., Scanlon, M. J. & Martin, J. L. Targeting virulence not viability in the search for future

antibacterials. Br. J. Clin. Pharmacol. 79, 208–215 (2015).

307. Miki, T., Okada, N. & Danbara, H. Two periplasmic bisulfide oxidoreductases, DsbA and SrgA,

target outer membrane protein SpiA, a component of the Salmonella pathogenicity island 2 type

III secretion system. J. Biol. Chem. 279, 34631–34642 (2004).

308. Dacheux, D., Epaulard, O., De Groot, A., et al. Activation of the Pseudomonas aeruginosa type

III secretion system requires an intact pyruvate dehydrogenase aceAB operon. Infect. Immun.

70, 3973–3977 (2002).

309. Ha, U., Wang, Y. & Jin, S. DsbA of Pseudomonas aeruginosa Is Essential for Multiple

Virulence Factors. Infect. Immun. 71, 1590–1595 (2003).

310. Sinha, S., Langford, P. R. & Kroll, J. S. Functional diversity of three different DsbA proteins

from Neisseria meningitidis. Microbiology 150, 2993–3000 (2004).

197

311. Grimshaw, J. P. A., Stirnimann, C. U., Brozzo, M. S., et al. DsbL and DsbI Form a Specific

Dithiol Oxidase System for Periplasmic Arylsulfate Sulfotransferase in Uropathogenic

Escherichia coli. J. Mol. Biol. 380, 667–680 (2008).

312. Bouwman, C. W., Kohli, M., Killoran, A., et al. Characterization of SrgA, a Salmonella enterica

Serovar typhimurium Virulence Plasmid-Encoded Paralogue of the Disulfide Oxidoreductase

DsbA, Essential for Biogenesis of Plasmid-Encoded Fimbriae. Microbiology 185, 991–1000

(2003).

313. Matias, V. R. F. & Beveridge, T. J. Cryo-electron microscopy reveals native polymeric cell wall

structure in Bacillus subtilis 168 and the existence of a periplasmic space. Mol. Microbiol. 56,

240–251 (2005).

314. Zuber, B., Haenni, M., Ribeiro, T., et al. Granular layer in the periplasmic space of gram-positive

bacteria and fine structures of Enterococcus gallinarum and Streptococcus gordonii septa

revealed by cryo-electron microscopy of vitreous sections. J. Bacteriol. 188, 6652–6660 (2006).

315. Walden, P., Halili, M., Kurth, F., et al. Rv2969c, essential for optimal growth in Mycobacterium

tuberculosis, is a DsbA-like enzyme that interacts with VKOR-derived peptides and has atypical

features of DsbA-like disulfide oxidases research papers. Acta Crystallogr. Sect. D D6, 1981–

1994 (2013).

316. Goulding, C. W., Apostol, M. I., Gleiter, S., et al. Gram-positive DsbE Proteins Function

Differently from Gram-negative DsbE homologs. J. Biol. Chem. 279, 3516–3524 (2004).

317. Chim, N., Riley, R., The, J., et al. An extracellular disulfide bond forming protein (DsbF) from

Mycobacterium tuberculosis: Structurel, biochemical, and gene expression analysis. J. Mol.

Biol. 396, 1211–1226 (2011).

318. Sassetti, C. M. & Rubin, E. J. Genetic requirements for mycobacterial survival during infection.

Proc. Natl. Acad. Sci. U.S.A. 100, 12989–12994 (2003).

319. Bolhuis, A., Venema, G., Quax, W. J., et al. Functional Analysis of Paralogous Thiol-disulfide

Oxidoreductases in Bacillus subtilis. J. Biol. Chem. 274, 24531–24538 (1999).

320. Heras, B., Kurz, M., Jarrott, R., et al. Staphylococcus aureus DsbA does not have a destabilizing

disulfide: A new paradigm for bacterial oxidative folding. J. Biol. Chem. 283, 4261–4271

(2008).

321. Yu, J. Inactivation of DsbA, but not DsbC and DsbD, affects the intracellular survival and

virulence of Shigella flexneri. Infect. Immun. 66, 3909–3917 (1998).

322. Totsika, M., Vagenas, D., Paxman, J. J., et al. Inhibition of Diverse DsbA Enzymes in Multi-

DsbA Encoding Pathogens. Antioxidants Redox Signal. 29, 653–666 (2018).

323. Denoncin, K. & Collet, J.-F. Disulfide Bond Formation in the Bacterial Periplasm: Major

198

Achievements and Challenges Ahead. Antioxid. Redox Signal. 19, 63–71 (2013).

324. Allen, R. C., Popat, R., Diggle, S. P., et al. Targeting virulence: Can we make evolution-proof

drugs? Nat. Rev. Microbiol. 12, 300–308 (2014).

325. Landeta, C., Blazyk, J. L., Hatahet, F., et al. Compounds targeting disulfide bond forming

enzyme DsbB of Gram-negative bacteria. Nat. Chem. Biol. 11, 292–298 (2015).

326. Smith, R. P., Paxman, J. J., Scanlon, M. J., et al. Targeting bacterial Dsb proteins for the

development of anti-virulence agents. Molecules 21, (2016).

327. Duprez, W., Premkumar, L., Halili, M. A., et al. Peptide inhibitors of the Escherichia coli DsbA

oxidative machinery essential for bacterial virulence. J. Med. Chem. 58, 577–587 (2015).

328. Früh, V., Zhou, Y., Chen, D., et al. Application of fragment-based drug discovery to membrane

proteins: Identification of ligands of the integral membrane enzyme DsbB. Chem. Biol. 17, 881–

891 (2010).

329. Halili, M. A., Bachu, P., Lindahl, F., et al. Small molecule inhibitors of disulfide bond formation

by the bacterial DsbA-DsbB dual enzyme system. ACS Chem. Biol. 10, 957–964 (2015).

330. Adams, L. A., Sharma, P., Mohanty, B., et al. Application of fragment-based screening to the

design of inhibitors of Escherichia coli DsbA. Angew. Chemie - Int. Ed. 54, 2179–2184 (2015).

331. Duprez, W., Bachu, P., Stoermer, M. J., et al. Virtual screening of peptide and peptidomimetic

fragments targeted to inhibit bacterial dithiol oxidase DsbA. PLoS One 10, 1–16 (2015).

332. Kurth, F., Duprez, W., Premkumar, L., et al. Crystal structure of the dithiol oxidase DsbA

enzyme from Proteus mirabilis bound non-covalently to an active site peptide ligand. J. Biol.

Chem. 289, 19810–19822 (2014).

333. Landeta, C., McPartland, L., Tran, N. Q., et al. Inhibition of Pseudomonas aeruginosa and

Mycobacterium tuberculosis disulfide bond forming enzymes. Mol. Microbiol. 111, 918–937

(2019).

334. Meehan, B. M., Landeta, C., Boyd, D., et al. The disulfide bond formation pathway is essential

for anaerobic growth of Escherichia coli. J. Bacteriol. 199, 1–9 (2017).

335. Altmeyer, R. M., McNern, J. K., Bossio, J. C., et al. Cloning and molecular characterization of

a gene involved in Salmonella adherence and invasion of cultured epithelial cells. Mol.

Microbiol. 7, 89–98 (1993).

336. Yu, J., Webb, H. & Hirst, T. R. A homologue of the Escherichia coli DsbA protein involved in

disulphide bond formation is required for enterotoxin biogenesis in Vibrio cholerae. Mol.

Microbiol. 6, 1949–1958 (1992).

337. Totsika, M., Heras, B., Wurpel, D. J., et al. Characterization of two homologous disulfide bond

systems involved in virulence factor biogenesis in uropathogenic Escherichia coli CFT073. J.

199

Bacteriol. 191, 3901–3908 (2009).

338. Furniss, R. C. D., Kadeřábková, N., Barker, D., et al. Breaking antimicrobial resistance by

disrupting extracytoplasmic protein folding. ELife (in Revis. 1–33.

339. Kim, J., Webb, A. M., Kershner, J. P., et al. A versatile and highly efficient method for scarless

genome editing in Escherichia coli and Salmonella enterica. BMC Biotechnol. 14, 1–13 (2014).

340. Mckenzie, G. J. & Craig, N. L. Fast, easy and efficient: site-specific insertion of transgenes into

Enterobacterial chromosomes using Tn 7 without need for selection of the insertion event. BMC

Microbiol. 6, 1–7 (2006).

341. Vasseur, P., Vallet-Gely, I., Soscia, C., et al. The pel genes of the Pseudomonas aeruginosa

PAK strain are involved at early and late stages of biofilm formation. Microbiology 151, 985–

997 (2005).

342. Kaniga, K., Delor, I. & Cornelis, G. R. A wide-host-range suicide vector for improving reverse

genetics in Gram-negative bacteria: inactivation of the blaA gene of Yersinia enterocolitica.

Gene 137–141 (1991).

343. Welker, E., Domfeh, Y., Tyagi, D., et al. Genetic manipulation of Stenotrophomonas

maltophilia. Curr. Protoc. Microbiol. 37, 6F.2.1-6F.2.14 (2015).

344. Hanahan, D. DNA Cloning: A Practical Approach 1: IRL. (McLean Press, Virginia, 1985).

345. Herrero, M., de Lorenzo, V. & Timmis, K. Transposon Vectors Containing Non-Antibiotic

Resistance Selection Markers for Cloning and Stable Chromosomal Insertion of Foreign Genes

in Gram-Negative Bacteria. J. Bacteriol. 172, 6557–6567 (1990).

346. Boyer, H. W. & Roulland-Dussoix, D. Complementation analysis of the restriction and

modification of DNA in Escherichia coli. J. Mol. Biol. 41, 459–472 (1969).

347. Casadaban, M. J. & Cohen, S. N. Analysis of gene control signals by DNA fusion and cloning

in Escherichia coli. J. Mol. Biol. 138, 179–207 (1980).

348. Blattner, F. R., Plunkett, G., Bloch, C. A., et al. The complete genome sequence of Escherichia

coli K-12. Science (80-. ). 277, 1453–1462 (1997).

349. Aubert, D., Poirel, L., Chevalier, J., et al. Oxacillinase-mediated resistance to cefepime and

susceptibility to ceftazidime in Pseudomonas aeruginosa. Antimicrob. Agents Chemother. 45,

1615–1620 (2001).

350. Dortet, L., Poirel, L. & Nordmann, P. Rapid detection of carbapenemase-producing

Pseudomonas spp. J. Clin. Microbiol. 50, 3773–3776 (2012).

351. Mugnier, P., Casin, I., Bouthors, A. T., et al. Novel OXA-10-derived extended-spectrum β-

lactamases selected in vivo or in vitro. Antimicrob. Agents Chemother. 42, 3113–3116 (1998).

352. El Garch, F., Bogaerts, P., Bebrone, C., et al. OXA-198, an acquired carbapenem-hydrolyzing

200

class D β-lactamase from Pseudomonas aeruginosa. Antimicrob. Agents Chemother. 55, 4828–

4833 (2011).

353. Poirel, L., Brinas, L., Verlinde, A., et al. BEL-1, a novel clavulanic acid-inhibited extended-

spectrum β-lactamase, and the class 1 integron In120 in Pseudomonas aeruginosa. Antimicrob.

Agents Chemother. 49, 3743–3748 (2005).

354. Emeraud, C., Escaut, L., Boucly, A., et al. Aztreonam plus clavulanate, tazobactam, or

avibactam for treatment of infections caused by metallo-lactamase-producing gram-negative

bacteria. Antimicrob. Agents Chemother. 63, 1–7 (2019).

355. Rahme, L. G., Stevens, E. J., Wolfort, S. F., et al. Common virulence factors for bacterial

pathogenicity in plants and animals. Science (80-. ). 268, 1899–1902 (1995).

356. Holloway, B. W. Genetics of Pseudomonas. Bacteriol. Rev. 33, 419–443 (1969).

357. Mavridou, D. A. I., Gonzalez, D., Clements, A., et al. The pUltra plasmid series: A robust and

flexible tool for fluorescent labeling of Enterobacteria. Plasmid 87–88, 65–71 (2016).

358. Kessler, B., de Lorenzo, V. & Timmis, K. N. A general system to integrate lacZ fusions into the

chromosomes of gram-negative eubacteria: regulation of the Pm promoter of the TOL plasmid

studied with all controlling elements in monocopy. Mol. Gen. Genet. 233, 293–301 (1992).

359. Paradis-Bleau, C., Kritikos, G., Orlova, K., et al. A Genome-Wide Screen for Bacterial Envelope

Biogenesis Mutants Identifies a Novel Factor Involved in Cell Wall Precursor Metabolism. PLoS

Genet. 10, (2014).

360. Kishigami, S., Akiyama, Y. & Ito, K. Redox states of DsbA in the periplasm of Escherichia coli.

FEBS Lett. 364, 55–58 (1995).

361. Helander, I. M. & Mattila-Sandholm, T. Fluorometric assessment of Gram-negative bacterial

permeabilization. J. Appl. Microbiol. 88, 213–219 (2000).

362. Adobe. Photoshop CS4 Extended v 11.0.

363. McCarthy, R. R., Mazon-Moya, M. J., Moscoso, J. A., et al. Cyclic-di-GMP regulates

lipopolysaccharide modification and contributes to Pseudomonas aeruginosa immune evasion.

Nat. Microbiol. 2, 1–10 (2017).

364. Lee, S., Park, Y. J., Kim, M., et al. Prevalence of Ambler class A and D β-lactamases among

clinical isolates of Pseudomonas aeruginosa in Korea. J. Antimicrob. Chemother. 56, 122–127

(2005).

365. Laudy, A. E., Róg, P., Smolinska-Król, K., et al. Prevalence of ESBL-producing Pseudomonas

aeruginosa isolates in Warsaw, Poland, detected by various phenotypic and genotypic methods.

PLoS One 12, 1–15 (2017).

366. Szarecka, A., Lesnock, K. R., Ramirez-Mondragon, C. A., et al. The Class D β-lactamase

201

family: Residues governing the maintenance and diversity of function. Protein Eng. Des. Sel.

24, 801–809 (2011).

367. Poirel, L., Naas, T. & Nordmann, P. Diversity, epidemiology, and genetics of class D β-

lactamases. Antimicrob. Agents Chemother. 54, 24–38 (2010).

368. Simakov, N., Leonard, D. A., Smith, J. C., et al. A Distal Disulfide Bridge in OXA-1 β-

Lactamase Stabilizes the Catalytic Center and Alters the Dynamics of the Specificity

Determining Ω Loop. J. Phys. Chem. B 121, 3285–3296 (2017).

369. Curley, K. & Pratt, R. F. The oxyanion hole in serine β-lactamase catalysis: Interactions of

thiono substrates with the active site. Bioorg. Chem. 28, 338–356 (2000).

370. Denoncin, K., Vertommen, D., Paek, E., et al. The protein-disulfide isomerase DsbC cooperates

with SurA and DsbA in the assembly of the essential β-barrel protein LptD. J. Biol. Chem. 285,

29425–29433 (2010).

371. Silvestro, L., Weiser, J. N., Axelsen, P. H., et al. Antibacterial and Antimembrane Activities of

Cecropin A in Escherichia coli. Antimicrob Agents Chemother. 44, 602–607 (2000).

372. Hizukuri, Y., Yakushi, T., Kawagishi, I., et al. Role of the intramolecular disulfide bond in FlgI,

the flagellar P-ring component of Escherichia coli. J. Bacteriol. 188, 4190–4197 (2006).

373. Nakae, T., Nakajima, A., Ono, T., et al. Resistance to β-lactam antibiotics in Pseudomonas

aeruginosa due to interplay between the MexAB-OprM efflux pump and β-lactamase.

Antimicrob. Agents Chemother. 43, 1301–1303 (1999).

374. Girlich, D., Naas, T. & Nordmann, P. Biochemical Characterization of the Naturally Occurring

Oxacillinase OXA-50 of Pseudomonas aeruginosa. Antimicrob. Agents Chemother. 48, 2043–

2048 (2004).

375. Arts, I. S., Ball, G., Leverrier, P., et al. Dissecting the machinery that introduces disulfide bonds

in Pseudomonas aeruginosa. MBio 4, (2013).

376. Wang, R., Van Dorp, L., Shaw, L. P., et al. The global distribution and spread of the mobilized

colistin resistance gene mcr-1. Nat. Commun. 9, 1–9 (2018).

377. Toleman, M. A., Bennett, P. M., Bennett, D. M. C., et al. Global emergence of

trimethoprim/sulfamethoxazole resistance in Stenotrophomonas maltophilia mediated by

acquisition of sul genes. Emerg. Infect. Dis. 13, 559–565 (2007).

378. Okazaki, A. & Avison, M. B. Induction of L1 and L2 β-lactamase production in

Stenotrophomonas maltophilia is dependent on an AmpR-type regulator. Antimicrob. Agents

Chemother. 52, 1525–1528 (2008).

379. Lemmen, S. W., Häfner, H., Reinert, R. R., et al. Comparison of serum bactericidal activity of

ceftazidime, ciprofloxacin and meropenem against Stenotrophomonas maltophilia. J.

202

Antimicrob. Chemother. 47, 113–124 (2001).

380. Wiersinga, W. J., van der Poll, T., White, N. J., et al. Melioidosis: Insights into the pathogenicity

of Burkholderia pseudomallei. Nat. Rev. Microbiol. 4, 272–282 (2006).

381. Cox, G. & Wright, G. D. Intrinsic antibiotic resistance: Mechanisms, origins, challenges and

solutions. Int. J. Med. Microbiol. 303, 287–292 (2013).

382. Queenan, A. M., Torres-Viera, C., Gold, H. S., et al. SME-type carbapenem-hydrolyzing class

A β-lactamases from geographically diverse Serratia marcescens strains. Antimicrob. Agents

Chemother. 44, 3035–3039 (2000).

383. Borgianni, L., De Luca, F., Thaller, M. C., et al. Biochemical characterization of the POM-1

metallo-β-lactamase from Pseudomonas otitidis. Antimicrob. Agents Chemother. 59, 1755–1758

(2015).

384. Antunes, N. T., Frase, H., Toth, M., et al. The class A β-lactamase FTU-1 is native to Francisella

tularensis. Antimicrob. Agents Chemother. 56, 666–671 (2012).

385. Ho, P. L., Cheung, T. K. M., Yam, W. C., et al. Characterization of a laboratory-generated

variant of BPS β-lactamase from Burkholderia pseudomallei that hydrolyses ceftazidime. J.

Antimicrob. Chemother. 50, 723–726 (2002).

386. Wachino, J. I., Yoshida, H., Yamane, K., et al. SMB-1, a novel subclass B3 metallo-β-lactamase,

associated with ISCR1 and a class 1 integron, from a carbapenem-resistant Serratia marcescens

clinical isolate. Antimicrob. Agents Chemother. 55, 5143–5149 (2011).

387. Thaller, M. C., Borgianni, L., Di Lallo, G., et al. Metallo-β-lactamase production by

Pseudomonas otitidis: A species-related trait. Antimicrob. Agents Chemother. 55, 118–123

(2011).

388. Donato, J. J., Moe, L. A., Converse, B. J., et al. Metagenomic analysis of apple orchard soil

reveals antibiotic resistance genes encoding predicted bifunctional proteins. Appl. Environ.

Microbiol. 76, 4396–4401 (2010).

389. Poirel, L., Magalhaes, M., Lopes, M., et al. Molecular Analysis of Metallo--Lactamase Gene.

Antimicrob. Agents Chemother. 48, 1406–1409 (2004).

390. Yong, D., Toleman, M. A., Giske, C. G., et al. Characterization of a new metallo-β-lactamase

gene, blaNDM-1, and a novel erythromycin esterase gene carried on a unique genetic structure

in Klebsiella pneumoniae sequence type 14 from India. Antimicrob. Agents Chemother. 53,

5046–5054 (2009).

391. Shakibaie, M. R. & Moradie, M. Detection of Carbenicillin Hydrolysing (CARB) Type of ESBL

Enzyme in Acinetobacter baumannii Strains Isolated from Bacterimic and UTI Patients. Int. J.

Biol. Life Sci. 1, 1–5 (2011).

203

392. Toleman, M. A., Bennett, P. M. & Walsh, T. R. ISCR Elements: Novel Gene-Capturing Systems

of the 21st Century? Microbiol. Mol. Biol. Rev. 70, 296–316 (2006).

393. Berg, G., Eberl, L. & Hartmann, A. The rhizosphere as a reservoir for opportunistic human

pathogenic bacteria. Environ. Microbiol. 7, 1673–1685 (2005).

394. Ju, L.-C., Cheng, Z., Fast, W., et al. The Continuing Challenge of Metallo-β-Lactamase

Inhibition: Mechanism Matters. Trends Pharmacol Sci 39, 635–647 (2018).

395. He, J., Baldini, R. L., Déziel, E., et al. The broad host range pathogen Pseudomonas aeruginosa

strain PA14 carries two pathogenicity islands harboring plant and animal virulence genes. Proc.

Natl. Acad. Sci. U.S.A. 101, 2530–2535 (2004).

396. Lodise, T. P., Lomaestro, B. & Drusano, G. L. Piperacillin-tazobactam for Pseudomonas

aeruginosa infection: Clinical implications of an extended-infusion dosing strategy. Clin. Infect.

Dis. 44, 357–363 (2007).

397. Godfrey, A. J., Wong, S., Dance, D. A. B., et al. Pseudomonas pseudomallei resistance to β-

lactam antibiotics due to alterations in the chromosomally encoded β-lactamase. Antimicrob.

Agents Chemother. 35, 1635–1640 (1991).

398. Tribuddharat, C., Moore, R. A., Baker, P., et al. Burkholderia pseudomallei class a β-lactamase

mutations that confer selective resistance against ceftazidime or clavulanic acid inhibition.

Antimicrob. Agents Chemother. 47, 2082–2087 (2003).

399. Nikaido, H. Structure and mechanism of RND-type multidrug efflux pumps. Adv. Enzymol.

Relat. Areas Mol. Biol. 77 1, 1–60 (2010).

400. Alav, I., Sutton, J. M. & Rahman, K. M. Role of bacterial efflux pumps in biofilm formation. J.

Antimicrob. Chemother. 73, 2003–2020 (2018).

401. Lomovskaya, O. & Lewis, K. Emr, an Escherichia coli locus for multidrug resistance. Proc.

Natl. Acad. Sci. U. S. A. 89, 8938–8942 (1992).

402. Wang, Z., Chen, M., Shi, X., et al. In situ structure of the AcrAB-TolC efflux pump at

subnanometer resolution. bioRxiv (preprint) (2020) doi:10.1101/2020.06.10.144618 .

403. Shi, X., Chen, M., Yu, Z., et al. In situ structure and assembly of the multidrug efflux pump

AcrAB-TolC. Nat. Commun. 10, 4–9 (2019).

404. Lee, A., Mao, W., Warren, M. S., et al. Interplay between efflux pumps may provide either

additive or multiplicative effects on drug resistance. J. Bacteriol. 182, 3142–3150 (2000).

405. Okusu, H., Ma, D. & Nikaido, H. AcrAB efflux pump plays a major role in the antibiotic

resistance phenotype of Escherichia coli multiple-antibiotic-resistance (Mar) mutants. J.

Bacteriol. 178, 306–308 (1996).

406. Wang, Z., Fan, G., Hryc, C. F., et al. An allosteric transport mechanism for the AcrAB-TolC

204

multidrug efflux pump. Elife 6, 1–19 (2017).

407. Hobbs, E. C., Yin, X., Paul, B. J., et al. Conserved small protein associates with the multidrug

efflux pump AcrB and differentially affects antibiotic resistance. Proc. Natl. Acad. Sci. U. S. A.

109, 16696–16701 (2012).

408. Luisi, B., Koronakis, V., Hughes, C., et al. Crystal structure of the bacterial membrane protein

TolC central to multidrug efflux and protein export. Nature 405, 914–919 (2002).

409. Weston, N., Sharma, P., Ricci, V., et al. Regulation of the AcrAB-TolC efflux pump in

Enterobacteriaceae. Res. Microbiol. 169, 425–431 (2018).

410. Martin, R. G., Jair, K., Wolf, R. E., et al. Autoactivation of the marRAB Multiple Antibiotic

Resistance Operon by the MarA Transcriptional Activator in Escherichia coli. J. Bacteriol. 178,

2216–2223 (1996).

411. Webber, M. A., Talukder, A. & Piddock, L. J. V. Contribution of mutation at amino acid 45 of

AcrR to acrB expression and ciprofloxacin resistance in clinical and veterinary Escherichia coli

isolates. Antimicrob. Agents Chemother. 49, 4390–4392 (2005).

412. Everett, M. J., Jin, Y. F., Ricci, V., et al. Contributions of individual mechanisms to

fluoroquinolone resistance in 36 Escherichia coli strains isolated from humans and animals.

Antimicrob. Agents Chemother. 40, 2380–2386 (1996).

413. Hayashi, S., Nakazawa, T., Kimoto, M., et al. The DsbA-DsbB Disulfide Bond Formation

System of Burkholderia cepacia Is Involved in the Production of Protease and Alkaline Metal

Resistance, and Multi-Drug Resistance. Microbiology 44, 41–50 (2000).

414. Manno, G., Dalmastri, C., Tabacchioni, S., et al. Epidemiology and Clinical Course of

Burkholderia cepacia Complex Infections, Particularly Those Caused by Different Burkholderia

cenocepacia Strains, among Patients Attending an Italian Cystic Fibrosis Center. J. Clin.

Microbiol. 42, 1491–1497 (2004).

415. Jones, A. L., DeShazer, D. & Woods, D. E. Identification and characterization of a two-

component regulatory system involved in invasion of eukaryotic cells and heavy-metal

resistance in Burkholderia pseudomallei. Infect. Immun. 65, 4972–4977 (1997).

416. Nies, D. H., Nies, A., Chu, L., et al. Expression and nucleotide sequence of a plasmid-

determined divalent cation efflux system from Alcaligenes eutrophus (heavy metal plasmid

resistance/cation transport system). Proc. Natl. Acad. Sci. U.S.A. 86, 7351–7355 (1989).

417. Rensing, C., Pribyl, T. & Nies, D. H. New functions for the three subunits of the CzcCBA cation-

proton antiporter. J. Bacteriol. 179, 6871–6879 (1997).

418. Nicoloff, H., Perreten, V., McMurry, L. M., et al. Role for tandem duplication and lon protease

in AcrAB-TolC-dependent multiple antibiotic resistance (Mar) in an Escherichia coli mutant

205

without mutations in marRAB or acrRAB. J. Bacteriol. 188, 4413–4423 (2006).

419. Abdi, S. N., Ghotaslou, R., Ganbarov, K., et al. Acinetobacter baumannii efflux pumps and

antibiotic resistance. Infect. Drug Resist. 13, 423–434 (2020).

420. Leiser, O. P., Charlson, E. S., Gerken, H., et al. Reversal of the degP Phenotypes by a Novel

rpoE Allele of Escherichia coli. PLoS One 7, (2012).

421. Gerken, H. & Misra, R. Genetic evidence for functional interactions between TolC and AcrA

proteins of a major antibiotic efflux pump of Escherichia coli. Mol. Microbiol. 54, 620–631

(2004).

422. Clausen, T., Southan, C. & Ehrmann, M. The HtrA family of proteases: Implications for protein

composition and cell fate. Mol. Cell 10, 443–455 (2002).

423. Hiniker, A. & Bardwell, J. C. A. In Vivo Substrate Specificity of Periplasmic Disulfide

Oxidoreductases. J. Biol. Chem. 279, 12967–12973 (2004).

424. Skorko-Glonek, J., Zurawa, D., Tanfani, F., et al. The N-terminal region of HtrA heat shock

protease from Escherichia coli is essential for stabilization of HtrA primary structure and

maintaining of its oligomeric structure. Biochim. Biophys. Acta 1649, 171–182 (2003).

425. Werner, J., Augustus, A. M. & Misra, R. Assembly of TolC, a structurally unique and

multifunctional outer membrane protein of Escherichia coli K12. J. Bacteriol. 185, 6540–6547

(2003).

426. Ariza, R. R., Cohen, S. P., Bachhawat, N., et al. Repressor mutations in the marRAB operon that

activate oxidative stress genes and multiple antibiotic resistance in Escherichia coli. J. Bacteriol.

176, 143–148 (1994).

427. Keeney, D., Ruzin, A., Mcaleese, F., et al. MarA-mediated overexpression of the AcrAB efflux

pump results in decreased susceptibility to tigecycline in Escherichia coli. J. Antimicrob.

Chemother. 61, 46–53 (2008).

428. European Committee on Antimicrobial Susceptibility Testing. Breakpoint tables for

interpretation of MICs and zone diameters. (2020).

429. Bloom, J. D., Labthavikul, S. T., Otey, C. R., et al. Protein stability promotes evolvability. Proc.

Natl. Acad. Sci. U. S. A. 103, 5869–5874 (2006).

430. Schreiber, G., Buckle, A. M. & Fersht, A. R. Stability and function: two constraints in the

evolution of barstar and other proteins. Structure 2, 945–951 (1994).

431. Zhang, X. jun, Baase, W. A., Shoichet, B. K., et al. Enhancement of protein stability by the

combination of point mutations in T4 lysozyme is additive. Protein Eng. Des. Sel. 8, 1017–1022

(1995).

432. Giakkoupi, P., Hujer, A. M., Miriagou, V., et al. Substitution of Thr for Ala-237 in TEM-17,

206

TEM-12 and TEM-26: Alterations in β-lactam resistance conferred on Escherichia coli. FEMS

Microbiol. Lett. 201, 37–40 (2001).

433. Cabot, G., Bruchmann, S., Mulet, X., et al. Pseudomonas aeruginosa ceftolozane-tazobactam

resistance development requires multiple mutations leading to overexpression and structural

modification of Ampc. Antimicrob. Agents Chemother. 58, 3091–3099 (2014).

434. Rosenkilde, C. E. H., Munck, C., Porse, A., et al. Collateral sensitivity constrains resistance

evolution of the CTX-M-15 β-lactamase. Nat. Commun. 10, 1–10 (2019).

435. Helfand, M. S., Bethel, C. R., Hujer, A. M., et al. Understanding resistance to β-lactams and β-

lactamase inhibitors in the SHV β-lactamase: Lessons from the mutagenesis of Ser-130. J. Biol.

Chem. 278, 52724–52729 (2003).

436. Wang, X., Minasov, G. & Shoichet, B. K. Evolution of an antibiotic resistance enzyme

constrained by stability and activity trade-offs. J. Mol. Biol. 320, 85–95 (2002).

437. Fröhlich, C., Sørum, V., Thomassen, A. M., et al. OXA-48-Mediated Ceftazidime-Avibactam

Resistance Is Associated with Evolutionary Trade-Offs. mSphere 4, 1–15 (2019).

438. Schultz, S. C., Farland, G. D. & Richards, J. H. Stability of Wild-Type and Mutant RTEM-1 ,

8-Lactamases: Effect of the Disulfide Bond. PROTEINS Struct. Funct. Genet. 2, 290–297

(1987).

439. Salverda, M. L. M., de Visser, J. A. G. M. & Barlow, M. Natural evolution of TEM-1 -

lactamase: experimental reconstruction and clinical relevance. FEMS Microbiol. Rev. 34, 1015–

1036 (2010).

440. Zhu, F., He, B., Gu, F., et al. Improvement in organic solvent resistance and activity of

metalloprotease by directed evolution. J. Biotechnol. 309, 68–74 (2020).

441. Matthew, M., Hedges, R. W. & Smith, J. T. Types of -Lactamase Determined by Plasmids in

Gram-Negative Bacteria. J. Bacteriol. 138, 657–662 (1979).

442. Arlet, G., Rouveau, M. & Philippon, A. Substitution of alanine for aspartate at position 179 in

the SHV-6 extended-spectrum -lactamase. FEMS Microbiol. Lett. 152, 163–167 (1997).

443. Rasheed, J. K., Jay, C., Metchock, B., et al. Evolution of Extended-Spectrum -Lactam

Resistance ( SHV-8 ) in a Strain of Escherichia coli during Multiple Episodes of Bacteremia.

Antimicrob. Agents Chemother. 41, 647–653 (1997).

444. Barlow, M. & Hall, B. G. Predicting Evolutionary Potential: In Vitro Evolution Accurately

Reproduces Natural Evolution of the TEM -Lactamase. Genetics 160, 823–832 (2002).

445. Billot-Klein, D., Gutmann, L. & Collatz, E. Nucleotide sequence of the SHV-5 β-lactamase gene

of a Klebsiella pneumoniae plasmid. Antimicrob. Agents Chemother. 34, 2439–2441 (1990).

446. Gniadkowski, M. Evolution of extended-spectrum β-lactamases by mutation. Clin. Microbiol.

207

Infect. 14, 11–32 (2008).

447. Matagne, A., Lamotte-Brasseur, J. & Frere, J.-M. Catalytic properties of class A β-lactamases:

efficiency and diversity. Biochem. J. 330, 581–598 (1998).

448. Liakopoulos, A., Mevius, D. & Ceccarelli, D. A review of SHV extended-spectrum β-

lactamases: Neglected yet ubiquitous. Front. Microbiol. 7, 1–27 (2016).

449. Gaibani, P., Campoli, C., Lewis, R. E., et al. In vivo evolution of resistant subpopulations of

KPC-producing Klebsiella pneumoniae during ceftazidime/avibactam treatment. J. Antimicrob.

Chemother. 73, 1525–1529 (2018).

450. Andersson, D. I. & Hughes, D. Antibiotic resistance and its cost: Is it possible to reverse

resistance? Nat. Rev. Microbiol. 8, 260–271 (2010).

451. Heras, B., Kurz, M., Shouldice, S. R., et al. The name’s bond......disulfide bond. Curr. Opin.

Struct. Biol. 17, 691–698 (2007).

452. Cassini, A., Högberg, L. D., Plachouras, D., et al. Attributable deaths and disability-adjusted

life-years caused by infections with antibiotic-resistant bacteria in the EU and the European

Economic Area in 2015: a population-level modelling analysis. Lancet Infect. Dis. 19, 56–66

(2019).

453. Pletzer, D., Mansour, S. C., Wuerth, K., et al. New mouse model for chronic infections by gram-

negative bacteria enabling the study of anti-infective efficacy and host-microbe interactions.

MBio 8, 1–16 (2017).

454. Peters, J. M., Koo, B. M., Patino, R., et al. Enabling genetic analysis of diverse bacteria with

Mobile-CRISPRi. Nat. Microbiol. 4, 244–250 (2019).

455. Zheng, W. D., Quan, H., Song, J. L., et al. Does DsbA have chaperone-like activity? Arch.

Biochem. Biophys. 337, 326–331 (1997).

456. Naas, T., Ouslati, S., Bonnin, R., et al. -lactamase database (BLDB) -structure and function. J

Enzym. Inhib Med Chem 32, 917–919 (2017).

457. Doublet, B., Robin, F., Casin, I., et al. Molecular and biochemical characterization of the natural

chromosome-encoded class A β-lactamase from Pseudomonas luteola. Antimicrob. Agents

Chemother. 54, 45–51 (2010).

458. Aubert, D., Poirel, L., Ali Ben, A., et al. OXA-35 is an OXA-10-related -lactamase from

Pseudomonas aeruginosa. J. Antimicrob. Chemother. 48, 717–721 (2001).

459. Milenkovic, D., Ramming, T., Muller, J. M., et al. Identification of the Signal Directing Tim9

and Tim10 into the Intermembrane Space of Mitochondria. Mol. Biol. Cell 20, 2530–2539

(2009).

208

9 APPENDIX I

Supplementary Table 1 Overview of the β-lactamase enzymes investigated in this thesis. All tested enzymes belong to

distinct phylogenetic clusters.338 The “Cysteine positions” column states the positions of cysteine residues after position 30

and hence, does not include amino acids that would be part of the periplasmic signal peptide which is cleaved after protein

translocation. All β-lactamase enzymes except L2-1 and LUT-1, which are used as negative controls in this study (top two

rows), have one or more disulfide bonds. The “Mob.” (mobilizable) column refers to the possibility for the β-lactamase gene

to be mobilized from the chromosome; “yes” indicates that the gene of interest is located on a mobile element, while “no”

refers to immobile chromosomally-encoded enzymes. The “Spectrum” column refers to the hydrolytic spectrum of each tested

enzyme; tested enzymes are narrow-spectrum β-lactamases (NS), extended spectrum β-lactamases (ESBL) or carbapenemases.

The “Inh.” (inhibition) column refers to classical inhibitor susceptibility i.e. susceptibility to inhibition by clavulanic acid,

tazobactam or sulbactam. The “Organism” column refers to the bacterial species that most commonly express the tested β-

lactamase enzymes.

Enzyme Cysteine

positions

Ambler

class Mob. Spectrum Inh. Organism

L2-1 C82, C136 C233 A no ESBL yes S. maltophilia

LUT-1 C54 C129 A no457 NS yes P. luteola

OXA-4 C43, C63 D yes ESBL yes P. aeruginosa

OXA-10 C44, C51 D yes ESBL no458 P. aeruginosa

OXA-18 C25, C39, C57 D yes ESBL yes P. aeruginosa

OXA-19 C12, C44, C51 D yes ESBL no351 P. aeruginosa

OXA-198 C116, C119 D yes carbapenemase no352 P. aeruginosa

BEL-1 C61 C231 A yes353 ESBL yes P. aeruginosa

BPS-1m C75 C83 C129 A no385 ESBL yes B. pseudomallei

CARB-2 C72 C118 A yes NS yes Pseudomonas. spp.

FTU-1 C60 C230 A no384 NS yes F. tularensis

AIM-1 C31 C56 C194

C199 C234 C274 B3 yes109 carbapenemase no62 P. aeruginosa

L1-1 C239 C265 B3 no62 carbapenemase no62 S. maltophilia

POM-1 C237 C265 B3 no387 carbapenemase no62 P. otitidis

SMB-1 C180 C 185 C226

C260 B3 yes386 carbapenemase no62 Serratia spp.

OXA-50 C208 C211 D no374 NS no374 Pseudomonas spp.

209

Supplementary Table 2 Deletion of dsbA lowers the β-lactam MIC values for E. coli MC1000 expressing diverse β-lactamases. In the absence of DsbA the β-lactam MICs for E. coli

MC1000 expressing disulfide-bond-containing β-lactamases are reduced. This table shows the MIC data used to generate Figure 3.1, Figure 3.2, Figure 4.1, and Figure 4.2. The aminoglycoside

antibiotic gentamicin and E. coli MC1000 strains harbouring pDM1 (vector alone), pDM1-blaL2-1 or pDM1-blaLUT-1 (cysteine-containing β-lactamases which lack disulfide bonds) serve as negative

controls. Combinations for which MICs were not recorded are marked with a dash (-). MIC values (µg/mL) show three independent experiments. The following abbreviations are used: GM,

gentamicin; AC, amoxicillin; AM, ampicillin; XM, cefuroxime; TZ, ceftazidime; IP, imipenem; AT, aztreonam.

Strain (MC1000) GM AC AM XM TZ IP AT

pdM1

dsbA pdM1

1.00, 1.00, 0.50

1.00, 1.00, 0.50

4.00, 6.00, 6.00

6.00, 6.00, 6.00 -

6.00, 6.00, 6.00

4.00, 6.00, 6.00

0.38, 0.38, 0.19

0.19, 0.19, 0.19

0.19, 0.19, 0.19

0.19, 0.25, 0.19

0.125, 0.125, 0.19

0.125, 0.125, 0.19

pDM1-blaL2-1 CF

dsbA pDM1-blaL2-1 CF

0.38, 0.25, 0.75

0.50, 038, 1.00 -

2000, 2000, 2000

2000, 2000, 2000 -

12.0, 8.00, 16.0

8.00, 8.00, 12.0 -

500, 500, 500

500, 1000, 500

pDM1-blaLUT-1

dsbA pDM1-blaLUT-1

0.50, 0.75, 0.38

0.75, 0.75, 0.38 -

2000, 2000, 2000

1000, 2000, 4000 -

1.50, 1.50, 0.75

1.00, 1.00, 0.75 -

32.0, 16.0, 24.0

32.0, 16.0, 12.0

pDM1-blaOXA-4

dsbA pDM1-blaOXA-4

0.25, 0.50, 1.00

0.50, 0.75, 1.00 -

1000, 1000, 1000

250, 250, 250

24.0, 16.0, 16.0

4.00, 6.00, 4.00 - - -

pDM1-blaOXA-10

dsbA pDM1-bla OXA-10

0.50, 1.00, 0.75

0.75, 1.00, 1.00 -

8000, 8000, 8000

2000, 2000, 2000

256, 256, 256

16.0, 24.0, 32.0 - -

4.00, 6.00, 4.00

0.75, 1.00, 0.75

pDM1-bla OXA-18

dsbA pDM1-bla OXA-18

0.75, 0.75, 1.00

0.75, 0.75, 1.00 -

1000, 1000, 1000

500, 500, 500 -

1000, 1000, 1000

500, 500, 500 -

2000, 2000, 2000

1000, 1000, 1000

pDM1-bla OXA-198

dsbA pDM1-bla OXA-198

0.19, 0.75, 1.00

0.38, 1.00, 1.00 -

4000, 4000, 4000

1000, 2000, 1000

16.0, 12.0, 12.0

6.00, 4.00, 6.00 -

32.0, 32.0, 32.0

3.00, 2.00, 4.00 -

pDM1-blaBEL-1

dsbA pDM1-blaBEL-1

0.75, 0.75, 1.00

1.00, 1.00, 1.50 -

2000, 4000, 2000

500, 2000, 1000 -

12.0, 6.00, 8.00

4.00, 2.00, 1.50 -

24.0, 12.0, 16.0

6.00, 4.00, 4.00

pDM1-blaBPS-1m

dsbA pDM1-blaBPS-1m

1.00, 1.00, 0.75

0.75, 0.75, 0.75 -

500, 500, 500

250, 250, 250 -

256, 256, 256

24.0, 24.0, 12.0 -

3.00, 2.00, 1.50

0.75, 0.25, 0.38

210

pDM1-blaCARB-2

dsbA pDM1-blaCARB-2

0.50, 0.75, 1.00

0.75, 0.75, 0.75 -

16000, 16000,

16000

8000, 8000, 8000

12.0, 12.0, 12.0

4.00, 4.00, 4.00 - - -

pDM1-blaFTU-1

dsbA pDM1-blaFTU-1

1.00, 1.00, 075

1.00, 1.00, 1.00 -

500, 500, 500

250, 250, 250 - - - -

pDM1-blaAIM-1

dsbA pDM1-blaAIM-1

0.50, 0.75, 1.00

0.50, 0.75, 1.00 -

4000, 4000, 4000

250, 250, 500 -

6.00, 6.00, 6.00

0.50, 0.75, 0.5

1.00, 2.00, 1.50

0.25, 0.38, 0.75 -

pDM1-blaL1-1CF

dsbA pDM1-blaL1-1CF

0.125, 0.380, 0.75

0.19, 0.380, 1.00

256, 256, 256

16.0, 16.0, 96.0 - -

128, 48, 48

0.38, 0.75, 2

1.00, 1.50, 0.75

0.25, 0.19, 0.19

0.19, 0.125, 0.125

0.094, 0.50, 0.125

pDM1-blaPOM-1

dsbA pDM1-blaPOM-1

1.00, 0.75, 1.00

1.00, 0.75 1.50 -

4000, 4000, 4000

2000, 2000, 2000

256, 256, 256

64.0, 48.0, 32.0 -

4.00, 6.00, 4.00

1.50, 3.00, 2.00 -

pDM1-blaSMB-1

dsbA pDM1-blaSMB-1

1.00, 0.75, 1.00

1.50, 0.75, 1.50 -

2000, 4000, 4000

250, 1000, 1000

6.00, 4.00, 4.00

2.00, 1.50, 1.50 -

1.00, 1.00, 1.00

0.25, 0.38, 0.38 -

pDM1-blaOXA-50

dsbA pDM1-blaOXA-50

1.00, 0.75, 0.75

0.75, 0.75, 1.00

12.0, 8.00, 8.00

6.00, 6.00, 4.00 -

16.0, 24.0, 16.0

3.00, 8.00, 3.00 - - -

CF – Courtesy of Dr R. Christopher D. Furniss.

211

Supplementary Table 3 Chemical inhibition of the DSB system reduces the MIC values of representative -lactam

antibiotics for E. coli MC1000 expressing disulfide-bond-containing class D β-lactamases in a similar manner to the

deletion of dsbA. This table shows the MIC data used to generate Figure 3.8. The aminoglycoside antibiotic gentamicin and

the E. coli MC1000 strain, harbouring the pDM1 empty vector, serve as negative controls. Combinations for which MICs were

not recorded are marked with a dash (-). MIC values (µg/mL) show three independent experiments. The following

abbreviations are used: GM, gentamicin; XM, cefuroxime; IP, imipenem.

Strain (MC1000) Media GM XM IP

pdM1

MHA

M63 + DMSO

M63 +inhibitor

1.00

1.00, 2.00, 1.50

1.00, 1.00, 1.00

6.00

2.00, 2.00, 3.00

2.00, 2.00, 2.00

0.19

0.19, 0.19, 0.19

0.19, 0.19, 0.19

pDM1-blaOXA-4

MHA

M63 + DMSO

M63 +inhibitor

1.00

1.50, 2.00, 2.00

1.50, 1.50, 1.50

16.0

16.0, 12.0, 12.0

3.00, 2.00, 4.00

-

pDM1-blaOXA-10

MHA

M63 + DMSO

M63 +inhibitor

0.50

1.50, 2.00, 1.50

1.50, 1.50, 1.50

256

12.0, 16.0, 24.0

1.00, 2.00, 2.00

-

pDM1-bla OXA-198

MHA

M63 + DMSO

M63 +inhibitor

1.00

1.50, 1.50, 1.50

1.50, 1.50, 1.50

12.0

12.0, 16.0, 8.00

6.00, 6.00, 3.00

32.0

32.0, 16.0, 32.00

4.00, 3.00, 6.00

212

Supplementary Table 4 Antibiotic resistance profiles (MIC values in µg/mL) of the clinical isolates and laboratory

strains tested in this study for -lactam compounds. Values highlighted in pink indicate resistance, as defined by the

EUCAST clinical breakpoint guidelines, whilst values highlighted in light blue indicate antibiotics for which there is no

EUCAST clinical breakpoint. The remaining values (white cells) indicate sensitivity to the tested antibiotic compound.

Combinations for which MIC values were not recorded (yellow cells) are marked with a dash (-). The following abbreviations

are used: AC, amoxicillin; PP, piperacillin; PT, piperacillin/tazobactam; XM, cefuroxime; TZ, ceftazidime; IP, imipenem; AT,

aztreonam.

Strain AC PP PT XM TZ IP AT

Pseudomonas aeruginosa SOF-1

(blaOXA-4) >256 - - >256 2 3 8

Pseudomonas aeruginosa PU21

(blaOXA-10) >256 - - >256 3 6 16

Pseudomonas aeruginosa

(blaOXA-19) >256 >256 - >256 >256 8 >256

Pseudomonas aeruginosa PA41437

(blaOXA-198) >256 24 32CF >256 2 >32 6

Pseudomonas aeruginosa PA14

(blaOXA-50) >256 8 8 >256 1.5 0.38 6

Pseudomonas aeruginosa PAO1 LA

(blaOXA-50) - 4 4 >256 1 2 3

Pseudomonas aeruginosa PAO1 LD

(blaOXA-50) - 4 4 >256 1.5 2 4

Pseudomonas aeruginosa G4R7

(blaOXA-50 blaAIM-1) >256 >256 >256 >256 24 >32 6

Pseudomonas aeruginosa G6R7

(blaOXA-50 blaAIM-1) >256 32 32 >256 24 >32 6

Stenotrophomonas maltophilia GUE

(blaL2-1 blaL1-1) >256 32 8 >256 12 >32 >256

CF – Courtesy of Dr R. Christopher D. Furniss.

213

Supplementary Table 5 Antibiotic resistance profiles (MIC values in µg/mL) of the clinical isolates and laboratory

strains tested in this study for a range of commonly used non--lactam antibiotics. Values highlighted in pink indicate

resistance, as defined by the EUCAST clinical breakpoint guidelines, whilst values highlighted in light blue indicate antibiotics

for which there is no EUCAST clinical breakpoint. The remaining values (white cells) indicate sensitivity to the tested

antibiotic compound. Combinations for which MIC values were not recorded (yellow cells) are marked with a dash (-). The

following abbreviations are used: GM, gentamicin; CO, colistin; CI, ciprofloxacin; TR, trimethoprim; TS,

trimethoprim/sulfamethoxazole.

Strain GM CO CI TR TS

Pseudomonas aeruginosa SOF-1

(blaOXA-4) >256 - >32 >32 -

Pseudomonas aeruginosa PU21

(blaOXA-10) >256 - 0.38 >32 -

Pseudomonas aeruginosa

(blaOXA-19) >256 1 >32 >32 -

Pseudomonas aeruginosa PA41437

(blaOXA-198) 16 1CF >32 >32 -

Pseudomonas aeruginosa PA14

(blaOXA-50) 1.5 1 1 - -

Pseudomonas aeruginosa PAO1 LA

(blaOXA-50) 1.5 2 0.125 - -

Pseudomonas aeruginosa PAO1 LD

(blaOXA-50) 1.5 1 0.125 - -

Pseudomonas aeruginosa G4R7

(blaOXA-50 blaAIM-1) >256 0.5 >32 - -

Pseudomonas aeruginosa G6R7

(blaOXA-50 blaAIM-1) >256 2 >32 - -

Stenotrophomonas maltophilia GUE

(blaL2-1 blaL1-1) 1 2 - >32 0.06

CF – Courtesy of Dr R. Christopher D. Furniss

214

Supplementary Table 6 Deletion of dsbA does not decrease the β-lactam MIC values for E. coli MC1000 expressing the narrow-spectrum β-lactamases TEM-1 and SHV-1 at either

37°C or 42°C. This table shows the MIC data used to generate Figure 6.3, Figure 6.4, and Figure 6.5. The aminoglycoside antibiotic gentamicin and the E. coli MC1000 strains harbouring pDM2

(vector alone) serve as controls. Strains in this table are referred to as Background strains throughout Chapter 6, MIC values highlighted in bold were used to calculate fold change values presented

in Three independent experiments are shown for MIC values recorded at 37°C and one independent experiment is shown for MIC values recorded 42°C. MIC values (µg/mL), abbreviations used:

GM, gentamicin; AC, amoxicillin; XM, cefuroxime; TZ, ceftazidime; IP, imipenem; AT, aztreonam.

Strain (MC1000) Temp (°C) GM AC XM TZ IP AT

pDM2

dsbA pDM2 37

1.00, 0.75, 1.00

1.00, 0.75, 0.75

6.00, 6.00, 6.00

4.00, 4.00, 4.00

6.00, 6.00, 6.00

6.00, 6.00, 6.00

0.25, 0.19, 0.25

0.19, 0.25, 0.25

0.38, 0.38, 0.38

0.38, 0.38, 0.38

0.19, 0.19, 0.19

0.19, 0.19, 0.19

pDM2

dsbA pDM2 42

0.75

1.00

6.00

4.00

6.00

4.00

0.25

0.25

0.38

0.38

0.25

0.19

pDM2-blaSHV-1

dsbA pDM2- blaSHV-1

pDM2- blaSHV-1 C54

37

1.00, 0.75, 1.00

1.00, 0.75, 0.75

0.50, 1.00, 1.00

>256, >256, >256

>256, >256, >256

>256, >256, >256

24.0, 12.0, 12.0

12.0, 8.00, 8.00

12.0, 8.00, 8.00

4.0, 4.0, 4.0

2.0, 2.0, 2.0

2.0, 2.0, 2.0

0.50, 0.38, 0.50

0.50, 0.38, 0.38

0.38, 0.38, 0.38

1.0, 0.75, 0.75

0.75, 0.75, 0.75

1.0, 0.75, 0.75

pDM2-blaSHV-1

dsbA pDM2- blaSHV-1

pDM2- blaSHV-1 C54

42

1.00

1.00

0.50

>256

>256

>256

24.0

16.0

12.0

4.0

3.0

2.0

0.38

0.50

0.50

0.75

1.0

1.0

pDM2-blaTEM-1

dsbA pDM2-blaTEM-1

pDM2-blaTEM-1 C86A

37

0.75, 1.00, 1.00

1.00, 0.75, 1.00

0.75, 1.00, 1.00

>256, >256, >256

>256, >256, >256

>256, >256, >256

12.0, 12.0, 12.0

6.00, 4.00, 6.00

8.00, 6.00, 6.00

0.25, 0.25, 0.25

0.25, 0.50, 0.25

0.50, 0.25, 0.50

0.38, 0.25, 0.25

0.25, 0.25, 0.25

0.25, 0.25, 0.25

0.19, 0.25, 0.19

0.19, 0.125, 0.125

0.25, 0.094, 0.19

pDM2-blaTEM-1

dsbA pDM2-blaTEM-1

pDM2-blaTEM-1 C86A

42

0.75

0.75

0.75

>256

>256

>256

12.0

6.00

8.00

0.25

0.094

0.50

0.38

0.38

0.38

0.19

0.25

0.25

215

Supplementary Table 7 β-lactam MIC values and MIC fold changes (FC) recorded in evolved and original backgrounds, after experimental evolution of E. coli MC1000 strains

expressing the narrow-spectrum β-lactamase SHV-1. This table shows the MIC data used to generate Figure 6.3, Figure 6.4, and Figure 6.5. The fold changes were calculate using Background

strain MIC values (bold font, Supplementary Table 6). The aminoglycoside antibiotic gentamicin serves as a control and shows no changes in MIC values for any of the tested strains. MIC fold

changes (MIC fold changes: > 2, fold change defined as Evolved or Original MC1000 MIC (µg/mL) / median background MC1000 (µg/mL); MIC values (µg/mL); abbreviations used: GM,

gentamicin; AC, amoxicillin; XM, cefuroxime; TZ, ceftazidime; IP, imipenem; AT, aztreonam.

Construct Colony GM AC XM TZ IP AT

MIC FC MIC FC MIC FC MIC FC MIC FC MIC FC

Evolved MC1000 pDM2-blaSHV-1

#1 0.75 0.75 >256 1.0 24.0 2.0 16.0 4.0 0.25 0.50 2.00 2.6

#2 1.00 1.0 >256 1.0 16.0 1.3 8.00 2.0 0.25 0.50 1.00 1.3

#3 1.00 1.0 >256 1.0 16.0 1.3 16.0 4.0 0.25 0.50 1.50 2.0

Original MC1000 evolved pDM2-blaSHV-1

#1 0.75 0.75 >256 1.0 16.0 1.3 160 4.0 0.25 0.50 1.00 1.3

#2 0.75 0.75 >256 1.0 16.0 1.3 8.00 2.0 0.25 0.50 1.00 1.3

#3 0.50 0.50 >256 1.0 12.0 1.0 120. 3.0 0.25 0.50 1.00 1.3

Evolved MC1000 dsbA pDM2-blaSHV-1

#1 0.75 1.0 >256 1.0 48.0 6.0 16.0. 8.0 0.25 0.66 3.00 4.0

#2 1.00 1.0 >256 1.0 48.0 6.0 16.0 8.0 0.38 1.0 4.00 5.3

#3 1.00 1.3 >256 1.0 64.0 8.0 32.0 16 0.50 1.3 6.00 8.0

Original MC1000 dsbA pDM2-blaSHV-1

#1 0.50 0.66 >256 1.0 12.0 1.5 8.00 4.0 0.25 0.66 1.50 2.0

#2 0.75 1.0 >256 1.0 8.00 1.0 16.0 8.0 0.25 0.66 1.50 2.0

#3 0.50 0.66 >256 1.0 12.0 1.5 12.0 6.0 0.25 0.66 1.00 1.3

Original MC1000 pDM2-blaSHV-1

#1 0.75 1.0 >256 1.0 12.0 1.5 6.0 3.0 0.25 0.66 1.00 1.3

#2 0.50 0.66 >256 1.0 12.0 1.5 4.0 2.0 0.25 0.66 1.00 1.3

#3 0.50 0.66 >256 1.0 12.0 1.5 16.0 8.0 0.25 0.66 1.00 1.3

Evolved MC1000 pDM2-blaSHV-1 C54A

#1 1.00 1.0 >256 1.0 24.0 3.0 24.0 12 0.19 0.50 0.75 1.0

#2 0.75 0.75 >256 1.0 16.0 2.0 16.0 8.0 0.19 0.50 0.50 0.66

#3 0.50 0.50 >256 1.0 16.0 2.0 16.0 8.0 0.19 0.50 0.50 0.66

216

Original MC1000 evolved pDM2-blaSHV-1 C54A

#1 0.75 0.75 >256 1.00 8.00 1.0 12.0 6.0 0.19 0.50 0.38 0.50

#2 0.50 0.50 >256 1.00 16.0 2.0 8.00 4.0 0.25 0.66 0.38 0.50

#3 0.50 0.50 >256 1.00 16.0 2.0 12.0 6.0 0.19 0.50 0.38 0.50

217

10 APPENDIX II

Breaking antimicrobial resistance by disrupting extracytoplasmic protein folding

R. Christopher D. Furniss2,†, Nikol Kadeřábková2,†, Declan Barker2, Patricia Bernal3, Evgenia

Maslova4, Amanda A.A. Antwi2, Helen E. McNeil5, Hannah L. Pugh5, Laurent Dortet2,6,7,8, Jessica M.A.

Blair5, Gerald Larrouy-Maumus2, Ronan R. McCarthy4, Diego Gonzalez9, Despoina A.I. Mavridou1,2,*

1Department of Molecular Biosciences, University of Texas at Austin, Austin, 78712, Texas, USA 2MRC Centre for Molecular Bacteriology and Infection, Department of Life Sciences, Imperial College

London, London, SW7 2AZ, UK 3Department of Biology, Faculty of Sciences, Universidad Autónoma de Madrid, Madrid, 28049, Spain 4Division of Biosciences, Department of Life Sciences, College of Health and Life Sciences, Brunel

University London, Uxbridge, UB8 3PH, UK 5Institute of Microbiology and Infection, College of Medical and Dental Sciences, University of

Birmingham, Birmingham, B15 2TT, UK 6Department of Bacteriology-Hygiene, Bicêtre Hospital, Assistance Publique - Hôpitaux de Paris, Le

Kremlin-Bicêtre, 94270, France 7EA7361 “Structure, Dynamics, Function and Expression of Broad-spectrum β-lactamases", Paris-Sud

University, LabEx Lermit, Faculty of Medicine, Le Kremlin-Bicêtre, 94270, France 8French National Reference Centre for Antibiotic Resistance, Le Kremlin-Bicêtre, 94270, France 9Laboratoire de Microbiologie, Institut de Biologie, Université de Neuchâtel, Neuchâtel, 2000,

Switzerland

*Correspondence: despoina.mavridou@austin.utexas.edu †These authors have contributed equally to this work

ABSTRACT

Antimicrobial resistance in Gram-negative bacteria is one of the greatest threats to global health. New

antibacterial strategies are urgently needed, and the development of antibiotic adjuvants that either

neutralize resistance proteins or compromise the integrity of the cell envelope is of ever-growing

interest. Most available adjuvants are only effective against specific resistance proteins from the same

class. Here we demonstrate that disruption of cell envelope protein homeostasis simultaneously

incapacitates three major classes of resistance determinants. In particular, we find that impairing DsbA-

mediated disulfide bond formation incapacitates β-lactamases and mobile colistin resistance enzymes,

whilst also compromising Resistance-Nodulation-Division efflux pumps. Furthermore, we show that

chemical inhibition of DsbA sensitizes multidrug-resistant clinical isolates to existing antibiotics and

that absence of DsbA allows clearance of a multidrug-resistant Pseudomonas aeruginosa strain from

the Galleria mellonella infection model. This work lays the foundation for the development of novel

antibiotic adjuvants that function as broad-acting resistance breakers.

IMPACT STATEMENT: Disruption of disulfide bond formation sensitizes resistant Gram-negative

bacteria expressing β-lactamases, mobile colistin resistance enzymes and efflux pumps to currently

available antibiotics.

218

INTRODUCTION

Antimicrobial resistance (AMR) is one of the most important public health concerns of our time (1).

With few new antibiotics in the pharmaceutical pipeline and multidrug-resistant bacterial strains

continuously emerging, it is more important than ever to develop novel antibacterial strategies and find

alternative ways to break resistance. The development of new treatments for Gram-negative bacteria is

considered critical by the WHO (2), however identifying novel approaches to target these organisms is

particularly challenging due to their unique double-membrane permeability barrier and the vast range

of AMR determinants they produce. For this reason, rather than targeting cytoplasmic processes,

antimicrobial strategies that inhibit cell-envelope components or impair the activity of resistance

determinants are being increasingly pursued (3-7).

The Gram-negative cell envelope is home to many different AMR determinants, with β-lactamase

enzymes currently posing a seemingly insurmountable problem. Almost 4,000 unique enzymes capable

of degrading β-lactam compounds have been identified to date (File S1). Despite the development of

more advanced β-lactams, for example carbapenems and monobactams, resistance has continued to

emerge through the evolution of many broad-acting β-lactamases (8). This constant emergence of

resistance not only threatens β-lactams, the most commonly prescribed antibiotics worldwide (9, 10),

but increases the use of last-resort agents like the polymyxin antibiotic colistin for the treatment of

multidrug-resistant infections (11). As a result, resistance to colistin is also on the rise, due in part to

the spread of alarming novel cell-envelope colistin resistance determinants. These proteins, called

mobile colistin resistance (MCR) enzymes, represent the only mobilizable mechanism of polymyxin

resistance reported to date (12). Since their discovery in 2015, nine families of MCR proteins have been

identified and these enzymes are quickly becoming a major threat to the longevity of colistin (13).

Alongside β-lactamases and MCR enzymes, Resistance-Nodulation-Division (RND) efflux pumps

further enrich the repertoire of AMR determinants in the cell envelope. These multi-protein assemblies

span the periplasm and remove many antibiotics (14, 15) rendering Gram-negative bacteria inherently

resistant to important existing antimicrobials.

Inhibition of AMR determinants has traditionally been achieved through the development of antibiotic

adjuvants. These molecules impair the function of resistance proteins and are used in combination with

existing antibiotics to eliminate challenging infections (4). Whilst this approach has proven successful

and has led to the deployment of several β-lactamase inhibitors that are used clinically (4), it cannot be

used to incapacitate multiple different AMR determinants. This is because antibiotic adjuvants bind to

the active site of a resistance enzyme and thus are only effective for specific protein families. To disrupt

AMR more broadly, new strategies have to be developed that target the biogenesis or stability, rather

than the activity, of resistance determinants. In this way, the formation of multiple resistance proteins

can be inhibited at once, instead of developing specific compounds that inactivate individual AMR

enzymes after they are already in place.

In extracytoplasmic environments protein stability largely relies on the formation of disulfide bonds

between cysteine residues (16, 17). Notably, in the cell envelope of Gram-negative bacteria this process

is performed by a single pathway, the DSB system, and more specifically by a single protein, the thiol

oxidase DsbA (18-22). DsbA has been shown to assist the folding of hundreds of proteins in the

periplasm (21, 23, 24) (Figure 1A), including a vast range of virulence factors (25, 26). As such,

inhibition of DSB proteins has been proposed as a promising broad-acting strategy to target bacterial

pathogenesis without impairing bacterial viability (19, 25-27). Nonetheless, the role of oxidative protein

folding in AMR has never been examined. Since several cell envelope AMR determinants contain

multiple cysteines (18, 28), we hypothesized that interfering with the function of DsbA, would not only

compromise bacterial virulence (27), but might also offer a broad approach to break resistance across

different mechanisms by affecting the stability of resistance proteins. Here we test this hypothesis by

219

investigating the contribution of disulfide bond formation to three of the most important resistance

mechanisms in the cell envelope of Enterobacteria: the breakdown of β-lactam antibiotics by β-

lactamases, polymyxin resistance arising from the production of MCR enzymes and intrinsic resistance

to multiple antibiotic classes due to RND efflux pumps. We find that all these resistance mechanisms

depend on DsbA and we demonstrate that when DsbA activity is chemically inhibited, resistance is

abrogated. Our findings prove that it is possible to simultaneously incapacitate multiple classes of AMR

determinants and therefore hold great promise for the development of next-generation therapeutic

approaches that would abolish resistance.

220

RESULTS

The activity of cysteine-containing resistance determinants is dependent on DsbA

DsbA has been shown to assist the folding of numerous periplasmic and surface-exposed proteins in

Gram-negative bacteria (Figure 1A) (25-27). As many AMR determinants also transit through the

periplasm, we postulated that inactivation of the DSB system may affect their folding, and therefore

impair their function. To test this, we first focused on resistance proteins that are present in the cell

envelope and contain two or more cysteine residues, since they may depend on the formation of

disulfide bonds for their stability and folding (18, 28). We selected a panel of twelve clinically important

β-lactamases from different Ambler classes (classes A, B and D), most of which are encoded on

plasmids (Table S1). The chosen enzymes represent different protein structures, belong to discrete

phylogenetic families (File S1) and have distinct hydrolytic activities ranging from the degradation of

penicillins and first, second and third generation cephalosporins (extended spectrum β-lactamases,

ESBLs) to the inactivation of last-resort β-lactams (carbapenemases). In addition to β-lactamases, we

selected five representative phosphoethanolamine transferases from throughout the MCR phylogeny

(Figure S1) to gain a comprehensive overview of the contribution of DsbA to the activity of these

colistin-resistance determinants.

We expressed our panel of 17 discrete resistance enzymes in an Escherichia coli K-12 strain and its

isogenic dsbA mutant and recorded minimum inhibitory concentration (MIC) values for β-lactam or

polymyxin antibiotics, as appropriate. We found that the absence of DsbA resulted in a substantial

decrease in MIC values (>2-fold) for all but one of the tested β-lactamases (Figure 1B, Figure S2). In

addition, deletion of dsbA led to clinically meaningful decreases in colistin MIC values for all MCR

enzymes (Figure 1C), especially given the narrow therapeutic window of this antibiotic (29, 30). More

specifically, expression of all MCR enzymes in our wild-type E.coli K-12 strain resulted in colistin

resistance, however in most cases absence of DsbA caused re-sensitization of the strain as defined by

the EUCAST breakpoint (E. coli strains with an MIC of 2 μg/mL or below are classified as susceptible)

(Figure 1C). For the only enzyme included in this screen that seemed unaffected by the absence of

DsbA, the SHV-27 β-lactamase, we performed the same experiment under temperature stress conditions

(at 42 °C rather than 37 °C). Under these conditions the lack of DsbA also resulted in a noticeable drop

in the cefuroxime MIC value (Figure S3).

Wild-type MIC values could be restored for all tested enzymes by complementation of dsbA (Figures

S4, S5). Moreover, since DsbA acts on its substrates post-translationally, we performed a series of

control experiments designed to assess whether the recorded effects were specific to the interaction of

the resistance proteins with DsbA, and not a result of a general inability of the dsbA mutant strain to

resist antibiotic stress. We found that no decreases in MIC values were observed for the aminoglycoside

antibiotic gentamicin, which is not affected by the tested enzymes (Figure 1B, Figure S6). Furthermore,

the β-lactam MIC values of strains harboring the empty-vector alone, or a plasmid encoding L2-1

(Figure 1B), a β-lactamase containing three cysteine residues, but no disulfide bond (PDB ID: 5NE1)

remained unchanged. Finally, to rule out the possibility that deletion of dsbA caused changes in

membrane permeability that might confound our results, we measured the permeability of the outer and

inner membrane of the dsbA mutant using the fluorescent dyes 1-N-phenylnaphthylamine (NPN) and

propidium iodide (PI), respectively and found it to be no different from that of the parental strain (Figure

S7).

Together, these results indicate that many cell envelope AMR determinants that contain more than one

cysteine residue are substrates of DsbA and that the process of disulfide bond formation is essential for

their activity.

221

The function of E. coli RND efflux pumps is compromised in the absence of DsbA

Unlike β-lactamases and MCR enzymes, none of the components of the six E. coli RND efflux pumps

contain periplasmic cysteine residues (31), thus they are not substrates of the DSB system. Nonetheless,

as DsbA assists the folding of approximately 300 extracytoplasmic proteins and, hence plays a central

role in maintaining the homeostasis of the cell envelope proteome (21, 23, 24), we wanted to assess

whether changes in periplasmic proteostasis that occur in its absence could indirectly influence efflux

pump function. To do this we determined the MIC values of three antibiotics that are RND efflux pump

substrates using E. coli MG1655, a model strain for efflux studies, its dsbA mutant and a mutant lacking

acrA, an essential component of the major E. coli RND pump AcrAB-TolC. MIC values for the dsbA

mutant were lower than for the parental strain for all tested substrate antibiotics, but not for the non-

substrate gentamicin (Figure 1D). As before, the observed phenotype could be reversed by

complementation of dsbA (Figure S8) and the recorded effects were not due to changes in membrane

permeability (Figure S9). The indirect effect of DsbA absence on efflux efficiency, although less

substantial than that measured for a mutant lacking acrA (Figure 1D), is robust and in agreement with

previous studies reporting that deletion of dsbA increases the sensitivity of E. coli to dyes like acridine

orange and pyronin Y (18), which are known substrates of AcrAB-TolC.

β-lactamases and MCR enzymes degrade or misfold in the absence of DsbA

To understand the underlying mechanisms that result in the decreased MIC values observed for the

dsbA mutant strains, we assessed the protein levels of a representative subset of β-lactamases (GES-1,

L1-1, KPC-3, FRI-1, OXA-4, OXA-10, OXA-198) and all tested MCR enzymes by immunoblotting.

When expressed in the dsbA mutant, all Ambler class A and B β-lactamases (Table S1), except GES-1

which we were not able to visualize by immunoblotting, exhibited drastically reduced protein levels

whilst the amount of the control enzyme L2-1 remained unaffected (Figure 2A). This suggests that

when these enzymes lack their disulfide bond, they are unstable and ultimately are degraded. We did

not detect any decrease in protein amounts for Ambler class D enzymes (Table S1, Figure 2B).

However, the hydrolytic activity of these β-lactamases was significantly lower in the dsbA mutant

(Figure 2C), suggesting a folding defect that leads to loss of function.

Like with class A and B β-lactamases, MCR enzymes were undetectable when expressed in a dsbA

mutant (Figure 3A) suggesting that their stability is severely compromised when they lack their

disulfide bonds. We further confirmed this by directly monitoring the lipid A profile of all MCR-

expressing strains with substantial MIC drops (i.e. strains expressing MCR-3, -4, -5 and -8, Figure 1C)

using MALDI-TOF mass spectrometry (Figure 3BC). MCR activity leads to the addition of

phosphoethanolamine to the lipid A portion of bacterial lipopolysaccharide (LPS), resulting in reduced

binding of colistin to LPS and thus resistance. In E. coli the major lipid A peak detected by mass

spectrometry is present at m/z 1796.2 (Figure 3B, first spectrum). This peak corresponds to hexa-acyl

diphosphoryl lipid A (native lipid A). The lipid A profile of E. coli MC1000 dsbA was identical to that

of the parental strain (Figure 3B, second spectrum). In the presence of MCR enzymes two additional

peaks were observed, at m/z 1821.2 and 1919.2 (Figure 3B, third spectrum). The peak at m/z 1919.2

corresponds to the addition of a phosphoethanolamine moiety to the phosphate group at position 1 of

native lipid A, and the peak at m/z 1821.2 corresponds to the addition of a phosphoethanolamine moiety

to the 4ˊ phosphate of native lipid A and the concomitant loss of the phosphate group at position 1 (32).

For dsbA mutants expressing MCR-3, -5 and -8 (Figure 3C), the peaks at m/z 1821.2 and m/z 1919.2

could no longer be detected, whilst the native lipid A peak at m/z 1796.2 remained unchanged (Figure

222

3B, fourth spectrum); dsbA mutants expressing MCR-4 retain some basal lipid A-modifying activity,

nonetheless this is not sufficient for this strain to efficiently evade colistin treatment (Figure 1C).

Together these data suggest that in the absence of DsbA, MCR enzymes are unstable (Figure 3A) and

therefore no longer able to efficiently catalyze the addition of phosphoethanolamine to native lipid A

(Figure 3BC); as a result, they cannot confer resistance to colistin (Figure 1C).

The data presented above validate our initial hypothesis. Absence of DsbA affects the stability and

folding of cysteine-containing resistance proteins and in most cases leads to drastically reduced levels

of the tested enzyme. As a result, and in agreement with the recorded decreases in MIC values (Figure

1BC), these folding defects impede the ability of AMR determinants to confer resistance. Therefore, by

compromising cell envelope protein folding we can impair a broad range of AMR proteins and abrogate

resistance across multiple mechanisms.

RND efflux pump function is compromised in the absence of DsbA due to altered periplasmic

proteostasis

As RND efflux pump proteins do not contain any disulfide bonds, the decreases in MIC values for

pump substrates in the absence of dsbA (Figure 1D) are likely mediated by additional cell-envelope

components. The protease DegP, previously found to be a DsbA substrate (20), seemed a promising

candidate for linking DsbA to efflux pump function. DegP degrades a range of misfolded

extracytoplasmic proteins including, but not limited to, subunits of higher order protein complexes and

proteins lacking their native disulfide bonds (33). We hypothesized that in a dsbA mutant the substrate

burden on DegP would be dramatically increased, whilst DegP itself would not function optimally due

to absence of its disulfide bond (34). Consequently, protein turn over in the cell envelope would not

occur efficiently. Since the essential RND efflux pump component AcrA needs to be cleared by DegP

when it becomes misfolded or nonfunctional (35), we expected that the reduced DegP efficiency in a

dsbA mutant would result in accumulation of nonfunctional AcrA in the periplasm, which would then

interfere with pump function. In agreement with our hypothesis we found that in the absence of DsbA

degradation of DegP occurred, reducing the pool of active enzyme (Figure 4A) (34). In addition, AcrA

accumulated to the same extent in a dsbA and in a degP mutant (Figure 4B), suggesting that in both

these strains AcrA was not efficiently cleared. Finally, no accumulation was detected for the outer-

membrane protein TolC, which is not a DegP substrate (Figure 4C) (36). Thus, in the absence of DsbA,

inefficient DegP-mediated periplasmic proteostasis impacts RND efflux pump function (Figure 1D)

through accumulation of AcrA that should have been degraded and removed from the cell envelope.

These results demonstrate that changes in cell envelope protein homeostasis have a profound effect on

protein function in this compartment, which could be exploited in the design of antibacterial strategies.

Here, disrupting periplasmic proteostasis by preventing disulfide bond formation indirectly impairs

efflux pump activity in addition to incapacitating direct substrates of DsbA, like β-lactamases and MCR

proteins. This allows us to simultaneously abrogate three distinct resistance mechanisms (Figure 4D).

DsbA is a tractable AMR target

DsbA is essential for the folding of many virulence factors. As such, inhibition of the DSB system has

been proposed as a promising anti-virulence strategy (25-27) and efforts have been made to develop

inhibitors for DsbA (37, 38), its redox partner DsbB (Figure 1A) (39) or both (40). These studies have

made the first steps towards the production of chemical compounds that inhibit the function of the DSB

proteins, providing us with a laboratory tool to test our approach against AMR.

223

4,5-dichloro-2-(2-chlorobenzyl)pyridazin-3-one, termed “compound 12” in Landeta et al. (39) is a

potent laboratory inhibitor of E. coli DsbB and its analogues from closely related organisms. Using this

molecule, we could chemically inhibit the function of the DSB system. We first tested the motility of

E. coli MC1000 in the presence of the inhibitor and found that cell were significantly less motile (Figure

S10), consistent with the fact that impairing DSB function prevents the formation of the flagellar P-ring

component FlgI (41, 42). Furthermore, we directly assessed the redox state of DsbA in the presence of

the compound to probe whether it was being re-oxidized by DsbB, a necessary step that occurs after

each round of oxidative protein folding and allows DsbA to remain active (Figure 1A). Under normal

growth conditions, DsbA was in its active oxidized form in the bacterial periplasm (i.e. C30 and C33

form a disulfide bond), showing that it was efficiently regenerated by DsbB (43) (Figure 5A). By

contrast, addition of the inhibitor to growing E. coli MC1000 cells resulted in accumulation of inactive

reduced DsbA, thus confirming that DsbB function was impeded (Figure 5A).

After testing the efficacy of the DsbB inhibitor, we proceeded to examine whether chemical inhibition

of the DSB system could be used to broadly impair the function of AMR determinants. We found that

addition of the compound during MIC testing phenocopied the effects of a dsbA deletion on β-lactamase

and MCR activity (Figure 5BCD) by determining MIC values for the latest generation β-lactam that

each β-lactamase can hydrolyze, and colistin, as appropriate The observed effects are not a result of

altered cell growth, as addition of the molecule does not affect the growth profile of the bacteria (Figure

S11A), in agreement with the fact that deletion of dsbA does not affect cell viability (Figure S11B).

Furthermore, the changes in the recorded MIC values are due solely to inhibition of the DSB system as

no additive effects on MIC values were observed when the dsbA mutant harboring a β-lactamase or mcr

gene was exposed to the compound (Figure S12).

Sensitization of clinical isolates to existing antibiotics can be achieved by chemical inhibition of DsbA

activity

Having shown that the DSB system is a tractable target in the context of AMR, we examined the effect

of chemical inhibition on several species of β-lactamase- and MCR-expressing Enterobacteria (Table

S2). Tested clinical isolates from four different species, including multidrug-resistant E. coli and

Citrobacter freundii strains, showed clinically relevant decreases in their MIC values to last-resort

antibiotics when their DSB system was chemically inhibited. In all but one case this led to sensitization

as defined by EUCAST breakpoints (Figure 6AB, Figure S13). In the one case where sensitization was

not achieved, chemical inhibition of the DSB system of an Enterobacter cloacae isolate expressing

FRI-1 caused a drastic reduction in the aztreonam MIC value by over 180 µg/mL, resulting in

intermediate resistance as defined by EUCAST. These results obtained using clinical strains provide

further validation of the significance of our data from heterologously expressed β-lactamase and MCR

enzymes in E. coli K-12 strains (Figure 1BC), and showcase the potential of this approach for clinical

applications.

Regarding intrinsic resistance mediated by RND efflux pumps, chloramphenicol is the only antibiotic

from the efflux pump substrates that were tested in this study that has a EUCAST breakpoint for Gram-

negative bacteria (E. coli strains with an MIC of 8 μg/mL or below are classified as sensitive). It is

notable that the MIC drop for this pump substrate (Figure 1D) caused by deletion of dsbA sensitized E.

coli to chloramphenicol (Figure 6C), showing that even the indirect effects of compromising disulfide

bond formation are potentially clinically important. Since mutations in marR that derepress MarA and

cause constitutive expression of AcrAB (44, 45) are observed in clinical isolates with increased efflux

(46), we recorded the chloramphenicol MIC for the dsbA mutant of an E. coli MG1655 marR strain and

found that sensitization to chloramphenicol occurred (Figure 6C) even when efflux pump components

were overexpressed (Figure S14).

224

To determine if our approach for Enterobacteria would be appropriate for other multidrug-resistant

pathogens we tested it on Pseudomonas aeruginosa. This bacterium has two DsbB analogues which are

functionally redundant (47). The chemical inhibitor used in this study has been shown to be effective

against DsbB1, but less effective against DsbB2 of P. aeruginosa PA14 (39), making it unsuitable for

MIC assays on P. aeruginosa clinical isolates. Nonetheless, deletion of dsbA1 in the multidrug-resistant

P. aeruginosa PA43417 clinical isolate expressing OXA-198, led to sensitization of this strain to the

antipseudomonal β-lactam piperacillin (Figure 6D), suggesting that targeting disulfide bond formation

could be useful for the sensitization of many more clinically important Gram-negative species.

Finally, to test our approach in an infection context we performed in vivo clearance assays using the

wax moth model Galleria mellonella (Figure 6E). Larvae were infected with the P. aeruginosa

PA43417 clinical isolate producing OXA-198 and its dsbA1 mutant and infections were treated once

with piperacillin at a final concentration below the EUCAST breakpoint. Neither deletion of dsbA1 nor

treatment with piperacillin was sufficient to clear the infection reliably when applied alone, although

the former led to a significant decrease in the recovered bacterial load due to the fact that absence of

the principal DsbA likely affects the virulence of the pathogen (48). However, treatment of the dsbA1

mutant with piperacillin resulted in a drastic (> 99% on average) reduction in bacterial load in the

infected larvae, in agreement with the fact that in the absence of DsbA the ability of OXA-198 to

hydrolyze β-lactams is impaired (Figure 1B, 2C). As OXA-198, in this case produced by a multi-drug

resistant clinical strain (Table S2 and Figure 6DE), is a broad-spectrum β-lactamase that cannot be

neutralized by classical β-lactamase inhibitors (Table S1) and piperacillin is a first-line antibiotic, these

results further highlight the promise of our approach for future clinical applications.

225

DISCUSSION

This work is the first report of a strategy capable of simultaneously impairing multiple types of AMR

determinants by compromising the function of a single target. By inhibiting DsbA, a non-essential cell

envelope protein which is unique to bacteria, we can overcome three entirely distinct resistance

mechanisms and sensitize critically important pathogens to multiple classes of existing antibiotics. This

proof of principle opens a new avenue to reverse AMR in Gram-negative organisms, through the

development of DsbA inhibitors that would function as broad-acting resistance breakers.

We have shown that targeting DsbA incapacitates broad-spectrum β-lactamases from three of the four

Ambler classes (class A, B and D, Figure 1B). This includes enzymes that are not susceptible to classical

β-lactamase inhibitors (Table S1), such as members of the KPC and OXA families, as well as metallo-

β-lactamases like L1-1 from the often pan-resistant organism Stenotrophomonas maltophilia. The

function of these proteins is impaired without a small molecule binding to their active site, unlike

currently-used β-lactamase inhibitors which often generate resistance (4). As DsbA dependence is

conserved within phylogenetic groups (see Figure S2, GES-1, -2, -11 and KPC-2, -3), based on the

number of enzymes belonging to the same phylogenetic family as the β-lactamases tested in this study

(File S1), we anticipate that a total of 149 discrete enzymes rely on DsbA for their stability and function.

DsbA is widely conserved (25), thus targeting the DSB system should not only compromise β-

lactamases in Enterobacteria, but as demonstrated by our experiments using a P. aeruginosa clinical

isolate (Figure 6DE), could also be a promising avenue for impairing the function of AMR determinants

expressed by other highly-resistant Gram-negative organisms. As such, together with the fact that

approximately 25% of β-lactamases found in pathogens and organisms capable of causing opportunistic

infections contain two or more cysteines (File S1), we expect many more clinically relevant β-

lactamases, beyond those already tested in this study, to depend on DsbA.

MCR enzymes are rapidly becoming a grave threat to the use of colistin (13), a drug of last resort often

needed for the treatment of multidrug-resistant infections (11). Currently, experimental inhibitors of

these proteins are sparse and poorly characterized (49). As all MCR members contain multiple disulfide

bonds, inhibition of the DSB system provides a broadly applicable solution for combating MCR-

mediated colistin resistance (Figure 1C and Figure 6B) that would likely extend to novel MCR proteins

emerging in the future. As for MCR enzymes, no clinically applicable efflux pump inhibitors have been

identified to date (50) despite many efforts to use these macromolecular assemblies as targets against

intrinsic resistance. Deletion of dsbA sensitizes the tested E. coli strain to chloramphenicol (Figure 6C)

offering a novel way of targeting RND efflux pumps. Efflux pumps containing AcrA-like components

are dependent on DegP for their homeostasis in the periplasm, and hence their function would be

compromised if DsbA was inhibited. This means that the generation of clinically useful DsbA inhibitors

to combat β-lactamase- and MCR-mediated resistance offers a new platform for exploring the potential

of efflux pump inhibition.

More generally, our findings demonstrate that cell envelope proteostasis pathways have significant yet

untapped potential for the development of novel antibacterial strategies. The example of the DSB

system presented here is particularly telling. This pathway, initially considered merely a housekeeping

system (51), plays a major role in clinically relevant bacterial niche adaptations. In addition to assisting

the folding of 40% of the cell-envelope proteome (23, 24), the DSB system is essential for virulence

(25, 26), has a key role in the formation and awakening of bacterial persister cells (52) and, as seen in

this work, is required for bacterial survival in the presence of a broad range of antibiotic compounds.

As shown in our in vivo experiments (Figure 6E), targeting such a system in Gram-negative pathogens

could lead to adjuvant approaches that inactivate multiple AMR determinants whilst simultaneously

incapacitating an arsenal of virulence factors. Therefore, this study not only lays the groundwork for

future clinical applications, such as the development of broad-acting antibiotic adjuvants, but also

226

serves as a paradigm for exploiting other accessible cell envelope proteostasis processes for the design

of next-generation therapeutic strategies.

227

MATERIALS AND METHODS

Reagents and bacterial growth conditions. Unless otherwise stated, chemicals and reagents were

acquired from Sigma Aldrich, growth media were purchased from Oxoid and antibiotics were obtained

from Melford Laboratories. Lysogeny broth (LB) (10 g/L NaCl) and agar (1.5% w/v) were used for

routine growth of all organisms at 37 °C with shaking at 220 RPM, as appropriate. Unless otherwise

stated, Mueller-Hinton (MH) broth and agar (1.5% w/v) were used for Minimum Inhibitory

Concentration (MIC) assays. Growth media were supplemented with the following, as required: 0.25

mM Isopropyl β-D-1-thiogalactopyranoside (IPTG) (for strains harboring β-lactamase-encoding pDM1

plasmids), 0.5 mM IPTG (for strains harboring MCR-encoding pDM1 plasmids), 12.5 μg/mL

tetracycline, 100 μg/mL ampicillin, 50 μg/mL kanamycin, 10 μg/mL gentamicin, 33 μg/mL

chloramphenicol and 50 μg/mL streptomycin.

Construction of plasmids and bacterial strains. Bacterial strains, plasmids and oligonucleotides used

in this study are listed in Tables S3, S4 and S5, respectively. DNA manipulations were conducted using

standard methods. KOD Hot Start DNA polymerase (Merck) was used for all PCR reactions according

to the manufacturer’s instructions, oligonucleotides were synthesized by Sigma Aldrich and restriction

enzymes were purchased from New England Biolabs. All constructs were DNA sequenced and

confirmed to be correct before use.

Genes for β-lactamase and MCR enzymes were amplified from genomic DNA extracted from clinical

isolates (Table S6) with the exception of mcr-3 and mcr-8, which were synthesized by GeneArt Gene

Synthesis (ThermoFisher Scientific). β-lactamase and MCR genes were cloned into the IPTG-inducible

plasmid pDM1 using primers P1-P36. pDM1 (GenBank accession number MN128719) was constructed

from the p15A-ori plasmid pACYC184 (53) to contain the Lac repressor, the Ptac promoter, an

optimized ribosome binding site and a multiple cloning site (NdeI, SacI, PstI, KpnI, XhoI and XmaI)

inserted into the NcoI restriction site of pACYC184. All StrepII-tag fusions of β-lactamase and MCR

enzymes (constructed using primers P1, P3, P9, P11, P13, P15, P17, P21, P23, P25, P27, P29, P37, P38

and P41-P50) have a C-terminal StrepII tag (GSAWSHPQFEK) except for OXA-4, where an N-

terminal StrepII tag was inserted between the periplasmic signal sequence and the body of the protein

using the primer pairs P7/P40, P39/P9 and P7/P8. Plasmids encoding ges-1, kpc-3 and mcr-3.2 were

obtained by performing QuickChange mutagenesis on pDM1 constructs encoding ges-5, kpc-2 and mcr-

3, respectively (primers P31-P36).

E. coli gene mutants were constructed using a modified lambda-Red recombination method, as

previously described (54) (primers P53-P62). To complement the dsbA mutant, a DNA fragment

consisting of dsbA preceded by the Ptac promoter was inserted into the NotI/XhoI sites of pGRG25

(primers P51/P52) and was reintroduced into the E. coli chromosome at the attTn7 site, as previously

described (55). The dsbA1 mutant of the Pseudomonas aeruginosa PA43417 clinical isolate was

constructed by allelic exchange, as previously described (56). Briefly, the dsbA1 gene area of P.

aeruginosa PA43417 (including the dsbA1 gene and 600 bp on either side of this gene) was amplified

(primers P63/P64) and the obtained DNA was sequenced to allow for accurate primer design for the

ensuing cloning step. Subsequently, 500-bp DNA fragments upstream and downstream of the dsbA1

gene were amplified using P. aeruginosa PA43417 genomic DNA (primers P65/P66 (upstream) and

P67/P68 (downstream)). A fragment containing both of these regions was obtained by overlapping PCR

(primers P65/P68) and inserted into the XbaI/BamHI sites of pKNG101. The suicide vector pKNG101

(57) is not replicative in P. aeruginosa; it was maintained in E. coli CC118λpir and mobilized into P.

aeruginosa PA43417 by triparental conjugation.

Minimum inhibitory concentration (MIC) assays. Unless otherwise stated, antibiotic MIC assays were

carried out in accordance with the EUCAST recommendations using E-test strips (BioMérieux).

228

Briefly, overnight cultures of each strain to be tested were standardized to OD600 0.063 in 0.85% NaCl

(equivalent to McFarland standard 0.5) and distributed evenly across the surface of MH agar plates. E-

test strips were placed on the surface of the plates, evenly spaced, and the plates were incubated for 18-

24 hours at 37 °C. MICs were read according to the manufacturer’s instructions. β-lactam MICs were

also determined using the Broth Microdilution (BMD) method, as required. Briefly, a series of antibiotic

concentrations was prepared by two-fold serial dilution in MH broth in a clear-bottomed 96-well

microtiter plate (Corning). When used, tazobactam was included at a fixed concentration of 4 μg/mL in

every well, in accordance with the EUCAST guidelines. The strain to be tested was added to the wells

at approximately 5 x 104 colony forming units (CFU) per well and plates were incubated for 18-24 hours

at 37 °C. The MIC was defined as the lowest antibiotic concentration with no visible bacterial growth

in the wells. All colistin sulphate MIC assays were performed using the BMD method as described

above except that instead of two-fold serial dilutions, the following concentrations of colistin (Acros

Organics) were prepared individually in MH broth: 4 μg/mL, 3.5 μg/mL, 3 μg/mL, 2.5 μg/mL, 2 μg/mL,

1.5 μg/mL, 1 μg/mL, 0.5 μg/mL.

The covalent DsbB inhibitor 4,5-dichloro-2-(2-chlorobenzyl)pyridazin-3-one (39) was used to

chemically impair the function of the DSB system. Inactivation of DsbB results in abrogation of DsbA

function (43) only in media free of small-molecule oxidants (41). Therefore, MIC assays involving

chemical inhibition of the DSB system were performed using M63 broth (15.1 mM (NH4)2SO4, 100

mM KH2PO4, 1.8 mM FeSO4.7H2O, adjusted to pH 7.2 with KOH) and agar (1.5% w/v) supplemented

with 1 mM MgSO4, 0.02% w/v glucose, 0.005% w/v thiamine, 31 µM FeCl3.6H2O, 6.2 μM ZnCl2, 0.76

µM CuCl2.2H2O, 1.62 µM H3BO3, 0.081 µM MnCl2.4H2O, 84.5 mg/L alanine, 19.5 mg/L arginine, 91

mg/L aspartic acid, 65 mg/L glutamic acid, 78 mg/L glycine, 6.5 mg/L histidine, 26 mg/L isoleucine,

52 mg/L leucine, 56.34 mg/L lysine, 19.5 mg/L methionine, 26 mg/L phenylalanine, 26 mg/L proline,

26 mg/L serine, 6.5 mg/L threonine, 19.5 mg/L tyrosine, 56.34 mg/L valine, 26 mg/L tryptophan, 26

mg/L asparagine and 26 mg/L glutamine. CaCl2 was also added at a final concentration of 0.223 mM

for colistin sulfate MIC assays. Either DMSO (vehicle control) or the covalent DsbB inhibitor 4,5-

dichloro-2-(2-chlorobenzyl)pyridazin-3-one (final concentration of 50 μM) (Enamine) (39) were added

to the M63 medium, as required. The strain to be tested was added at an inoculum that recapitulated the

MH medium MIC values obtained for that strain.

SDS-PAGE analysis and immunoblotting. Samples for immunoblotting were prepared as follows.

Strains to be tested were grown on LB or MH agar plates as lawns in the same manner as for MIC

assays described above. Bacteria were collected using an inoculating loop and resuspended in 0.85%

NaCl or LB to OD600 2.0 (except for strains expressing OXA-4, where OD600 6.0 was used). For strains

expressing β-lactamase enzymes, the cell suspensions were spun at 10,000 x g for 10 minutes and

bacterial pellets were lysed by addition of BugBuster Master Mix (Merck Millipore) for 25 minutes at

room temperature with gentle agitation. Subsequently, lysates were spun at 10,000 x g for 10 minutes

at 4 °C and the supernatant was added to 4 x Laemmli buffer. For strains expressing MCR enzymes cell

suspensions were directly added to 4 x Laemmli buffer, while for E. coli MG1655 and its mutants, cells

were lysed as above and lysates were added to 4 x Laemmli buffer. All samples were boiled for 5

minutes before separation by SDS-PAGE.

Unless otherwise stated, SDS-PAGE analysis was carried out using 10% BisTris NuPAGE gels

(ThermoFisher Scientific) using MES/SDS running buffer prepared according to the manufacturer’s

instructions and including pre-stained protein markers (SeeBlue Plus 2, ThermoFisher Scientific).

Proteins were transferred to Amersham Protran nitrocellulose membranes (0.45 µm pore size, GE Life

Sciences) using a Trans-Blot Turbo transfer system (Bio-Rad) before blocking in 3% w/v Bovine Serum

Albumin (BSA)/TBS-T (0.1 % v/v Tween 20) or 5% w/v skimmed milk/TBS-T and addition of primary

and secondary antibodies. The following primary antibodies were used in this study: Strep-Tactin-HRP

conjugate (Iba Lifesciences) (dilution 1:3,000 in 3 w/v % BSA/TBS-T), Strep-Tactin-AP conjugate (Iba

229

Lifesciences) (dilution 1:3,000 in 3 w/v % BSA/TBS-T), rabbit anti-DsbA antibody (dilution 1:1,000

in 5 w/v % skimmed milk/TBS-T), rabbit anti-AcrA antibody (dilution 1:10,000 in 5 w/v % skimmed

milk/TBS-T), rabbit anti-TolC antibody (dilution 1:5,000 in 5 w/v % skimmed milk/TBS-T), rabbit

anti-HtrA1 (DegP) antibody (Abcam) (dilution 1:1,000 in 5 w/v % skimmed milk/TBS-T) and mouse

anti-DnaK 8E2/2 antibody (Enzo Life Sciences) (dilution 1:10,000 in 5% w/v skimmed milk/TBS-T).

The following secondary antibodies were used in this study: goat anti-rabbit IgG-AP conjugate (Sigma

Aldrich) (dilution 1:6,000 in 5% w/v skimmed milk/TBS-T), goat anti-rabbit IgG-HRP conjugate

(Sigma Aldrich) (dilution 1:6,000 in 5% w/v skimmed milk/TBS-T), goat anti-mouse IgG-AP conjugate

(Sigma Aldrich) (dilution 1:6,000 in 5% w/v skimmed milk/TBS-T) and goat anti-mouse IgG-HRP

conjugate (Sigma Aldrich) (dilution 1:6,000 in 5% w/v skimmed milk/TBS-T). Membranes were

washed three times for 5 minutes with TBS-T prior to development. Development for AP conjugates

was carried out using a SigmaFast BCIP/NBT tablet, while HRP conjugates were visualized with the

Novex ECL HRP chemiluminescent substrate reagent kit (ThermoFisher Scientific) or the Luminata

Crescendo chemiluminescent reagent (Merck) using a Gel Doc XR+ Imager (Bio-Rad).

β-lactam hydrolysis assay. β-lactam hydrolysis measurements were carried out using the chromogenic

β-lactam nitrocefin (Abcam). Briefly, overnight cultures of strains to be tested were centrifugated,

pellets were weighed and resuspended in 150 μL of 100 mM sodium phosphate buffer (pH 7.0) per 1

mg of wet-cell pellet, and cells were lysed by sonication. For strains harboring pDM1, pDM1-blaL2-1,

pDM1-blaOXA-10 and pDM1-blaGES-1, lysates corresponding to 0.34 mg of bacterial pellet were

transferred into clear-bottomed 96-well microtiter plates (Corning). For strains harboring pDM1-

blaOXA-4 and pDM1-blaOXA-198, lysates corresponding to 0.2 mg and 0.014 mg of bacterial pellet were

used, respectively. In all cases, nitrocefin was added at a final concentration of 400 μM and the final

reaction volume was made up to 100 μL using 100 mM sodium phosphate buffer (pH 7.0). Nitrocefin

hydrolysis was monitored at 25 °C by recording absorbance at 490 nm at 60-second intervals for 15

minutes using an Infinite M200 Pro microplate reader (Tecan). The amount of nitrocefin hydrolyzed by

each lysate in 15 minutes was calculated using a standard curve generated by acid hydrolysis of

nitrocefin standards.

NPN uptake assay. 1-N-phenylnaphthylamine (NPN) (Acros Organics) uptake assays were performed

as described by Helander & Mattila-Sandholm (58). Briefly, mid-log phase cultures of strains to be

tested were diluted to OD600 0.5 in 5 mM HEPES (pH 7.2) before transfer to clear-bottomed 96-well

microtiter plates (Corning) and addition of NPN at a final concentration of 10 μM. Colistin sulphate

(Acros Organics) was included at a final concentration of 0.5 μg/mL, as required. Immediately after the

addition of NPN, fluorescence was measured at 60-second intervals for 10 minutes using an Infinite

M200 Pro microplate reader (Tecan); the excitation wavelength was set to 355 nm and emission was

recorded at 405 nm.

PI uptake assay. Exponentially-growing (OD600 0.4) E. coli strains harboring pUltraGFP-GM (59) were

diluted to OD600 0.1 in phosphate buffered saline (PBS) (pH 7.4) and cecropin A was added to a final

concentration of 20 μM, as required. Cell suspensions were incubated at room temperature for 30

minutes before centrifugation and resuspension of the pellets in PBS. Propidium iodide (PI) was then

added at a final concentration of 3 μM. Suspensions were incubated for 10 minutes at room temperature

and analyzed on a two-laser, four color BD FACSCalibur flow cytometer (BD Biosciences). 50,000

events were collected for each sample and data were analyzed using FlowJo v.10.0.6 (Treestar).

MALDI-TOF Mass spectrometry. Lipid A profiles of strains to be tested were determined using intact

bacteria, as previously described (60). The peak for E. coli native lipid A is detected at m/z 1796.2,

whereas the lipid A profiles of strains expressing functional MCR enzymes have two additional peaks,

at m/z 1821.2 and 1919.2. These peaks result from MCR-mediated modification of native lipid A

through addition of phosphoethanolamine moieties (32). The ratio of modified to unmodified lipid A

230

was calculated by summing the intensities of the peaks at m/z 1821.2 and 1919.2 and dividing this value

by the intensity of the native lipid A peak at m/z 1796.2.

Motility assay. 500 μL of overnight culture of each strain to be tested were centrifuged and the pellets

were washed in M63 broth before resuspension in the same medium to achieve a final volume of 25

μL. Bacterial motility was assessed by growth in M63 medium containing 0.25% w/v agar

supplemented as described above. DMSO (vehicle control) or the covalent DsbB inhibitor 4,5-dichloro-

2-(2-chlorobenzyl)pyridazin-3-one (final concentration of 50 μM) (Enamine) were added to the

medium, as required. 1 μL of the washed cell suspension was inoculated into the center of a 90 mm

diameter agar plate, just below the surface of the semi-solid medium. Plates were incubated at 37 °C in

a humidified environment for 16-18 hours and growth halo diameters were measured.

AMS labelling. Bacterial strains to be tested were grown for 18 hours in M63 broth supplemented as

described above. DMSO (vehicle control) or the covalent DsbB inhibitor 4,5-dichloro-2-(2-

chlorobenzyl)pyridazin-3-one (final concentration of 50 μM) (Enamine) were added to the medium, as

required. Cultures were standardized to OD600 2.0 in M63 broth, were spun at 10,000 x g for 10 minutes

and bacterial pellets were lysed by addition of BugBuster Master Mix (Merck Millipore) for 25 minutes

at room temperature with gentle agitation. Subsequently, lysates were spun at 10,000 x g for 10 minutes

at 4 °C prior to reaction with 4-acetamido-4ˊ-maleimidyl-stilbene-2,2ˊ-disulfonic acid (AMS)

(ThermoFisher Scientific). AMS alkylation was performed by vortexing the lysates in 15 mM AMS,

50 mM Tris-HCl, 3% w/v SDS and 3 mM EDTA (pH 8.0) for 30 minutes at 25 °C, followed by

incubation at 37 °C for 10 minutes. SDS-PAGE analysis and immunoblotting was carried out as

described above, except that 12% BisTris NuPAGE gels (ThermoFisher Scientific) and MOPS/SDS

running buffer were used. DsbA was detected using a rabbit anti-DsbA primary antibody and an AP-

conjugated secondary antibody, as described above.

Bacterial growth assays. To assess the effect of DSB system inhibition of the growth of E. coli,

overnight cultures of the strains to be tested were centrifuged and the pellets were washed in M63 broth

before transfer to clear-bottomed 96-well microtiter plates (Corning) at approximately 5 x 107 CFU/well

(starting OD600 ~ 0.03). M63 broth supplemented as described above was used as a growth medium.

DMSO (vehicle control) or the covalent DsbB inhibitor 4,5-dichloro-2-(2-chlorobenzyl)pyridazin-3-

one (final concentration of 50 μM) (Enamine) were added to the medium, as required. Plates were

incubated at 37 °C with orbital shaking (amplitude 3 mm, equivalent to ~ 220 RPM) and OD600 was

measured at 900-second intervals for 18 hours using an Infinite M200 Pro microplate reader (Tecan).

The same experimental setup was also used for recording growth curves of E. coli strains and their

isogenic mutants, except that overnight cultures of the strains to be tested were diluted 1:100 into clear-

bottomed 96-well microtiter plates (Corning) (starting OD600 ~ 0.01) and that LB was used as the growth

medium.

In vivo clearance assay. The wax moth model Galleria mellonella was used for in vivo clearance assays

(61). Individual G. mellonella larvae were randomly allocated to experimental groups; no masking was

used. Overnight cultures of the strains to be tested were standardized to OD600 1.0. Suspensions were

centrifuged and the pellets were washed three times in PBS and serially diluted. 10 μl of the 10–5 dilution

of each bacterial suspension were injected into the last right abdominal proleg of 3 to 5 G. mellonella

larvae per condition; a second, equal-size group of larvae were injected with PBS as negative control.

3 hours after infection, larvae were injected with 13 μl of piperacillin to a final concentration of 12

μg/mL in the last left abdominal proleg. 24 hours after infection larvae were euthanized and macerated

individually in 1 ml of PBS by vortexing for 15 minutes. The larval suspension was then serially diluted

and 20 μl of each dilution plated on Pseudomonas Isolation Agar. Plates were incubated at 37 °C for 16

hours before CFU counting.

231

Statistical analysis of experimental data. The total numbers of performed biological experiments and

technical repeats are mentioned in the figure legend of each display item. Biological replication refers

to completely independent repetition of an experiment using different biological and chemical

materials. Technical replication refers to independent data recordings using the same biological sample.

For MIC assays, according to common practice, we show representative MIC values from one of the

performed biological experiments. For all other assays, statistical analysis was performed in GraphPad

PRISM v8.0.2 using an unpaired T-test with Welch’s correction, a one-way ANOVA with correction

for multiple comparisons, or a Kruskal-Wallis test with correction for multiple comparisons, as

appropriate. Statistical significance was defined as p < 0.05. Outliers were defined as any technical

repeat >2 SD away from the average of the other technical repeats within the same biological

experiment. Such data were excluded and all remaining data were included in the analysis. Detailed

information for each figure is provided below:

Figure 2C: unpaired T-test with Welch’s correction; n=3; 3.621 degrees of freedom, t-value=0.302,

p=0.7792 (non-significance) (for pDM1 strains); 3.735 degrees of freedom, t-value=0.4677, p=0.666

(non-significance) (for pDM1-blaL2-1 strains); 2.273 degrees of freedom, t-value=5.069, p=0.0281

(significance) (for pDM1-blaGES-1 strains); 2.011 degrees of freedom, t-value=6.825, p=0.0205

(significance) (for pDM1-blaOXA-4 strains); 2.005 degrees of freedom, t-value=6.811, p=0.0208

(significance) (for pDM1-blaOXA-10 strains); 2.025 degrees of freedom, t-value=5.629, p=0.0293

(significance) (for pDM1-blaOXA-198 strains)

Figure 3C: one-way ANOVA with Tukey’s multiple comparison test; n=4; 24 degrees of freedom; F

value=21.00; p=0.000000000066 (for pDM1-mcr-3 strains), p=0.0004 (for pDM1-mcr-4 strains),

p=0.000000000066 (for pDM1-mcr-5 strains), p=0.00066 (for pDM1-mcr-8 strains)

Figure 6E: Kruskal-Wallis test with Dunn’s multiple comparisons test; n=8; Kruskal-Wallis H=25.24,

3 degrees of freedom; p=<0.0001. For multiple comparisons, p=0.0029 (P. aeruginosa versus P.

aeruginosa dsbA1), p=<0.0001 (P. aeruginosa versus P. aeruginosa dsbA1 treated with piperacillin),

p=0.0369 (P. aeruginosa treated with piperacillin versus P. aeruginosa dsbA1 treated with piperacillin)

Figure S7A: one-way ANOVA with Bonferroni’s multiple comparison test; n=3; 6 degrees of freedom;

F value=39.22; p=0.0007 (significance), p=0.99 (non-significance)

Figure S7B: one-way ANOVA with Bonferroni’s multiple comparison test; n=3; 6 degrees of freedom;

F value=61.84; p=0.0002 (significance), p=0.99 (non-significance)

Figure S9A: one-way ANOVA with Bonferroni’s multiple comparison test; n=3; 6 degrees of freedom;

F value=261.4; p=0.00000055 (significance), p=0.0639 (non-significance)

Figure S9B: one-way ANOVA with Bonferroni’s multiple comparison test; n=3; 6 degrees of freedom;

F value=77.49; p=0.0001 (significance), p=0.9999 (non-significance)

Figure S10B: one-way ANOVA with Bonferroni’s multiple comparison test; n=3; 6 degrees of freedom;

F value=1878; p=0.000000002 (significance)

Bioinformatics. The following bioinformatics analyses were performed in this study. Short scripts and

pipelines were written in perl (version 5.18.2) and executed on macOS Sierra 10.12.5.

β-lactamase enzymes. All available protein sequences of β-lactamases were downloaded from

http://www.bldb.eu (62). Sequences were clustered using the ucluster software with a 90% identity

threshold and the cluster_fast option (usearch v.7.0 (63)); the centroid of each cluster was used as a

cluster identifier for every sequence. All sequences were searched for the presence of cysteine residues

using a perl script. Proteins with two or more cysteines after the first 30 amino acids of their primary

sequence were considered potential substrates of the DSB system for organisms where oxidative protein

folding is carried out by DsbA and provided that translocation of the β-lactamase outside the cytoplasm

is performed by the Sec system. The first 30 amino acids of each sequence were excluded to avoid

considering cysteines that are part of the signal sequence mediating the translocation of these enzymes

outside the cytoplasm. The results of the analysis can be found in File S1.

232

MCR enzymes. E. coli MCR-1 (AKF16168.1) was used as a query in a blastp (64) search limited to

Proteobacteria on the NCBI Reference Sequence (RefSeq) proteome database (21-04-2019) (evalue <

10e-5). 17,503 hit sequences were retrieved and clustered using the ucluster software with a 70%

identity threshold and the cluster_fast option (usearch v.7.0 (63)). All centroid sequences were retrieved

and clustered again with a 20% identity threshold and the cluster_fast option. Centroid sequences of all

clusters comprising more than five sequences (809 sequences retrieved) along with the sequences of

the five MCR enzymes tested in this study were aligned using muscle (65). Sequences which were

obviously divergent or truncated were manually eliminated and a phylogenetic tree was built from a

final alignment comprising 781 sequences using fasttree with the wag substitution matrix and default

parameters (63). The assignment of each protein sequence to a specific group was done using

hmmsearch (HMMER v.3.1b2)(66) with Hidden Markov Models built from confirmed sequences of

MCR-like and EptA-like proteins.

Data availability. All data generated during this study that support the findings are included in the

manuscript or in the Supplementary Information. All materials are available from the corresponding

author upon request.

233

ACKNOWLEDGEMENTS: We thank J. Rowley for assistance with flow cytometry, IHMA Inc.

Schaumburg for the kind gift of the E. coli 1144230 isolate and J. Beckwith, F. Alcock and V. Koronakis

for the kind gifts of the anti-DsbA, the anti-AcrA and the anti-TolC antibodies, respectively. This study

was funded by the MRC Career Development Award MR/M009505/1 (to D.A.I.M.), the institutional

BBSRC-DTP studentships BB/M011178/1 (to N.K.) and BB/M01116X/1 (to H.L.P.), the BBSRC

David Philips Fellowship BB/M02623X/1 (to J.M.A.B.), the ISSF Wellcome Trust grant

105603/Z/14/Z (to G.L.-M.), the Brunel Research Innovation and Enterprise Fund, Innovate UK and

British Society for Antimicrobial Chemotherapy grants 2018-11143, 37800 and BSAC-2018-0095,

respectively (to R.R.MC) and the Swiss National Science Foundation Postdoc Mobility and Ambizione

Fellowships P300PA_167703 and PZ00P3_180142, respectively (to D.G.).

AUTHOR CONTRIBUTIONS: R.C.D.F. and D.A.I.M. designed the research. R.C.D.F. and N.K.

performed most of the experiments. D.B. performed colistin MIC assays and prepared samples for

MALDI-TOF analysis. P.B. provided genetic tools and advice on P. aeruginosa molecular biology.

A.A.A.A. performed β-lactam MIC assays. L.D. provided laboratory materials and strains. H.E.M.,

H.L.P. and J.M.A.B. constructed strains and provided advice on RND efflux pump biology and

experimental design. G.L.-M. performed MALDI-TOF experiments and analyzed the data. E.M and

R.R.MC performed in vivo clearance assays. D.G. performed in silico analyses and advised on several

aspects of the project. R.C.D.F. and D.A.I.M. wrote the manuscript with input from all authors.

D.A.I.M. directed the project.

DECLARATION OF INTERESTS: The authors declare no competing interests.

234

FIGURES

Figure 1. Several antimicrobial resistance mechanisms depend on disulfide bond

formation. (A) DsbA introduces disulfide bonds into extracytoplasmic proteins containing two

or more cysteine residues. After each round of oxidative protein folding, DsbA is regenerated

by the quinone (Q)-containing protein DsbB, which in turn transfers the reducing equivalents

to the respiratory chain (RC) (51). DsbA substrates (in dark blue) are distributed throughout

the extracytoplasmic space of Gram-negative bacteria. Disulfides are introduced to 1) soluble

periplasmic proteins (e.g. alkaline phosphatase, β-lactamases (18)), 2) periplasmic domains of

inner-membrane proteins (e.g LptA-like enzymes (28)), 3) periplasmic domains of outer-

235

membrane proteins (e.g. RcsF (19)), 4) outer-membrane proteins (e.g. OmpA, LptD (19, 25)),

5) secreted proteins (e.g. toxins or enzymes (25)), 6-9) protein components of macromolecular

assemblies like secretion systems, pili or flagella (25) (e.g. 6) GspD, 7) EscC, 8) BfpA, 9)

FlgI); all examples are E. coli proteins with the exception of LptA. (B) β-lactam minimum

inhibitory concentration (MIC) values for E. coli MC1000 expressing diverse disulfide-bond-

containing β-lactamases (Ambler classes A, B and D) are substantially reduced in the absence

of DsbA (MIC fold changes: > 2, fold change of 2 is indicated by the black dotted lines); no

effect is observed for SHV-27, which is further discussed in Figure S3. DsbA dependence is

conserved within phylogenetic groups (see Figure S2, GES-1, -2, -11 and KPC-2, -3). No

changes in MIC values are observed for the aminoglycoside antibiotic gentamicin (white bars)

confirming that absence of DsbA does not compromise the general ability of this strain to resist

antibiotic stress. No changes in MIC values are observed for strains harboring the empty vector

control (pDM1) or those expressing the class A β-lactamase L2-1, which contains three

cysteines but no disulfide bond (top row). Graphs show MIC fold changes for β-lactamase-

expressing E. coli MC1000 and its dsbA mutant and are representative of three biological

experiments each conducted as a single technical repeat; the MIC values used to generate this

panel are presented in Figure S2. (C) Colistin MIC values for E. coli MC1000 expressing

diverse MCR enzymes (Figure S1) are substantially reduced in the absence of DsbA. Graphs

show MIC values (µg/mL) from four biological experiments, each conducted in technical

quadruplicate, to demonstrate the robustness of the observed effects. Gentamicin control data

are presented in Figure S6. (D) Deletion of dsbA reduces the erythromycin, chloramphenicol

and nalidixic acid MIC values for E. coli MG1655, but no effects are detected for the non-

substrate antibiotic gentamicin. The essential pump component AcrA serves as a positive

control. Graphs show MIC values (µg/mL) and are representative of three biological

experiments, each conducted as a single technical repeat.

236

Figure 2. β-lactamase enzymes from most classes become unstable in the absence of DsbA.

(A) Protein levels of disulfide-bond-containing Ambler class A and B β-lactamases are

drastically reduced when these enzymes are expressed in E. coli MC1000 dsbA; the amount of

the control enzyme L2-1 is unaffected. (B) Protein levels of Class D disulfide-bond-containing

β-lactamases are unaffected by the absence of DsbA. OXA-4 is detected as two bands at ~ 28

kDa. For panels (A) and (B) protein levels of StrepII-tagged β-lactamases were assessed using

a Strep-Tactin-AP conjugate or a Strep-Tactin-HRP conjugate. A representative blot from three

biological experiments, each conducted as a single technical repeat, is shown; molecular weight

markers (M) are on the left, DnaK was used as a loading control and solid black lines indicate

where the membrane was cut. (C) The hydrolytic activities of the tested Class D β-lactamases

and of the Class A enzyme GES-1, which could not be detected by immunoblotting, are

237

significantly reduced in the absence of DsbA. The hydrolytic activities of strains harboring the

empty vector or expressing the control enzyme L2-1 show no dependence on DsbA. n=3 (each

conducted in technical triplicate), table shows means ±SD, significance is indicated by * = p <

0.05, ns = non-significant.

238

Figure 3. MCR enzymes become unstable in the absence of DsbA. (A) The amounts of

MCR proteins are drastically reduced when they are expressed in E. coli MC1000 dsbA; the

red arrow indicates the position of the MCR-specific bands. Protein levels of StrepII-tagged

MCR enzymes were assessed using a Strep-Tactin-AP conjugate. A representative blot from

three biological experiments, each conducted as a single technical repeat, is shown; molecular

weight markers (M) are on the left, DnaK was used as a loading control and solid black lines

indicate where the membrane was cut. (B) The ability of MCR enzymes to transfer

phoshoethanolamine to the lipid A portion of LPS is either entirely abrogated or significantly

reduced in the absence of DsbA. This panel shows representative MALDI-TOF mass spectra

of unmodified and MCR-modified lipid A in the presence and absence of DsbA. In E. coli

MC1000 and MC1000 dsbA the major peak for native lipid A peak is detected at m/z 1796.2

(first and second spectrum, respectively). In the presence of MCR enzymes (E. coli MC1000

expressing MCR-3 is shown as a representative example), two additional peaks are observed,

at m/z 1821.2 and 1919.2 (third spectrum). For dsbA mutants expressing MCR enzymes (E.

coli MC1000 dsbA expressing MCR-3 is shown), these additional peaks are not present, whilst

the native lipid A peak at m/z 1796.2 remains unchanged (fourth spectrum). Mass spectra are

representative of the data generated from four biological experiments each conducted as a

technical duplicate. (C) Quantification of the intensities of the lipid A peaks recorded by

MALDI-TOF mass spectrometry for all tested MCR-expressing strains. n=4 (each conducted

239

in technical duplicate), table shows means ±SD, significance is indicated by *** = p < 0.001

or **** = p <0.0001.

240

Figure 4. (A,B, C) RND efflux pump function is impaired in the absence of DsbA due to

accumulation of unfolded AcrA resulting from insufficient DegP activity. (A) In the

absence of DsbA the pool of active DegP is reduced. In E. coli MG1655 (lane 1), DegP is

detected as a single band, corresponding to the intact active enzyme. In E. coli MG1655 dsbA

(lane 2), an additional lower molecular weight band of equal intensity is present, indicating

that DegP is degraded in the absence of its disulfide bond (20, 34). DegP protein levels were

assessed using an anti-DegP primary antibody and an HRP-conjugated secondary antibody. E.

coli MG1655 degP was used as a negative control for DegP detection (lane 3); the red arrow

indicates the position of intact DegP. (B) The RND pump component AcrA accumulates to the

same extent in the E. coli MG1655 dsbA and degP strains, indicating that in both strains protein

clearance is affected. AcrA protein levels were assessed using an anti-AcrA primary antibody

and an HRP-conjugated secondary antibody. E. coli MG1655 acrA was used as a negative

control for AcrA detection; the red arrow indicates the position of the AcrA band. (C) TolC,

the outer-membrane channel of the AcrAB pump, does not accumulate in a dsbA or a degP

mutant. TolC is not a DegP substrate (36), hence similar TolC protein levels are detected in E.

coli MG1655 (lane 1) and its dsbA (lane 2) and degP (lane 3) mutants. TolC protein levels were

assessed using an anti-TolC primary antibody and an HRP-conjugated secondary antibody. E.

coli MG1655 tolC was used as a negative control for TolC detection (lane 4); the red arrow

indicates the position of the bands originating from TolC. For all panels a representative blot

241

from three biological experiments, each conducted as a single technical repeat, is shown;

molecular weight markers (M) are on the left, DnaK was used as a loading control and solid

black lines indicate where the membrane was cut. (D) Impairing disulfide bond formation

in the cell envelope simultaneously incapacitates three distinct classes of AMR

determinants. (Left) When DsbA is present, i.e. when disulfide bond formation occurs,

degradation of β-lactam antibiotics by β-lactamases (marked “bla”), modification of lipid A by

MCR proteins and active efflux of RND pump substrates lead to resistance. The major E. coli

RND efflux pump AcrAB-TolC is depicted in this schematic as a characteristic example.

(Right) In the absence of DsbA, i.e. when the process of disulfide bond formation is impaired,

most cysteine-containing β-lactamases as well as MCR proteins are unstable and degrade,

making bacteria susceptible to β-lactams and colistin. Absence of DsbA also affects

proteostasis in the cell envelope which results in reduced clearance of nonfunctional AcrA-like

proteins (termed “AcrA'” and depicted in dark red color) by periplasmic proteases. Insufficient

clearance of these damaged AcrA components from the pump complex makes efflux

ineffective and offers a way to bypass intrinsic resistance.

242

Figure 5. Chemical inhibition of the DSB system impedes DsbA re-oxidation in E. coli

MC1000 and phenocopies the β-lactam and colistin MIC changes that were observed

using a dsbA mutant. (A) Addition of the reducing agent DTT to E. coli MC1000 bacterial

lysates allows the detection of DsbA in its reduced form (DsbAred) during immunoblotting; this

redox state of the protein, when labelled with the cysteine-reactive compound AMS, shows a

1 kDa size difference (lane 2) compared to oxidized DsbA as found in AMS-labelled but not

reduced lysates of E. coli MC1000 (lane 3). Addition of a small-molecule inhibitor of DsbB to

growing E. coli MC1000 cells also results in accumulation of reduced DsbA (lane 4). E. coli

MC1000 dsbA was used as a negative control for DsbA detection (lane 1). A representative

blot from two biological experiments, each conducted as a single technical repeat, is shown;

DsbA was visualized using an anti-DsbA primary antibody and an AP-conjugated secondary

antibody. Molecular weight markers (M) are shown on the left. (B) MIC experiments using

representative β-lactam antibiotics show that chemical inhibition of the DSB system reduces

the MIC values for E. coli MC1000 expressing disulfide-bond-containing β-lactamases in a

243

similar manner to the deletion of dsbA (compare with Figure 1B). Graphs show MIC fold

changes (i.e. MC1000 MIC (µg/mL) / MC1000 + DSB system inhibitor MIC (µg/mL)) for β-

lactamase-expressing E. coli MC1000 with and without addition of a DSB system inhibitor to

the culture medium and are representative of two biological experiments, each conducted as a

single technical repeat. Black dotted lines indicate an MIC fold change of 2. The

aminoglycoside antibiotic gentamicin serves as a control for all strains; gentamicin MIC values

(white bars) are unaffected by chemical inhibition of the DSB system (MIC fold changes: < 2).

No changes in MIC values (MIC fold changes: < 2) are observed for strains harboring the

empty vector control (pDM1) or expressing the class A β-lactamase L2-1, which contains three

cysteines but no disulfide bond (PDB ID: 5NE1) (top row). (C) Colistin MIC experiments

show that chemical inhibition of the DSB system reduces the MIC values for E. coli MC1000

expressing MCR enzymes in a similar manner to the deletion of dsbA (compare with Figure

1C). Colistin MIC values for strains harboring the empty vector control (pDM1) are unaffected

by chemical inhibition of the DSB system. Graphs show MIC values (µg/mL) from four

biological experiments, each conducted in technical quadruplicate, to demonstrate the

robustness of the observed effects. (D) The aminoglycoside antibiotic gentamicin serves as a

control for all strains tested in panel (C); gentamicin MIC values are unaffected by chemical

inhibition of the DSB system. Graphs show MIC values (µg/mL) from two biological

experiments, each conducted as a single technical repeat.

244

Figure 6. Chemical inhibition of the DSB system sensitizes multidrug-resistant clinical

isolates to currently available antibiotics. (A) Addition of a small-molecule inhibitor of

DsbB results in sensitization of Klebsiella pneumoniae, E. coli and Citrobacter freundii clinical

isolates to imipenem. Chemical inhibition of the DSB system of an Enterobacter cloacae

clinical isolate harboring blaFRI-1 results in reduction of the aztreonam MIC value by over 180

245

µg/mL, resulting in intermediate resistance as defined by EUCAST. MIC values determined

using Mueller-Hinton agar (MHA) in accordance with the EUCAST guidelines (light blue bars)

are comparable to the values obtained using defined media (M63 agar, white bars); use of

growth media lacking small-molecule oxidants is required for the DSB system inhibitor to be

effective. Graphs show MIC values (μg/ml) representative of two biological experiments, each

conducted as a single technical repeat. (B) Application of the same chemical inhibitor to

colistin-resistant E. coli expressing MCR enzymes results in sensitization of all tested clinical

isolates to colistin. Graphs show MIC values (µg/mL) from four biological experiments, each

conducted in technical quadruplicate, to demonstrate the robustness of the observed effects.

(C) Deletion of dsbA sensitizes the efflux-active E. coli MG1655 strain to chloramphenicol;

the data presented in the blue and light blue bars were also used to generate part of Figure 1D.

Sensitization is also observed for the dsbA mutant of the deregulated E. coli MG1655 marR

strain (chloramphenicol MIC of 6 μg/mL). The graph shows MIC values (μg/ml) from 2

biological experiments, each conducted as a single technical repeat. (D) Absence of the

principal pseudomonal DsbA analogue (DsbA1) sensitizes the P. aeruginosa PA43417 clinical

isolate expressing OXA-198 to the first-line antibiotic piperacillin (piperacillin MIC of 12

μg/mL). The graph shows MIC values (μg/ml) from 2 biological experiments, each conducted

as a single technical repeat. For all panels, red dotted lines indicate the EUCAST clinical

breakpoint for each antibiotic. (E) Absence of the principal DsbA analogue (DsbA1) from

a P. aeruginosa clinical isolate expressing OXA-198 allows it to be cleared from infected

G. mellonella larvae by piperacillin. Neither deletion of dsbA1, nor treatment with

piperacillin (at a concentration of 12 μg/mL) is sufficient to clear P. aeruginosa PA4317 from

infected G. mellonella larvae, but the combination of both results in an average reduction in

bacterial load that is greater than 99%. The graph shows the average number of colony forming

units (CFU) recovered from infected larvae for each condition relative to the CFU recovered

for the untreated P. aeruginosa PA4317 strain. n = 8 groups infected on eight different days;

each group contains five G. mellonella larvae per condition except for one group which

contains three G. mellonella larvae per condition. Graph shows means ±SD, significance is

indicated by * = p < 0.05, ** = p < 0.01, *** = p <0.001.

246

REFERENCES

1. Rochford C, Sridhar D, Woods N, Saleh Z, Hartenstein L, Ahlawat H, et al. Global

governance of antimicrobial resistance. Lancet. 2018;391(10134):1976-8.

2. Tacconelli E, Carrara E, Savoldi A, Harbarth S, Mendelson M, Monnet DL, et al.

Discovery, research, and development of new antibiotics: the WHO priority list of antibiotic-

resistant bacteria and tuberculosis. Lancet Infect Dis. 2018;18(3):318-27.

3. Hart EM, Mitchell AM, Konovalova A, Grabowicz M, Sheng J, Han X, et al. A small-

molecule inhibitor of BamA impervious to efflux and the outer membrane permeability barrier.

Proc Natl Acad Sci U S A. 2019;116(43):21748-57.

4. Laws M, Shaaban A, Rahman KM. Antibiotic resistance breakers: current approaches

and future directions. FEMS Microbiol Rev. 2019;43(5):490-516.

5. Luther A, Urfer M, Zahn M, Muller M, Wang SY, Mondal M, et al. Chimeric

peptidomimetic antibiotics against Gram-negative bacteria. Nature. 2019;576(7787):452-8.

6. Nicolas I, Bordeau V, Bondon A, Baudy-Floc'h M, Felden B. Novel antibiotics

effective against Gram-positive and -negative multi-resistant bacteria with limited resistance.

PLoS Biol. 2019;17(7):e3000337.

7. Srinivas N, Jetter P, Ueberbacher BJ, Werneburg M, Zerbe K, Steinmann J, et al.

Peptidomimetic antibiotics target outer-membrane biogenesis in Pseudomonas aeruginosa.

Science. 2010;327(5968):1010-3.

8. Bush K. Past and present perspectives on β-lactamases. Antimicrob Agents Chemother.

2018;62(10).

9. Meletis G. Carbapenem resistance: overview of the problem and future perspectives.

Ther Adv Infect Dis. 2016;3(1):15-21.

10. Versporten A, Zarb P, Caniaux I, Gros MF, Drapier N, Miller M, et al. Antimicrobial

consumption and resistance in adult hospital inpatients in 53 countries: results of an internet-

based global point prevalence survey. Lancet Glob Health. 2018;6(6):e619-e29.

11. Li J, Nation RL, Turnidge JD, Milne RW, Coulthard K, Rayner CR, et al. Colistin: the

re-emerging antibiotic for multidrug-resistant Gram-negative bacterial infections. Lancet

Infect Dis. 2006;6(9):589-601.

12. Poirel L, Jayol A, Nordmann P. Polymyxins: antibacterial activity, susceptibility

testing, and resistance mechanisms encoded by plasmids or chromosomes. Clin Microbiol Rev.

2017;30(2):557-96.

13. Sun J, Zhang H, Liu YH, Feng Y. Towards understanding MCR-like colistin resistance.

Trends Microbiol. 2018;26(9):794-808.

14. Blair JMA, Richmond GE, Piddock LJV. Multidrug efflux pumps in Gram-negative

bacteria and their role in antibiotic resistance. Future Microbiol. 2014;9(10):1165-77.

247

15. Cox G, Wright GD. Intrinsic antibiotic resistance: mechanisms, origins, challenges and

solutions. Int J Med Microbiol. 2013;303(6-7):287-92.

16. Goemans C, Denoncin K, Collet JF. Folding mechanisms of periplasmic proteins.

Biochim Biophys Acta. 2014;1843(8):1517-28.

17. Heras B, Kurz M, Shouldice SR, Martin JL. The name's bond......disulfide bond. Curr

Opin Struct Biol. 2007;17(6):691-8.

18. Bardwell JC, McGovern K, Beckwith J. Identification of a protein required for disulfide

bond formation in vivo. Cell. 1991;67(3):581-9.

19. Denoncin K, Collet JF. Disulfide bond formation in the bacterial periplasm: major

achievements and challenges ahead. Antioxid Redox Signal. 2013;19(1):63-71.

20. Hiniker A, Bardwell JC. In vivo substrate specificity of periplasmic disulfide

oxidoreductases. J Biol Chem. 2004;279(13):12967-73.

21. Kadokura H, Tian H, Zander T, Bardwell JC, Beckwith J. Snapshots of DsbA in action:

detection of proteins in the process of oxidative folding. Science. 2004;303(5657):534-7.

22. Martin JL, Bardwell JC, Kuriyan J. Crystal structure of the DsbA protein required for

disulphide bond formation in vivo. Nature. 1993;365(6445):464-8.

23. Dutton RJ, Boyd D, Berkmen M, Beckwith J. Bacterial species exhibit diversity in their

mechanisms and capacity for protein disulfide bond formation. Proc Natl Acad Sci U S A.

2008;105(33):11933-8.

24. Vertommen D, Depuydt M, Pan J, Leverrier P, Knoops L, Szikora JP, et al. The

disulphide isomerase DsbC cooperates with the oxidase DsbA in a DsbD-independent manner.

Mol. Microbiol. 2008;67(2):336-49.

25. Heras B, Shouldice SR, Totsika M, Scanlon MJ, Schembri MA, Martin JL. DSB

proteins and bacterial pathogenicity. Nat Rev Microbiol. 2009;7(3):215-25.

26. Landeta C, Boyd D, Beckwith J. Disulfide bond formation in prokaryotes. Nature

Microbiol. 2018;3(3):270-80.

27. Heras B, Scanlon MJ, Martin JL. Targeting virulence not viability in the search for

future antibacterials. Br J Clin Pharmacol. 2014;79(2):208-15.

28. Piek S, Wang Z, Ganguly J, Lakey AM, Bartley SN, Mowlaboccus S, et al. The role of

oxidoreductases in determining the function of the neisserial lipid A phosphoethanolamine

transferase required for resistance to polymyxin. PloS One. 2014;9(9):e106513.

29. Nation RL, Garonzik SM, Li J, Thamlikitkul V, Giamarellos-Bourboulis EJ, Paterson

DL, et al. Updated US and European dose recommendations for intravenous colistin: how do

they perform? Clin Infect Dis. 2016;62(5):552-8.

30. Plachouras D, Karvanen M, Friberg LE, Papadomichelakis E, Antoniadou A, Tsangaris

I, et al. Population pharmacokinetic analysis of colistin methanesulfonate and colistin after

intravenous administration in critically ill patients with infections caused by Gram-negative

bacteria. Antimicrob Agents Chemother. 2009;53(8):3430-6.

248

31. Wang Z, Fan G, Hryc CF, Blaza JN, Serysheva, II, Schmid MF, et al. An allosteric

transport mechanism for the AcrAB-TolC multidrug efflux pump. eLife. 2017;6.

32. Dortet L, Bonnin RA, Pennisi I, Gauthier L, Jousset AB, Dabos L, et al. Rapid detection

and discrimination of chromosome- and MCR-plasmid-mediated resistance to polymyxins by

MALDI-TOF MS in Escherichia coli: the MALDIxin test. J Antimicrob Chemother.

2018;73(12):3359-67.

33. Clausen T, Southan C, Ehrmann M. The HtrA family of proteases: implications for

protein composition and cell fate. Mol Cell. 2002;10(3):443-55.

34. Skorko-Glonek J, Zurawa D, Tanfani F, Scire A, Wawrzynow A, Narkiewicz J, et al.

The N-terminal region of HtrA heat shock protease from Escherichia coli is essential for

stabilization of HtrA primary structure and maintaining of its oligomeric structure. Biochim

Biophys Acta. 2003;1649(2):171-82.

35. Gerken H, Misra R. Genetic evidence for functional interactions between TolC and

AcrA proteins of a major antibiotic efflux pump of Escherichia coli. Mol Microbiol.

2004;54(3):620-31.

36. Werner J, Augustus AM, Misra R. Assembly of TolC, a structurally unique and

multifunctional outer membrane protein of Escherichia coli K-12. J Bacteriol.

2003;185(22):6540-7.

37. Duprez W, Premkumar L, Halili MA, Lindahl F, Reid RC, Fairlie DP, et al. Peptide

inhibitors of the Escherichia coli DsbA oxidative machinery essential for bacterial virulence.

J Med Chem. 2015;58(2):577-87.

38. Totsika M, Vagenas D, Paxman JJ, Wang G, Dhouib R, Sharma P, et al. Inhibition of

diverse DsbA enzymes in multi-DsbA encoding pathogens. Antioxid Redox Signal.

2018;29(7):653-66.

39. Landeta C, Blazyk JL, Hatahet F, Meehan BM, Eser M, Myrick A, et al. Compounds

targeting disulfide bond forming enzyme DsbB of Gram-negative bacteria. Nat Chem Biol.

2015;11(4):292-8.

40. Halili MA, Bachu P, Lindahl F, Bechara C, Mohanty B, Reid RC, et al. Small molecule

inhibitors of disulfide bond formation by the bacterial DsbA-DsbB dual enzyme system. ACS

Chem Biol. 2015;10(4):957-64.

41. Dailey FE, Berg HC. Mutants in disulfide bond formation that disrupt flagellar

assembly in Escherichia coli. Proc Natl Acad Sci U S A. 1993;90(3):1043-7.

42. Hizukuri Y, Yakushi T, Kawagishi I, Homma M. Role of the intramolecular disulfide

bond in FlgI, the flagellar P-ring component of Escherichia coli. J Bacteriol.

2006;188(12):4190-7.

43. Kishigami S, Akiyama Y, Ito K. Redox states of DsbA in the periplasm of Escherichia

coli. FEBS Lett. 1995;364(1):55-8.

249

44. Ariza RR, Cohen SP, Bachhawat N, Levy SB, Demple B. Repressor mutations in the

marRAB operon that activate oxidative stress genes and multiple antibiotic resistance in

Escherichia coli. J Bacteriol. 1994;176(1):143-8.

45. Okusu H, Ma D, Nikaido H. AcrAB efflux pump plays a major role in the antibiotic

resistance phenotype of Escherichia coli multiple-antibiotic-resistance (Mar) mutants. J

Bacteriol. 1996;178(1):306-8.

46. Keeney D, Ruzin A, McAleese F, Murphy E, Bradford PA. MarA-mediated

overexpression of the AcrAB efflux pump results in decreased susceptibility to tigecycline in

Escherichia coli. J Antimicrob Chemother. 2008;61(1):46-53.

47. Arts IS, Ball G, Leverrier P, Garvis S, Nicolaes V, Vertommen D, et al. Dissecting the

machinery that introduces disulfide bonds in Pseudomonas aeruginosa. mBio. 2013;4(6).

48. Landeta C, McPartland L, Tran NQ, Meehan BM, Zhang Y, Tanweer Z, et al. Inhibition

of Pseudomonas aeruginosa and Mycobacterium tuberculosis disulfide bond forming

enzymes. Mol Microbiol. 2019;111(4):918-37.

49. Zhou YL, Wang JF, Guo Y, Liu XQ, Liu SL, Niu XD, et al. Discovery of a potential

MCR-1 inhibitor that reverses polymyxin activity against clinical mcr-1-positive

Enterobacteriaceae. J Infect. 2019;78(5):364-72.

50. Sharma A, Gupta VK, Pathania R. Efflux pump inhibitors for bacterial pathogens: From

bench to bedside. Indian J Med Res. 2019;149(2):129-45.

51. Kadokura H, Katzen F, Beckwith J. Protein disulfide bond formation in prokaryotes.

Annu Rev Biochem. 2003;72:111-35.

52. Wilmaerts D, Dewachter L, De Loose PJ, Bollen C, Verstraeten N, Michiels J. HokB

monomerization and membrane repolarization control persister awakening. Mol Cell.

2019;75(5):1031-42.

53. Chang AC, Cohen SN. Construction and characterization of amplifiable multicopy

DNA cloning vehicles derived from the P15A cryptic miniplasmid. J Bacteriol.

1978;134(3):1141-56.

54. Kim J, Webb AM, Kershner JP, Blaskowski S, Copley SD. A versatile and highly

efficient method for scarless genome editing in Escherichia coli and Salmonella enterica. BMC

Biotechnol. 2014;14:84.

55. McKenzie GJ, Craig NL. Fast, easy and efficient: site-specific insertion of transgenes

into enterobacterial chromosomes using Tn7 without need for selection of the insertion event.

BMC Microbiol. 2006;6:39.

56. Vasseur P, Vallet-Gely I, Soscia C, Genin S, Filloux A. The pel genes of the

Pseudomonas aeruginosa PAK strain are involved at early and late stages of biofilm formation.

Microbiology. 2005;151:985-97.

57. Kaniga K, Delor I, Cornelis GR. A wide-host-range suicide vector for improving

reverse genetics in Gram-negative bacteria: inactivation of the blaA gene of Yersinia

enterocolitica. Gene. 1991;109(1):137-41.

250

58. Helander IM, Mattila-Sandholm T. Fluorometric assessment of Gram-negative

bacterial permeabilization. J Appl Microbiol. 2000;88(2):213-9.

59. Mavridou DA, Gonzalez D, Clements A, Foster KR. The pUltra plasmid series: a robust

and flexible tool for fluorescent labeling of Enterobacteria. Plasmid. 2016;87-88:65-71.

60. Larrouy-Maumus G, Clements A, Filloux A, McCarthy RR, Mostowy S. Direct

detection of lipid A on intact Gram-negative bacteria by MALDI-TOF mass spectrometry. J

Microbiol Methods 2016;120:68-71.

61. McCarthy RR, Mazon-Moya MJ, Moscoso JA, Hao Y, Lam JS, Bordi C, et al. Cyclic-

di-GMP regulates lipopolysaccharide modification and contributes to Pseudomonas

aeruginosa immune evasion. Nat Microbiol. 2017;2:17027.

62. Naas T, Oueslati S, Bonnin RA, Dabos ML, Zavala A, Dortet L, et al. β-lactamase

database (BLDB) - structure and function. J Enzyme Inhib Med Chem. 2017;32(1):917-9.

63. Edgar RC. Search and clustering orders of magnitude faster than BLAST.

Bioinformatics. 2010;26(19):2460-1.

64. Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ. Basic local alignment search

tool. J Mol Biol. 1990;215(3):403-10.

65. Edgar RC. MUSCLE: multiple sequence alignment with high accuracy and high

throughput. Nucleic Acids Res. 2004;32(5):1792-7.

66. Finn RD, Clements J, Arndt W, Miller BL, Wheeler TJ, Schreiber F, et al. HMMER

web server: 2015 update. Nucleic Acids Res. 2015;43(W1):W30-8.

251

SUPPLEMENTARY INFORMATION FOR

Breaking antimicrobial resistance by disrupting extracytoplasmic

protein folding

R. Christopher D. Furniss2,†, Nikol Kadeřábková2,†, Declan Barker2, Patricia Bernal3, Evgenia

Maslova4, Amanda A.A. Antwi2, Helen E. McNeil5, Hannah L. Pugh5, Laurent Dortet2,6,7,8,

Jessica M.A. Blair5, Gerald Larrouy-Maumus2, Ronan R. McCarthy4, Diego Gonzalez9,

Despoina A.I. Mavridou1,2,*

1Department of Molecular Biosciences, University of Texas at Austin, Austin, 78712, Texas,

USA 2MRC Centre for Molecular Bacteriology and Infection, Department of Life Sciences, Imperial

College London, London, SW7 2AZ, UK 3Department of Biology, Faculty of Sciences, Universidad Autónoma de Madrid, Madrid,

28049, Spain 4Division of Biosciences, Department of Life Sciences, College of Health and Life Sciences,

Brunel University London, Uxbridge, UB8 3PH, UK 5Institute of Microbiology and Infection, College of Medical and Dental Sciences, University

of Birmingham, Birmingham, B15 2TT, UK 6Department of Bacteriology-Hygiene, Bicêtre Hospital, Assistance Publique - Hôpitaux de

Paris, Le Kremlin-Bicêtre, 94270, France 7EA7361 “Structure, Dynamics, Function and Expression of Broad-spectrum β-lactamases",

Paris-Sud University, LabEx Lermit, Faculty of Medicine, Le Kremlin-Bicêtre, 94270, France 8French National Reference Centre for Antibiotic Resistance, Le Kremlin-Bicêtre, 94270,

France 9Laboratoire de Microbiologie, Institut de Biologie, Université de Neuchâtel, Neuchâtel, 2000,

Switzerland

*Correspondence: despoina.mavridou@austin.utexas.edu

†These authors have contributed equally to this work

This PDF file includes:

Figures S1 toS14

Tables S1 to S6

Legend for File S1

Supplementary references

252

SUPPLEMENTARY FIGURES

Figure S1. Phylogenetic analysis of MCR- and EptA-like enzymes found in

Proteobacteria. A phylogenetic tree was built based on the alignment of 781 sequences from

Proteobacteria. The assignment of each sequence to a specific group was done using Hidden

Markov Models built from confirmed sequences of MCR- and EptA-like proteins; EptA-like

enzymes are chromosomal phosphoethanolamine transferases that belong to the same extended

protein superfamily as MCR enzymes, but do not give rise to polymyxin resistance in

Enterobacteria (1). The different MCR groups are broadly indicated in different colors,

however it should be noted that there is significant overlap between groups. Open circles mark

the enzymes tested in this study which are distributed throughout the MCR phylogeny.

253

Figure S2. Deletion of dsbA lowers the β-lactam MIC values for E. coli MC1000

expressing diverse β-lactamases. In the absence of DsbA the β-lactam MICs (for multiple

classes of β-lactam antibiotics) for E. coli MC1000 expressing disulfide-bond-containing β-

lactamases are drastically reduced. This figure shows the MIC data used to generate Figure 1B,

and the MIC values for the additional GES and KPC family members GES-2, GES-11 and

KPC-2. The aminoglycoside antibiotic gentamicin serves as a control for all strains. (A)

MC1000 pDM1 (vector alone) (B) MC1000 pDM1-blaL2-1 (cysteine-containing β-lactamase

which lacks a disulfide bond) (C) MC1000 pDM1-blaGES-1, (D) MC1000 pDM1-blaGES-2, (E)

MC1000 pDM1-blaGES-11, (F) MC1000 pDM1-blaSHV-27, (G) MC1000 pDM1-blaOXA-4, (H)

MC1000 pDM1-blaOXA-10, (I) MC1000 pDM1-blaOXA-198, (J) MC1000 pDM1-blaFRI-1, (K)

MC1000 pDM1-blaL1-1, (L) MC1000 pDM1-blaKPC-2, (M) MC1000 pDM1-blaKPC-3, (N)

MC1000 pDM1 blaSME-1. Graphs show MIC values (µg/mL) and are representative of three

biological experiments, each conducted as a single technical repeat.

254

255

256

Figure S3. SHV-27 function is dependent on DsbA at temperatures higher than 37 °C.

The ESBL SHV-27 differs from the canonical SHV-1 enzyme by a single amino acid

substitution (D156G) (2). At 37 °C deletion of dsbA does not affect the cefuroxime MIC for E.

coli MC1000 harboring pDM1-blaSHV-27. However, at 42 °C the cefuroxime MIC for E. coli

MC1000 dsbA harboring pDM1-blaSHV-27 is notably reduced. Thus, in a similar manner to

TEM-1 (3), the SHV-27 disulfide bond becomes important for enzyme function under stress

conditions (temperature stress). As SHV-27 has the narrowest hydrolytic spectrum out of all

the enzymes tested in this study, this result suggests that there could be a correlation between

the hydrolytic spectrum of the β-lactamase and its dependence on DsbA for conferring

resistance. The graph shows MIC values (µg/mL) and is representative of three biological

experiments, each conducted as a single technical repeat.

257

Figure S4. Complementation of dsbA restores the β-lactam MIC values for E. coli

MC1000 dsbA expressing β-lactamases. Re-insertion of dsbA at the attTn7 site of the

chromosome restores the β-lactam MIC values for E. coli MC1000 dsbA harboring (A) pDM1-

blaGES-1 (ceftazidime MIC), (B) pDM1-blaOXA-4 (cefuroxime MIC), (C) pDM1-blaOXA-10

(aztreonam MIC), (D) pDM1-blaOXA-198 (imipenem MIC), (E) pDM1-blaL1-1 (ceftazidime

MIC), (F) pDM1-blaFRI-1 (aztreonam MIC) and (G) pDM1-blaKPC-3 (ceftazidime MIC). Graphs

show MIC values (µg/mL) and are representative of two biological experiments, each

conducted as a single technical repeat.

258

Figure S5. Complementation of dsbA restores the colistin MIC values for E. coli MC1000

dsbA expressing MCR enzymes. Re-insertion of dsbA at the attTn7 site of the chromosome

restores the colistin MIC values for E. coli MC1000 dsbA harboring (A) pDM1-mcr-1 (B)

pDM1-mcr-3 (C) pDM1-mcr-4 (D) pDM1-mcr-5 (E) pDM1-mcr-8. Graphs show MIC values

(µg/mL) from four biological experiments, each conducted in technical quadruplicate, to

demonstrate the robustness of the observed effects.

259

Figure S6. Gentamicin MIC values for E. coli MC1000 strains expressing MCR enzymes.

Deletion of dsbA does not affect the gentamicin MIC values for E. coli MC1000 strains

expressing MCR enzymes, confirming that absence of DsbA does not compromise the general

ability of this strain to resist antibiotic stress. Graphs show MIC values (µg/mL) and are

representative of two biological experiments, each conducted as a single technical repeat.

260

Figure S7. Deletion of dsbA has no effect on membrane permeability in E. coli MC1000.

(A) The bacterial outer membrane acts as a selective permeability barrier to hydrophobic

molecules. Deletion of dsbA has no effect on the outer membrane integrity of E. coli MC1000,

as the hydrophobic fluorescent dye NPN crosses the outer membrane of E. coli MC1000 and

its dsbA mutant to the same extent. Conversely, exposure to the outer-membrane-

permeabilizing antibiotic colistin results in a significant increase in NPN uptake. (B) PI is a

cationic hydrophilic dye that fluoresces upon intercalation with nucleic acids. Under normal

conditions PI freely crosses the outer membrane but is unable to cross the inner membrane.

Deletion of dsbA does not result in damage to the bacterial inner membrane, as no difference

in basal PI uptake is seen between E. coli MC1000 and its dsbA mutant. Both strains express

superfolder GFP (sfGFP), and fluorescence was used to distinguish live from dead cells.

Addition of the inner-membrane-permeabilizing antimicrobial peptide cecropin A (4) to E. coli

MC1000 induces robust inner-membrane permeabilization in the sfGFP-positive population

indicating that the inner membrane becomes compromised. For both experiments n=3 (each

conducted in technical triplicate), graph shows means ±SD, significance is indicated by *** =

p < 0.001, ns = non-significant.

261

Figure S8. Complementation of dsbA restores efflux-pump substrate MIC values for E.

coli MG1655 dsbA. Re-insertion of dsbA at the attTn7 site of the chromosome restores (A)

erythromycin, (B) chloramphenicol and (C) nalidixic acid MIC values for MG1655 dsbA.

Graphs show MIC values (µg/mL) and are representative of two biological experiments, each

conducted as a single technical repeat.

262

Figure S9. Deletion of dsbA has no effect on membrane permeability in E. coli MG1655.

(A) Deletion of dsbA has no effect on the outer membrane integrity of E. coli MG1655, as the

hydrophobic fluorescent dye NPN crosses the outer membrane of E. coli MG1655 and its dsbA

mutant to the same extent. Conversely, exposure to the outer-membrane-permeabilizing

antibiotic colistin results in a significant increase in NPN uptake. (B) Deletion of dsbA does

not result in damage to the bacterial inner membrane, as no difference in basal PI uptake is

seen between E. coli MG1655 and its dsbA mutant. Both strains express sfGFP, and

fluorescence was used to distinguish live from dead cells. Addition of the inner-membrane-

permeabilizing antimicrobial peptide cecropin A (4) to E. coli MG1655 induces robust inner

membrane permeabilization in the sfGFP-positive population indicating that the inner

membrane becomes compromised. For both experiments n=3 (each conducted in technical

triplicate), graph shows means ±SD, significance is indicated by **** = p < 0.0001, ns = non-

significant.

263

Figure S10. Chemical inhibition of the DSB system impedes flagellar motility in E. coli

MC1000. (A) A functional DSB system is necessary for flagellar motility in E. coli because

folding of the P-ring component FlgI requires DsbA-mediated disulfide bond formation (5). In

the absence of DsbA, or upon addition of a chemical inhibitor of the DSB system, the motility

of E. coli MC1000 is significantly impeded. Representative images of motility plates are

shown. (B) Quantification of the growth halo diameters in the motility assays shown in panel

(A). n=3 (each conducted as a single technical repeat), graph shows means ±SD, significance

is indicated by **** = p < 0.0001.

264

Figure S11. Chemical inhibition of the DSB system or deletion of dsbA does not

compromise the growth of E. coli MC1000. Growth curves of (A) E. coli MC1000 with and

without chemical inhibition of the DSB system and (B) E. coli MC1000 and its dsbA mutant

show that bacterial growth remains unaffected by the DSB system inhibitor compound used in

this study, or by the absence of DsbA. n=3 (each conducted as a technical triplicate), solid lines

indicate mean values, shaded areas indicate SD.

265

Figure S12. Changes in MIC values observed using the DSB system inhibitor are due

solely to inhibition of the DSB system. (A) E. coli MC1000 harboring pDM1-blaKPC-3 has an

imipenem MIC value of 24 μg/mL. Upon chemical inhibition of the DSB system the imipenem

MIC for this strain drops to 4 μg/mL, and accordingly the imipenem MIC for E. coli MC1000

dsbA harboring pDM1-blaKPC-3 is 2 μg/mL. The imipenem MIC for E. coli MC1000 dsbA

harboring pDM1-blaKPC-3 when exposed to the chemical inhibitor of the DSB system is also 2

μg/mL, indicating that the chemical compound used in this study does not have any off-target

effects and only affects the function of the DSB system proteins. (B) Chemical inhibition of

the DSB system does not lead to any cumulative effects when tested on an E. coli MC1000

strain expressing MCR-5. The colistin MIC for E. coli MC1000 harboring pDM1-mcr-5 is 3

μg/mL and it drops to 1 μg/mL when the DSB system is chemically inhibited or dsbA is deleted.

The same drop in colistin MIC is observed when the E. coli MC1000 dsbA strain harboring

pDM1-mcr-5 is exposed to the chemical inhibitor of the DSB system. Data shown in both

panels are representative of two biological experiments, each conducted as a single technical

repeat.

266

Figure S13. Deletion of dsbA results in reduced MIC values for E. coli MC1000 expressing

MCR-3.2. When cloned into pDM1 and expressed in E. coli MC1000, MCR-3.2 confers

colistin resistance as expected (MIC of 3.0-3.5 μg/ml). Deletion of dsbA reduces the colistin

MIC values for E. coli MC1000 expressing MCR-3.2 (MIC ≤ 2 μg/mL). Graphs show MIC

values (µg/mL) from four biological experiments, each conducted in technical quadruplicate,

to demonstrate the robustness of the observed effects.

267

Figure S14. Deletion of marR results in increased expression of the AcrAB pump (lane 2)

compared to the parental strain (lane 1) even though the chloramphenicol MIC for both

strains is the same (Figure 6C). E. coli MG1655 has an already high level of efflux activity

and therefore deletion of marR likely does not result in a drastic change in the observed

chloramphenicol MIC value. Expression of the AcrAB pump was assessed using an anti-AcrA

primary antibody and an HRP-conjugated secondary antibody. E. coli MG1655 acrA was used

as a negative control for AcrA detection (lane 3); the red arrow indicates the position of the

AcrA band. A representative blot from two biological experiments, each conducted as a single

technical repeat, is shown; molecular weight markers (M) are shown on the left and DnaK was

used as a loading control.

268

SUPPLEMENTARY TABLES

Table S1. Overview of the β-lactamase enzymes investigated in this study. Enzymes GES-1, -

2 and -11 as well as KPC-2 and -3 belong to the same phylogenetic cluster (GES-27 and KPC-

9, respectively, see File S1). All other tested enzymes belong to distinct phylogenetic clusters

(File S1). The “Cysteine positions” column states the positions of cysteine residues after

position 30 and hence, does not include amino acids that would be part of the periplasmic signal

sequence. All β-lactamase enzymes except L2-1 (shaded in grey; PDB ID 1O7E) have one

disulfide bond. The “Mobile” column refers to the genetic location of the β-lactamase gene;

“yes” indicates that the gene of interest is located on a plasmid, while “no” refers to

chromosomally-encoded enzymes. All tested enzymes have a broad hydrolytic spectrum and

are either Extended Spectrum β-Lactamases (ESBLs) or carbapenemases. The “Inhibition”

column refers to classical inhibitor susceptibility i.e. susceptibility to inhibition by clavulanic

acid, tazobactam or sulbactam.

Enzyme Ambler

class

Cysteine

positions Mobile Spectrum Inhibition

L2-1 A C82 C136 C233 no ESBL yes

GES-1 A C63 C233 yes ESBL yes

GES-2 A C63 C233 yes ESBL yes

GES-11 A C63 C233 yes Carbapenemase yes

SHV-27 A C73 C119 no ESBL yes

OXA-4 D C43 C63 yes ESBL yes

OXA-10 D C44 C51 yes ESBL no (6)

OXA-198 D C116 C119 yes Carbapenemase no (7)

FRI-1 A C68 C238 yes Carbapenemase no (8)

L1-1 B3 C239 C267 no Carbapenemase no (9)

KPC-2 A C68 C237 yes Carbapenemase no (10)

KPC-3 A C68 C237 yes Carbapenemase no (10)

SME-1 A C72 C242 no Carbapenemase yes

269

Table S2. Antibiotic resistance profiles of the clinical isolates tested in this study. The table

shows MIC values (µg/mL) for a range of commonly used antibiotics. Values highlighted in

pink indicate resistance, as defined by the EUCAST clinical breakpoint guidelines, whilst

values highlighted in light blue indicate antibiotics for which there is no EUCAST clinical

breakpoint. The remaining values (white cells) indicate sensitivity to the tested antibiotic

compound. Strains shaded in yellow are multidrug resistant. The following abbreviations are

used: AC, amoxicillin; XM, cefuroxime; TZ, ceftazidime; IP, imipenem; AT, aztreonam; PT,

piperacillin/tazobactam; GM, gentamicin; CO, colistin; CI, ciprofloxacin; NF, nitrofurantoin;

TR, trimethoprim.

Strain AC XM TZ IP AT PT GM CO CI NF TR

E. coli BM16

(blaTEM-1b blaKPC-

2)

>256 >256 192 12 >256 >64 8 1 >32 >512 >32

E. coli LIL-1

(blaTEM-1 blaOXA-9

blaKPC-2)

>256 >256 8 3 192 >64 1.5 2 >32 6 >32

E. coli CNR1790 (blaTEM-15 mcr-1)

>256 >256 32 0.5 16 <2 1 4 >32 16 >32

E. coli

CNR20140385

(blaOXA-48 mcr-1)

>256 96 1 0.25 0.38 32 2 4 >32 8 >32

E. coli WI2

(blaOXA-48 blaKPC-

28 mcr-1)

>256 >256 >256 1.5 32 >64 1.5 4 0.016 6 0.25

E. coli 27841 (blaCTX-M-55 mcr-

3.2)

>256 >256 16 0.19 64 <2 32 3 >32 8 >32

E. coli 1144230 (blaCMY-2 mcr-5)

>256 96 12 0.5 6 8 1.5 4 0.025 48 1

K. pneumoniae

ST234

(blaSHV-27 blaKPC-

2)

>256 >256 48 16 128 >64 0.38 2 0.047 96 1

C. freundii

BM19

(blaKPC-2)

>256 >256 128 4 64 >64 24 2 >32 8 >32

E. cloacae DUB

(blaFRI-1) >256 >256 16 12 >256 >64 1.5 >4 0.016 48 0.75

P. aeruginosa

PA43417

(blaOXA-198)

>256 >256 2 >32 6 32 16 1 >32 >256 >32

270

Table S3. Bacterial strains used in this study. All listed isolates are clinical strains except for

Escherichia coli 27841 (ST744), which is an environmental strain. For clinical and

environmental isolates, the multi-locus sequence types (ST) are given in parenthesis, where

available.

Name Description Source

Escherichia coli

DH5α

F– endA1 glnV44 thi-1 recA1 relA1

gyrA96 deoR nupG purB20

φ80dlacZ∆M15 ∆(lacZYA-argF)U169

hsdR17(rK–mK

+) λ–

(11)

CC118λpir araD Δ(ara, leu) ΔlacZ74 phoA20

galK thi-1 rspE rpoB argE recA1 λpir (12)

HB101 supE44 hsdS20 recA13 ara-14 proA2

lacY1 galK2 rpsL20 xyl-5 mtl-1 (13)

MC1000 araD139 ∆(ara, leu)7697 ∆lacX74

galU galK strA (14)

MC1000 dsbA dsbA::aphA, KanR (15)

MC1000 dsbA attTn7::Ptac-dsbA dsbA::aphA attTn7::dsbA, KanR This study

MG1655 K-12 F– λ– ilvG– rfb-50 rph-1 (16)

MG1655 dsbA dsbA::aphA, KanR This study

MG1655 dsbA attTn7::Ptac-dsbA dsbA::aphA attTn7::dsbA, KanR This study

MG1655 acrA acrA This study

MG1655 tolC tolC This study

MG1655 degP degP::strAB, SmR This study

MG1655 marR marR::accC, GentR This study

MG1655 dsbA marR dsbA::aphA marR::accC, KanR, GentR This study

Clinical / environmental isolates

Escherichia coli BM16 blaTEM-1b blaKPC-2 (17)

Escherichia coli LIL-1 blaTEM-1 blaOXA-9 blaKPC-2 (17)

Escherichia coli CNR1790 blaTEM-15 mcr-1 (18)

Escherichia coli CNR20140385 blaOXA-48 mcr-1 (18)

Escherichia coli WI2 (ST1288) blaOXA-48 blaKPC-28 mcr-1 (19)

Escherichia coli 27841 (ST744) blaCTX-M-55 mcr-3.2 (20)

Escherichia coli 1144230 (ST641) blaCMY-2 mcr-5 (21)

Klebsiella pneumoniae (ST234) blaSHV-27 blaKPC-2 (22)

Citrobacter freundii BM19 blaKPC-2 (17)

Enterobacter cloacae DUB blaFRI-1 (8)

Pseudomonas aeruginosa PA43417 blaOXA-198 (7)

Pseudomonas aeruginosa PA43417

dsbA1 dsbA1 blaOXA-198 This study

271

Table S4. Plasmids used in this study.

Name Description Source

pDM1 pDM1 vector (GenBank MN128719), p15A

ori, Ptac promoter, MCS, TetR Lab stock

pDM1-blaL2-1 blaL2-1 cloned into pDM1, TetR This study

pDM1-blaGES-1 blaGES-1 cloned into pDM1, TetR This study

pDM1-blaGES-2 blaGES-2 cloned into pDM1, TetR This study

pDM1-blaGES-11 blaGES-11 cloned into pDM1, TetR This study

pDM1-blaSHV-27 blaSHV-27 cloned into pDM1, TetR This study

pDM1-blaOXA-4 blaOXA-4 cloned into pDM1, TetR This study

pDM1-blaOXA-10 blaOXA-10 cloned into pDM1, TetR This study

pDM1-blaOXA-198 blaOXA-198 cloned into pDM1, TetR This study

pDM1-blaFRI-1 blaFRI-1 cloned into pDM1, TetR This study

pDM1-blaL1-1 blaL1-1 cloned into pDM1, TetR This study

pDM1-blaKPC-2 blaKPC-2 cloned into pDM1, TetR This study

pDM1-blaKPC-3 blaKPC-3 cloned into pDM1, TetR This study

pDM1-blaSME-1 blaSME-1 cloned into pDM1, TetR This study

pDM1-mcr-1 mcr-1 cloned into pDM1, TetR This study

pDM1-mcr-3 mcr-3 cloned into pDM1, TetR This study

pDM1-mcr-3.2 mcr-3.2 cloned into pDM1, TetR This study

pDM1-mcr-4 mcr-4 cloned into pDM1, TetR This study

pDM1-mcr-5 mcr-5 cloned into pDM1, TetR This study

pDM1-mcr-8 mcr-8 cloned into pDM1, TetR This study

pDM1-blaL2-1-StrepII blaL2-1 encoding L2-1 with a C-terminal

StrepII tag cloned into pDM1, TetR This study

pDM1-blaGES-1-StrepII blaGES-1 encoding GES-1 with a C-terminal

StrepII tag cloned into pDM1, TetR This study

pDM1-StrepII-blaOXA-4 blaOXA-4 encoding OXA-4 with an N-terminal

StrepII tag cloned into pDM1, TetR This study

pDM1-blaOXA-10-StrepII blaOXA-10 encoding OXA-10 with a C-terminal

StrepII tag cloned into pDM1, TetR This study

pDM1-blaOXA-198-StrepII blaOXA-198 encoding OXA-198 with a C-

terminal StrepII tag cloned into pDM1, TetR This study

pDM1-blaFRI-1-StrepII blaFRI-1 encoding FRI-1 with a C-terminal

StrepII tag cloned into pDM1, TetR This study

pDM1-blaL1-1-StrepII blaL1-1 encoding L1-1 with a C-terminal

StrepII tag cloned into pDM1, TetR This study

pDM1-blaKPC-3-StrepII blaKPC-3 encoding KPC-3 with a C-terminal

StrepII tag cloned into pDM1, TetR This study

pDM1-mcr-1-StrepII blaMCR-1 encoding MCR-1 with a C-terminal

StrepII tag cloned into pDM1, TetR This study

pDM1-mcr-3-StrepII blaMCR-3 encoding MCR-3 with a C-terminal

StrepII tag cloned into pDM1, TetR This study

pDM1-mcr-4-StrepII blaMCR-4 encoding MCR-4 with a C-terminal

StrepII tag cloned into pDM1, TetR This study

272

pDM1-mcr-5-StrepII blaMCR-5 encoding MCR-5 with a C-terminal

StrepII tag cloned into pDM1, TetR This study

pDM1-mcr-8-StrepII blaMCR-8 encoding MCR-8 with a C-terminal

StrepII tag cloned into pDM1, TetR This study

pGRG25

Encodes a Tn7 transposon and tnsABCD under

the control of ParaB, thermosensitive pSC101

ori, AmpR

(23)

pGRG25-Ptac::dsbA

Ptac::dsbA fragment cloned within the Tn7 of

pGRG25; when inserted into the chromosome

and the plasmid cured, the strain expresses

DsbA upon IPTG induction, AmpR

This study

pSLTS Thermosensitive pSC101ori, ParaB for λ-Red,

PtetR for I-SceI, AmpR (24)

pUltraGFP-GM

Constitutive sfGFP expression from a strong

Biofab promoter, p15A ori, (template for the

accC cassette), GentR

(25)

pKD4 Conditional oriRγ ori, (template for the aphA

cassette), AmpR (26)

pKNG101

Gene replacement suicide vector, oriR6K,

oriTRK2, sacB, (template for the strAB

cassette), StrR

(27)

pKNG101-dsbA1

PCR fragment containing the regions upstream

and downstream P. aeruginosa dsbA1 cloned

in pKNG101; when inserted into the

chromosome the strain is a merodiploid for

dsbA1 mutant, StrR

This study

pRK600 Helper plasmid, ColE1 ori, mobRK2, traRK2,

CamR (28)

pMA-T mcr-3 GeneArt® cloning vector containing mcr-3,

ColE1 ori, (template for mcr-3), AmpR This study

pMK-T mcr-8 GeneArt® cloning vector containing mcr-8,

ColE1 ori, (template for mcr-8), KanR This study

273

Table S5. Oligonucleotide primers used in this study. The “Brief description” column provides

basic information on the primer design (restriction enzyme used for cloning, encoded protein

or gene replaced by antibiotic resistance cassette, forward or reverse orientation of the primer

(F or R); QC stands for QuickChange primers and SQ stands for sequencing primers).

Numbe

r

Brief description Sequence (5ˊ-3ˊ)

P1 SacI.L2.F ctggagctcctcgcccgtcgccgatt

P2 XmaI.L2.R ctgcccgggtcatccgatcaaccggtcggca

P3 SacI.GES.F ctggagctccgcttcattcacgcac

P4 XmaI.GES.R ctgcccgggctatttgtccgtgctcaggatg

P5 SacI.SHV.F ctggagctccgttatattcgcctgtg

P6 XmaI.SHV.R ctgcccgggttagcgttgccagtgctcga

P7 SacI.OXA-4.F ctggagctcaaaaacacaatacatataacttcgc

P8 KpnI.OXA-4.R cagggtaccttataaatttagtgtgtttagaatggtg

P9 SacI.OXA-10.F ctggagctcaaaacatttgccgcatatgtaattatcgc

P10 KpnI.OXA-10.R cagggtaccttagccaccaatgatgccctc

P11 NdeI.OXA-198.F actgcatatgcataaacacatgagtaagctcttc

P12 KpnI.OXA-198.R ctgggtaccttattcgatgatcccctttgctt

P13 SacI.FRI-1.F ctggagctctttttttttaaaaaaggtgcaagtac

P14 XmaI.FRI-1.R ctgcccgggttatttataacttccataaactgcctttatagc

P15 SacI.L1.F ctggagctccgttctaccctgctcgc

P16 XhoI.L1.R actgagctctcagcgggccccggccgt

P17 SacI.KPC.F ctggagctctcactgtatcgccgtc

P18 KpnI.KPC.R ctgccatggttactgcccgttgacgccca

P19 SacI.SME-1.F ctggagctctcaaacaaagtaaattttaaaacgg

P20 XmaI.SME-1.R ctgcccgggttaatcaattgcctgaattgcaatacg

P21 SacI.MCR-1.F ctggagctcatgcagcatacttctgtgtggtac

P22 XmaI.MCR-1.R ctgcccgggtcagcggatgaatgcggtgc

P23 NdeI.MCR-3.F ctgatacatatgccttcccttataaaaataaaaattgttccg

P24 XmaI.MCR-3.R cagcccgggttattgaacattacgacattgactgaaaatatctag

P25 SacI.MCR-4.F ctggagctccgtgctgacgagatttaaaaccc

P26 XmaI.MCR-4.R ctgcccgggttaaccgcggcagcgggcaaaaatatc

P27 SacI.MCR-5.F ctggagctccggttgtctgcatttatcac

P28 XmaI.MCR-5.R ctgcccgggtcattgtggttgtccttttctg

P29 SacI.MCR-8.F ctggagctcttcaagtatcttttatctttcaaact aacc

P30 XmaI.MCR-8.R ctgcccgggctaaccattcccatctgttttctc

P31 QC.GES5-GES1.F aaagagccggagatgggcgacaacacacctg

P32 QC.GES5-GES1.R caggtgtgttgtcgcccatctccggctcttt

P33 QC.KPC2-KPC3.F ctaacaaggatgacaagtacagcgaggccgtcatc

P34 QC.KPC2-KPC3.R gatgacggcctcgctgtacttgtcatccttgttag

P35 QC.MCR3-3.2.F caacgcctttctctttgataaatccaggtgacatcc

P36 QC.MCR3-3.2.R ggatgtcacctggatttatcaaagagaaaggcgttg

P37 XmaI.StrepII.L2.R ctgcccgggttatttttcaaattgcggatggctccaagcgctccctccg

atcaaccggtcggca

274

P38 XmaI.StrepII.GES.R ctgcccgggctatttttcaaattgcggatggctccaagcgctccctttg

tccgtgctcaggatgag

P39 OXA-4.body.F tcaacagatatctctactgttgca

P40 OXA-4.StrepII.R tgcaacagtagagatatctgttgatttttcaaattgcggatggctccaa

gcgctccctgcactggcgctgctgta

P41 KpnI.StrepII.OXA-10.R cagggtaccttatttttcaaattgcggatggctccaagcgctcccgcc

accaatgatgccctcacttg

P42 KpnI.StrepII.OXA-198.R ctgggtaccttatttttcaaattgcggatggctccaagcgctcccttcga

tgatcccctttgcttg

P43 XmaI.StrepII.FRI-1.R ctgcccgggttatttttcaaattgcggatggctccaagcgctccctttat

aacttccataaactgcctttatagc

P44 KpnI.StrepII.L1.R gggggtacctcatttttcaaattgcggatggctccaagcgctcccgcg

ggccccggccgtttccttggccaactgc

P45 KpnI.StrepII.KPC.R ctgccatggttatttttcaaattgcggatggctccaagcgctcccctgc

ccgttgacgcccaatc

P46 XmaI.StrepII.MCR-1.R cagcccgggttatttttcaaattgcggatggctccaagcgctcccgcg

gatgaatgcggtgcggt

P47 XmaI.StrepII.MCR-3.R cagcccgggttatttttcaaattgcggatggctccaagcgctcccttga

acattacgacattgactgaaaatatctag

P48 XmaI.StrepII.MCR-4.R ctgcccgggctatttttcaaattgcggatggctccaagcgctcccacc

gcggcagcgggcaaaaatatc

P49 XmaI.StrepII.MCR-5.R ctgcccgggctatttttcaaattgcggatggctccaagcgctcccttgt

ggttgtccttttctgca

P50 XmaI.StrepII.MCR-8.R ctgcccgggctatttttcaaattgcggatggctccaagcgctcccacc

attcccatctgttttctctcttac

P51 NotI.Ptac.EcDsbA.F

ctggcggccgctgacaattaatcatcggctcgtataatgtgtggaatt

gtgactagtcgaggtccaggacctcggatcgctaagataggatgatt

gtatgaaaaagatttggctggc

P52 XhoI.EcDsbA.R ctgctcgagttattttttctcggacagatatttc

P53 EcdsbA::aphA.F atgaaaaagatttggctggcgctggctggtttagttttagcgtttagcg

cgtgtaggctggagctgcttc

P54 EcdsbA::aphA.R ttattttttctcggacagatatttcactgtatcagcatactgctgaacaag

ggaattagccatggtccat

P55 EcacrA::aphA.F atgaacaaaaacagagggtttacgcctctggcggtcgttctggtgtag

gctggagctgcttc

P56 EcacrA::aphA.R ttaagacttggactgttcaggctgagcaccgcttgcggcttggggaat

tagccatggtccat

P57 EctolC::aphA.F ttttacagtttgatcgcgctaaatactgcttcaccacaaggaatgcaag

tgtaggctggagctgcttc

P58 EctolC::aphA.R tcgtcgtcatcagttacggaaagggttatgatgggaattagccatggt

cc

P59 EcdegP::strAB.F atgaaaaaaaccacattagcactgagtgcactggctctgagtttaggt

ttggaactgcacattcgggatatttctc

P60 EcdegP::strAB.R ttactgcattaacaggtagatggtgctgtcgccgcgctgaatgttgagt

gccaggccggatctagatatctagtatga

275

P61 EcmarR::accC.F

atggttaatcagaagaaagatcgcctgcttaacgagtatctgtctccg

ctggtgaagttcctatactttctagagaataggaacttcaagatcccct

g

P62 EcmarR::accC.R

ttacggcaggactttcttaagcaaatactcaagtgttgccacttcgtcc

gcgaagttcctattctctagaaagtataggaacttcacttactcaatgg

aattctagatcg

P63 SQ.dsbA1.Paeruginosa.F tacctgctcaagcagatgcatg

P64 SQ.dsbA1.Paeruginosa.R ggtgttcatgtcgcccatca

P65 XbaI.dsbA1.F ggttcctctagagcctacttcgccagccagaa

P66 dsbA1.body.R ctacttcttgttacgcatcgttcactc

P67 dsbA1.body.F atgcgtaacaagaagtaggcaaggtga

P68 BamHI.dsbA1.R aattaaggatcctcatcactaccaccagcgcg

276

Table S6. Sources of genomic DNA used for amplification of β-lactamase and MCR genes in

this study.

Strain Gene(s) Source

Stenotrophomonas maltophilia ATCC 13637 blaL2-1 blaL1-1 ATCC

Pseudomonas aeruginosa GW-1 blaGES-2 (29)

Enterobacter cloacae CHE-2 blaGES-5 (30)

Acinetobacter baumannii K45 blaGES-11 (31)

Klebsiella pneumoniae ST234 blaSHV-27 blaKPC-2 (22)

Pseudomonas aeruginosa SOF1 blaOXA-4 (32)

Pseudomonas aeruginosa PU21 blaOXA-10 (33)

Pseudomonas aeruginosa PA41437 blaOXA-198 (7)

Enterobacter cloacae DUB blaFRI-1 (8)

Serratia marcescens blaSME-1 (22)

Escherichia coli CNR1790 mcr-1 (18)

Shewanella bicestrii JAB-1 mcr-4 (34)

Escherichia coli 1144230 mcr-5 (21)

277

LEGENDS FOR SUPPLEMENTARY DATA FILES

File S1. Analysis of the cysteine content and phylogeny of all identified β-lactamases.

3,947 unique β-lactamase protein sequences were clustered with a 90% identity threshold and

the centroid of each cluster was used as a phylogenetic cluster identified for each sequence

(“Phylogenetic cluster” column). All sequences were searched for the presence of cysteine

residues (“Total number of cysteines” and “Positions of all cysteines” columns). Proteins with

two or more cysteines after the first 30 amino acids of their primary sequence (cells shaded in

grey in the “Number of cysteines after position 30” column) are potential substrates of the DSB

system for organisms where oxidative protein folding is carried out by DsbA and provided that

translocation of the β-lactamase outside the cytoplasm is performed by the Sec system. The

first 30 amino acids of each sequence were excluded to avoid considering cysteines that are

part of the signal sequence mediating the translocation of these enzymes outside the cytoplasm.

Cells shaded in grey in the “Reported in pathogens” column mark β-lactamases that are found

in pathogens or organisms capable of causing opportunistic infections. The Ambler class of

each enzyme is indicated in the “Ambler class column” and each class (A, B1, B2, B3, C and

D) is highlighted with a different color.

278

SUPPLEMENTARY REFERENCES

1. Zhang H, Srinivas S, Xu Y, Wei W, Feng Y. Genetic and biochemical mchanisms for

bacterial lipid A modifiers associated with polymyxin resistance. Trends Biochem Sci.

2019;44(11):973-88.

2. Corkill JE, Cuevas LE, Gurgel RQ, Greensill J, Hart CA. SHV-27, a novel cefotaxime-

hydrolysing β-lactamase, identified in Klebsiella pneumoniae isolates from a Brazilian

hospital. J Antimicrob Chemother. 2001;47(4):463-5.

3. Schultz SC, Dalbadie-McFarland G, Neitzel JJ, Richards JH. Stability of wild-type and

mutant RTEM-1 β-lactamases: effect of the disulfide bond. Proteins. 1987;2(4):290-7.

4. Silvestro L, Weiser JN, Axelsen PH. Antibacterial and antimembrane activities of

cecropin A in Escherichia coli. Antimicrob Agents Chemother. 2000;44(3):602-7.

5. Dailey FE, Berg HC. Mutants in disulfide bond formation that disrupt flagellar

assembly in Escherichia coli. Proc Natl Acad Sci U S A. 1993;90(3):1043-7.

6. Aubert D, Poirel L, Ali AB, Goldstein FW, Nordmann P. OXA-35 is an OXA-10-

related β-lactamase from Pseudomonas aeruginosa. J Antimicrob Chemother. 2001;48(5):717-

21.

7. El Garch F, Bogaerts P, Bebrone C, Galleni M, Glupczynski Y. OXA-198, an acquired

carbapenem-hydrolyzing class D β-lactamase from Pseudomonas aeruginosa. Antimicrob

Agents Chemother. 2011;55(10):4828-33.

8. Dortet L, Poirel L, Abbas S, Oueslati S, Nordmann P. Genetic and biochemical

characterization of FRI-1, a carbapenem-hydrolyzing class A β-lactamase from Enterobacter

cloacae. Antimicrob Agents Chemother. 2015;59(12):7420-5.

9. Palzkill T. Metallo-β-lactamase structure and function. Ann N Y Acad Sci.

2013;1277:91-104.

10. Papp-Wallace KM, Bethel CR, Distler AM, Kasuboski C, Taracila M, Bonomo RA.

Inhibitor resistance in the KPC-2 β-lactamase, a preeminent property of this class A β-

lactamase. Antimicrob Agents Chemother. 2010;54(2):890-7.

11. Hanahan D. In: Glover DM, editor. DNA Cloning: A Practical Approach. 1: IRL Press,

McLean, Virginia; 1985. p. 109.

12. Herrero M, de Lorenzo V, Timmis KN. Transposon vectors containing non-antibiotic

resistance selection markers for cloning and stable chromosomal insertion of foreign genes in

Gram-negative bacteria. J Bacteriol. 1990;172(11):6557-67.

13. Boyer HW, Roulland-Dussoix D. A complementation analysis of the restriction and

modification of DNA in Escherichia coli. J Mol Biol. 1969;41(3):459-72.

14. Casadaban MJ, Cohen SN. Analysis of gene control signals by DNA fusion and cloning

in Escherichia coli. J Mol Biol. 1980;138(2):179-207.

279

15. Kadokura H, Tian H, Zander T, Bardwell JC, Beckwith J. Snapshots of DsbA in action:

detection of proteins in the process of oxidative folding. Science. 2004;303(5657):534-7.

16. Blattner FR, Plunkett G, 3rd, Bloch CA, Perna NT, Burland V, Riley M, et al. The

complete genome sequence of Escherichia coli K-12. Science. 1997;277(5331):1453-62.

17. Dortet L, Brechard L, Poirel L, Nordmann P. Rapid detection of carbapenemase-

producing Enterobacteriaceae from blood cultures. Clin Microbiol Infect. 2014;20(4):340-4.

18. Dortet L, Bonnin RA, Pennisi I, Gauthier L, Jousset AB, Dabos L, et al. Rapid detection

and discrimination of chromosome- and MCR-plasmid-mediated resistance to polymyxins by

MALDI-TOF MS in Escherichia coli: the MALDIxin test. J Antimicrob Chemother.

2018;73(12):3359-67.

19. Beyrouthy R, Robin F, Lessene A, Lacombat I, Dortet L, Naas T, et al. MCR-1 and

OXA-48 in vivo acquisition in KPC-producing Escherichia coli after colistin treatment.

Antimicrob Agents Chemother. 2017;61(8).

20. Haenni M, Beyrouthy R, Lupo A, Chatre P, Madec JY, Bonnet R. Epidemic spread of

Escherichia coli ST744 isolates carrying mcr-3 and blaCTX-M-55 in cattle in France. J Antimicrob

Chemother. 2018;73(2):533-6.

21. Wise MG, Estabrook MA, Sahm DF, Stone GG, Kazmierczak KM. Prevalence of mcr-

type genes among colistin-resistant Enterobacteriaceae collected in 2014-2016 as part of the

INFORM global surveillance program. PloS One. 2018;13(4):e0195281.

22. Nordmann P, Poirel L, Dortet L. Rapid detection of carbapenemase-producing

Enterobacteriaceae. Emerg Infect Dis. 2012;18(9):1503-7.

23. McKenzie GJ, Craig NL. Fast, easy and efficient: site-specific insertion of transgenes

into enterobacterial chromosomes using Tn7 without need for selection of the insertion event.

BMC Microbiol. 2006;6:39.

24. Kim J, Webb AM, Kershner JP, Blaskowski S, Copley SD. A versatile and highly

efficient method for scarless genome editing in Escherichia coli and Salmonella enterica. BMC

Biotechnol. 2014;14:84.

25. Mavridou DA, Gonzalez D, Clements A, Foster KR. The pUltra plasmid series: a robust

and flexible tool for fluorescent labeling of Enterobacteria. Plasmid. 2016;87-88:65-71.

26. Datsenko KA, Wanner BL. One-step inactivation of chromosomal genes in Escherichia

coli K-12 using PCR products. Proc Natl Acad Sci U S A. 2000;97(12):6640-5.

27. Kaniga K, Delor I, Cornelis GR. A wide-host-range suicide vector for improving

reverse genetics in Gram-negative bacteria: inactivation of the blaA gene of Yersinia

enterocolitica. Gene. 1991;109(1):137-41.

28. Kessler B, Delorenzo V, Timmis KN. A general system to integrate lacZ fusions into

the chromosomes of Gram-negative eubacteria: regulation of the Pm Promoter of the TOL

plasmid studied with all controlling elements in monocopy. Mol Gen Genet. 1992;233(1-

2):293-301.

280

29. Poirel L, Weldhagen GF, Naas T, De Champs C, Dove MG, Nordmann P. GES-2, a

class A β-lactamase from Pseudomonas aeruginosa with increased hydrolysis of imipenem.

Antimicrob Agents Chemother. 2001;45(9):2598-603.

30. Poirel L, Carrer A, Pitout JD, Nordmann P. Integron mobilization unit as a source of

mobility of antibiotic resistance genes. Antimicrob Agents Chemother. 2009;53(6):2492-8.

31. Bonnin RA, Rotimi VO, Al Hubail M, Gasiorowski E, Al Sweih N, Nordmann P, et al.

Wide dissemination of GES-type carbapenemases in Acinetobacter baumannii isolates in

Kuwait. Antimicrob Agents Chemother. 2013;57(1):183-8.

32. Aubert D, Poirel L, Chevalier J, Leotard S, Pages JM, Nordmann P. Oxacillinase-

mediated resistance to cefepime and susceptibility to ceftazidime in Pseudomonas aeruginosa.

Antimicrob Agents Chemother. 2001;45(6):1615-20.

33. Dortet L, Poirel L, Nordmann P. Rapid detection of carbapenemase-producing

Pseudomonas spp. J Clin Microbiol. 2012;50(11):3773-6.

34. Jousset AB, Dabos L, Bonnin RA, Girlich D, Potron A, Cabanel N, et al. CTX-M-15-

producing Shewanella species clinical isolate expressing OXA-535, a chromosome-encoded

OXA-48 variant, putative progenitor of the plasmid-encoded OXA-436. Antimicrob Agents

Chemother. 2018;62(1).