University of Dundee

DOCTOR OF PHILOSOPHY

Investigations into the roles of bacterial TorD family chaperone proteins

Connelly, Katherine

Award date: 2014

Link to publication

General rights Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

• Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal Take down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Download date: 04. Oct. 2021 Investigations into the roles of bacterial TorD

family chaperone proteins

by

Katherine Connelly

September 2014

Thesis submitted to the University of Dundee in partial fulfilment of the

requirements for the degree of Doctor of Philosophy

ii

Copyright © Katherine Connelly, September 2014.

All rights reserved. This copy of the thesis has been supplied on condition that anyone, who consults it, is understood to recognise that its copyright rests with the author and that no quotation from the thesis, nor any information derived therefrom, may be published without the author‘s prior, written consent. iii

Declaration

I declare that I am the author of this thesis and that, unless otherwise stated, all references cited have been consulted; that the work of which this thesis is a record of has been performed by me, and that it has not been previously accepted for a higher degree: where the thesis is based upon joint research, the nature and extent of my individual contribution is defined.

Katherine Connelly

iv

Abstract

The twin-arginine translocase (Tat) is a highly specialised protein transport system, present in prokaryotes and plant chloroplasts. This translocase functions to transport proteins across the cytoplasmic membrane in a fully folded state. Substrates of the Tat system are targeted to the system by N-terminal signal peptides, which bear the consensus motif (S/T)RRxFLK. These proteins often require redox centres for function that must be incorporated prior to transport, with many utilising the molybdenum cofactor Mo-bis-MGD for enzymatic activity during anaerobic respiration. Cofactor loading and protein maturation of these substrates is often coordinated by a ‘Tat proofreading’ process, involving the binding of cytoplasmic chaperone proteins to the substrate signal peptide to prevent premature translocation before maturation is complete. A large group of these proteins form the TorD family of chaperone proteins, of which two examples are DmsD and TorD. DmsD is known to aid maturation of three Tat substrates in Salmonella enterica subsp. enterica serovar Typhimurium; YnfE - the catalytic subunit of selenate reductase, YnfF, and DmsA - the catalytic subunit of dimethyl sulphoxide (DMSO) reductase. TorD assists in the assembly of the catalytic subunit of trimethylamine N- oxide (TMAO) reductase, TorA.

Work presented in this thesis has demonstrated that the interaction between S. Typhimurium DmsD and YnfE requires a hydrophobic stretch of residues on the N-terminal signal peptide of YnfE, the sequence of which is highly conserved amongst signal peptide sequences of all three DmsD target proteins. Genetic and biochemical analysis also revealed residues of importance on the DmsD protein, with a proposed binding mechanism being discussed involving a hydrophobic cleft on the protein surface. The possibility that DmsD is involved in activities other than signal peptide binding was also touched upon. TorA/TorD binding interactions of Escherichia coli were also investigated, and again highlighted the prospect of dual functionality of these chaperone proteins, with TorD amino acid residues being implicated in TorA signal peptide interactions and TMAO reductase activity. High resolution microscopy was employed to enable imaging of TorD within the cellular environment, and super resolution microscopy was utilized to elucidate the interplay between Tat substrates and the Tat translocase. Finally, broad metabolic phenotype screening technology was used to gain an understanding of the broader function of dmsD and other genes in bacterial cell metabolism.

v

Table of Contents

Declaration ...... iii

Abstract ...... iv

Table of Contents ...... v

Table of Figures ...... xii

Table of Tables ...... xvi

Conference poster presentations ...... xvii

List of abbreviations ...... xvii

Table of Figures ...... xi

Acknowledgments ...... xxi

1 Introduction ...... 1

1.1 Protein translocation in Gram-negative organisms ...... 2

1.1.1 Escherichia coli and Salmonella enterica subsp. enterica serovar Typhimurium ...... 2

1.1.2 Structure of the Gram-negative bacterial cell membranes ...... 2

1.1.3 Protein translocation in Gram-negative bacteria ...... 4

1.1.4 The general secretory (Sec) pathway ...... 5

1.1.4.1 Overview of Sec translocation ...... 5

1.1.4.2 Components of the Sec translocon ...... 7

1.1.5 Co- and post-translational protein targeting ...... 8

1.1.6 YidC ...... 11

1.1.7 The twin-arginine translocation (Tat) pathway ...... 12 vi

1.1.7.1 Overview of Tat translocation ...... 12

1.1.7.2 Components of the Tat translocon ...... 13

1.1.7.3 TatA family proteins ...... 14

1.1.7.4 TatC proteins ...... 17

1.1.7.5 Formation of the Tat translocase ...... 19

1.1.7.6 The Tat signal peptide ...... 22

1.1.7.7 Substrates of Tat ...... 24

1.1.7.8 The translocation cycle ...... 29

1.1.7.9 Tat quality control ...... 33

1.1.7.10 Tat proofreading ...... 36

1.1.7.11 The TorD family of chaperone proteins ...... 37

1.1.7.12 The role of TorD chaperones ...... 39

1.1.7.13 TorD chaperone structures ...... 42

1.1.7.14 Other Tat proofreading chaperones ...... 43

1.1.8 General chaperones involved with Tat substrates ...... 45

1.2 Respiratory capability of E. coli and Salmonella ...... 46

1.2.1 Quinones and respiration ...... 48

1.2.2 The mononuclear molybdenum enzymes ...... 52

1.2.3 N and S-oxide reductases ...... 55

1.2.4 Nitrate reductases ...... 59

1.2.5 Hydrogenases ...... 62

1.3 Biosynthesis of molybdoenzymes ...... 64

1.3.1 Biosynthesis of the molybdenum cofactor ...... 64 vii

1.3.2 Biosynthesis of Fe-S clusters ...... 68

1.4 Aims...... 71

2 Tat-dependent selenate metabolism in S. Typhimurium: the role of the twin-arginine signal peptide ...... 72

2.1 Introduction ...... 73

2.2 Aims...... 74

2.3 Results ...... 76

2.3.1 Identification of the DmsD binding site on the YnfE signal peptide ...... 76

2.3.2 Signal peptide amino acid substitutions affect selenate reduction in vivo ...... 79

2.3.3 Tools for probing YnfE assembly: antibodies to YnfE ...... 83

2.4 Discussion ...... 88

2.4.1 A conserved DmsD binding epitope ...... 88

2.4.2 The structure of the YnfE signal peptide...... 95

3 Tat-dependent selenate metabolism in S. Typhimurium: the role of DmsD ...... 101

3.1 Introduction ...... 102

3.2 Aims...... 104

3.3 Results ...... 105

3.3.1 Genetic analysis of the YnfE – DmsD interaction...... 105

3.3.1.1 Isolating DmsD variants affecting selenate reductase activity ...... 105

3.3.1.2 Analysis of DmsD variants for effects on selenate reductase activity ...... 109

3.3.1.3 Analysis of DmsD variants for interactions with spYnfE ...... 110

3.3.1.4 Alternative roles for DmsD in YnfE maturation ...... 111

3.3.2 Biochemical investigations of the spYnfE interaction with DmsD ...... 113 viii

3.3.2.1 Overproduction and purification of DmsD ...... 113

3.3.2.2 Crosslinking analysis of purified DmsD and synthetic spYnfE ...... 115

3.3.2.3 Investigating the location of the crosslinks between DmsD and ssYnfE ...... 118

3.3.2.4 DmsD can be co-purified in solution with synthetic spYnfE ...... 122

3.3.2.5 Does nucleotide binding affect peptide crosslinking? ...... 126

3.4 Discussion ...... 128

3.4.1 New insight into the Tat proofreading function of DmsD using a genetic screen .... 128

3.4.2 Chemical crosslinking as a tool to study the signal peptide binding site on DmsD .. 132

3.4.3 Evidence for DmsD biochemical activities other than signal peptide binding ...... 137

4 Understanding the biosynthesis of E. coli TorA and its relationship with the Tat pathway

141

4.1 Introduction ...... 142

4.2 Aims...... 146

4.3 Results ...... 147

4.3.1 Molecular dissection of TorD activity ...... 147

4.3.1.1 Assembly of TMAO reductase by TorD ...... 149

4.3.1.2 Tat proofreading by TorD...... 150

4.3.2 Imaging TorD-TorD interactions in live cells ...... 152

4.3.2.1 Confocal microscopy of mCer-TorD and YFP-TorD ...... 156

4.3.3 Super-resolution imaging of the E. coli cell ...... 158

4.3.3.1 Super-resolution imaging of chromosomally tagged TatA-YFP...... 158

4.3.3.2 Effects of substrate overexpression on visible TatA-YFP spots ...... 159

4.3.3.3 Imaging TatA using fluorescent antibodies ...... 161 ix

4.4 Discussion ...... 163

4.4.1 TorD Pro-122 is important for TorD function ...... 163

4.4.2 Imaging TorD and TatA in cells ...... 164

5 Discovery of novel Tat-related phenotypes using phenotype microarrays ...... 169

5.1 Introduction ...... 170

5.2 Aim ...... 171

5.3 Results ...... 172

5.3.1 Metabolic abilities of S. Typhimurium strains were investigated using high- throughput phenotype microarray analysis...... 172

5.3.1.1 Aerobic metabolism with carbon sources ...... 173

5.3.1.2 Metabolism of other substrates ...... 178

5.4 Discussion ...... 183

5.4.1 A requirement of a functional Tat translocase for maltose metabolism? ...... 183

5.4.2 Further involvement of the Tat pathway in cell wall synthesis and remodelling ..... 185

6 Final Conclusions and Future Perspectives ...... 187

6.1 Final conclusions ...... 188

6.2 Future Perspectives...... 190

6.2.1 How are these Tat substrate enzymes and their chaperones involved in aiding cell survival, for example during host infection? ...... 190

6.2.2 Could Tat substrates and chaperones be utilised to the advantage of human health?

196

7 Materials and methods ...... 199

7.1 Bacterial strains ...... 200 x

7.2 Media, additives and growth conditions ...... 201

7.3 Buffers and solutions ...... 202

7.4 Genetic manipulations ...... 204

7.4.1 Plasmid DNA preparation ...... 204

7.4.2 Amplification of DNA by Polymerase Chain Reaction (PCR) ...... 212

7.4.3 Site-directed mutagenesis by QuikchangeTM PCR ...... 212

7.4.4 Agarose gel electrophoresis ...... 212

7.4.5 DNA digestion and preparation for cloning ...... 213

7.4.6 DNA ligation ...... 213

7.4.7 DNA sequencing ...... 213

7.4.8 Preparation of competent cells and transformation with plasmid DNA ...... 214

7.4.9 In-frame chromosomal integration using pMAK705 vector ...... 215

7.5 Protein methods ...... 216

7.5.1 Overproduction of proteins ...... 216

7.5.2 Purification of hexa-histidine tagged proteins using immobilised metal ion affinity chromatography ...... 216

7.5.3 Purification of tobacco etch virus protease ...... 217

7.5.4 TEV cleavage of hexa-histidine tags ...... 218

7.5.5 Purification of proteins using size-exclusion chromatography ...... 218

7.5.6 Protein concentration determination ...... 219

7.5.7 Purified protein crystallisation trials ...... 219

7.5.8 Polyclonal antibody production ...... 219

7.5.9 Affinity purification of polyclonal antiserum ...... 220 xi

7.5.10 SDS-PAGE ...... 220

7.5.11 Semi-dry western immunoblot ...... 221

7.5.12 Chemical crosslinking of proteins ...... 222

7.6 Protein assays ...... 224

7.6.1 BV-linked oxidoreductase assays ...... 224

7.6.1.1 TMAO reductase activity assay ...... 224

7.6.1.2 Hydrogenase-2 (Tat proofreading) activity assay ...... 225

7.6.2 Bacterial two-hybrid system and β-galactosidase assay ...... 226

7.7 Cellular fractionation ...... 227

7.7.1 Preparation of cytoplasmic/membrane/periplasmic fractions ...... 227

7.8 Growth curves ...... 228

7.8.1 Set-up of 96-well plates ...... 228

7.9 Microscopy methods ...... 229

7.9.1 Confocal microscopy ...... 229

7.9.1.1 Slide preparation ...... 229

7.9.1.2 Imaging ...... 229

7.9.2 OMX microscopy ...... 229

7.9.2.1 Slide preparation ...... 229

7.9.2.2 Imaging ...... 230

8 Appendix ...... 231

8.1 Biolog plates ...... 232

9 Bibliography ...... 236 xii

Table of Figures

Figure 1.1: The Gram-negative cell envelope...... 3

Figure 1.2: General structural features of Sec signal peptides...... 7

Figure 1.3: 3D crystal structure of the SecYEG complex from Methanococcus jonnaschii...... 8

Figure 1.4: Post- and co-translational export through the Sec pathway...... 9

Figure 1.5: Organisation of tat genes in E. coli...... 13

Figure 1.6: Solution NMR structures of monomeric TatA...... 15

Figure 1.7: Solution NMR structure of E. coli TatB...... 17

Figure 1.8: Three dimensional crystal structure of TatC from A. aeolicus...... 18

Figure 1.9: Modelling the TatA oligomeric pore...... 21

Figure 1.10: General features of the Tat signal peptide...... 23

Figure 1.11: Proposed interaction sites on TatC...... 31

Figure 1.12: The cycle of protein translocation through Tat...... 30

Figure 1.13: The control mechanisms of Tat in place to prevent premature translocation...... 35

Figure 1.14: The TorD family of cytoplasmic chaperone proteins...... 38

Figure 1.15: 3D crystal structures of TorD family proteins show conserved α-helical structure.

...... 43

Figure 1.16: Solution NMR structure of E. coli NapD...... 45

Figure 1.17: Respiratory pathways in E. coli...... 48

Figure 1.18: Active site structures of molybdenum-containing enzymes...... 54

Figure 1.19: The structure of the molybdopterin component of the molybdenum cofactor. ... 65

Figure 1.20: Biosynthesis of the molybdenum cofactor...... 66

Figure 2.1: The selenate reductase from S. Typhimurium...... 73

Figure 2.2: The principles of the bacterial two-hybrid system...... 77

Figure 2.3: Tat signal peptide of S. Typhimurium YnfE...... 78

Figure 2.4: Interaction study of cytoplasmic chaperone protein DmsD with spYnfE and spYnfE amino acid variants, measured in vivo using a bacterial two-hybrid system...... 79 xiii

Figure 2.5: pMAK protocol for targeted chromosomal integration of spYnfE amino acid substitutions...... 80

Figure 2.6: Selenate reductase ability of chromosomal spYnfE variants...... 82

Figure 2.7: IMAC purification of S. Typhimurium YnfE...... 83

Figure 2.8: Tryptic mass spectrometry analysis of purified YnfE samples shows two forms of

YnfE...... 84

Figure 2.9: Testing the quality of antibody raised against purified YnfE...... 85

Figure 2.10: YnfE antibody test with S. Typhimurium strains LT2, NTCT12023 and SL1344...... 87

Figure 2.11: Multiple sequence alignment of Tat signal sequences from S. Typhimurium (stm) and E. coli (ec) DmsA, YnfF and YnfE...... 89

Figure 2.12: Predicted cytoplasmic chaperone binding epitopes of Tat substrate signal peptides...... 92

Figure 2.13: Alignment of a number of E. coli and S. Typhimurium bona fide and remnant Tat signal peptides...... 94

Figure 2.14: Modelling of spYnfE...... 97

Figure 3.1: Three dimensional crystal structure of S. Typhimurium DmsD...... 102

Figure 3.2: ‘Hot pocket’ for signal peptide interactions shown on 3D crystal structure of E. coli

DmsD ...... 103

Figure 3.3: An spYnfE A28Q phenotype can be rescued by excess DmsD...... 106

Figure 3.4: Identification of amino acid substitutions in DmsD that disrupt selenate reductase activity...... 108

Figure 3.5: DmsD residues V16, W91 and G100 are required for selenate reductase assembly and activity...... 110

Figure 3.6: DmsD V16Q variant is unable to recognise the YnfE signal peptide...... 111

Figure 3.7: DmsD interacts preferentially with the twin-arginine signal peptide...... 112

Figure 3.8: Purification of recombinant DmsD...... 114

Figure 3.9: DSS or formaldehyde, but not EDC, can crosslink a peptide ligand to DmsD ...... 116 xiv

Figure 3.10: SDS-PAGE and Western blot analysis of DmsD crosslinked by DSS with synthetic peptide 1 and 2...... 117

Figure 3.11: Ability of synthetic peptide 3 to form crosslinks with DmsD...... 118

Figure 3.12: Evidence that peptide ligands are crosslinking to the N-terminus of DmsD...... 120

Figure 3.13: Analysis of purified DmsDhis interactions with synthetic ssYnfE peptides...... 121

Figure 3.14: Elution profile for SEC of DmsD purified protein, with or without synthetic peptide

4...... 123

Figure 3.15: ESI mass spectrometry analysis of DmsD subjected to size exclusion chromatography (SEC) with and without synthetic peptide 4...... 125

Figure 3.16: Analysis of crosslinking interactions of DmsD and peptide 2 when pre-incubated with various nucleotides...... 127

Figure 3.17: Potential amino acid residues on DmsD involved in signal peptide binding...... 129

Figure 3.18: Multiple alignment of DmsD/TorD-like family chaperones showing possible signal peptide interacting residues...... 130

Figure 3.19: Locations of the two lysines of DmsD...... 134

Figure 3.20: Residues of DmsDSt equivalent to those identified in E. coli TorD crosslinking experiments...... 137

Figure 3.21: W91 and G100 residues are conserved amongst DmsD proteins...... 138

Figure 4.1 Organisation of TMAO reductase in E. coli...... 143

Figure 4.2: The TorAD binding interaction...... 145

Figure 4.3: Production of TorD variants...... 148

Figure 4.4: TMAO reductase activity directed by TorD variants...... 150

Figure 4.5: Tat proofreading ability of TorD variants...... 151

Figure 4.6: Diagram of fluorescent TorD fusions...... 153

Figure 4.7: TMAO reductase activity in the presence of fluorescently-tagged TorD...... 154

Figure 4.8: Cellular localisation of TorD-mCer and TorD-YFP fusions ...... 155

Figure 4.9: Co-localisation of mCer-TorD and YFP-TorD fusions within E. coli...... 157 xv

Figure 4.10: Fluorescence imaging of TatA in E. coli...... 159

Figure 4.11: The effects of Tat substrate overexpression on TatA-YFP foci...... 160

Figure 4.12: Imaging TatA using fluorescent antibodies...... 162

Figure 5.1: Genetic organisation of the tat genes of S. Typhimurium LT2...... 170

Figure 5.2: Phenotype microarray data for carbon sources...... 174

Figure 5.3: The respiratory ability of S. Typhimurium strains in minimal media with maltose as the single carbon source...... 175

Figure 5.4: Growth of a ΔtatABC mutant is defective in M9 media containing maltose...... 176

Figure 5.5: Location of maltose binding protein in S. Typhimurium strains...... 177

Figure 5.6: Phenotypic microarray data for nitrogen sources...... 179

Figure 5.6: Phenotypic microarray data for nitrogen sources (contd)...... 180

Figure 5.7: Phenotype microarray data for phosphorus and sulphur sources...... 182

Figure 5.8: Phenotype microarray data for respiration rates of S. Typhimurium strains on nutrient sources...... 183

Figure 6.1: TorD family chaperone proteins did not display interactions with spTtrA or spTtrB.

...... 192

Figure 6.2: Specific interactions between spTtrB and TatC were analysed...... 193

Figure 6.3: Salmonella is able to utilise host inflammatory response during infection for competitive advantage in the intestine...... 195

Figure 7.1: Chemical structures of crosslinkers DSS, EDC and formaldehyde...... 223

xvi

Table of Tables

Table 1.1: The thirty seven known and predicted Tat substrates of E. coli K-12...... 25

Table 3.1 Details of synthetic peptides used in this study...... 115

Table 7.1: Bacterial strains used in this study...... 200

Table 7.2: Growth media used in this study...... 201

Table 7.3: Additives used in this study, with their respective stock and working concentrations.

...... 202

Table 7.4: Antibiotics used in this study, with their stock and working concentrations...... 202

Table 7.5: General buffers and solutions used in this study...... 202

Table 7.6: Buffers used during the purification of proteins by FPLC...... 203

Table 7.7: Plasmids used throughout this study...... 205

Table 7.8: Oligonucleotides used throughout this study...... 207

Table 7.9: Antibodies used in this work...... 222

xvii

Conference poster presentations

• Gordon Research Conference (Mount Snow Resort, West Dover, Vermont, USA). Bacterial Cell Surfaces: Building, Splitting, and Traversing the Cell Surface, June 2014. Poster presented entitled ‘Protein trafficking on the bacterial twin-arginine translocation pathway’.

• Joint FEBS-EMBO Advanced Lecture Course. Biomembranes: Molecular Architecture, Dynamics and Function, June 2013 (Cargese, Corsica, France). Poster presented entitled ‘Tat-dependent selenate reduction in bacteria’.

• The Society of General Microbiology Autumn Conference, September 2011 (York, UK). Presented a poster entitled ‘Imaging protein traffic on the Tat pathway’.

xviii

List of abbreviations

Å Ångstrom (10-10m; 0.1 nm) A Absorbance Amp Ampicillin APH Amphipathic helix Apra Apramycin ATP Adenosine triphosphate bp base pair BSA Bovine serum albumin BV Benzyl viologen Cml Chloramphenicol C-terminus Carboxy terminus DMSO Dimethyl sulphoxide DNA Deoxyribonucleic acid DNase Deoxyribonuclease dNTP Deoxynucleoside triphosphate DTT Dithiothreitol EDTA Ethylenediaminetetraacetate Fe-S Iron-sulphur x g Relative centrifugal force g gram GFP Green fluorescent protein HRP Horse radish peroxidase Hyd Hydrogenase Hyd-1 Hydrogenase-1 Hyd-2 Hydrogenase-2 IPTG Isopropyl β-D-1-thiogalactopyranoside Kan Kanamycin kb kilobase (1000 bp) kDa kilodalton L Litre LB Luria Bertani μ micro M Molar m milli MCD Molybdopterin cytosine dinucleotide mCer mCerulean MGD Molybdopterin guanine dinucleotide min minute MoCo Molybdenum cofactor mRNA messenger ribonucleic acid NAD Nicotinamide adenine dinucleotide xix

NADP Nicotinamide adenine dinucleotide phosphate nm nanometre NMR Nuclear magnetic resonance N-terminus Amino-terminus OD Optical density PAGE Polyacrylamide gel electrophoresis PCR Polymerase chain reaction pmf (Δp) Proton motive force psi Pounds per square inch PFE Protein film electrochemistry ROS Reactive oxygen species rpm Rotations per minute SDS Sodium dodecyl sulphate Sec Secretory Tat Twin arginine translocase TEMED N, N, N’, N’-tetramethylethylenediamine TMA Trimethylamine TMAO Trimethylamine N-oxide Tris Tris(hydroxymethyl)aminomethane UV Ultraviolet [v/v] volume per volume [w/v] weight per volume YFP Yellow fluorescent protein Δψ Electrical field gradient

xx

Amino acids and their abbreviations

Amino-acid Three-letter abbreviation One-letter abbreviation Alanine Ala A Arginine Arg R Asparagine Asn N Aspartate Asp D Cysteine Cys C Glutamate Glu E Glutamine Gln Q Glycine Gly G Histidine His H Isoleucine Ile I Leucine Leu L Lysine Lys K Methionine Met M Phenylalanine Phe F Proline Pro P Serine Ser S Threonine Thr T Tryptophan Trp W Tyrosine Tyr Y Valine Val V Any amino acid - X

xxi

Acknowledgments

Firstly I would like to thank my supervisor, Frank Sargent, for the opportunity to undertake this project in his research group. His encouragement and guidance, throughout the years and throughout the writing of this thesis, has been deeply appreciated. Thank you so much for the great experience the past four years has been.

Thanks must also go to Tracy Palmer, members of the TP/FS group, and the whole of the

Molecular Microbiology division, past and present. Thank you for help and advice in the lab as well as outside of it, I couldn’t have got through it without you and I’ve made so many wonderful memories from the many fun times we’ve had. To those I had the pleasure of sharing an office with over the years, thank you for the company and conversation, whether science-related or not, it all helped.

In particular I must thank those who have advised, helped and collaborated with me in some way. David Guymer, Holger Kneuper and Sarah Coulthurst for the use of vectors and strains they created. Rory Duncan and Colin Rickman at Heriot-Watt University for their great advice and for generously allowing me to use their microscopes. Kelly Hodge for protein crosslinking advice. In addition, thanks to Markus Posch, the scientific officer for the OMX microscope here in the College of Life Science, for teaching me about OMX microscopy. Grant, thank you for picking up my cloning when it was all going wrong for me, as well as for always making the coffee break chat that little bit more interesting. And Jen, thank you for teaching me how to purify proteins, and not to be scared of FPLC machines. Also thanks to Richard Owen for helping me massively with crystal trials, and Nicola Stanley-Wall and Adam Ostrowski for microscopy advice. xxii

I would also like to express my gratitude to those who have funded my over the years. SULSA, for this PhD studentship that enabled me to undertake this research. SGM and the Biochemical

Society, for providing me with grants to allow me to attend some great national and international conferences to present my work. And the Euro-Bioimaging Proof of Concept initiative, for enabling me to use the OMX microscope here in Dundee.

Last but not least I must mention those closest to me. To my family, thank you for always being interested in how my work was going, and for the love and support that has got me to this point. And to my husband Paul, who has barely had a wife since we were married due to the writing of this thesis. Thank you for the constant love and encouragement, and for always believing in me, even when I didn’t. I couldn’t have done it without you. 1

1 Introduction 2

1.1 Protein translocation in Gram-negative organisms

1.1.1 Escherichia coli and Salmonella enterica subsp. enterica serovar

Typhimurium

Escherichia coli K-12 is a typical member of the gut microflora and is normally harmless to the host. However, certain strains do exist that can cause disease, including enterohaemorrhagic E. coli (EHEC) and enteropathogenic E. coli (EPEC). These virulent strains have arisen through the acquisition of virulence genes such as those encoding Shiga toxin, and are well known for their ability to cause diarrhoeal disease and other serious illnesses (Johnson & Nolan, 2009). Due to its genetic tractability and relatively simple growth requirements under laboratory conditions

E. coli K-12 has become a model bacterium for biochemical and cellular studies.

The Salmonella genus contains over 2000 serovars, with some causing systemic infections such as meningitis and others affecting the GI tract causing localised alimentary infections. The bacterium Salmonella enterica subsp. enterica serovar Typhimurium is a facultative intracellular pathogen and a major cause of human gastroenteritis. S. Typhimurium strain LT2 was first isolated in the 1940s and is now a standard laboratory research strain, with the entire chromosome having been sequenced (McClelland et al., 2001).

1.1.2 Structure of the Gram-negative bacterial cell membranes

E. coli and S. Typhimurium are both Gram-negative bacteria. Gram-negative bacteria can be characterised by the presence of a double layered cell envelope; the inner membrane enclosing the cytoplasm and the outer membrane surrounding this inner membrane (Figure

1.1). The ‘space’ between the two membranes is known as the periplasm and is an aqueous and highly viscous environment that contains the peptidoglycan cell wall. The contents of the periplasm are able to act as a buffer against the extracellular environment, and this environment is highly reactive with many proteins being present involved in a great many 3 cellular processes. These include cell wall biogenesis and division, nutrient uptake, motor protein assembly and anaerobic respiration (Hall et al., 1997, Kojima et al., 2009, Richardson,

2000, Tarry et al., 2009a, Tsui et al., 1994).

Figure 1.1: The Gram-negative cell envelope. The cell envelope enclosing the cytoplasm of Gram-negative organisms comprises two layers, the inner cytoplasmic membrane (IM) and the outer membrane (OM). The inner membrane consists of a phospholipid bilayer with two types of protein: integral inner membrane proteins (IMPs) that span the membrane bilayer through α-helical transmembrane domains, and inner membrane lipoproteins that are anchored to the membrane by a lipid moiety. Soluble proteins and the peptidoglycan layer reside in the aqueous periplasm. The asymmetric outer membrane is composed of a phospholipid inner layer and a lipopolysaccharide (LPS) outer layer. Outer membrane proteins (OMPs) and OM lipoproteins reside within this layer. Figure based on (Ruiz et al., 2009)

The cytoplasmic and outer cell membranes are significantly different in their compositions due to the differing requirements of their roles. The cytoplasmic membrane separates the cytoplasm from the periplasm. It consists of a phospholipid bilayer studded with integral membrane proteins, and functions as a permeability barrier. The phospholipid composition of the bilayer is generally 70-80 % phosphatidylethanolamine (PE), 20-25 % phosphatidylglycerol 4

(PG) and up to 5 % cardiolipin (CL) (Dowhan, 1997). CL levels have been shown to rise with a related decrease in PG upon entry to the stationary phase (Hiraoka et al., 1993). The hydrophobic nature of the internal portion of the cytoplasmic membrane causes it to be impermeable to ions, thereby enabling the establishment of a charge separation and electrochemical gradient across the cis (cytoplasmic) and trans (periplasmic) sides of the membrane. This transmembrane electrochemical gradient is utilised by the bacterial cell for a large number of cellular functions, for example protein translocation (Driessen, 1992), flagellar rotation (Berry, 1993), active solute transport (Ramos et al., 1976) and ATP synthesis (Maloney et al., 1974).

The outer membrane of the Gram-negative cell also consists of a phospholipid bilayer and proteins, however in addition to these components are polysaccharides, which in combination with lipids form lipopolysaccharides (LPS). LPS is arranged on the extracellular face of the outer membrane, and can be structurally described in three components: the hydrophobic lipid A portion, the core polysaccharide and the O-specific polysaccharide (or O-antigen). The lipid A region of LPS is associated with causing toxicity to animals, whilst the O-antigen is involved in bacterial resistance to antibiotics and is a key target of the complement pathway of the immune system (Raetz & Whitfield, 2002).

1.1.3 Protein translocation in Gram-negative bacteria

Transport of proteins across the cytoplasmic membrane is an essential capability of bacteria, with approximately 20% of synthesised proteins requiring translocation partially or entirely across the cytoplasmic membrane (Pugsley, 1993). The cytoplasmic membrane plays an important role in energy production by bacteria, allowing the generation of proton-motive force required in respiratory and photosynthetic electron transfer chains (Berks et al., 2000). A number of redox proteins require an extracellular location for this proton motive force generation to occur (Berks et al., 2000). The cell envelope presents a physical barrier which all 5 enzymes requiring extra-cytoplasmic localisation must cross. There is also the issue of the differing environments created by this compartmentalisation that proteins must be able to overcome in order to assemble correctly for function.

In order to address these challenges, specialised transport systems have developed to target proteins from their site of synthesis in the cytoplasm to their site of function. These systems are able to distinguish between proteins requiring translocation and those that must remain cytoplasmic, and have mechanisms by which proteins are targeted to their cognate transport system to ensure correct and efficient transport. This transport may be termed export or secretion, depending upon the end location of the protein. Whilst protein export involves the transport of proteins from the cytoplasm across the cytoplasmic membrane to the periplasm, protein secretion defines the translocation of proteins through both cytoplasmic and outer membranes to the extracellular environment or into another target cell. A wide range of translocation systems have evolved to conduct these roles and by working alone or in cooperation enable translocation of proteins to every cellular compartment and extracellularly

(Dalbey & Kuhn, 2012).

1.1.4 The general secretory (Sec) pathway

1.1.4.1 Overview of Sec translocation

In Gram-negative bacteria such as Escherichia coli many proteins undergo transport across the cytoplasmic membrane into the periplasm via the general secretory (Sec) pathway (Rusch &

Kendall, 2007). The Sec pathway is the only protein translocation pathway known to be universally conserved across all domains, and is the route by which the majority of proteins reach the periplasm (Pohlschröder et al., 2005). During early studies of the Sec system in E. coli, the Sec components were identified through genetic screening. In this work, a number of genetic mutants were isolated; either general secretion (sec) mutants that displayed general secretion defects, or protein localisation dominant suppressor (prl) mutations that were able 6 to restore the export of precursor proteins with altered signal peptides (Bieker & Silhavy, 1990,

Danese & Silhavy, 1998, Trun et al., 1988). Through the use of these mutants, two Sec protein complexes were generated (Bieker-Brady & Silhavy, 1992) and this led to the reconstitution in vitro of the Sec translocase and further characterisation of the translocase (Natale et al., 2008).

The Sec translocase exports proteins in an unfolded state, either co- or post-translocationally

(Pohlschröder et al., 2005).

Proteins destined for export by Sec are synthesized as precursor proteins with N-terminal signal peptides that target them to this translocase. Whilst these signal peptides do not share specific sequence homology, they do have a general organization with clearly definable features (Figure 1.2). Normally between 18 and 26 amino acid residues in length, the peptides consist of an n-region with one or more positively charged residues, a core h-region of hydrophobic residues, and a c-region with a signal peptidase cleavage site (commonly AxA), the only conserved motif amongst these signal peptides (Cristóbal et al., 1999a, Izard &

Kendall, 1994, Stanley et al., 2000). Efficient transport is dependent upon conservation of these general features of Sec signal peptides, for example the positive charge of the n-region, although not dependent on amino acid-type, must be present for translocation to occur (Izard

& Kendall, 1994). Increasing the positive charge through amino acid substitutions can in fact increase efficiency of Sec transport (Sasaki et al., 1990). Mutations in the AxA motif inhibited cell growth and protein processing, thereby preventing maturation of Sec substrates (Barkocy-

Gallagher & Bassfor Jr., 1992).

7

Figure 1.2: General structural features of Sec signal peptides. Signal peptides of the Sec translocase range in length from 18 to 26 amino acids. They have a conserved tripartite structure with a positive n-region, a hydrophobic h-region and a c-region that contains the AxA signal peptidase cleavage site.

1.1.4.2 Components of the Sec translocon

The heterotrimeric protein conducting channel of Sec consists of two universally conserved transmembrane proteins, named SecE and SecY in prokaryotes, and also a third protein subunit that varies between domains. In bacteria this third subunit is named SecG (Natale et al., 2008, Pohlschröder et al., 2005, Pugsley, 1993). Whilst SecE and SecY are highly conserved and essential for protein export, and it has been shown that these protein components are sufficient for transport, SecG is not necessary for function and only makes limited contact with

SecEY (Berg et al., 2004, Brundage et al., 1990, Natale et al., 2008), but may play a role in translocation with its presence appearing to improve efficiency, for example at low temperature (Nishiyama et al., 1994). The crystal structure of the SecYEG complex from the

Archaea Methanococcus jannaschii was solved by van den Berg and colleagues in 2004 (Figure

1.3), and suggests that a single heterotrimer can serve as a translocation unit. SecY appears to consist of two halves (TM1-5 and TM6-10) linked by a flexible loop, with SecE acting to clamp these two domains together. The two halves of SecY create an hourglass shaped pore through their arrangement of an inverted two-fold rotation pseudo-symmetry. The pore is thought to exist in a closed inactive state with the distorted short helix TM22 of SecY acting as a plug (van den Berg et al., 2004). The pore has been predicted to open via the ‘lateral gate’ region between TM2 and TM7 upon binding of a substrate signal peptide (du Plessis et al., 2011).

8

Figure 1.3: 3D crystal structure of the SecYEG complex from Methanococcus jonnaschii. Images taken from (van den Berg et al., 2004). A – View from the cytoplasm. SecA helices 1-10 are coloured blue to red, with SecE in magenta and SecG in pink. B – View from the back of the SecYEG pore.

Peripherally associated with the SecYEG pore is cytoplasmic protein SecA, an ATPase involved in coupling ATP hydrolysis to protein translocation, thereby energizing export (Lill et al., 1989,

Pohlschröder et al., 2005). In addition to interacting with the Sec pore, SecA appears to bind to the Sec signal peptide, with both n- and h-regions being important for interactions, and also demonstrates reversible interactions with the cytoplasmic membrane (Cunningham &

Wickner, 1989, Lill et al., 1989). Whilst SecA is required for energy production for translocation, the transmembrane proton motive force also appears to play a role in Sec transport, possibly in prevention of reverse translocation (Arkowitz et al., 1993, Driessen,

1992).

1.1.5 Co- and post-translational protein targeting

There are two distinct targeting routes by which signal peptide-bearing precursor proteins

(‘preproteins’) arrive at the Sec translocon – post-translational and co-translational (Figure

1.4). The route a preprotein follows is determined immediately upon emergence from the ribosome, and appears to be related to signal peptide hydrophobicity with the signal 9 recognition particle (SRP) involved in co-translational export tending to recognise more

hydrophobic signal peptides than SecA (De Gier et al., 1997, Kebir & Kendall, 2002).

Figure 1.4: Post- and co-translational export through the Sec pathway. Post-translational export (left) is mediated by the SecB chaperone (light blue), that binds the unfolded Sec substrate as it exits the ribosome. SecB then shuttles the substrate to the Sec machinery where the SecA ATPase (green) drives export through the SecYEG pore. Co-translational export occurs with the signal recognition particle (SRP – red) binding to the signal sequence of the Sec substrate (orange). The ribosome/substrate complex is then taken to the Sec machinery and translocation can occur as the protein is synthesised. Folding of the substrate then takes place in the periplasm.

The majority of proteins are exported post-translationally, and this route involves primary interactions with the molecular chaperone protein SecB (Figure 1.4). SecB is a small cytoplasmic protein of 12 kDa that exists as a tetramer (Dekker et al., 2003, Xu et al., 2000), 10 and does not appear to be essential for export (Hartl et al., 1990, Trun et al., 1988), however it is thought to be involved in shuttling preproteins to SecA as well as maintaining proteins in an unfolded state to promote Sec export (Natale et al., 2008). Indeed SecB recognition of preproteins appears to be dependent upon them being in a non-native conformation (Hardy &

Randall, 1991), and it is thought to interact with multiple hydrophobic regions of the preprotein via hydrophobic grooves (Dekker et al., 2003, Randall et al., 1998a, Randall et al.,

1998b, Xu et al., 2000). SecB interacts with SecA at its extreme carboxyl region (Crane et al.,

2005), causing a conformational change enabling passing of the preprotein to SecA from SecB in coordination with SecB release (Fekkes et al., 1997, Joly & Wickner, 1993). It is thought that binding of a preprotein peptide to SecA causes conformational changes that repress the

ATPase DEAD/DEXH ATPase motor in the C domain. As the preprotein is passed to the SecYEG pore through an interaction with SecY, the DEAD motor becomes active and ATPase activity is initiated, which aids insertion of the preprotein into the SecYEG pore. This is thought to be a type of safety mechanism, ensuring ATPase activity does not occur when SecA is not associated with the SecYEG pore (Baud et al., 2002, Papanikou et al., 2005). SecA subsequently penetrates SecYEG, and preprotein is threaded through the SecYEG pore (Economou et al.,

1995, Economou & Wickner, 1994). The signal peptide is cleaved as translocation is completed and the mature protein then folds in the periplasm.

Co-translational export, facilitated by the SRP, is the route of choice for the majority of membrane proteins (Figure 1.4). The SRP of E. coli consists of two components, a 4.5S RNA and the 48 kDa GTPase called Ffh for “fifty four homolog” (Valent et al., 1998). The SRP recognises highly hydrophobic stretches such as those found on Sec signal peptides, whilst proteins with less hydrophobic signal sequences are unable to stably bind and so are instead dependent upon SecB (Valent et al., 1998, Zhou et al., 2014). Newly synthesised polypeptides emerge from the ribosome and are immediately picked up by SRP via their signal peptide, with less hydrophobic signal peptides thought to be prevented through competition with another ribosome-bound protein Trigger Factor (TF) (Ullers et al., 2003). Ribosome-bound SRP is 11 recognised by the cytoplasmic membrane SRP receptor, another GTPase (Zhang et al., 2008,

Zhang et al., 2009). Upon interactions with FtsY, the ribosome binds to the SecYEG complex in a high affinity interaction with the large ribosomal subunit (Prinz et al., 2000). Interactions of the FtsY-SRP complex with the membrane and SecYEG are thought to induce conformational changes activating GTP hydrolysis of both FtsY and Ffh, leading to the unloading of the preprotein from the SRP. The release of preproteins from the SRP and their association with

SecYEG are interlinked processes (Valent et al., 1998). GTP is eventually hydrolysed once the preprotein is successfully transferred to SecYEG, and appears to drive disassembly of the

SRP/FtsY complex for a new cycle of targeting to begin (Connolly & Gilmore, 1989). Preproteins are threaded through the Sec pore in a precursory unfolded state, with folding occurring in the periplasm once translocation is complete (Berks et al., 2000, Bogsch et al., 1998).

1.1.6 YidC

A large number of proteins are required to be inserted in the cytoplasmic membrane for function. The Sec translocase is able to insert a number of proteins, however YidC is also thought to be important in inserting proteins into the cytoplasmic membrane, both independently and in conjunction with the Sec translocase (Scotti et al., 2000, van der Laan et al., 2005). The YidC mechanism of membrane insertion is highly important for the maintenance of the proton-motive force, with many membrane proteins involved in energy transduction being inserted by YidC (van der Laan et al., 2003).

YidC is predicted to consist of six transmembrane regions plus a large periplasmic domain between regions 1 and 2. The crystal structure of the periplasmic part of YidC has been determined and forms a β-sandwich motif of 18 β-strands and three α-helices (Oliver &

Paetzel, 2008). The structure of the entire YidC protein has been predicted using modelling and appears to induce membrane ‘thinning’, which may be relevant to the membrane insertion of proteins (Wickles et al., 2014). 12

When inserting proteins with the Sec pathway, YidC interacts with the lateral gate of SecY involved in membrane insertion, perhaps involved in the coordination of lateral gate movement (Sachelaru et al., 2013). SecF and SecD of the accessory complex SecDFyajC have also been observed in interactions with YidC, perhaps important as a scaffold aiding to hold

YidC with SecY (Nouwen & Driessen, 2002).

When inserting membrane proteins independently of the Sec system, YidC appears to recognise a small number specific preproteins such as the phage Pf3 coat protein, the mechanosensitive MscL protein and other small membrane proteins (Chen et al., 2002, Facey et al., 2007). Some of these proteins also require SRP for insertion (Wang & Dalbey, 2011).

1.1.7 The twin-arginine translocation (Tat) pathway

1.1.7.1 Overview of Tat translocation

Many proteins functioning in the periplasm require cofactors that are synthesised in the cell cytoplasm. In order to secure such cofactors the protein must fold, however this then renders them incompatible with the Sec translocase (Berks et al., 2000). In some cases the protein and cofactors are transported separately into the periplasm where cofactor insertion and protein maturation then occurs (Berks et al., 2000). The more common mechanism, however, involves protein maturation in the cytoplasm with the fully folded, cofactor-containing protein then being transported out of the cytoplasm and into the periplasm. This Sec-independent pathway found in many bacteria and archaea, as well as being present in the thylakoid membranes of plant chloroplasts (Pohlschröder et al., 2005), and is named the Tat (twin-arginine translocation) pathway due to the consensus twin-arginine amino acid sequence motif that is apparent in signal peptides targeting proteins to this pathway (Berks et al., 2000). The Tat pathway differs from the Sec system in that it transports a very small subset of proteins in their fully folded form across the cytoplasmic membrane, with many of these proteins playing roles in many important cellular functions including energy metabolism, cell division and 13 motility, resistance to heavy metals and antimicrobials, and virulence (Palmer & Berks, 2012).

Since its discovery in the late 1990’s, the Tat system has been studied intensely and has revealed an intriguing and unique mechanism of protein export.

1.1.7.2 Components of the Tat translocon

In E. coli the fully assembled Tat translocase consists of a 600 kDa complex of four related membrane proteins named TatA, TatB, TatC and TatE (Bogsch et al., 1998, Sargent et al., 1998,

Weiner et al., 1998). The tatABCD operon encodes for TatA, TatB and TatC proteins, as well as for TatD; a cytoplasmic protein with no apparent function in protein translocation, thought instead to have DNase activity (Figure 1.5) (Bolhuis et al., 2001, Sargent et al., 1998, Wexler et al., 2000). The fourth Tat component is TatE, found as a completely separate monocistronic gene tatE (Figure 1.5) (Bolhuis et al., 2001) and is said to be a gene duplication encoding a functional homologue of TatA (Sargent et al., 1998).

Figure 1.5: Organisation of tat genes in E. coli. Genes encoding for components of the Tat system are indicated in green. Nearby genes are shown in white. Base numbers on the E. coli MG1655 genome are indicated above the illustrated genes. A – The essential genes of the Tat system are found in a single operon alongside tatD, a gene that is cotranscribed but not involved with protein translocation (Wexler et al., 2000). B - Gene tatE, a duplication of the tatA gene found as a monocistronic gene in another region of the genome (Chanal et al., 1998). 14

Expression studies have shown that tatABC and tatE are expressed constitutively, therefore indicating the importance of the translocase system in both aerobic and anaerobic growth conditions (Jack et al., 2001). It appears that tatE is expressed at significantly lower levels than tatA, with the cellular molar ratio of the main Tat components having been determined as

TatA:B:C 40:2:1 (Jack et al., 2001, Sargent et al., 1998). The TatA and TatB proteins are similar in structure, whilst TatC belongs to a separate family.

1.1.7.3 TatA family proteins

The TatA family encompasses proteins TatA, TatB and TatE. These proteins have been shown to be membrane spanning proteins consisting of one transmembrane α-helix at the N- terminus connected to a C-terminal amphipathic helix via a short hinge region (Hu et al., 2010).

Whilst both TatA and TatB are required for Tat transport in E. coli, suggesting distinct roles for each protein, some bacteria such as Gram-positives have minimal Tat systems consisting of just TatA and TatC (Freudl, 2013, Robinson et al., 2011). It has been shown that TatA from

Bacillus subtilis is able to complement both tatA and tatB deletions in E. coli (Barnett et al.,

2008). It appears that early gene duplication led to the emergence of TatA and TatB proteins, with the TatB proteins then diverging in sequence from TatA leading them to take on differing roles in the Tat translocase (Yen et al., 2002).

Encoded at minute 14 of the E. coli genome, TatA is the smallest of the main Tat components, being just 9.6 kDa protein 89 amino acids in length (Sargent et al., 1998). The structure and membrane alignment of monomeric TatA has been studied using both solid state and solution

NMR, and is shown to display an L-shaped structure composed of a membrane-bound helix and an amphipathic helix with a long unstructured C-region linked by a highly conserved hinge region (Figure 1.6). This hinge region contains the only amino acid residue absolutely 15 conserved in both TatA and TatB, Gly-21, and this amino acid in TatA is essential for Tat translocation (Hicks et al., 2003).

Figure 1.6: Solution NMR structures of monomeric TatA. A - Cartoon representation of a single TatAd molecule from Bacillus subtilis, solved by Hu et al. (2010). B – Cartoon representation of a TatA monomer from E. coli, solved by Rodriguez et al. (2013). The grey rectangles represent the cytoplasmic membrane in order to give an indication of protein orientation. TMH – transmembrane helix. APH – amphipathic helix.

In addition Phe-20 is also strictly conserved, and it is likely that these two residues are important for the formation of this hinge region, particularly due to the flexibility conferred by

Gly-21 (Hicks et al., 2003). The amphipathic helix appears to lie on a steep tilt along the face of the cytoplasmic membrane, causing it to be partially buried in the membrane (Chan et al., 16

2011, Hu et al., 2010, Walther et al., 2010, Zhang et al., 2014a). Numerous studies over several years have been undertaken to determine the exact orientation of TatA within the membrane.

A number of years ago it was suggested that TatA adopted an ‘N-in, C-out’ topology in the membrane (Chan et al., 2007). However more recently it has been categorically shown that

TatA does in fact adopt an ‘N-out, C-in’ topology (Koch et al., 2012).

The 7 kDa protein TatE shares 53 % sequence identity with TatA and is thought to play a similar role in Tat translocation. In-frame deletion mutants of tatA or tatE are not sufficient to completely abolish Tat export of proteins, only with the deletion of both genes was transport completely disrupted. This indicates that TatE and TatA have overlapping functions (Sargent et al., 1998, Sargent et al., 1999). However, whilst TatE has been shown to play a role in translocation it is thought to be to a lesser extent, due to the 50 to 100-fold lower expression levels of TatE and also the lack of severe transport defects induced by a tatE gene deletion compared to that of tatA (Jack et al., 2004, Jack et al., 2001, Sargent et al., 1998).

TatB is a specialist member of the wider TatA family and is a slightly larger protein than TatA at

18.4 kDa, containing 181 amino acids with 20 % shared sequence identify to TatA. TatB is also structurally similar to TatA although it has been shown to have a number of structural differences, in particular in its C-region which has been shown to be much longer than that of

TatA (Sargent et al., 1999). A new structure for E. coli TatB has recently been published, showing an extended L-shape conformation with four α-helices: a single transmembrane helix, plus three helices in the extended C-region comprising an amphipathic helix and a two helices that are hydrophilic in nature (Figure 1.7) (Zhang et al., 2014b). Truncation analysis has shown that TatB can still show significant transport with even 70 amino acids of the C-terminal domain missing, whilst interactions of TatB with TatC can still occur when the entire C-region truncated, and it is suggested that the C-terminal tail region is not essential for protein function (Lee et al., 2002, Lee et al., 2006a, Maldonado et al., 2011). The hinge region of TatB contains the conserved Gly-21, whilst amino acids around this essential residue have been 17 shown to differ to those in TatA (Hicks et al., 2003). In E. coli TatB is known to be functionally different to TatA and TatE, with complementation experiments with plasmids expressing tatA and tatE genes unable to resume translocation in a tatB deletion mutant. Only upon the re- introduction of tatB on a plasmid was the block on Tat export corrected (Sargent et al., 1999).

Figure 1.7: Solution NMR structure of E. coli TatB. Cartoon representation of a TatB monomer from E. coli. Comprises four helices; a single transmembrane helix (TMH), an amphipathic helix (APH) plus two hydrophilic helices (Zhang et al., 2014b).

1.1.7.4 TatC proteins

TatC proteins are the most highly conserved of the Tat protein subunits (Yen et al., 2002). In E. coli TatC is composed of 258 amino acids and are 28.9 kDa in molecular weight. This protein has been shown to be an essential component for Tat translocation (Bogsch et al., 1998).

Structural studies show that TatC consists of six transmembrane helices, whilst both N- and C- termini of the protein are located in the cytoplasm (Behrendt et al., 2004, Punginelli et al.,

2007). Recently the structure of TatC from Aquifex aeolicus, a hyperthermophile, was solved to a resolution of 3.5 Å (Figure 1.8) (Rollauer et al., 2012). This TatC protein shares 40 % sequence 18 identity with E. coli TatC, allowing previous data from E. coli TatC to be accurately mapped onto this new structure.

Figure 1.8: Three dimensional crystal structure of TatC from A. aeolicus. A – Cartoon representation of TatC structure, with helices coloured from blue (N-terminus) to red (C-terminus). Helix numbers are labelled in corresponding colour. B – Surface representation of TatC molecule. Left and right hand views in A and B are in the same orientation. Structure solved by Rollauer et al. (2012).

Using molecular dynamics (MD) simulations Aquifex TatC was shown to have a number of

helices spanning the membrane at a steep angle with relation to the membrane plane, thereby

preventing any large amount of protrusion from the membrane and limiting solvent exposure 19

(Figure 1.8). Transmembrane helix 1 has a severe kink towards the periplasmic face forming an interfacial helix. These characteristics, plus the presence of helices 5 and 6 that are extremely short and unable to fully span the membrane, cause the membrane bilayer to become distorted (Rollauer et al., 2012). It is thought that the protruding periplasmic loops of TatC may help to stabilise helix positions within the membrane, potentially important for TatC function.

Indeed, this corresponds with the tendency of amino acid mutations in these loop regions to lead to inactivation of Tat transport (Kneuper et al., 2012).

1.1.7.5 Formation of the Tat translocase

The major conundrum with the ability of the Tat translocase to transport folded substrates with such a large range of sizes is how the system is able to maintain ionic stability throughout the translocation process. Whilst most other transport systems translocate unfolded proteins of similar diameters and so are able to keep a permanent channel across the membrane, Tat substrates vary hugely in size from 20 – 70 Å in diameter and so it is not possible to follow a similar mechanism (Richter et al., 2007). This issue may explain the lack of homology between

Tat components and those of other transport systems. In order to produce a translocon system able to perform to these requirements, it appears that the Tat system forms from two functional complexes, the TatA complex and the TatBC complex (Oates et al., 2005, Orriss et al., 2007).

Electron microscopy of TatA purified protein complexes has revealed ring-like structures of varying size, with an apparent pore in the centre with a lid-like structure on the cytoplasmic face. Diameters of the pore ranged between approximately 30 and 70 Å, which would accommodate the entire size range of Tat substrates (Gohlke et al., 2005). Using blue native

PAGE it has been shown that purified TatA complexes exist in varying sizes from below 100 kDa to over 600 kDa, increasing in size in a stepwise fashion with estimated increments of 40 kDa

(Oates et al., 2005). Similar observations were made in vivo and at native expression levels 20

(Leake et al., 2008). In addition, purified TatA monomers were able to self-assemble into complexes and again display this ladder of complexes on blue native PAGE (Oates et al., 2005).

Higher oligomers of TatA were also present in the absence of TatB and TatC (De Leeuw et al.,

2001), although the size of these complexes considerably increase in the presence of the other component proteins (Leake et al., 2008). This suggests the functional TatA units are tetrameric protomers that interact and oligomerise to form larger complexes depending on the substrate requiring transport. This idea is supported by studies into the homologous transport system in plants, where it has been shown that oligomerisation of Tha4 – the TatA homologue found in the thylakoid membranes of chloroplasts – can be influenced by the presence of substrates

(Dabney-Smith et al., 2006).

The amphipathic helix of TatA has been shown to be essential for Tat function, with amino acid substitutions on the hydrophobic face of this helix inactivating transport (Greene et al., 2007).

A number of these residues have also been highlighted in the TatA dimeric crystal structure as being involved in the dimeric interface that may come into play at higher oligomeric numbers

(Zhang et al., 2014a). It appears that these amphipathic helices form a lid on the cytoplasmic side of the TatA channel that may work as a gating mechanism to maintain the ionic seal across the membrane (Gohlke et al., 2005). Using solution NMR, and utilising the knowledge that by adjusting the detergent levels the amount of TatA oligomerisation can be controlled, a high- resolution structure of the TatA oligomeric complex was obtained (Figure 1.9). This shows the pore is created through interactions of the transmembrane helices (Figure 1.9D), whilst amphipathic helices extend out and away from the pore. This would leave them accessible by the TatBC/substrate complex during translocation. Modelling has indicated that such a pore would lead to extreme distortions of the cytoplasmic membrane, an effect that would create a thinned membrane and likely aid creation of a translocation pathway through eventual and transient rupture of the membrane (Rodriguez et al., 2013). 21

Figure 1.9: Modelling the TatA oligomeric pore. A – Structural models of TatA oligomeric complex containing 16 or 25 subunits. B – Structural diagrams of oligomeric TatA, both from the cytoplasmic face (left) and side-on from within the membrane (right). C – Surface plot of TatA oligomer as seen from the cytoplasmic side (left) and periplasmic side (right) of membrane. Atoms are coloured according to residue type: Lys (blue), Asp or Glu (red), Phe or Trp (pink), Ser or Thr (cyan), Gly (light orange), Pro (green), Gln (magenta), and Leu, Ile, Val, Ala, or Met (light grey). For clarity, residues 1-3 are omitted due to their flexibility. D – A comprehensive view of the packing interface between residues 4-21 of transmembrane helices. E – Surface plot of interior of TatA pore. Atoms are coloured according to type: nitrogen (blue), oxygen (red), sulphur (yellow), and carbon (light grey). The carboxamide of Gln8 plus aliphatic side-chains of the pore-lining Ile11, Ile15, and Leu19 are indicated. From (Rodriguez et al., 2013). 22

There have been numerous studies highlighting the ability of TatB and TatC to interact with high affinity, and these proteins are known to form a complex in vivo (Bolhuis et al., 2001). A number of studies have suggested TatC in fact depends upon the presence of TatB for its stability, becoming undetectable in the cell when expressed alone in a complete tat deletion strain, perhaps highlighting the necessity of TatB for translocase formation (De Leeuw et al.,

2001, Sargent et al., 1999). However, recent work has shown it is possible to purify TatB and

TatC oligomers in the absence of all other Tat proteins (Orriss et al., 2007). TatBC complexes have been purified by various techniques, with complexes showing molecular masses of varying size depending upon the purification technique (Bolhuis et al., 2001, Oates et al., 2005,

Tarry et al., 2009b). These complexes have a 1:1 TatB:TatC stoichiometry with an estimated 7 copies of each protein per complex (Bolhuis et al., 2001, Tarry et al., 2009b). This is in agreement with data from the homologous plant thylakoid Tat system, where it has been demonstrated that TatB and TatC homologues form a signal peptide-binding complex independent of TatA (Cline & Mori, 2001).

1.1.7.6 The Tat signal peptide

Substrates of the Tat translocase are similar to substrates of the Sec pathways, with signal peptides present at the N-terminal end of the precursor protein used for targeting for translocation (Berks, 1996, Sargent, 2007a). Tat signal peptides have been identified to belong to two classes according to whether or not they have a role in biosynthesis in addition to targeting. Those encompassing the dual roles of Tat targeting and chaperone binding belong to class 1 or A-class whilst those with a single function in targeting belong to class 2 or B-class

(Sargent et al., 2006). It is also likely that a third class of signal peptides may exist involved in chaperone binding only, the putative C-class (Sargent, 2007b).

Tat signal peptides have a similar organisation to signal peptides of the Sec pathway with both composing a tripartite structure. This structure has an amino (n-) region, a hydrophobic (h-) 23 central region and a carboxyl (c-) region in both Sec and Tat signal peptides (Figure 1.10)

(Berks, 1996, Rusch & Kendall, 2007). Tat signal peptides are in general somewhat longer than

Sec signal peptides due to a slightly longer n-region of variable length (Cristóbal et al., 1999a).

Tat signal peptides have a consensus sequence (S/T)RRxFLK, where the two arginine residues are almost completely invariant (Berks, 1996). This twin arginine motif, for which the Tat translocase is named, is thought to be essential for Tat transport with substitutions of either or both arginines inhibiting translocation. However in some cases there has been some evidence of translocation still occurring with a single conservative arginine to lysine amino acid substitution (Hinsley et al., 2001, Stanley et al., 2000). All other amino acid residues of the twin-arginine motif are conserved approximately 10-50 % of the time (Bendtsen et al., 2005,

Berks, 1996, Stanley et al., 2000).

Figure 1.10: General features of the Tat signal peptide. The N-terminal Tat signal peptide is split in a tripartite fashion into the basic n-region (blue), the central hydrophobic h-region (grey) and the polar c-region (turquoise). The highly conserved twin-arginine motif is found on the border between the n- and h-regions, whilst the signal peptidase cleavage site AxA is located at the C- terminus.

Tat signal peptides contain more glycine and less leucine residues in their h-regions than Sec signal peptides, thereby reducing the hydrophobicity of their h-region in comparison. The overall hydrophobicity of Tat signal peptides appears to play an important role in Sec avoidance, as by increasing the hydrophobicity of the signal peptide h-region it was shown to be able to be redirected through the Sec system (Cristóbal et al., 1999a). Secondary structure predictions, and other studies, of the signal peptide suggest that the h-region is likely to have a 24 helical organisation, with the helix starting close by the end of the twin-arginine motif (San

Miguel et al., 2003).

The c-region of Tat signal peptides contains the site of signal peptidase cleavage, usually an

AxA motif. There is also thought to be a ‘Sec avoidance signal’ at the C-terminal region through the presence of a high number of basic residues. This feature in addition to the presence twin arginine motif appears to enable effective targeting to Tat and avoidance of Sec transport

(Blaudeck et al., 2003, Bogsch et al., 1997). In both Sec and Tat translocase systems, the signal peptide travels through in a hairpin conformation. In this way, the c-region is actually the first part of the protein to enter the Tat pore, explaining the importance of this region for preventing incorrect translocation through Sec.

1.1.7.7 Substrates of Tat

There are currently 27 experimentally confirmed twin-arginine signal peptide-bearing substrates in E. coli, making up approximately 6 % of the E. coli exportome (Berks et al., 2005,

Lee et al., 2006b, Palmer & Berks, 2012, Palmer et al., 2005). Evaluation of the current list of substrates (Table 1.1) reveals that the Tat system is utilised for transport of many proteins containing metal cofactors unavailable in the periplasm. In addition this specialised system enables transport of hetero-oligomeric protein complexes and helps proteins avoid metal ions that may be in higher abundance and so compete with required cofactors for insertion into their active site during maturation (Tottey et al., 2008). 25

Table 1.1: The thirty seven known and predicted Tat substrates of E. coli K-12. In some cases a Tat substrate does not have its own signal peptide, instead being exported as a complex with a signal peptide-containing partner. The signal peptide each substrate uses for export is listed, along with the signal peptide sequence. Assembly chaperones often aid insertion of bound cofactors, and for each substrate for which this is the case these are also shown below. (Bendtsen et al., 2005, Berks et al., 2005, Thomas et al., 1999, Tullman-Ercek et al., 2007).

The thirty seven known or predicted Tat substrates in E. coli K-12. Protein Description Signal peptide Bound Potential Tat Signal peptide sequence used for export cofactors(s) assembly chaperone(s) Proteins with complex multiatom cofactors DmsA DMSO reductase α- DmsA MGD, [4Fe-4S] DmsD MKTKIPDAVLAAEVSRRGLVKTTAIGGLAMASSALTLPFSRIAHA subunit DmsB DMSO reductase β- DmsA 4 x [4Fe-4S] DmsD MKTKIPDAVLAAEVSRRGLVKTTAIGGLAMASSALTLPFSRIAHA subunit FdnG Nitrate-inducible formate FdnG MGD, [4Fe-4S] FdhE MDVSRRQFFKICAGGMAGTTVAALGFAPKQALAQA dehydrogenase subunit FdnH Nitrate-inducible formate FdnG 4 x [4Fe-4S] FdhE MDVSRRQFFKICAGGMAGTTVAALGFAPKQALAQA dehydrogenase subunit FdoG Formate dehydrogenase- FdoG MGD, [4Fe-4S] FdhE MQVSRRQFFKICAGGMAGTTAAALGFAPSVALA O α-subunit FdoH Formate dehydrogenase- FdoG 4 x [4Fe-4S] FdhE MQVSRRQFFKICAGGMAGTTAAALGFAPSVALA O β-subunit HyaA Hydrogenase-1 small HyaA 2 x [4Fe-4S], HyaE, HyaF MNNEETFYQAMRRQGVTRRSFLKYCSLAATSLGLGAGMAPKIAWA subunit [3Fe-4S] HyaB Hydrogenase-1 large HyaA [Ni- HyaE, HyaF MNNEETFYQAMRRQGVTRRSFLKYCSLAATSLGLGAGMAPKIAWA subunit Fe(CN)2CO] HybA Electron transfer to HybA 4 x [4Fe-4S] - MNRRNFIKAASCGALLTGALPSVSHAAA hydrogenase-2

26

HybC Hydrogenase-2 large HybO [Ni- HybE MTGDNTLIHSHGINRRDFMKLCAALAATMGLSSKAAA subunit Fe(CN)2CO] HybO Hydrogenase-2 small HybO 2 x [4Fe-4S], HybE MTGDNTLIHSHGINRRDFMKLCAALAATMGLSSKAAA subunit [3Fe-4S] NapA Periplasmic nitrate NapA MGD, [4Fe- NapD MKLSRRSFMKANAVAAAAAAAGLSVPGVARAVVGQQ reductase 4S] NapG Electron transfer to NapG 4 x [4Fe-4S] - MSRSAKPQNGRRRFLRDVVRTAGGLAAVGVALGLQQQTARA periplasmic nitrate reductase NrfC Electron transfer to NrfC 4 x [4Fe-4S] - MTWSRRQFLTGVGVLAAVSGTAGRVVA periplasmic nitrite reductase TorA TMAO reductase TorA MGD TorD MNNNDLFQASRRRFLAQLGGLTVAGMLGPSLLTPRRATA TorZ TorA homologue TorZ MGD TorD/YcdY MTLTRREFIKHSGIAAGALVVTSAAPLPAWA YagR Xanthine dehydrogenase YagT MCD? YagQ MSNQGEYPEDNRVGKHEPHDLSLTRRDLIKVSAATAATA family domain homologue YagS Xanthine dehydrogenase YagT FAD YagQ MSNQGEYPEDNRVGKHEPHDLSLTRRDLIKVSAATAATA family domain homologue YagT Xanthine dehydrogenase YagT 2 x [2Fe-2S] YagQ MSNQGEYPEDNRVGKHEPHDLSLTRRDLIKVSAATAATA family domain homologue YdhX Iron-sulfur protein YdhX 4 x [4Fe-4S] - MSFTRRKFVLGMGTVIFFTGSASSLLA YedY Sulfite oxidase family YedY Molybdopterin - MKKNQFLKESDVTAESVFFMKRRQVLKALGISATALSLPHAAHA homologue YnfE Selenate reductase YnfE MGD, [4Fe- DmsD MSKNERMVGISRRTLVKSTAIGSLALAAGGFSLPFTLRNAAA 4S] YnfF Selenate reductase YnfF MGD, [4Fe- DmsD MMKIHTTEALMKAEISRRSLMKTSALGSLALASSAFTLPFSQMVRA 4S] YnfG DmsB homologue YnfE/YnfF 4 x [4Fe-4S] DmsD MSKNERMVGISRRTLVKSTAIGSLALAAGGFSLPFTLRNAAA

27

Proteins that appear to lack complex multiatom cofactors AmiA N-acetylmuramoyl-l- AmiA Metal ions? - MSTFKPLKTLTSRRQVLKAGLAALTLSGMSQAIA alanine amidase AmiC N-acetylmuramoyl-l- AmiC Metal ions? - MSGSNTAISRRRLLQGAGAMWLLSVSQVSLA alanine amidase CueO Multicopper oxidase CueO Cu ions - MQRRDFLKYSVALGVASALPLWSRAVFA linked to copper resistance FhuD Ferrichrome-binding FhuD Fe(III) - MSGLPLISRRRLLTAMALSPLLWQMNTAHAAA protein hydroxamates MdoD Periplasmic glucan MdoD ? - MDRRRFIKGSMAMAAVCGTSGIASLFSQAAFA (OpgD) biosynthesis SufI Multicopy suppressor of SufI None? - MSLSRRQFIQASGIALCAGAVPLKASA an ftsI mutation YaeI Probable metallo- YaeI Metal ions? - MISRRRFLQATAATIATSSGFGYMHYCEPGWFELIRH phosphoesterase RLAFFKDNAAPFKILFLADLHYSRFVPLSLISDAIA

YahJ Probable metallo- YahJ Metal ions? - MKESNSRREFLSQSGKMVTAAALFGTSVPLAHA hydrolase YcbK Unknown YcbK ? - MDKFDANRRKLLALGGVALGAAILPTPAFA YcdB Unknown; structural gene YcdB ? - MQYKDENGVNEPSRRRLLKVIGALALAGSCPVAHA overlaps that of YcdO YcdO Unknown; structural gene YcdO ? - MTINFRRNALQLSVAALFSSAFMANA overlaps that of YcdB WcaM Coded in colonic acid WcaM ? - MPFKKLSRRTFLTASSALAFLHTPFARA biosynthetic gene cluster; possible lyase YfhG Unknown YfhG ? - MRHIFQRLLPRRLWLAGLPCLALLG 28

There are also a number of cofactor-less Tat substrates, requiring transport via Tat rather than

Sec perhaps due to folding taking place too fast after leaving the ribosome to allow transport by Sec, or for those whose activity depends upon folding in the cytoplasm to be biologically active, such as has been observed with TorA signal peptide-GFP fusion protein (Feilmeier et al.,

2000, Santini et al., 2001, Thomas et al., 2001). The majority of Tat substrates have non- covalently bound metal cofactors, and by inserting these prior to transport the need for additional mechanisms to transport cofactors and additional accessory proteins required for cofactor insertion is removed.

One interesting way by which proteins can be translocated through the Tat system without requiring a signal peptide of their own is through a mechanism term ‘hitchhiking’ (Rodrigue et al., 1999). This involves complex formation of the signal peptide-less substrate with a partner protein that does have a Tat signal peptide to direct the complex to the Tat translocase.

Processing and transport of these proteins relies upon them being folded and able to form a complex, and therefore for these proteins folding in the cytoplasm is essential.

Another challenge some bacteria may use the Tat system to overcome is the environmental conditions in which they live. For example, halophilic archaea have a greatly increased number of Tat substrates when compared to E. coli. This is likely because of the high extracellular salt conditions in which they exist creating difficulties in protein folding outside of the cytoplasm without the presence of additional chaperone proteins. Because of these hostile conditions it appears that these species have adapted and transport a much larger proportion of proteins that include lipoproteins and other non-redox proteins through Tat than bacteria living in more accommodating environments (Dilks et al., 2003, Rose et al., 2002, Storf et al., 2010).

Of the 27 Tat substrates in E. coli, 16 contain metal cofactors, most commonly Fe-S clusters and the molybdenum based cofactor molybdopterin guanine dinucleotide, also termed guanylyl molybdenum cofactor (MGD) (Table 1.1). These substrates play an important role in metabolism of the cell, specifically catalysing redox reactions during anaerobic respiration. For 29 example, the DMSO reductase family are involved in transferring electrons from quinols to terminal electron acceptors, whilst hydrogenases couple hydrogen oxidation with quinone reduction and transfer electrons to the quinone pool. Other substrates of Tat are known to be involved in completely different roles, in cellular process such as formation of biofilms, cell wall metabolism and iron acquisition (Ize et al., 2004, Ize et al., 2003, Stanley et al., 2001).

A number of Tat substrates have been exploited as reporters of Tat functionality. For example, the cell wall amidases AmiA and AmiC were identified after characterization of a defective outer membrane phenotype displayed by E. coli tat mutants. Mutant cells were shown to be defective in cell separation following division (the cells form long chains) and also displayed leaky or permeable envelopes, rendering them sensitive to growth in the presence of SDS (Ize et al., 2003). The Tat-dependent cell wall amidase proteins are involved in cleavage of amide bonds from the stem peptide in peptidoglycan, and therefore play an important role in cell separation after cell division (Vollmer et al., 2008b). SDS sensitivity is now commonly used as a reporter of Tat function.

1.1.7.8 The translocation cycle

When inactive, the Tat translocase is unassembled and found as two complexes in the membrane. These are TatA, thought to exist under resting conditions as a tetramer, and the

TatBC complex, comprising approximately 7 copies of each protein. During Tat system operation, up to four steps have been defined: (1) Tat signal-peptide recognition; (2) protein transport; (3) proton anti-transport; and (4) Tat signal peptide cleavage (Figure 1.11).

30

Figure 1.11: The cycle of protein translocation through Tat. The Tat translocase is made up of three main proteins, TatA, TatB and TatC. Initiation of the translocation cycle occurs upon interaction of the signal peptide of a fully folded Tat substrate with the TatBC complex. 1. Signal peptide recognition by the TatBC complex. 2. Upon substrate binding to TatBC, TatA polymerisation occurs, forming a pore across the cytoplasmic membrane of the cell. 3. The substrate is then able to cross the membrane through the TatA channel. 4. After translocation, the signal peptide is cleaved by a signal peptidase to leave the substrate free in the periplasm. TatA subunits then dissociate from TatBC and depolymerize back to free protomers.

Dynamic changes in the topology of the Tat machinery have been observed, thought to have a role in the translocation of proteins across the cytoplasmic membrane (Gouffi et al., 2004). It has been hypothesized that protein translocation via the Tat pathway occurs by a cyclical mechanism, with initiation of transport involving primary sequence recognition of the 31 substrate protein by TatC and TatB (Figure 1.12 & Figure 1.11) (Alami et al., 2003, Berks et al.,

2014, Cline & Mori, 2001, Gérard & Cline, 2006, Xiong et al., 2007).

Figure 1.12: Proposed interaction sites on TatC. Surface representation of TatC from Aquifex aeolicus as solved by Rollauer et al. (2012). Shown in red are two areas of highly conserved surface residues. The conserved residue patch on the cytoplasmic face of the protein is predicted to be the binding site for Tat substrate signal peptides, interacting with their n-regions. The other conserved region near the periplasmic side of the membrane is proposed to be involved in interactions with the transmembrane helix of TatB. The position of this helix upon interactions with TatC is indicated by the semi- transparent cylinder. Figure taken from Berks et al. (2014).

The TatC component of the complex is essential and sufficient for initial signal peptide binding, whilst TatB becomes involved later during transfer of the substrate from the TatBC complex to the TatA pore. Interestingly, TatB and TatC proteins appear to interact with signal peptides in different ways, with TatC being involved in a specific interaction localised to the area of the RR motif whilst TatB is capable of recognising a much longer stretch of the peptide encompassing the RR motif and hydrophobic region (Alami et al., 2003). It is possible that TatB acts as an intermediate between initial signal peptide binding and translocation, aiding in movement of the substrate to the translocation pore. It has been suggested that TatB helps to prevent premature signal peptide cleavage during translocation (Fröbel et al., 2012). There is no 32 requirement for TatA during this step in translocation, as binding of substrates to the TatBC complex in the absence of TatA has been observed in both bacterial and plant Tat systems

(Cline & Mori, 2001, Tarry et al., 2009b). Three dimensional maps for TatBC-substrate complexes show that only one or occasionally two substrates are bound per TatBC complex. It was also noticed that on the event of two substrates binding, they would bind to TatBC subunits adjacent to one another, consistent with the idea that one subunit of the TatBC complex is in a high affinity state compared to all others within the complex. Indeed, the binding of one substrate induces structural changes in the complex that do not change further upon the addition of a second substrate (Tarry et al., 2009b). This suggests that secondary substrate binding to TatBC is not mechanistically relevant, and that the structural modification may be involved in substrate affinity.

Once the substrate is bound to the TatBC complex, TatA then interacts with this TatBC- substrate complex, stabilizing TatB and leading to the formation of an active translocon (Figure

1.11) (Mangels et al., 2005). This is a transient interaction that is dependent upon the transmembrane electrochemical gradient across the cytoplasmic membrane; whilst this gradient is maintained translocation of substrates is allowed to occur (Alami et al., 2003,

Bogsch et al., 1998, Dabney-Smith et al., 2006, Mori & Cline, 2001, Yahr & Wickner, 2001).

TatA polymerisation has been visualised using TatA-YFP fusion proteins, a process that has been shown to require TatB and TatC (Leake et al., 2008). Interactions of the substrate with

TatA are thought to occur in a membrane potential-dependent manner, with experiments using inside-out cytoplasmic membrane vesicles showing that removal of this H+ gradient appeared to eradicate any detectable substrate-TatA interactions (Alami et al., 2003). It has been suggested that translocation is able to occur through the creation of membrane distortions, caused by a combination of the TatA pore itself and also TatC, that aid to weaken the membrane and eventually allow for temporary membrane rupture (Rodriguez et al., 2013,

Rollauer et al., 2012). As the substrate crosses the membrane pore the signal peptide remains bound to the TatBC complex (Panahandeh et al., 2008). It is unknown whether this part of the 33 translocation cycle requires the proton motive force. Following translocation the signal peptide is cleaved at the AxA site within the C-region by the enzyme signal peptidase LepB

(Lüke et al., 2009), with the Tat components then going on to dissociate. Dissociation of TatA protomers may be induced simultaneously with the substrate passing through the TatA pore due to the predicted membrane rupturing involved in translocation (Rodriguez et al., 2013).

The entire translocation process can be completed in a matter of minutes, and it is thought that there are at least two steps within this cycle that require the electrical potential (Δψ) of the proton motive force (PMF), although the exact nature of the energetic driving force for this system remains to be determined (Bageshwar & Musser, 2007, Whitaker et al., 2012).

The Tat translocase is a highly specialised system, highly selective in that it only exports fully folded proteins across the membrane. Therefore, it is of great importance that bacterial cells should have control over proteins leaving the cytoplasm through Tat so as to prevent wasteful translocation of immature proteins, as well as avoidance of competition between fully folded and immature proteins at the Tat complex (Jack et al., 2004). The Tat system has two mechanisms by which such events are prevented, namely ‘quality control’ and ‘Tat proofreading’ (Figure 1.13).

1.1.7.9 Tat quality control

The Tat ‘quality control’ mechanism is suggested to be an inherent property of the TatABCE components within the membrane. Proteins that are incorrectly folded or unable to bind cofactors are thought to remain in the cell cytoplasm and become degraded (Figure 1.13)

(DeLisa et al., 2003, Matos et al., 2008). An initial study by (DeLisa et al., 2003) used Tat signal peptide fusions to the enzyme alkaline phosphatase (PhoA), an enzyme dependent upon disulphide bond formation under oxidising conditions for activity, in conjunction with bacteria manipulated to have an oxidising cytoplasm to investigate Tat transport of folded and unfolded proteins. It was found that PhoA was only active and exported in the strain 34 engineered to have an oxidising cytoplasm whilst in strains with a reducing cytoplasm (and therefore unfolded PhoA) the inactive PhoA fusions were found localised to the cytoplasm.

From this it was concluded that the Tat translocase must be capable of detecting the folded state of the substrate prior to transport. This work was followed up by investigations using a clever reporter system using export of a family of proteins displaying various levels of protein folding ranging from unfolded and molten globules to well-ordered stable proteins (Rocco et al., 2012). It was noted that whilst those proteins exhibiting unfolded and molten globule states remained in the cytoplasm of cells, those displaying native-like stable structures were transported to the periplasm in a Tat-dependent manner. This result fits in with the previous hypothesis that the Tat translocase is involved in structural proofreading.

35

Figure 1.13: The control mechanisms of Tat in place to prevent premature translocation. 1. Cytoplasmic chaperone proteins bind to their substrate partners via the Tat signal peptide and occasionally on a region of the apoprotein itself. 2. Cofactor insertion occurs. 3. Completion of cofactor insertion triggers the release of the substrate. 4a. The substrate goes on to interact with the Tat system or 4b. the substrate is subjected to quality control.

Using this reporter system a number of TatABC mutants were identified that supress this apparent proofreading function and were able to transport these unfolded proteins that before remained in the cytoplasm, as well as the aforementioned PhoA. Remarkably, a number of these mutants were identified as clustering in TatB and TatC proteins, suggesting a secondary role of the signal peptide recognition complex. A role of the TatBC complex in folded state sensing has also been suggested in other work (Panahandeh et al., 2008)

However, it must also be mentioned that whilst amino acid substitutions such as TatC F94, P97 36 and E103 imparted a gain of function phenotype in this study, they have previously been identified as disrupting export (Buchanan et al., 2002). It has been suggested that contact of

TatBC with exposed hydrophobic patches on unfolded proteins causes termination of Tat transport (Richter et al., 2007). Conflicting evidence has implied that turnover of malformed proteins by quality control is thought to be a function of TatA and TatE specifically, as a double knockout mutant of these Tat constituents in E. coli resulted in a lack of degradation ability

(Matos et al., 2008). Recent stochastic modelling studies published have suggested that specific ‘quality control’ does not exist, with an alternative passive model being proposed where substrate-dependent association and disassociation rates instead play a role in translocation (Chitra et al., 2014).

1.1.7.10 Tat proofreading

In addition to quality control, it is apparent that a large proportion of Tat substrates are subjected to control by cytoplasmic chaperone proteins during maturation (Figure 1.13), thereby undergoing a two-tier system of transport regulation. This second mechanism by which premature translocation is prevented is known as ‘Tat proofreading’. This exists earlier in the protein biosynthesis pathway, and involves cytoplasmic substrate-specific chaperone proteins, acting to prevent targeting to the Tat complex before the substrate is fully matured

(Dubini & Sargent, 2003b, Jack et al., 2004, Sargent, 2007a). These chaperone proteins are substrate specific and play a number of important roles in substrate maturation, including binding and masking of the signal peptide, cofactor(s) insertion, protection from protease degradation (Turner et al., 2004), and oligomer assembly (Berks et al., 2005, Santini et al.,

1998, Sargent, 2007a). Cofactor-containing proteins are often monitored by this control mechanism. For example, the maturation of the Tat-dependent periplasmic redox enzyme trimethylamine N-oxide reductase (TorA), a molybdenum binding protein, is monitored by its chaperone protein TorD (Hatzixanthis et al., 2005, Jack et al., 2004, Matos et al., 2008). It has 37 been suggested that these chaperones escort their partner substrate proteins to the Tat translocase, with interactions of these chaperones with Tat thought to have been observed

(Papish et al., 2003, Turner et al., 2004), although the requirements of chaperones for Tat interactions has been disputed (Ray et al., 2003).

These chaperone proteins have become well-studied proteins, with a large number having been identified either through experimental data displaying interactions with known Tat substrates or prediction through sequence analysis and homology (Pommier et al., 1998,

Redelberger et al., 2011) . These chaperone proteins are highly important in the biogenesis of their substrates.

1.1.7.11 The TorD family of chaperone proteins

As already mentioned, a considerable number of Tat substrates contain molybdenum as a key catalytic cofactor at the active site. Most of these enzymes contain the common form of molybdenum cofactor (MoCo) known as the bis-molybdopterin guanine dinucleotide (Mo- bisMGD). The coordination of these molybdoenzyme active sites, alongside phylogenetic studies, has led to the classification of these bis-MGD-containing enzymes into three types:

Type I, Type II and Type III. Type I bis-MGD enzymes normally have a cysteine or selenocysteine sidechain as a direct ligand to molybdenum, and comprises enzymes such as E. coli FdnG and

NapA (periplasmic nitrate reductase) (Jormakka et al., 2004, McDevitt et al., 2002, Sargent,

2007a), and Salmonella TtrA (tetrathionate reductase) (Hensel et al., 1999). Type II bis-MGD- containing enzymes have an aspartate molybdenum ligand, and include E. coli NarG plus its many homologues (Jormakka et al., 2004), Rhodovulum sulfidophilum DdhA (the catalytic subunit of dimethyl sulphide dehydrogenase) (McDevitt et al., 2002) and Thauera selenatis

SerA (selenate reductase) (Dridge et al., 2007). The Type III bis-MGD group display a serine ligand to molybdenum, including E. coli TMAO reductases TorA and TorZ and DmsA (DMSO 38 reductase), plus TMAO and DMSO reductases of Rhodobacter sp. and BisC (biotin sulfoxide reductase) (McDevitt et al., 2002, Sargent, 2007a).

A great number of Type II and Type III molybdoenzymes interact with cytoplasmic chaperone proteins during their maturation. These important chaperone proteins all display the same general structure, and comprise a large protein family named the TorD family. This protein family can be split into three clades: the TorD, the DmsD/YcdY clade and the NarJ clade (Figure

1.14) (Turner et al., 2004). Although structurally these proteins are very similar, they are all highly specific to their substrates (Ilbert et al., 2004).

Figure 1.14: The TorD family of cytoplasmic chaperone proteins. Phylogenetic tree of the TorD family of chaperones, reproduced from (Turner et al., 2004). This family is separated into three clades, namely the NarJ, TorD and YcdY and DmsD clades.

In E. coli there have been five TorD-like chaperones identified to date: TorD, responsible for

TorA assembly (Ilbert et al., 2003, Pommier et al., 1998); DmsD, unusual in that it is involved in 39 maturation of three Tat substrates DmsA, YnfE and YnfF (Guymer et al., 2009, Oresnik et al.,

2001); NarJ and NarW, homologous proteins required for maturation of the nitrate reductases

NarG and NarZ respectively; and most recently identified YcdY, predicted to be involved in maturation of YcdX, a zinc-containing protein with a putative involvement in bacterial cell swarming (Redelberger et al., 2011). A number of specific signature sequence motifs have been identified amongst these proteins, with most appearing to be confined to a specific clade. The existence of a highly conserved ‘E(Q)PxDH’ motif across all sub-domains of the TorD family was also identified (Turner et al., 2004).

1.1.7.12 The role of TorD chaperones

The chaperone DmsD was discovered to bind tightly to DmsA in affinity chromatography experiments, and from this work was deemed to be a signal peptide binding chaperone protein

(Oresnik et al., 2001). DmsA is a Tat-targeted protein forming the active subunit of the heterotrimeric molybdoenzyme DMSO reductase (Bilous & Weiner, 1988). DMSO reductase is encoded by the dmsABC operon, where DmsA contains the MGD cofactor and DmsB has four

[4Fe-4S] clusters (Rothery et al., 1995, Sambasivarao et al., 1990, Sambasivarao & Weiner,

1991). It has been shown that, although essential for DMSO reductase activity and despite interactions with the signal peptide being observed (Oresnik et al., 2001), DmsD is not required for correct targeting of the DmsA signal peptide to the Tat translocase and therefore must be involved instead in enzyme biogenesis (Ray et al., 2003). Specific interactions of DmsD with the hydrophobic region of the DmsA signal peptide have been observed (Winstone et al., 2013a,

Winstone et al., 2006). It may be that DmsD exerts a transport blocking effect during binding as well as being involved in enzyme maturation. More recently it has been found that in addition to DmsA proofreading, DmsD is also involved in interactions with an enzyme responsible for selenate reduction in both E. coli and S. Typhimurium encoded by the operon ynfFEGHI; interestingly DmsD is in fact encoded by the gene ynfI in this operon (Guymer et al., 40

2009, Lubitz & Weiner, 2003, Oresnik et al., 2001). YnfE, the catalytic subunit of selenate reductase, is similar to DmsA in that it contains an MGD cofactor, and this enzyme is described in further detail in Section 2.1.

It was the identification of DmsD as a cytoplasmic chaperone that led to the suggestion of, and investigations into, the possibility of TorD belonging to the same family of chaperone proteins.

TorD shares sequence homology with DmsD and both chaperone proteins have been found to contain (Y/F/W)xxLF and E(PX or XP)D(H/Y) motifs common to members of the DmsD family clade (Turner et al., 2004). This chaperone protein has been identified as playing a role in the assembly of TorA, the active subunit of the molybdoenzyme TMAO reductase (Ilbert et al.,

2003, Jack et al., 2004, Pommier et al., 1998). Recent studies into TorD interactions with TorA have clearly identified two binding sites for TorD, one on the signal peptide and another on the

‘mature’ region of the TorA apoprotein (Dow et al., 2013, Jack et al., 2004). Although not essential for TorA export, TorD is thought to interact prior to cofactor loading and aid this maturation process by maintaining apo-TorA in the ideal conformation for insertion (Ilbert et al., 2003, Pommier et al., 1998). Remarkably, using TorA signal peptide fusions to the hydrogenase 2 subunit HybO (ssTorA::HybO) it was shown that interactions with TorD were maintained, with TorD appearing to restore correct targeting of HybO and its partner protein

HybC, suggesting that TorD was able to perform Tat proofreading functions on a non-native substrate (hydrogenase 2) closely enough to rescue assembly of a “non-molybdo” enzyme

(Jack et al., 2004). This result suggests TorD function is divorced, at least in part, from cofactor loading since hydrogenase 2 does not have the same cofactor as TorA. TorD also appears to bind GTP, with binding increasing when in the presence of the TorA signal peptide leading to suggestions that it may bind TorA in a GTP-dependent manner (Hatzixanthis et al., 2005). Later evidence to support this idea was provided by the findings of (Guymer et al., 2010) that TorD in its dimeric form was indeed able to hydrolyse GTP, albeit at a very low level. 41

In vitro characterisation of the ssTorA-TorD interaction was performed using isothermal titration calorimetry. Interactions of TorD to a ssTorA peptide encompassing residues 10-36 generated negative enthalpy values and a sigmoidal binding curve that reached a clear saturation point, giving an initial reported apparent dissociation constant (Kd) of around 1 µM.

Binding to a shorter peptide of ssTorA2-22 resulted in a considerably weaker Kd of 192 µM. This is likely due to the requirement of secondary structural features of the signal peptide that would not be present in an aqueous environment for the correct interactions with TorD to occur (Hatzixanthis et al., 2005, Sargent, 2007a). Subsequent work with maltose binding protein fusions suggested an apparent Kd of 59 nM for the ssTorA-TorD interaction (Buchanan et al., 2008). It is thought that Tat signal peptides adopt a helical structure from the twin- arginine motif onwards that would be relevant for chaperone protein binding (San Miguel et al., 2003), and this theory has been further confirmed by NMR analysis of signal peptides from

E. coli NarJ and the phosphodiesterase PhoA of Bacillus subtilis, where results of both studies suggested the formation of helical structures (Klein et al., 2012, Zakian et al., 2010).

As well as binding the signal peptide of their substrate partners TorD family chaperones have also been shown to interact with the ‘mature’ region of the apoprotein, for example a truncated version of NarG missing its first 40 residues is still able to interact with NarJ in bacterial two-hybrid studies (Vergnes et al., 2006). This work complements previous data using the TorA/TorD pair, where TorD is shown to still interact with mature TorA protein (therefore lacking its signal peptide) forced to unfold by heat treatment (Pommier et al., 1998).

An interesting fact to be mentioned is that not all chaperone proteins in the TorD family aid maturation of proteins that go on to be transported out of the cytoplasm by the Tat system. In fact, the previously mentioned NarG, chaperoned by NarJ, is part of the multi-subunit nitrate reductase that functions in the cytoplasm (Blasco et al., 1998). In experiments using NarG and chaperone NarJ it has been shown that NarJ maintains NarG in a soluble state for cofactor insertion (Vergnes et al., 2006), and it is thought that proteins such as this have remnant signal 42 peptides that are still capable of being recognised by their chaperone proteins. Sequence analysis suggests that they do have an N-terminal signal sequence related to the twin-arginine motif, with the major difference being that the second arginine has been replaced by an aromatic residue. This hints at the possibility that these enzymes were once periplasmic, a theory that has been suggested in the past (Turner et al., 2004). Substitutions of certain key residues in these remnant signal peptides has been shown to ‘reactivate’ them, once again leading to their redirection to and transport via the Tat system (Ize et al., 2009b). In addition, the first 15 amino acids of NarG adopts an α-helical conformation in solution, as is the case for genuine signal peptides (Zakian et al., 2010).

1.1.7.13 TorD chaperone structures

The 3D structures of a number of TorD family proteins have been solved so far, and all appear to show high structural homology despite relatively low levels of similarity on a sequence level of around 20 % for most proteins (Figure 1.15) (Ilbert et al., 2004). Initially through structure prediction it was thought that these proteins displayed a high level of α-helixes, and this was later confirmed for a number of chaperones using circular dichroism spectroscopy (Genest et al., 2008, Ilbert et al., 2004, Sarfo et al., 2004, Tranier et al., 2002). 3D crystal structures of several TorD chaperone proteins have been solved to date: DmsD from E. coli and Salmonella

Typhimurium, TorD from Shewanella massilia, and NarJ and TtrD, both from Archeoglobus fulgidus (Coulthurst et al., 2012, Kirillova et al., 2007, Ramasamy & Clemons Jr, 2009, Tranier et al., 2003). These proteins all share an all α-helical structure.

43

Figure 1.15: 3D crystal structures of TorD family proteins show conserved α-helical structure. A – S. massilia TorD dimer (1N1C). B – E. coli DmsD (3CW0). C – S. Typhimurium DmsD (1S9U). D – A. fulgidus NarJ (2o9x). E – A. fulgidus TtrD (2XOL).

Whilst TorD family proteins are primarily monomeric, dimers and higher oligomeric forms have also been identified. For example purified DmsD has been observed as a number of folding forms when analysed by Western blot (Sarfo et al., 2004). In addition TorD has been found in various oligomeric forms that are variable depending upon the acidity of the environment

(Tranier et al., 2002); indeed the crystal structure of TorD from S. massilia was solved as a dimer displaying a unique structure with extreme domain swapping (Tranier et al., 2003). TorD has been shown to have chaperone capabilities as both a monomer and a dimer (Tranier et al.,

2002), however due to the lack of other TorD family chaperone 3D crystal structures in dimeric form doubts have been raised over the relevance of this protein dimer.

1.1.7.14 Other Tat proofreading chaperones

Whilst the TorD family of chaperones cover assembly of type II and type III molybdoenzymes, additional chaperones must be involved in similar processes with type I molybdoenzymes as 44 well as FeS cluster proteins. These chaperones are important in the maturation of Tat substrates such as NapA (nitrate reductase), Hyd-1 and Hyd-2 (hydrogenases), and FdnG and

FdoG (formate dehydrogenases) (Sargent, 2007a).

E. coli periplasmic nitrate reductase, the catalytic subunit of which is Tat substrate NapA, is encoded in the operon napFDAGHBC. NapA contains a Mo-bis-MGD cofactor plus a [4e-4S] cluster (Jepson et al., 2007, Potter & Cole, 1999), and has been shown to reach its functional location in the periplasm through the Tat system due to the presence of an N-terminal signal peptide (Berks et al., 2005, Berks et al., 1995). NapD, one of the essential protein products of this operon predicted to be cytoplasmic in its localisation, was shown to be required for the production of active NapA hinting at a role in post-translational modifications (Berks et al.,

1995, Potter & Cole, 1999). Later work identified that NapD was able to interact with the signal peptide of NapA in vivo and also in vitro with a NapA fusion to MalE (Maillard et al., 2007), with the signal peptide binding epitope appearing to localise within the n-region of the sequence (Grahl et al., 2012). Through the use of high-resolution heteronuclear NMR the solution structure of NapD has been determined, and showed that the protein consists of two

α-helices and four β-strands, adopting a β-α-β-β-α-β ferredoxin like fold where the β-strands form one face as an antiparallel β-sheet and the α-helices form the opposite face (Figure 1.16)

(Maillard et al., 2007). This structure is indeed dissimilar from the TorD family all α-helical structure, highlighting differences between these families of chaperone proteins.

45

Figure 1.16: Solution NMR structure of E. coli NapD. NapD (2JSX) has a ferredoxin fold, with two α-helices and four β-strands forming opposite faces of the molecule (Maillard et al., 2007).

In E. coli, the membrane bound formate dehydrogenase isoenzymes formate dehydrogenase-N and formate dehydrogenase-O, consist of FdnGHI and FdoGHI, respectively. These enzymes require a functional Tat system for transport, and are exported as heterodimers using the Tat hitchhiker mechanism where only one subunit contains a signal peptide (Berks et al., 2005,

Sargent, 2007a, Stanley et al., 2002). FdnG and FdoG each contain the Mo-bis-MGD cofactor and a single [4Fe-4S] cluster as well as the Tat signal peptide for the complex, and associate with their respective FdnH and FdoH partners, containing 4[4Fe-4S] clusters, prior to transport.

The chaperone protein FdhE is essential for activity of both enzymes, and interactions have been shown between this chaperone and FdnG and FdoG (Lüke et al., 2008, Mandrand-

Berthelot et al., 1988 ). Differing from TorD-like chaperone proteins, there is no evidence for

FdhE binding to either formate dehydrogenase signal peptide, and therefore it is likely that these chaperones participate in a different mechanism of proofreading (Lüke et al., 2008).

1.1.8 General chaperones involved with Tat substrates

In addition to the many specific Tat chaperones involved in the Tat pathway, it is also known that a number of general chaperone proteins also appear to be involved in Tat substrate 46 assembly. Originally it was in the homologous Tat system in plant thylakoids that general chaperone proteins were identified as playing a role. Here chaperones Cpn60 and Hsp70 were identified in complex with the Rieske protein prior to transport, with the suggestion that they helped prevention of formation of assembly incompetent forms of the Rieske protein

(Madueño et al., 1993, Molik et al., 2001).

Evidence for a similar role of general chaperones in bacteria is perhaps less robust (Fisher &

DeLisa, 2004). For example, the trigger factor (TF) protein has been shown to interact with Tat substrates TorA and SufI via their signal peptides. However the deletion or overexpression of

TF did not alter Tat export of these substrates (Jong et al., 2004). GroEL appears to influence hydrogenase-1 assembly but not that of hydrogenase-2 (Rodrigue et al., 1996), whilst general chaperones including DnaK, PspA and IbpA have been shown to aid stabilisation and export efficiency of Tat substrates (DeLisa et al., 2004, Pérez-Rodríguez et al., 2007). It has been hypothesised that these chaperones are important for stabilising Tat substrates that do not undergo Tat proofreading (Graubner et al., 2007). This would certainly comply with observations that these proteins bind to extended strands of hydrophobic residues (Rudiger et al., 1997, Young et al., 2004, Zhu et al., 1996) .

1.2 Respiratory capability of E. coli and Salmonella

Respiration is essential in almost all cells, and is the process by which the cell is energised through ATP synthesis. In animals and plants the presence of oxygen is essential for respiratory oxidative phosphorylation to generate ATP. However, bacteria such as E. coli display considerable flexibility in their respiratory capability, both in use of electron donors and acceptors, and are able to produce ATP in processes generally termed anaerobic respiration.

Bacterial respiratory chains begin by donation of electrons from low redox-potential electron donors to higher redox potential electron acceptors. This process continues with transfer of electrons through a range of successively higher redox potential molecules forming a 47 continuous electron transport chain. This process ends with the reduction of a terminal electron acceptor, and the free energy is conserved in the form of an electrochemical gradient

(i.e. the proton motive force – PMF) across the membrane. Oxygen is the most energetically favourable electron acceptor, with the highest redox potential of +0.82 V, and so if available this is always used as the terminal electron acceptor. In the absence of oxygen, anaerobic respiration can be undertaken, but is significantly less productive than aerobic respiration

(Richardson, 2000).

E. coli is also capable of producing a large number of structurally and functionally diverse redox enzymes and also has the ability to adjust expression levels of these enzymes depending upon the environmental conditions (Figure 1.17). Amongst these enzymes are 15 dehydrogenases, able to act as quinone reductases and the first step in anaerobic respiration. Reduced quinones, termed quinols, enter the quinone pool where they can go on to be oxidised by any one of 10 terminal reductases, for example DMSO reductase or TMAO reductase (Unden &

Bongaerts, 1997). The respiratory enzymes are highly variable in both the subunit composition and prosthetic groups. Many of these enzymes contain the molybdenum-based cofactor Mo- bis-MGD. The molybdenum atom is greatly utilised in respiratory enzymes due to its ability to cycle between two oxidation states of Mo(IV) and Mo(VI), enabling two electron redox reactions to be performed (Hille, 1996, Hille, 2002). Many of these enzymes are periplasmic in location and are substrates of the Tat system. 48

Figure 1.17: Respiratory pathways in E. coli. Figure adapted from (Sargent, 2007a). Electron flow is from left to right, from electron donors to electron acceptors via the membrane associated quinone pool of ubiquinone (UQ), menaquinone (MK) and demethylmenaquinone (DMK). Respiratory enzymes (blue boxes) are indicated by their gene names whilst electron donor/acceptor substrates are also indicated. Enzymes with a periplasmic active site are highlighted red, whilst those with a subunit containing a Mo-bis-MGD cofactor are indicated with an asterisk. Those known to require a Tat chaperone for correct assembly are outlined light blue.

1.2.1 Quinones and respiration

Bacterial electron transport chains consist of a number of dehydrogenases and terminal reductases, and these are linked together by 2-electron carriers known as quinones. There are three quinones in E. coli: ubiquinone (UQ), menaquinone (MK) and demethylmenaquinone

(DMK), with the redox potentials +0.11, -0.08 and +0.04 respectively (Unden & Bongaerts,

1997). Under aerobic growth conditions, UQ acts as the major quinone, with this major role switching during anaerobic growth to MK in the presence of fumarate or DMSO acceptors, as well as nitrate (Magalon et al., 1997, Rothery et al., 2001, Unden & Bongaerts, 1997). DMK 49 appears to be utilised in the presence of succinate (Sharma et al., 2012). The additional adaptability this offers also counts towards the great flexibility in bacterial respiration.

In facultative anaerobic bacteria such as E. coli, there is usually a specific hierarchy of electron acceptor usage. This is due to the differing energy yields of acceptors based on the redox potential so the reactions and the theoretical ATP yields when redox reactions are coupled together. Through the use of transcriptional regulators, E. coli ensures that the most energetically favourable acceptors available in the environment are rapidly preferentially utilised, thereby maximising growth under all circumstances (Unden & Bongaerts, 1997). In E. coli, transcriptional regulation is mainly influenced by the presence of O2 and nitrate, although there are other terminal electron acceptors, such fumarate and TMAO, that may also impart regulatory effects. O2 sensing regulators ArcA/B and FNR are responsible for O2 regulation.

ArcA is most significantly utilised under microaerobic conditions, whilst FNR is normally inactive during microaerobic conditions and plays a major role during strictly anaerobic conditions (Liu & De Wulf, 2004, Shalel-Levanon et al., 2005, Unden & Guest, 1985). In addition nitrate is a major regulator, detected by homologous sensor regulator pairs NarX/L and NarP/Q

(Unden & Bongaerts, 1997, Yamamoto et al., 2005).

ArcB/ArcA is a two-component system of signal transduction, involved in the sensing of oxygen availability and coordinating oxidative and fermentative catabolism in response. ArcB is a membrane-bound sensor kinase that becomes activated under anaerobic conditions and causes the transphosphorylation of ArcA, the cytoplasmic response regulator containing an unusual very short periplasmic section with two canonical transmembrane segments (Alvarez et al., 2013, Iuchi et al., 1990). It appears that the kinase ability of ArcB is silenced during aerobic growth by ubiquinone, and upon a shift to anaerobic growth becomes reactivated involving menaquinone (Alvarez et al., 2013, Bekker et al., 2010, Sharma et al., 2013). ArcA-P directly represses and activates a number of operons, and has been shown to have over 170 chromosomal binding sites in E. coli (Park et al., 2013b). These binding sites have been shown 50 to have at least two but up to five direct repeat sequences, each of 18 base pairs in length

(Park et al., 2013b). Through examination of genome-wide mRNA expression profiles it was revealed that ArcA is involved in the direct regulation of eighty five operons, repressing seventy four in its predominant role as a global repressor of carbon oxidation, and activating eleven. In addition to this, ArcA also appears to act as an indirect regulator of a further 135 operons (Park et al., 2013b). This massive range of regulation highlights the importance of this two-component system in control of gene expression.

The FNR (fumarate nitrate reduction) is a one-component regulatory system discovered through studies showing the mutations at a locus encompassing the fnr gene negatively affected the synthesis of nitrate, nitrite and fumarate reductases. This led to the finding that this protein has a strong DNA affinity and is involved in regulation of anaerobic gene expression (Lin & Iuchi, 1991, Unden & Guest, 1985). FNR is a member of the cAMP receptor protein (CRP)-FNR superfamily of homodimeric transcriptional regulators, which consist of an

N-terminal sensory domain and a C-terminal DNA-binding domain. The FNR regulator protein in an active state is dimeric, with each monomer containing a [4Fe-4S] cluster (Moore & Kiley,

2001). FNR undergoes a switch from monomer to dimer between unbound and bound forms, with the binding of the [4Fe-4S] cluster initiating dimerization (Crack et al., 2014). As oxygen levels increase, the [4Fe-4S] cluster rapidly converts to a [2Fe-2S] cluster, causing the reversal of FNR back to monomeric form lacking the ability to specifically bind DNA. The interconversion between [4Fe-4S] and [2Fe-2S] clusters and subsequent dimer-monomer interchange is thought to enable FNR to be highly responsive to changes in oxygen levels

(Crack et al., 2008, Zhang et al., 2012). The conversion of the [4Fe-4S] cluster to the [2Fe-2S] cluster in the presence of oxygen appears to be enhanced two-fold when FNR is bound to DNA.

This observation has led to suggestions that DNA-bound [4Fe-4S] FNR is the preferred target of

O2 over the non-DNA-bound equivalent, thereby increasing the sensitivity of the system to oxygen (Crack et al., 2014). FNR appears to modulate ArcA expression, and it has also been 51 suggested that both FNR and ArcA are required for full expression of ArcA (Compan & Touati,

1994).

Nitrate and nitrite are the preferred electron acceptors in anaerobic environments. In order to allow maximum utilisation, E. coli has two systems for regulation with nitrate. These are the two-component systems NarX/L and NarP/Q, which work in conjunction to control gene expression anaerobically. NarX and NarQ are the membrane sensor kinases with NarL and

NarP acting as the regulatory subunits. Each of the sensor proteins responds to both nitrate and nitrite, although the NarX/L system mainly acts as a nitrate regulator. NarX appears to be able to discriminate between ligands, with the protein acting primarily as a NarL kinase in response to nitrate, but as a phospho-NarL phosphatase in the presence of nitrite. This is different to the NarQ sensor, that is thought to acts in a similar fashion as a kinase in response to both nitrate and nitrite (Stewart & Bledsoe, 2003). These systems work in a similar fashion to the ArcAB system, with the sensor protein autophosphorylating and inducing transphosphorylation of the response regulator. The NarX/L and NarL/P systems function to both activate nitrate and nitrite catabolic genes whilst at the same time repressing expression of genes such as those of other anaerobic respiratory systems for example fumarate reductase

(Lin & Iuchi, 1991, Unden & Bongaerts, 1997). Interestingly, NarX/L and NarQ/P regulatory systems have been shown to demonstrate crosstalk, and are thought to form a cross- regulation network. This cross-regulation network appears to asymmetric in interactions, with the NarX sensor being more specialised for NarL phosphotransfer, whilst the NarQ sensor is more generalised and displays similar phosphotransferase activities whether with interactions with NarP or NarL (Noriega et al., 2010).

NarL and its paralogue NarP are known to consist of two domains, the N-terminal receiver domain and the C-terminal DNA-binding domain. Crystal structures for the C-terminal domain of E. coli NarL in complex with a DNA fragment show an extensive NarL-DNA contact surface

(Maris et al., 2005, Maris et al., 2002). Crystal structures of full-length and the receiver domain 52 of the putative NarL (Spr1814) of Streptococcus pneumonia have more recently been obtained

(Park et al., 2012, Park et al., 2013a). These Spr1814 structures show that the receiver domains appear to demonstrate reorganisation upon phosphorylation, leading to the suggestion that there is an intermediate conformation between inactive and active states (Park et al., 2013b).

Dimerization of NarL upon phosphorylation is thought to promote DNA binding and transcription activation, and has been observed in other response regulator proteins (Lewis et al., 2002, Maris et al., 2002, Park et al., 2013a). DNA binding is thought to be initiated upon phosphorylation through a conformational change in the hinge region of NarL, thought to separate the N-terminal receiver domain from the C-terminal DNA-binding domain, enabling transcription regulation by NarL to occur (Zhang et al., 2003). Regulation of DNA binding is thought to occur in a similar fashion in NarP (Lin & Stewart, 2010).

1.2.2 The mononuclear molybdenum enzymes

The mononuclear molybdenum enzyme family is a large class of enzymes that catalyse a great range of basic metabolic processes. The majority of these enzymes coordinates molybdenum at their active site, however a subset of these are also able to bind tungsten as an alternative.

Although these enzymes can utilise both molybdenum and tungsten in similar fashions, the majority of the characterisation of these enzymes to date has been performed using molybdenum-containing enzymes. These enzymes can be broadly separated into two large classes of enzyme based on the chemical reaction they catalyse, giving the hydroxylases and the oxotransferases. These families can be further classified according to the cofactor coordination at the active site, resulting in three subfamilies (Hille, 1996, Hille, 2002, Magalon et al., 2011). These families are: Family I of the molybdoenzymes is the xanthine oxidase family, typified by the xanthine oxidase of Rhodobacter capsulatus, coordinating a Moco with a single pterin molecule (Truglio et al.); Family II, the dimethyl sulphoxide reductase (DMSOR) family, is a large well characterised family widespread in prokaryotes that coordinates a Mo- 53 bis-MGD; and Family III, the sulphite oxidase family, that bind a single Mo-MPT cofactor (Hille,

2002). In addition to these three families is an archaeal family of aldehyde oxidoreductases.

These enzymes bind a tungsten-bis-MPT cofactor that is analogous to the molybdenum cofactor found in oxotransferases (Chan et al., 1995).

The majority of these enzymes catalyse the transfer of an oxygen atom coupled to electron transfer during which the metal ion switches between redox states IV and VI, whilst also using additional cofactors such as Fe-S clusters. The structures of the molybdenum centre of the enzyme forms a basis for the reaction type performed and enzyme family the enzyme belongs to. The molybdenum atom is usually coordinated to the sidechain of a cysteine, selenocysteine, serine or aspartate residue, or may involve terminal oxo, hydroxo or sulphide groups (Figure 1.18).

The first of these families, the xanthine dehydrogenase family, contains enzymes involved in hydroxylation of carbon centres. This family can be characterised by their oxidised

LMoVIOS(OH) core, with one equivalent of the pterin cofactor (Figure 1.18). Enzymes of the xanthine dehydrogenase family characteristically catalyse the hydroxylation of carbon centres.

VI The sulphite oxidase family contains an oxidised LMo O2(S-Cys) core, with a cysteine ligand to the Mo centre plus a single equivalent of the pterin cofactor (Figure 1.18). These catalyse transfer of an atom of oxygen to or from the substrate. Important to note here is the fact that these enzymes, unlike the other molybdoenzyme families, do not contain or interact with any

Fe-S clusters.

The largest family of enzymes is the DMSO reductase family, with both structure and function of these enzymes being much more diverse than other families, but in general they catalyse oxygen atom transfer. These enzymes contain two equivalents of the pterin cofactor bound to the centre, with the Mo-bis-MGD version of the molybdenum cofactor. The core consists of

VI L2Mo O(X), with four dithiolene sulphur atoms acting as ligands to the Mo atom in addition to 54 an oxo group. A sixth ligand is usually a serine, cysteine selenocysteine, hydroxide or water

(Figure 1.18) (Hille, 2002, Leimkühler, 2007, Magalon et al., 2011) .

Figure 1.18: Active site structures of molybdenum-containing enzymes. The three families of molybdoenzymes display different coordinations of their molybdocentres. Xanthine oxidase contains a single pterin bound plus additional coordination from oxygen, hydroxide and sulphur. The sulphite oxidase centre is similar but shows ligands donated by oxygen and sulphur from a cysteine. DMSO reductase are more diverse in nature, with binding of two pterins, oxygen, and coordination from an amino acid that can be either serine (DMSO reductase), cysteine (nitrate reductase), selenocysteine (formate dehydrogenase) or hydroxide (arsenite oxidase). Taken from (Hille, 2002). The biosynthesis and full structure of the molybdenum cofactor is shown in Figure 1.20 (Section 1.3.1).

In E. coli the majority of molybdoenzymes belong to the DMSO reductase family. All the enzymes in this family contain the Mo-bis-MGD dipteran cofactor, and all catalyse reactions following the general equation:

- + XO + 2e + 2H ↔ X + H2O

Although the general cofactor is the same throughout members of the family, differences in the Mo binding site do exist. These small variations are important in enzyme function and substrate specificity. Despite crystal structures of members of the DMSO reductase family show a high degree of homology in the overall fold, there are a great many subtle variations in active site structures. It appears that the amino acid involved in molybdenum atom coordination can vary amongst different enzymes, with cysteine, selenocysteine, aspartate and serine all having been observed binding Mo (Romao, 2009). These differences in complexity 55 allow for further subclassification of the DMSO reductase family (McEwan et al., 2002, Romao,

2009).

DMSO and TMAO reductase belong to a clade where the Mo atom is coordinated by both a terminal oxygen ligand from a serine residue found in the active site. Four sulphur atoms are also in interactions with the Mo, two from the dithiolene each MGD molecule. Whilst DMSO and TMAO reductase contain a single prosthetic group in the Mo-bis-MGD cofactor, Nap, FDH and arsenite oxidase enzymes contain an additional redox centre in the form of an Fe-S cluster of either the [4Fe-4S] or [3Fe-4S] types. A second clade is composed of formate dehydrogenases plus NasA and NapA nitrate reductases. In this clade enzymes contain an active site of four dithiolene sulphur ligands and a polypeptide ligand from a cysteine (nitrate reductase) or selenocysteine (formate dehydrogenase - FDH) to the Mo, plus a sixth ligand from a hydroxide or aquaoxide. The final clade contains the arsenite oxidase enzymes, a clade with similarities to the Nap/FDH clade. One prominent difference is the existence of a [3Fe-4S] cluster replacing the [4Fe-4S] cluster found in the previously described clade. IN addition, this third clade does not have a protein-based ligand at the active site, instead simply having four sulphur ligands and a hydroxo ligand coordinating the Mo atom (McEwan et al., 2002).

1.2.3 N and S-oxide reductases

E. coli encodes for a number of reductases required for the respiratory reduction of N- and S- oxides, including the enzymes trimethylamine N-oxide (TMAO) reductase and dimethylsulphoxide (DMSO) reductase. These enzymes are both members of the type III clade of the DMSO reductase family of molybdenum-containing enzymes, which is characterised by the presence of a serine side chain ligand to the molybdenum in their active site. In these enzymes the molybdenum cofactor is in the bisMGD form (Jormakka et al., 2004, Kisker et al.,

1997, McDevitt et al., 2002, Romao, 2009). 56

The organic compound TMAO is present in high levels in the tissues of marine fish where it is thought to play an important role in the osmotic equilibrium, and was intensively studied after suggestions that its reduction to trimethylamine (TMA) by bacteria may be a major indicator of fish spoilage (Barrett & Kwan, 1985). A small number of bacteria are able to utilise TMAO as a sole carbon source (Raymond & Plopper, 2002), whilst the majority exploit this compound as a substrate for respiration, where TMAO can act as the sole electron acceptor for oxidative phosphorylation in the absence of oxygen (Kwan & Barrett, 1983). TMAO reduction to TMA can be catalysed by three systems: the periplasmic TorYZ (Gon et al., 2000), the membrane bound DmsABC, and the periplasmic TorCAD systems (Barrett & Kwan, 1985, Bilous & Weiner,

1985b, Bilous & Weiner, 1988).

The TorYZ system is the least well understood system. It appears to be expressed at low levels, with no obvious regulation so far identified. In particular, there does not appear to any up- regulation by TMAO or DMSO (Gon et al., 2000). This system is also unable to complement a dmsABC torAC double mutant strain (Sambasivarao & Weiner, 1991). However, the expression of the torCAD operon has been well studied, and interestingly despite its anaerobic expression is independent of both global regulators FNR and ArcA, as well as the nitrate regulator NarL.

Instead, TorCAD is regulated by a TMAO specific two-component regulatory system named the

TorRS system (Simon et al., 1994). The TorCAD system is highly specific for N-oxide reduction, in particular TMAO (Gon et al., 2000). In the absence of oxygen and the presence of TMAO

TorS, a histidine kinase sensor, autophosphorylates and in turn phosphorylates its partner protein TorR. This TorR protein in turn activates tor operon expression through binding to a decameric direct repeats called tor boxes in a region upstream in the promoter region (Gon et al., 2000, Jourlin et al., 1996a, Simon et al., 1995). In addition to TorRS and overlapping the torR gene is torT, encoding a periplasmic protein necessary for torCAD operon expression

(Jourlin et al., 1996b). TorT binds TMAO in the periplasm and subsequently binds the periplasmic domain of TorS. Upon TMAO binding to TorT it is thought to induce a structural change that is transduced to TorS, leading to TorS autophosphorylation and subsequent TorR 57 phosphorylation as previously mentioned (Baraquet et al., 2006, Jourlin et al., 1996b). In the absence of TMAO, TorS can also dephosphorylate TorR, therefore limiting unnecessary gene expression (Ansaldi et al., 2001). An additional level of regulation is performed by the small response regulator protein of phage origin named TorI, shown to act downstream of TorS by abolishing torCAD expression. TorI is thought to function by binding to the effector domain at the C-terminus of TorR and preventing recruitment of RNA polymerase, which in turn negatively regulates torCAD expression (Ansaldi et al., 2004). Another form of regulation seems to be through TorC in its apoprotein form, appearing to exert a negative autoregulatory effect on the torCAD operon. This is thought to also be through the TorS/R system, either directly inhibiting TorS through conformational changes or indirectly by preventing the binding of TMAO and/or TorT to TorS (Ansaldi et al., 1999).

The torCAD operon encodes three proteins, the first of which is TorC - a pentahaem c-type cytochrome that is anchored to the cytoplasmic membrane (Méjean et al., 1994). TorC is homologous to members of the NirT/NapC class of proteins, which contains multi-haem c-type cytochromes, with the TorC N-terminal region being homologous to NirT and NapC. TorC is anchored to the membrane by an N-terminal region that also contains four of its five haems, with the fifth being present in the C-terminal half of the protein (Gon et al., 2001). Intact TorC is required for TMAO reductase activity and is involved in the transfer of electrons to the catalytic subunit TorA by acting as the site of the menaquinol oxidation reaction involved in the TMAO reduction pathway, transferring electrons from MQ to TorA (Gon et al., 2001, Kwan

& Barrett, 1983). TorC is transported to the periplasm via the Sec translocase, with maturation being accomplished through the use of a periplasmic set of cytochrome assembly proteins

(Ansaldi et al., 1999). TorA, the active subunit of TMAO reductase, is a substrate of Tat that localises to the periplasm in its fully folded cofactor-containing state (Santini et al., 1998). The final gene of torCAD encodes for the cytoplasmic chaperone protein TorD which is a member of the TorD family of chaperones (Section 1.1.7.11). TorD is important for the correct assembly of TorA, and has been shown to enable insertion of the Mo-bis-MGD cofactor (Ilbert et al., 58

2003, Pommier et al., 1998). There are thought to be two binding sites for TorD on the TorA protein, one on the signal peptide and one in the ‘mature’ region of the protein (Dow et al.,

2013). However, whilst TorD is important in TorA assembly, it does not appear to be essential, with torD mutant strains still able to reduce TMAO at approximately 50 % wild-type levels

(Pommier et al., 1998). It is possible that in addition to TorA binding TorD is also involved in direct interactions with proteins of the molybdenum cofactor biosynthetic pathway during

TorA maturation (Genest et al., 2008).

The enzyme DMSO reductase is encoded in the operon dmsABC, an operon that is controlled by FNR and is constitutively expressed in anaerobic environments. It is also strongly repressed ten-fold by NarL in the presence of nitrate (Bearson et al., 2002). The localisation of DMSO reductase within the cell was debated for a number of years, with publications appearing to show a cytoplasmic location (Rothery & Weiner, 1993, Sambasivarao et al., 1990,

Sambasivarao et al., 2000). However, the location for both DmsA and DmsB has now been determined to be the periplasm (Stanley et al., 2002). Indeed this makes sense of the presence of an active Tat signal peptide on DmsA, with DmsB reaching the periplasm alongside DmsA through the use of the Tat hitchhiking mechanism (Berks, 1996, Berks et al., 2005, Stanley et al., 2002). DMSO reductase is a heterotrimeric enzyme, with DmsA being the active catalytic subunit. DmsA has a much broader spectrum of action than TorA, having been shown to be able to catalyse the reduction of a range of different N- and S-oxide compounds (Bilous &

Weiner, 1985a, Gon et al., 2000, Simala-Grant & Weiner, 1998, Weiner et al., 1988). In fact,

TMAO appears to be reduced at a much higher rate than DMSO, suggesting this may be the preferred substrate for DMSO reductase (Weiner et al., 1988). DmsB is a 4[4Fe-4S] containing protein that interacts closely with DmsC, the membrane anchor and site of menaquinol oxidation (Rothery et al., 1995, Weiner et al., 1988, Weiner et al., 1993a).

Another enzyme expressed in E. coli and S. Typhimurium is encoded by the operon ynfEFGHdmsD. YnfE and YnfF are homologous to one another and also share identity with 59

DmsA (Lubitz & Weiner, 2003). The ynfEFGHdmsD operon is regulated in a similar way to

DMSO reductase, with upregulation by FNR and negative regulation by NarL (Constantinidou et al., 2006, Xu et al., 2009). YnfE has been shown to catalyse reduction of selenate in both E. coli and S. Typhimurium, with YnfF also displaying similar capabilities, although perhaps to a lesser extent (Guymer et al., 2009). YnfF has also demonstrated DMSO reductase activity, although also to a poor extent compared to DmsA (Lubitz & Weiner, 2003). DmsD, a TorD-like cytoplasmic chaperone protein encoded for in this operon, is known to be required for the correct maturation of DmsA and DMSO reductase function (Oresnik et al., 2001, Ray et al.,

2003). DmsD has been shown to act in a similar way to TorD in that it only binds the unfolded

DmsA protein, with specific interactions observed with the DmsA signal peptide (Oresnik et al.,

2001). Although it has not yet been shown to bind elsewhere on DmsA, going by what is known for the TorD chaperone it is likely that such a binding event can take place.

1.2.4 Nitrate reductases

Nitrate is the electron acceptor of choice for cells during anaerobic respiration, due to its properties as the most energetically favourable of the known anaerobic electron acceptors.

Only when nitrate is depleted does the E. coli cell switch from using nitrate reductases to expression of the N- and S-oxide reductases. All nitrate reductases are molybdopterin enzymes, and E. coli encodes for three respiratory nitrate reductases containing Mo-bis-MGD in their catalytic subunits (Sparacino-Watkins et al., 2014). These are nitrate reductase-A (Nar-

A), nitrate reductase-Z (Nar-Z) and periplasmic nitrate reductase (Nap).

Nar-A is the main nitrate reductase in E. coli, encompassing 98 % of cytoplasmic nitrate reductase activity (Bonnefoy & DeMoss, 1994). The heterotrimeric complex is made up of:

NarG, the catalytic subunit containing Mo-bis-MGD in the active site; NarH, an [Fe-S] cluster- containing electron transfer subunit; and NarI, an integral haem-containing cytochrome b membrane anchor that links the enzyme to the quinone pool (Bertero et al., 2005, Rothery et 60 al., 1998, Sodergren et al., 1988). These are encoded in the narGHJI operon, alongside the gene for NarJ, a protein required for assembly of the active enzyme (Sodergren & DeMoss,

1988, Sodergren et al., 1988). NarJ has been identified as having homology with the cytoplasmic chaperone protein DmsD, surprising given the fact that Nar-A is a cytoplasmic protein, and NarG does not have a functional Tat signal peptide (Turner et al., 2004). NarJ appears to be pivotal to the activity of Nar-A, and is involved in many important stages of assembly. In addition to binding NarG and aiding cofactor insertion, NarJ is also important for

FeS cluster insertion in the NarGH dimer. Furthermore, it safeguards NarI maturation by preventing premature membrane anchoring (Lanciano et al., 2007). After maturation of these proteins is complete NarJ becomes dissociated from the active enzyme (Blasco et al., 1998).

The second nitrate reductase, Nar-Z, plays a less significant role than Nar-A as it is expressed at relatively low levels in high nitrate environments. The operon by which it is encoded, narZYWV, shares 73 % homology with narGHIJ and appears to have arisen from gene duplication (Blasco et al., 1990, Iobbi et al., 1987). However, despite similarities to Nar-A, the

Nar-Z operon does not appear to be regulated by the same mechanisms and is not induced by nitrate or FNR. In addition, oxygen does not appear to repress expression of the operon

(Blasco et al., 1990, Iobbi et al., 1987). Expression has also been shown to be up-regulated by the stress response regulator RpoS, suggesting this enzyme is involved in an adaptive response, perhaps during the change from aerobic to anaerobic respiration. RpoS is a stationary phase regulatory factor in addition to being involved in stress responses, and it appears to play an important role in respiratory metabolism through the increase of Nar-Z expression during stationary phase (Chang et al., 1999). By analogy with Nar-A, it is likely for

NarZ to be the catalytic Mo-bis-MGD subunit, and NarY to be an Fe-S cluster containing protein in the enzyme complex. NarW is also a member of the TorD family of cytoplasmic chaperones, and it is likely to function in a similar manner to NarJ (Turner et al., 2004). 61

Whilst these two nitrate reductase enzymes are cytoplasmic in orientation and require the transport or nitrate into the cell cytoplasm for function, the third nitrate reductase in E. coli is a periplasmic enzyme. Periplasmic nitrate reductase (Nap) is encoded by the operon napFDAGHBC in anoxic conditions, and has been implicated in nitrate scavenging in nitrate- limited conditions. Interestingly, Nap activity is not able to produce the same level of growth in the presence of nitrate compared to Nar-A, and it appears that expression of Nap is in part regulated by NarL (Constantinidou et al., 2006, Potter et al., 1999). Nitrate reduction by Nap has been shown to enable growth with nitrate as the sole respiratory electron acceptor in a respiratory chain with a proton translocating dehydrogenase (such as a formate dehydrogenase), but does not constitute a site for generating proton motive force itself

(Stewart et al., 2002). The NapA component of this nitrate reductase is the catalytic subunit, containing a Mo-bis-MGD cofactor and [4Fe-4S] cluster (Jepson et al., 2007). Unsurprisingly due to its characteristics, NapA is a Tat substrate and bears an N-terminal Tat signal peptide that is subsequently cleaved after transport (Thomas et al., 1999). NapA works in complex with the dihaem c-type cytochrome NapB to produce the active complex. NapC forms the membrane anchor and is a member of the tetrahaem c-type cytochromes and acts to shuttle electrons to the active complex (Roldán et al., 1998). None of the iron-sulphur proteins encoded in this operon, NapF, NapG and NapH, are essential for nitrate reduction (Potter &

Cole, 1999). However these components are thought to play roles in electron transfer from quinol to the NapAB complex (NapG/F), and [4Fe-4S] cofactor assembly of NapA (NapF)

(Brondijk et al., 2004, Olmo-Mira et al., 2004, Potter & Cole, 1999). The NapD protein is extremely important for the function of this Nap enzyme, with NapA maturation dependent upon the presence of NapD. NapD is a Tat signal peptide-binding chaperone protein, however unlike the chaperones for Nar-A and Nar-Z this chaperone does not belong to the TorD family.

Instead it is thought to represent a new family of cytoplasmic chaperones, adopting a ferredoxin-type fold (Maillard et al., 2007).

62

1.2.5 Hydrogenases

A wide range of bacteria are known to be able to metabolize hydrogen for energy production, with hydrogen oxidation also implicated in pathogenesis of a number of species (Maier, 2005,

Vignais & Billoud, 2007). Hydrogen metabolising enzymes are termed hydrogenases, and are typically involved in this simple and reversible chemical reaction:

+ - 2H + 2e ↔ H2

The direction of this reaction is dependent upon the presence of either an electron acceptor with a high redox potential or an electron donor with a low redox potential. The organisation of the active site of hydrogenases enable the classification of these enzymes into three main groups: the [NiFe]-hydrogenases, containing one nickel and one iron atom; the [FeFe]- hydrogenases, containing two iron atoms; and the [Fe]-hydrogenases, containing a single iron atom (Vignais & Billoud, 2007, Vignais & Colbeau, 2004).

[Fe]-hydrogenases are unique to methanogens, and as such have not yet been extensively studied (Vignais & Billoud, 2007). This enzyme, formerly named H2-forming methylene- tetrahydromethanopterin dehydrogenase (Hmd), is involved in CO2 reduction with H2 to methane, catalysing an intermediate step by which methenyltetrahydromethanopterin

+ + (methenyl-H4MPT ) is reduced with H2 to methylene-H4MPT and H (Shima & Thauer, 2007,

Thauer et al., 2010). A number of crystal structures have been solved for [Fe]-hydrogenases, and appear to show a homodimer composed of a central unit composed of the C-terminal sections of each subunit, plus two peripheral globular units consisting of the N-terminal domains (Pilak et al., 2006). The structure of the active site has been further investigated and is shown to contain biologically unusual carbon monoxide molecules coordinating the iron centre (Shima et al., 2008, Thauer et al., 2010).

[FeFe]-hydrogenases are found in anaerobic prokaryotes, and are also the only hydrogenase class to be found in eukaryotes, being present in chloroplasts and hydrogenosomes. These 63 enzymes are generally monomeric, although dimeric, trimeric and tetrameric enzymes have also been observed (Vignais & Billoud, 2007). The [FeFe]-hydrogenase active site contains two iron atoms, with one of these irons existing bridged to an Fe-S cluster through a bridging cysteine thiolate and is termed the H cluster (Peters et al., 2014, Vignais & Billoud, 2007).

Oxygen is known to inactivate [FeFe]-hydrogenases, presumably through interaction and destruction of the H-cluster (De Lacey et al., 2007).The [NiFe]-hydrogenases are by far the most numerous and best studied of the hydrogenase classes. The core enzyme consists of two subunits; a large subunit (also termed the α-subunit) of approximately 60 kDa that contains the [NiFe] active site, and the small subunit (or β-subunit) of 30 kDa in size that contains up to three Fe-S clusters. The [NiFe] active site is buried deep within the large subunit and is coordinated by four cysteines, with the iron atom also ligated by three non-protein and unusual ligands: two cyanide molecules and one carbon monoxide molecule. The Fe-S clusters in the small subunit are arranged linearly and act to transfer electrons to or from the active site depending upon the presence of al electron donor or acceptor (Vignais & Billoud, 2007).

Based on sequence alignments, the [NiFe]-hydrogenases could be classified into four groups that also are in agreement with cellular functions of these enzymes. These groups are the uptake [NiFe]-hydrogenases, the cytoplasmic H2 sensors and cyanobacterial uptake [NiFe]- hydrogenases, the bidirectional heteromultimeric cytoplasmic [NiFe]-hydrogenases, and the

H2-evolving, energy-conserving, membrane-associated [NiFe]-hydrogenases (Vignais et al.,

2001, Vignais & Colbeau, 2004). These enzymes are also known to be inactivated by trace amounts of oxygen (De Lacey et al., 2007). The Group 1 [NiFe]-hydrogenases are synthesised in the cytoplasm and display N-terminal signal peptides on their small subunits (Table 1.1,

Section 1.1.7.7) (Sargent et al., 1998). These enzymes are located on the cytoplasmic membrane with the active sites facing the periplasm and so rely on the Tat system for their correct localisation. As these enzymes contain two subunits, but only the small subunits have

Tat signal peptides, tight coordination between the subunits must be necessary to ensure successful translocation and localisation. 64

1.3 Biosynthesis of molybdoenzymes

Molybdenum is utilised by a wide range of animals, plants and microorganisms due to its high levels in the environment, particularly in the oceans where it is the most abundant transition metal. Molybdenum is a second row transition metal, unusual in that it is the only metal of these that most living organisms require; indeed those that do not depend on molybdenum use tungsten instead, directly below molybdenum on the periodic table. There are over 40 identified molybdoenzymes identified in prokaryotic and eukaryotic life so far, where molybdenum is used as the cofactor in redox enzymes involved in many reactions of the carbon, sulphur and nitrogen cycles (Leimkühler, 2007). In prokaryotic enzymes with the exception of nitrogenase, molybdenum is coordinated in a mononuclear form with an organic pyranopterin cofactor (Romao, 2009). Enzymes using this cofactor are oxidoreductases, involved in catalysing the transfer of an oxo group to or from a substrate in a two electron reaction. Molybdenum is extremely useful as it is able to cycle between oxidation states VI and

IV, thereby allowing the catalysis of two-electron redox reactions (Leimkühler, 2007).

1.3.1 Biosynthesis of the molybdenum cofactor

The molybdenum cofactor (MoCo) biosynthesis pathway is a highly conserved ancient and ubiquitous pathway. The redox enzymes containing MoCo are extremely diverse, covering many phylogenetic families and varying greatly in their architectures. Molybdenum is coordinated to dithiolene group on the 6-alkyl side chain of the pterin named molybdopterin

(Figure 1.19). Several variants of the molybdopterin-containing MoCo have been identified, including the molybdopteringuanine dinucleotide (MGD) form present in a large number of E. coli molybdoenzymes. 65

In E. coli, six genetic loci have been implicated in MoCo biosynthesis – moa, mob, moc, mod, moe and mog. These encode at least 17 genes involved in MoCo biosynthesis, and the proteins they encode are highly conserved in other organisms (Schwarz, 2005). Prior to the biosynthesis of the MoCo, molybdenum must first be imported into the cell in its soluble molybdate oxyanion form. This action is performed by an ABC transporter composed of proteins ModA,

ModB and ModC. ModA appears to be targeted to the periplasm and acts as the high affinity periplasmic molybdate binding protein. Two copies of ModB and ModC act together to form the membrane domains and the ATP-binding cassette of the transporter (Maupin-Furlow et al.,

1995). Expression of the ModABC system is regulated by the sensor protein ModE in a molybdate-dependent manner. ModE can act as a repressor of mod genes and an activator of moa genes (McNicholas et al., 1997). ModE appears to bind to the modA promoter upon binding molybdate, in this way repressing modABC expression under high molybdate availability (Anderson et al., 2000, McNicholas et al., 1997). ModE also appears to play a regulatory role in expression of a number of molybdoenzymes themselves, including DMSO reductase and the nitrate reductases Nar-A and Nap, plus the molybdenum cofactor biosynthesis operon moaABCDE (Anderson et al., 2000, McNicholas et al., 1998, Self et al.,

1999).

Figure 1.19: The structure of the molybdopterin component of the molybdenum cofactor. The molybdopterin component of the molybdenum cofactor is composed of a pyran ring fused to a pterin moiety. Upon formation of the molybdenum cofactor, Mo is coordinated with a dithiolene group present in the pyran ring. 66

In prokaryotes such as E. coli, the biosynthesis of molybdenum cofactor (Figure 1.20) can be divided into four steps: 1) formation of precursor Z from GTP; 2) formation of molybdopterin

(MPT) from Precursor Z; 3) insertion of molybdenum into MPT to form MoCo; and 4) GMP or

CMP attachment modifications to MoCo to form MGD and MCD respectively.

Figure 1.20: Biosynthesis of the molybdenum cofactor. The biosynthetic steps involved in the formation of Mo-bis-MGD. Precursor Z is formed from GTP, molybdopterin (MPT) from precursor Z, MoCo from MPT and Mo-bis-MGD form MoCo. Enzymatic processes at each step are displayed. Adapted from (Leimkühler, 2007).

67

The first stable intermediate of the MoCo biosynthetic pathway is precursor Z, also named cyclic pyranopterin monophosphate (cPMP). Precursor Z, an oxygen sensitive 6-alkyl pterin containing a cyclic phosphate group, is thought to be produced from GTP through the actions of enzymes MoaA and MoaC. MoaA is a S-adenosylmethionine (SAM)-dependent enzyme that contains two [4Fe-4S] clusters, and is involved in the initial opening of the guanine imidazole ring of GTP and successive rearrangements required to give the formyl-diaminopyrimidine nucleotide intermediate (Sofia et al., 2001). MoaC then cleaves the phosphate group of this intermediate to form the cyclic ring of precursor Z (Hänzelmann et al., 2004).

During the second step of biosynthesis precursor Z is converted to MPT through the incorporation of two sulphur atoms at positions C1’ and C2’, a reaction catalysed by the MPT synthase MoaDE. Purification of this enzyme showed it to be a heterotetrameric complex with two copies of both MoeA and MoeD (Pitterle & Rajagopalan, 1989). In this enzyme it is thought that the MoaD carries the sulphur for the reaction and MoaE acts to bind precursor Z

(Daniels et al., 2007, Gutzke et al., 2001). In E. coli it has been shown that MoeB and the L- cysteine desulphurase protein IscS are involved in assisting this reaction through the donation of sulphur (Leimkühler et al., 2001, Zhang et al., 2010).

The next step involves the insertion of molybdenum into the MPT, a step requiring moeA and mogA gene products (Johnson & Rajagopalan, 1987, Joshi et al., 1996, Xiang et al., 2001). The crystal structure of MogA, the first MoCo biosynthetic protein structure to be solved, shows a trimeric protein with a high binding affinity for MPT (Liu et al., 2000). Meanwhile MoeA has four distinct domains, with one being similar in structure to MogA (Schrag et al., 2001, Xiang et al., 2001). The current model for this step involves MogA catalysing the adenylation of MPT to form an MPT-AMP intermediate. MoeA then cleaves AMP from MPT and specifically inserts molybdate into the molecule, which binds to the MPT dithiolene sulphurs (Leimkühler et al.,

2011, Neumann & Leimkühler, 2008). 68

MoCo in E. coli is further modified through the attachment of either GMP or CMP. Only one enzyme, a sulphite oxidase homologue with an unknown function, is known to use the simple

Mo-MPT form of the MoCo (Loschi et al., 2004). This final step encompasses the covalent addition of GMP or CMP to the C4’ atom of MPT by the MobA or MocA proteins respectively, forming either Mo-bis-MGD or Mo-bis-MCD (Neumann et al., 2011). The Mo-bis-MGD cofactor, that used by the DMSO reductase enzymes, is produced from MoCo by the MobA protein

(Iobbi-Nivol et al., 1995, Palmer et al., 1996). This enzyme catalyses the transfer of the GMP moiety to the C4’ atom of MPT. Whilst it is understood that MobA is essential for this reaction the role of a second protein of the mob locus, MobB, is not yet understood. It has been suggested that MobB works alongside MobA, perhaps to enhance MoCo to MGD conversion

(McLuskey et al., 2003, Palmer et al., 1996).

It is clear that for this elaborate biosynthetic pathway, MoCo must be synthesised and assembled into apoenzymes within the regulated environment of the cytoplasm, and that extracellular MoCo-containing enzymes must be targeted via the Tat pathway.

1.3.2 Biosynthesis of Fe-S clusters

As well as the molybdenum cofactor, many proteins that are Tat substrates also contain Fe-S clusters. As with molybdenum, iron was available in high levels in its soluble form, Fe2+, in the ancient oceans, and so it is unsurprising that bacteria have evolved the ability to utilise this. In these early times sulphur was also readily available, and so Fe-S clusters were developed by early life as one of the earliest types of prosthetic groups (Ayala-Castro et al., 2008).

Electron transfer is a major role for Fe-S clusters due to their physiologically relevant redox potentials of between -0.5 to 0.15 V. These cofactors can also be involved in other processes, for example substrate binding, enzyme activity control and gene expression (Fontecave &

Ollagnier-de-Choudens, 2008). They can exist in many different molar ratios of iron to 69 elemental sulphur, with [2Fe-2S], [3Fe-4S] and [4Fe-4S] being the most common arrangements. In these clusters the Fe ions are normally coordinated by cysteine residues, and link to each other by sulphide bridges.

The process of Fe-S assembly is an important one, as both Fe and S are highly reactive and toxic, and therefore biosynthetic pathways must be tightly controlled. There are three pathways so far identified for Fe-S cluster biosynthesis so far: the Isc (iron sulphur cluster) system, the Suf (sulphur formation) system, and the Nif (nitrogen fixation) system. It has been suggested that these pathways can be loosely divided into different physiological roles, where those used for housekeeping cluster assembly go through the Isc system, those under stress conditions are assembled by the Suf system and the Nif system makes those with complex and specialised clusters for specific enzymes (Ayala-Castro et al., 2008). The usage and relative importance of each pathway varies greatly between species, however in E. coli the primary system is the Isc system.

A number of functional components are common to Fe-S cluster assembly. At the first stage of assembly a cysteine desulphurase enzyme, in the case of the Isc system, IscS, catalyses the conversion of environmental L-cysteine to L-alanine and sulphane sulphur (Mihara & Esaki,

2002). Cysteine desulphurases characteristically have a cysteine residue that is essential for function, and it is to this that the sulphur atom is transferred once released (Fontecave &

Ollagnier-de-Choudens, 2008). The sulphur liberated in this reaction plus iron is then donated to a second protein that acts as a scaffold. The source of iron donation for Fe-S assembly is unknown, although it has been suggested that IscA is able to recruit and deliver intracellular iron for this system (Ding & Clark, 2004, Ding et al., 2004). The scaffold protein, for example

IscU, binds both the sulphur and iron, and acts as the platform for synthesis of [2Fe-2S] and

[4Fe-4S] cluster (Smith et al., 2005, Yuvaniyama et al., 2000). IscU has been shown to be able to mediate multiple rounds of [2Fe-2S] cluster formation (Agar et al., 2000, Ayala-Castro et al.,

2008, Fontecave & Ollagnier-de-Choudens, 2008). Scaffold proteins are also able to transfer 70 the iron sulphur cluster to the acceptor protein for insertion, presumably through formation of a transient complex. This appears to be an ATP-dependent reaction that is aided by a number of chaperones, in the Isc system HscA and HscB (Bonomi et al., 2008, Chandramouli & Johnson,

2006), plus A-type carriers ErpA and IscA (Pinske & Sawers, 2012, Pinske & Sawers, 2011).

The Isc system is regulated by the IscR repressor. This protein is encoded by the first gene of the iscRSUA operon, and acts as a direct negative regulator of the operon. IscR is a [2Fe-2S] protein that blocks transcription through binding of a promoter region upstream of the operon, regulating expression of isc genes as well as the downstream accessory protein hscBAfdx operon (Schwarz et al., 2001).

It is clear, then, due to the complex and carefully regulated systems at play for assembly of Fe-

S clusters that they must be assembled into apoenzymes within the regulated environment of the cytoplasm and that extracellular Fe-S cluster-containing enzymes must be targeted via the

Tat pathway.

71

1.4 Aims

The overall aim of this study was to gain a greater understanding of the nature of interactions occurring in the Tat transport pathway. This study focused on a number of Tat substrates and their cytoplasmic chaperone partners, both in S. Typhimurium and E. coli, namely YnfE/DmsD and TorA/TorD. Genetic and biochemical techniques were used to elucidate important regions involved in interactions during the Tat proofreading. Super and high resolution microscopy was utilised as a way of monitoring protein interactions and localization within the cell. The project had three broad objectives:

1. To understand the dual functions of some twin-arginine signal peptides

2. To probe the mechanism of signal peptide recognition by Tat chaperones

3. To explore the interactions and wider roles of Tat chaperones during Tat transport

72

2 Tat-dependent selenate metabolism in S.

Typhimurium: the role of the twin-arginine

signal peptide

73

2.1 Introduction

Many bacterial species have been found to display the ability to utilise selenium in a number of enzymatic reactions, with selenate reduction to elemental selenium via the intermediate selenite being a well-known environmental process (Stolz et al., 2006). The enzyme selenate reductase is responsible for the initial reduction of selenate to selenite, catalysing the reaction:

-2 - + -2 SeO4 + 2e + 2H → SeO3 + H2O

The best characterised bacterial selenate reductase is that of Thauera selenatis, which has been shown to be a molybdoenzyme localised to the periplasm and encoded by the serABCD operon (Dridge et al., 2007, Schröder et al., 1997).

Figure 2.1: The selenate reductase from S. Typhimurium. Genes encoding for selenate reductase components (STM1495-STM1499) are present in an operon together with the gene encoding DmsD – a TorD-like cytoplasmic chaperone protein. YnfF is a protein with sequence similarity to YnfE, but does not contribute to selenate reduction in S. Typhimurium (Guymer et al., 2009).

74

In S. Typhimurium, selenate reductase was recently identified as being encoded by the ynfEFGHdmsD operon (Guymer et al., 2009) (Figure 2.1A). Sequence analysis of S.

Typhimurium selenate reductase had shown similarities to the membrane-bound DMSO reductase rather than to the periplasmic selenate reductase of T. selenatis (Guymer et al.,

2009). Both YnfE and YnfF share sequence homology with DmsA, the catalytic subunit of DMSO reductase. Indeed E. coli YnfF has been shown to demonstrate low levels of DMSO reductase activity whilst interestingly also displaying the ability to reduce selenate (Guymer et al., 2009,

Lubitz & Weiner, 2003). In S. Typhimurium, YnfE has been shown to have selenate reductase activity, whilst strains expressing only YnfF only have trace levels of selenate reduction

(Guymer et al., 2009, Lubitz & Weiner, 2003). Continuing the comparison of selenate reductase to the DMSO reductase system, it is likely that YnfG and YnfH are associated with YnfE. YnfH likely acts as the membrane anchor and the site for menaquinol oxidation, passing electrons to a [4Fe-4S] cluster within YnfE via four [4Fe-4S] clusters in YnfG and onto the molybdenum- containing active site of YnfE (Figure 2.1B) (Hille, 1996, Sambasivarao & Weiner, 1991, Weiner et al., 1993b). Both YnfE and YnfF have N-terminal signal peptides for transport through the

Tat system, with a functional Tat translocase being essential for selenate reductase activity

(Guymer et al., 2009). Another protein encoded within the ynf operon is DmsD, a member of the TorD family of Tat cytoplasmic chaperone proteins, which interacts directly with Tat signal peptides of YnfE and YnfF (Guymer et al., 2009). The ability of the YnfE signal peptide to interact with both the Tat translocase and the DmsD protein places this signal into a special class of dual-functional peptides (Ize et al., 2009b).

2.2 Aims

The aim of this Chapter was to characterise the role of the YnfE Tat signal peptide in S.

Typhimurium selenate reductase. The YnfE signal peptide has two distinct biological functions -

(1) directing assembly of the enzyme prior to translocation, and (2) transmembrane 75 translocation. Here, molecular genetic approaches were taken to identify specific regions of the signal peptide involved in interactions with the biosynthetic chaperone DmsD, and relate those to physiological activity of the enzyme. 76

2.3 Results

2.3.1 Identification of the DmsD binding site on the YnfE signal peptide

In order to understand the role of the YnfE signal peptide in S. Typhimurium selenate reductase activity and function, the first step was to identify the binding epitope for DmsD on the signal peptide. To do this, a glutamine scanning mutagenesis approach was taken combined with a rapid bacterial two-hybrid genetic screen (Karimova et al., 1998). The principles of the bacterial two-hybrid system chosen are described in Figure 2.2. This system utilises the fact that the adenylate cyclase domain of the toxin from Bordetella pertussis can be separated into two complementary inactive fragments, namely T18 and T25, that can form an active enzyme when brought in close proximity. The two fragments can together convert ATP to cyclic AMP, which then binds a protein known as both catabolite activator protein (CAP) or cAMP receptor protein (CRP) (Figure 2.2A). CRP is a pleiotropic regulator of gene expression in

E. coli, and when in complex with cAMP is able to promote the expression of a number of resident genes, such as the lac or mal operons. This enables bacteria to utilize lactose or maltose as carbon sources, which can be easily detectable by growth on indicator plates or through enzymatic activity assays. When the T18 and T25 fragments are co-expressed as separate polypeptides, they do not interact and therefore no cAMP is produced and no gene activation occurs (Figure 2.2B). Creating fusions of T18 and T25 to proteins of interest enables interactions to be investigated (Figure 2.2C). In this work, plasmids encoding spYnfE-T18, plus individual signal peptide amino acid variants, and T25-DmsD were used, with interactions being quantified through the measurement of lac operon activation using a β-galactosidase activity assay. 77

Figure 2.2: The principles of the bacterial two-hybrid system. The adenylate cyclase domain from the Bordetella pertussis toxin catalyses the conversion of ATP to cAMP. Under normal conditions the two fragments are able to convert ATP to cyclic AMP (A). This cAMP then binds CRP, leading to the activation of several resident genes such as the lac or mal operons whose activity can be used as reporters for cAMP production. When the T18 and T25 fragments are co-expressed as separate polypeptides (B), they do not interact and therefore no cAMP is produced and no gene activation occurs. By fusing two potentially interacting polypeptides (X and Y) to fragments T18 and T25 (C), interactions can be observed through measurement of reporter gene activity. Adapted from (Karimova et al., 1998).

First, a plasmid was constructed that encoded a fusion between spYnfE and the T18 fragment of the two-hybrid system. The 42-residue spYnfE (Figure 2.3) was placed at the N-terminus of

T18, thus preserving the correct physiological context of the signal peptide. Next, a plasmid was constructed that encoded a T25-DmsD fusion protein, with DmsD covalently attached to the C-terminus of T25. Co-expression of both plasmids in the appropriate reporter strain demonstrated a strong interaction between spYnfE and DmsD (Figure 2.4).

The YnfE signal peptide is a classical tri-partite Tat signal peptide (Figure 2.3). It contains a conserved SRRTLVK twin-arginine motif flanked by an extended polar n-region and hydrophobic h-region (Figure 2.3). Scanning mutagenesis of the YnfE sequence involved the substitution of each individual amino acid (from Glu-3 to Ala-42, but excluding Gln-6 and Gln-7) 78 of the signal peptide with glutamine through site-directed mutagenesis. Glutamine is a polar amino acid rarely present in Tat signal peptides (Cristóbal et al., 1999b), and therefore is an ideal amino acid for attempting to disrupt Tat signal peptide function. Glutamine has also previously been used to successfully identify important amino acids in the Tat signal sequence of NapA, the catalytic subunit of periplasmic nitrate reductase from E. coli (Grahl et al., 2012), and TorA, the catalytic subunit of TMAO reductase (Buchanan et al., 2008). In the case of YnfE

Gln-6 and Gln-7, these side chains were substituted with tryptophan, which has proved useful in dissecting the role of the signal peptide in E. coli hydrogenase-1 (Bowman et al., 2013).

Glutamine variants were constructed using QuickchangeTM PCR and validated by DNA sequencing of the signal sequence (DNA Sequencing & Services, University of Dundee).

Figure 2.3: Tat signal peptide of S. Typhimurium YnfE. The twin arginine motif is shown in bold, with the essential twin arginines being underlined. Also highlighted are the n-, h- and c-regions

Site-directed mutagenesis produced 40 individual spYnfE variants for which interactions with chaperone DmsD could be investigated (Figure 2.4). Interestingly, the amino acids of the highly conserved twin-arginine motif did not appear to have any function in chaperone binding, with near native levels of reporter activity being observed (Figure 2.4). This was also true of the entire n-region of the signal peptide, with no significant disruption in interactions occurring for any of the variants tested (Figure 2.4). Similarly, the c-region of the peptide was deemed to have no importance in signal peptide interaction, as there was no significant changes in β- galactosidase activity observed (Figure 2.4). The data instead highlighted the hydrophobic stretch of the signal peptide as important in interactions with DmsD. The spYnfE L24Q, A28Q, 79

V31Q, L33Q and F35Q all displayed levels of β-galactosidase activity of approximately 10 % that observed for the native signal peptide (Figure 2.4), which are similar values to the background levels observed for the empty vector negative controls (Figure 2.4).

Figure 2.4: Interaction study of cytoplasmic chaperone protein DmsD with spYnfE and spYnfE amino acid variants, measured in vivo using a bacterial two-hybrid system. MG1655 ΔcyaA::Apra was co-transformed with the vectors T25-DmsD and UT18-spYnfE variants, and β- galactosidase was measured as an indicator of interactions. β-galactosidase activities are displayed relative to native spYnfE-DmsD activity (WT). Cells containing both empty vectors were used as the negative control (labelled ‘UT18 + T25’). Data expressed as means (n = 3) ± standard error of means (SEM).

2.3.2 Signal peptide amino acid substitutions affect selenate reduction in vivo

Amino acids Leu-23, Ala-28, Val-31, Leu-33 and Phe-35 within the hydrophobic stretch of the

YnfE signal peptide have been identified as playing a role in DmsD interactions. Next, the importance of these amino acids for normal selenate reduction in vivo was investigated. This required the introduction of DNA encoding the amino acid substitutions onto the chromosome of S. Typhimurium at the native ynfE locus. This was performed using the pMAK705 protocol, first described by (Hamilton et al., 1989), which utilises a temperature-sensitive plasmid that is able to replicate at 30 oC but not at 44 oC. A pMAK705 plasmid derivative was designed with homologous genetic sequence with a length of approximately 500 base pairs on either side of the chromosomal region to be altered, enabling integration of the plasmid into the host 80 chromosome through homologous recombination (Figure 2.5A). Selection for cells where this recombination has taken place is achieved by growth at temperatures of 44 oC in the presence of chloramphenicol, where expression of the plasmid and chloramphenicol resistance is not permitted unless integrated into the chromosome. With subsequent growth at 30 oC, a single recombination event occurs leading to resolution of the plasmid, leaving behind either the native or altered gene in the chromosome depending upon the site of recombination (Figure

2.5B). The presence of the amino acid substitutions could then be confirmed through PCR amplification of the chromosomal region and DNA sequencing.

Figure 2.5: pMAK protocol for targeted chromosomal integration of spYnfE amino acid substitutions. A- pMAK705 vector containing YnfE signal sequence with flanking sequences of approximately 500 bp up- and down-stream. × indicates amino acid substitution on YnfE signal sequence. B – Principles of the pMAK protocol. Integration of the plasmid into the chromosome occurs via homologous recombination, resulting in strains that are screenable for chloramphenicol resistance at 44 oC. Resolution of the plasmid in the cell occurs with growth at 30 oC, as this permissive temperature allows expression from the plasmids’ origin of replication, which is toxic to the cell. This occurs via a single recombination result that depending upon the site will lead to native or amino acid variant spYnfE on the chromosome. 81

For this work, S. Typhimurium strain DIG103 (∆ynfF) was used as the parental strain for introduction of YnfE signal peptide variants. This strain contains an in-frame deletion of the gene STM1498, coding for the protein YnfF, which shares 71 % sequence overall identity with

YnfE. The DIG103 strain expressing only ynfE (Guymer et al., 2009) will enable all activity observed in the following mutagenesis experiments to be attributed to YnfE-containing selenate reductase complexes.

Amino acid substitutions L24Q, A28Q and L33Q in the YnfE signal peptide were successfully introduced onto the S. Typhimurium chromosome using this protocol, giving rise to S.

Typhimurium strains KM01, KM02 and KM03, respectively. Unfortunately, due to time constraints, it was not possible to introduce the ynfE V31Q and F35Q alleles onto the chromosome.

It has been widely observed that E. coli and other prokaryotes are able to reduce selenate to elemental selenium, via selenite, which results in elemental selenium deposits that accumulate on the cell surface and in the supernatent and turn cells red in appearance (Butler et al., 2012,

Gerrard et al., 1974, Turner et al., 1998). This phenomenon permits facile comparisons of the activity of selenate reductase within wild-type and mutant strains for cells grown micro-

2- aerobically or anaerobically in the presence of 10 mM sodium selenate (Na2SeO4 ) (Guymer et al., 2009).

82

Figure 2.6: Selenate reductase ability of chromosomal spYnfE variants. Parental S. Typhimurium strain DIG103 (ΔSTM1498) and mutants DIG100 (Δtat), KM01 (ynfE L24Q), KM02 (ynfE A28Q), and KM03 (ynfE L33Q) were grown overnight in microaerobic conditions in LB + 10 mM sodium selenate. Activity of selenate reductase is visible through production of red deposits in the cell culture.

The three mutant strains (KM01 [ynfE L24Q], KM02 [ynfE A28Q] and KM03 [ynfE L33Q]), and appropriate controls, were cultured in rich media containing selenate. From this experiment it was observed that DIG103 (∆ynfF), the parental strain, was able to reduce sodium selenate and therefore produce elemental selenium, turning the culture red (Figure 2.6). DIG100, a tatABC deletion strain, was unable to reduce sodium selenate, as would be expected (Guymer et al., 2009). However, interestingly, the spYnfE variants L24Q and A28Q were devoid of selenate reductase activity, with cultures lacking any red colouration after overnight growth

(Figure 2.6). The spYnfE variant L33Q, on the other hand, appeared able to reduce selenate to elemental selenium, with the cell culture still appearing red after overnight growth (Figure

2.6). These data corroborate in part the signal peptide interaction experiments (Figure 2.4) since they indicate that Leu-24 and Ala-28 on the YnfE signal sequence are important for mature selenate reductase production in S. Typhimurium and as such the mutant strains phenocopy a ∆dmsD mutant (Guymer et al., 2009). In contrast, the Leu-33 substitution with glutamine had no gross effect on the ability of the living cells to reduce selenate, which may suggest the bacterial two-hybrid data collected for YnfE L33Q are a “false-negative”.

83

2.3.3 Tools for probing YnfE assembly: antibodies to YnfE

In order to further examine the assembly and subcellular localization of YnfE, it was decided to raise antibodies against this protein. The S. Typhimurium YnfE protein was overproduced from an E. coli strain harbouring the vector pQE70-stmYnfE, which produces C-terminally His-tagged

YnfE alone. Following overproduction, a crude extract was subjected to immobilised metal affinity chromatography (IMAC). Resultant fractions were analysed by SDS-PAGE and observed to contain a double band at around 90 kDa (Figure 2.7). These two bands were excised from the gel and analysed by tryptic peptide mass fingerprinting (Fingerprints Proteomics Facility,

University of Dundee). This confirmed that both bands were in fact YnfE, with the smaller band missing part of the N-terminal (Figure 2.8). This protein sample was therefore deemed pure enough to raise antibodies against, and so was sent to Eurogentec S.A.TM for polyclonal antibody production.

Figure 2.7: IMAC purification of S. Typhimurium YnfE. YnfE was purified by IMAC and fractions separated by SDS-PAGE (12 % [w/v] acrylamide gel). Two electrophoretic variants of YnfE were observed. Bands marked with * were excised and analysed by tryptic peptide mass fingerprinting. 84

Figure 2.8: Tryptic mass spectrometry analysis of purified YnfE samples shows two forms of YnfE. Analysis of bands excised from SDS-PAGE gel in Figure 2.7. Identified peptides are indicated in bold red. A – higher molecular weight band from Figure 1.7. B – lower molecular weight band from Figure 1.7.

Upon receiving the polyclonal antibody in the form of serum from two individual rabbits, the specificity of the antibody for YnfE was tested by Western blot (Figure 2.9A). Blotting against purified YnfE revealed an extremely strong signal, from which it was difficult to identify individual bands due to the high signals even at the shortest exposure times. However, the strongest signal did appear to be at approximately the correct size for 90 kDa YnfE. A number of unspecific bands at varying sizes were also present. Next tested were whole cell samples of several S. Typhimurium strains (Figure 2.9A). In all samples a nonspecific band of 23 kDa was seen. LT2a, the wild-type strain from which all other strains in this work were derived, displayed two bands of approximately 75 kDa and 60 kDa in size, which are smaller than would be expected for YnfE. There was also a faint band present at approximately 100 kDa, which is more likely to be YnfE. However, the same pattern of bands is present the sample of DIG101 cells, a strain devoid of the genes encoding for YnfE and YnfF proteins. DIG102 (ΔynfE) also displays the two bands around the 75 kDa marker, but no band of 100 kDa is visible. Western blotting of DIG103 (ΔynfF), KM01 (ΔynfF spYnfE L24Q) and KM02 (ΔynfF spYnfE A28Q) also revealed a similar picture, with two stronger bands at around 75 kDa and 60 kDa respectively, and a 100 kDa band visible at varying degrees of intensity. From these results, the antibody 85 against YnfE does not appear to be able to specifically recognise and show YnfE on a Western blot.

Figure 2.9: Testing the quality of antibody raised against purified YnfE. Western blots were all performed after 10 µl samples were separated by SDS-PAGE (12 % acrylamide gel). A – Initial tests with the antibody. Western blotting with anti-YnfE and anti-rabbit both at concentrations of 1:10,000 was performed. Samples tested were the purified YnfE protein used for antibody production, along with whole cell samples of wild-type S. Typhimurium strain LT2a and various mutant strains – DIG101 (ΔynfEF), DIG102 (ΔynfE), DIG103 (ΔynfF), KM01 (ΔynfF spYnfE L24Q), KM02 (ΔynfF spYnfE A28Q). B – Anti-YnfE was pre-incubated with nitrocellulose membrane containing whole cell samples of DIG101 (ΔynfEF) prior to use with Western blot of whole cell samples of DIG105 (ΔdmsA ΔtorA), DIG107 (ΔdmsA) and DIG101 (ΔynfEF). Anti-YnfE – 1:5,000. Anti-rabbit – 1:10,000. C – Western blot after affinity clean-up of anti-YnfE antibody. Anti-YnfE – 1:200. Anti-rabbit – 1:10,000. Samples analysed were purified YnfE plus whole cell samples of DIG101 (ΔynfEF), DIG103 (ΔynfF), KM01 (ΔynfF spYnfE L24Q) and KM02 (ΔynfF spYnfE A28Q).

Next, in an attempt to remove any unspecific interactions prior to use, a nitrocellulose membrane containing whole cell DIG101 (∆ynfEF) was pre-incubated with the antibody (Figure

2.9B). Due to the lack of genes for both YnfE and YnfF in this strain, performing this pre- 86 incubation step this should cause any antibody that recognises proteins other than YnfE that are present in whole cell S. Typhimurium samples to become immobilised on the membrane, therefore removing some of the unspecific bands observed. Western blotting was then performed using this pre-incubated antibody with whole cells of DIG105 (ΔdmsA, ΔtorA),

DIG107 (ΔdmsA) and DIG101 (ΔynfEF). Strains DIG105 and DIG107 were selected as they are unable to produce TorA and DmsA, therefore eliminating the risk of any cross reactivity of the antibody with those proteins. It would be expected that if the antibody was able to detect

YnfE, differences could be observed between these two strains and that of DIG101, where no

YnfE or YnfF is produced. However, the bands observed for all three strains were identical, and as previously seen, with two stronger bands at around 75 kDa and a fainter band at 100 kDa, in addition to the unspecific band at 23 kDa. Therefore, YnfE was still not being successfully detected.

In addition, three further S. Typhimurium strains were tested using the YnfE antibody. These strains were LT2, NCTC12023 and SL1344, and were tested alongside our own laboratory stock of LT2a for comparison. Cells were grown in both aerobic and anaerobic conditions, with cell cultures then being analysed by SDS-PAGE and Western blot using YnfE antiserum that had been affinity purified (Figure 2.10). Results showed bands in all samples at approximately 60 and 75 kDa in size, similar to other Western blots performed using this antibody. In addition there was a band of approximately 23 kDa, also present in all samples, although showing some level of variation in intensity between samples. There appeared to be lower levels of this unidentified band in NCTC12023 samples, plus the anaerobic cell sample from the donated strain LT2. This band was also picked up by the antibody in previous Western blots (Figure 2.9).

From this result it was not possible to elucidate any more information about YnfE expression in these strains using the YnfE antibody. 87

Figure 2.10: YnfE antibody test with S. Typhimurium strains LT2, NTCT12023 and SL1344. LT2a (Dundee) and LT2, NCTC12023 and SL1344 (London) were analysed using anti-YnfE serum after affinity clean- up. Strains were grown either aerobically in LB or anaerobically in LB supplemented with 0.5 % (w/v) glycerol and 0.4 % (w/v) fumarate. Whole cell samples were then resolved through SDS-PAGE (12 % acrylamide) before visualisation with Western blot. Anti-YnfE - 1:200. Anti-rabbit, 1:20,000.

As a final effort to successfully visualise YnfE using this antibody, it was affinity purified against purified YnfE protein, using a protocol as described in Materials & Methods. After purification and elution steps, the antibody was again tested against purified YnfE, DIG101, DIG103, KM01 and KM02. However, no differences could be seen amongst any of the S. Typhimurium strains

(Figure 2.9C). Unfortunately, it seems that the native expression levels of the ynfEFGHdmsD operon are so low that the antibody appears to be unable to detect YnfE and therefore was not used for any further work.

88

2.4 Discussion

2.4.1 A conserved DmsD binding epitope

The signal peptide of Tat substrates is extremely important for protein translocation. No protein is able to be transported via the Tat system without itself having such an N-terminal peptide, or being tightly associated with a partner protein that has a signal peptide (Sargent et al., 2006). Whilst the conserved twin-arginine motif is essential for successful translocation via the Tat system (Stanley et al., 2000), indeed the other regions of the signal peptide must have their own roles to play in the Tat pathway. Some signal peptides are known to have a second role and so be important in protein maturation prior to transport. Such Tat Proofreading is an important step in the biosynthesis of Tat substrates, for example during incorporation of redox cofactors, particularly in molybdenum-containing proteins. In this case cytoplasmic chaperone partner proteins are known to bind the signal peptide of these substrates (Jack et al., 2004,

Sargent, 2007a). A large amount of research has been carried out on the substrate-chaperone pair DmsA and DmsD, with E. coli DmsD in fact being one of the best characterised Tat proofreading chaperones (Chan et al., 2008). This work instead focussed on the characteristics of the signal sequence of YnfE and its importance in the YnfE-DmsD binding interaction in

Salmonella.

Glutamine scanning of the signal peptide of YnfE revealed key residues towards the C-terminus of the hydrophobic region, rather than the essential twin-arginine motif, were involved in interactions with chaperone DmsD. By two-hybrid analysis alone, five amino acid substitutions apparently completely disrupted interactions. As DmsD is known to interact with at least three

Tat substrates (YnfE, YnfF and DmsA) the signal sequences of these three proteins were compared (Figure 2.11A).

89

Figure 2.11: Multiple sequence alignments of Tat signal sequences from S. Typhimurium (stm) and E. coli (ec). A - Primary amino acid sequences of the signal sequences of the three binding partners of DmsD (DmsA, YnfE and YnfF) from both E. coli and S. Typhimurium were aligned using ClustalW (http://embnet.vital- it.ch/software/ClustalW.html), with conserved residues highlighted using BoxShade (http://embnet.vital- it.ch/software/BOX_form.html). The hydrophobic stretch of the signal peptides is labelled. Arrows indicate amino acids L24, A28, V31, L33 and F35, implicated in interactions with DmsD by bacterial two-hybrid studies. B - Primary amino acid sequences of the signal sequences of chaperone-dependent Tat substrates of E.coli and S. Typhimurium. Aligned using ClustalW and conserved residues highlighted using BoxShade. Arrows indicate amino acids L24, A28, V31, L33 and F35, implicated in interactions between ssYnfE and DmsD by bacterial two-hybrid studies.

When the primary amino acid sequences of E. coli and S. Typhimurium YnfE, YnfF and DmsA signal peptides were aligned, it was observed that a number of these highlighted amino acids 90 are conserved or conservatively differ between enzymes, with all bar one amino acid being conserved between YnfE from S. Typhimurium and E. coli, pointing to a critical role for the hydrophobic region. Certainly, the n-regions and c-regions of these peptides are much less conserved at amino acid level (Figure 2.11A), suggesting that they are less important for signal peptide function than the twin-arginine motif or the h-region, or perhaps have more substrate specific roles instead of being important for recognition either by the Tat translocase or Tat cytoplasmic chaperones.

The two key side-chains implicated here in DmsD binding are YnfE L24 and YnfE A28.

Substitution of these residues with glutamine impaired DmsD binding in vivo and abolished physiological selenium formation. It is important to note that L24 is completely conserved in all known DmsD binding peptide from both E. coli and S. Typhimurium (Figure 2.11A). The YnfE

A28 residue, on the other hand, is specific to YnfE in E. coli and S. Typhimurium, being substituted by serine in all other DmsD-binding peptides (Figure 2.11A). Sequence alignments of a range of other Tat substrate signal peptides do no show any significant level of conservation of equivalent residues, indicating a DmsD specific motif (Figure 2.11B)

As experiments such as growth in the presence of sodium selenate are not quantitative, rather allowing just facile comparisons of selenate reductase activity, it could be questioned as to why a more quantitative approach was not taken. It has previously been mentioned by

Guymer et al. (2009) that attempts to quantify selenate reductase activity using classical redox-dye based enzyme assays were unsuccessful, and this could be due to low levels of gene expression as well as difficulties in coupling of redox dyes such as benzyl viologen to the specific enzyme reaction. In addition, quantifying the amount of elemental selenium produced in cell cultures through spectrophotometric measurement of samples was not possible due to the particulate nature of the selenium deposits as well as the presence of bacterial cells containing selenium particles on their surfaces. For these reasons such quantitative routes of investigation were not taken. 91

It has been suggested that Tat chaperone proteins function to prevent premature translocation of Tat substrates by masking the signal peptide from the Tat translocase. So why is the binding epitope of DmsD located in the hydrophobic region of the peptide, several amino acids away from the twin-arginine motif, the recognition motif essential for transport?

Certainly the suggested importance of the hydrophobic region for chaperone binding is supported in the literature. In recent work probing the interactions of the E. coli DmsA signal peptide with DmsD by Shanmugham et al. (2012) it was proposed that it was the hydrophobic region that is required for interactions with DmsD. Through the construction/design of region- swapped chimeric signal peptides containing n-, h- and c-regions from either DmsA or TorA, interactions with DmsD were observed to occur only in signal sequences containing DmsA h- and c- regions. The n-region was shown to be interchangeable with that of TorA with no great effect on DmsD binding. Further studies using various synthesised portions of the DmsA signal sequence in isothermal titration calorimetry (ITC) with DmsD also demonstrated the importance of the hydrophobic stretch for interactions (Winstone et al., 2013b). Investigations into other TorD family chaperones have suggested the RR motif is not important in signal peptide interactions (Coulthurst et al., 2012, Hatzixanthis et al., 2005). This suggests that the signal peptide functions of Tat targeting and substrate proofreading are two separate processes. It may be that binding of DmsD to the hydrophobic region of the signal peptide indirectly prevents interactions with the Tat translocase without requiring an interaction with the Tat motif specifically, perhaps by inducing a change in the secondary structure of the peptide. Alternatively, it should be considered that the binding of DmsD to the signal peptide out with the twin-arginine motif means that DmsD is never in direct competition with TatC for signal peptide binding. Instead, attached DmsD would prevent translocation by steric hindrance until assembly of the selenate reductase is complete.

From the literature, signal peptide binding epitopes appear to differ between classes of cytoplasmic chaperone proteins. For example, work performed investigating the TorA signal peptide and its chaperone partner TorD in E. coli pointed towards residues in the hydrophobic 92 region and in particular the border between the hydrophobic and c-terminal regions of the peptide as comprising the binding epitope (Figure 2.12) (Buchanan et al., 2008). As TorD is closely related to DmsD this ties in nicely with the identification of residues in the hydrophobic region of YnfE as the binding epitope for DmsD in S. Typhimurium, and similarly of E. coli DmsA as described previously (Shanmugham et al., 2012, Winstone et al., 2013b). Interestingly, Tat substrate-chaperone pairs do not all follow the same rules. Another E. coli pair NapA and NapD have been shown to have a completely different binding interaction. In this case, residues in or near the n-region of the signal peptide were shown as encompassing the NapD binding epitope, including two in the conserved Tat motif - Arg-6, one of the invariant arginine residues, and Lys-10 (Figure 2.12) (Grahl et al., 2012). Whilst TorD and DmsD are closely related proteins and both adopt all α-helical structure, NapD represents a different family of chaperones and displays a ferredoxin-type fold (Maillard et al., 2007, Turner et al., 2004). This data suggests that different chaperone families have developed different binding mechanisms to achieve the same overall objective of binding the Tat signal peptides of their substrate partners, in order to prevent interactions with the Tat system until maturation is complete.

Figure 2.12: Predicted cytoplasmic chaperone binding epitopes of Tat substrate signal peptides. Amino acid sequences of E. coli NapA and TorA are shown alongside S. Typhimurium YnfE, with residues identified through glutamine scanning and bacterial two-hybrid studies as having a potential role in binding to their cognate chaperone partners highlighted in red (Buchanan et al., 2008, Grahl et al., 2012). The conserved Tat motif is shown in bold, with the transport essential twin arginine being underlined. The n-, h- and c-regions are indicated.

93

The properties of amino acid residues highlighted as important for interactions may shed some light on areas of DmsD that may be involved. All the amino acids found to be important on spYnfE are of a hydrophobic nature, unsurprising due to their location in the hydrophobic stretch of the signal peptide, and due to their exposed location on the signal peptide are likely to form interactions with a hydrophobic region on DmsD. Interestingly, the amino acids on the

TorA signal peptide involved in TorD interactions are also all hydrophobic, apart from glycine which is able to fit into any environment due to its extremely short side chain and therefore could also be involved in this hydrophobic interaction. The similarities in the types of amino acids involved in both spYnfE and spTorA interactions with their chaperone partners also indicates the potential similarities in the mechanism of interactions between these protein- chaperone pairs. It should be noted however, that despite appearing to have similar binding epitopes, TorD and DmsD may not function in exactly the same way. TorD has been shown to be involved in binding the mature region of TorA and aid cofactor loading, whilst it appears that DmsD does not bind YnfE in the same manner. NapA, with its cytoplasmic chaperone

NapD belonging to a different family of chaperone proteins, unsurprisingly has a number of different properties in its signal peptide binding site. Whilst there are a number of hydrophobic residues, the binding site appears to overlap the twin arginine motif and so also contains a number of positively charged residues. This implies that whilst interactions between spNapA and NapD require hydrophobic bonds, electrostatic interactions are also important for binding. Therefore it may be concluded that different chaperone families undertake different mechanisms of interaction involving different types of non-covalent bonding.

The DmsD chaperone protein is unusual in that it appears to be involved in interactions with at least three Tat substrates, rather than having just one substrate partner as is more often the case with TorD family chaperones. Because of this unusual characteristic, it may come as no surprise that the h-region of DmsA, YnfE and YnfF signal peptides are rather similar in amino acid sequence. Although the h-region as a whole is a common feature of Tat signal sequences, the individual amino acid sequence itself can vary greatly between Tat substrates (Figure 94

2.11B). By having a binding epitope in the h-region, interactions of DmsD are immediately limited to those Tat substrates with the specific h-region, in this case YnfE, YnfF and DmsA peptides.

Figure 2.13: Alignment of a number of E. coli and S. Typhimurium bona fide and remnant Tat signal peptides. Primary amino acid sequences of Tat signal peptides from TorA, NarG and YnfE of both E. coli and S. Typhimurium plus TorZ from E. coli were aligned using ClustalW (http://embnet.vital-it.ch/software/ClustalW.html), with conserved residues highlighted using BoxShade (http://embnet.vital-it.ch/software/BOX_form.html). NarG is the active subunit of the cytoplasmic enzyme respiratory nitrate reductase, and although it is not transported through Tat, a remnant signal peptide has been identified on the N-tail of the protein (Ize et al., 2009a).

What enables DmsD to differentiate its binding partners DmsA, YnfE and YnfE from other Tat substrates is not known. Analysis of N-terminal signal peptides from a number of Tat substrates, as well as the remnant signal peptide of NarG may help elucidate this (Figure 2.13).

Whilst a number of amino acids are conserved, particularly between YnfE and TorA signal peptides, they also show several differences. Out of the signal peptides compared, spYnfE appears to have the greatest number of hydrophobic residues and actually the longest hydrophobic stretch. The spTorZ and spNarG sequences have very few hydrophobic residues in comparison, also when compared to spTorA from both E. coli and S. Typhimurium. The spTorZ sequence also appears to have a shorter hydrophobic region than the other signal peptides and this may also deter binding by DmsD. It may be that a very hydrophobic region is required for DmsD interactions, and the lack of such a strong hydrophobic region in these signal peptides deters interactions. Indeed, in the region normally expected to be hydrophobic the remnant signal peptide NarG has a number of charged and polar amino acid residues, further supporting the suggested requirement of hydrophobic amino acids for DmsD binding. If a 95 hydrophobic interaction is involved in DmsD binding to the signal peptides of YnfE and DmsA, the different characteristics of other signal peptides may in fact repel DmsD, thereby preventing incorrect interactions from occurring. Certainly it has been shown that slight differences between signal peptides have potential in determining binding. Studies with spTorA were able to highlight residues required for TorD binding, and also show those that were not, through substitution of specific amino acids (Buchanan et al., 2008). Interestingly,

Leu-21 on spTorA was not required, but when the equivalent amino acid on spDmsA, Leu-28, was replaced by a glutamine interactions with DmsD were disrupted (Winstone et al., 2013a).

This lends to the idea that slight differences in signal peptide sequences are enough to enable recognition by specific chaperones.

Whilst the signal peptides of TorA and YnfE share a much greater sequence identity, there are still subtle differences that may affect DmsD binding ability. In place of spYnfE Gly-30, spTorA has a proline residue. In addition, at position 31, where in spYnfE the strongly hydrophobic valine has been shown as essential for DmsD binding in this work, spTorA has a serine residue.

With this region differing in hydrophobicity to that of spYnfE, it may deter interactions with

DmsD.

2.4.2 The structure of the YnfE signal peptide

With interactions between the signal peptide and DmsD disrupted by the amino acid substitutions described here (L24Q, A28Q, V31Q, L33Q, F35Q), it is hypothesised that these amino acids are in some way involved in the proofreading and therefore maturation of YnfE.

It is thought that some Tat signal peptides are perhaps able to adopt an alpha-helical conformation (San Miguel et al., 2003, Zakian et al., 2010). When the hydrophobic region of the YnfE signal sequence is plotted in an alpha-helical projection, four of the five amino acids highlighted in the bacterial two-hybrid study can be seen to be located in the same region of 96 the helix (Figure 2.14). Whilst Leu-24, Ala-28, Val-31 and Phe-35 appear to form this alpha- helical face, amino acid Leu-33 is shown on the opposite side of the helix. This makes it seem highly unlikely that Leu-33 can be involved in interactions with DmsD alongside the other amino acids highlighted. Indeed, upon introduction of amino acid substitutions L24Q and A28Q to chromosomal YnfE, the activity, and by implication binding, of DmsD appeared to be completely disrupted. Meanwhile, L33Q did not appear to disrupt selenate reductase activity to any great degree, this demonstrating that L33 is not critical to either assembly or targeting.

With L33 now discounted, the remaining data nicely complements the modelling of the apparent locations of these amino acids when the h-region sequence is organised in an α-helix

(Figure 2.14).

97

Figure 2.14: Modelling of spYnfE. A – Secondary structure prediction of spYnfE, performed using PSIPRED v3.3 (http://bioinf.cs.ucl.ac.uk/psipred/). Conf – confidence of prediction. Helix (H) is shown in pink, random coil (C) is displayed as a black line. B - spYnfE hydrophobic region displayed as an α-helical wheel projection. Amino acids Ala-19 to Glu-36 were mapped onto an α-helical wheel projection. Amino acids highlighted as important in DmsD interactions are circled red. Leu-24, Ala- 28, Val-31 and Phe-35 form a face on one side of the helix, whilst Leu-33 appears to be localised on the opposite side of the helix. Created using heliQuest (http://heliquest.ipmc.cnrs.fr/cgi-bin/ComputParamsV2.py).

These data suggest the binding epitope for DmsD is hydrophobic in nature. This indicates that there may be a hydrophobic groove on DmsD involved in binding spYnfE. Indeed, in other substrate-TorD family chaperone interactions studied, this appears to be the case. For example, the 3D structure of the TorD-like chaperone TtrD from Archaeoglobus fulgidus has been solved to a resolution of 1.35 Å and displays two hydrophobic patches close to the ‘hinge’ region containing the conserved EPxDH motif (Coulthurst et al., 2012). One of these 98 hydrophobic pockets, involving helix α5 and α7, is very similar to a hydrophobic area of DmsD and so it may be that this region is involved in signal peptide binding in this family of proteins.

Hydrophobic interactions have long been known to be important in protein folding and interactions between protein faces. It has been demonstrated that surface hydrophobicity of proteins often plays a role in protein-protein recognition (Young et al., 1994). Hydrophobic interactions are driven by entropy, and interactions between DmsD and spDmsA have been shown to be entropic in nature, indicating a role for the hydrophobic effect in signal peptide binding (Winstone et al., 2013a). It has been shown in other protein interactions that, whilst hydrophobic interactions are important, these interactions are stabilised by additional electrostatic hydrogen bonds forming and helping to maintain the interaction (Chang et al.,

2013, Zhang et al., 2007). It is therefore likely that in addition to the apparent hydrophobic interactions involved in signal peptide binding, there are additional stabilising interactions occurring at the interface with the DmsD chaperone. It has already been proposed that electrostatic interactions are involved in initial binding of the signal peptide to the lipid membrane through the positive n-region of the signal peptide and the negatively charged lipid headgroups (Shanmugham et al., 2006). Hydrophobic interactions are then thought to come into play and allow further binding of the signal peptide to the membrane to occur.

The DmsA signal peptide has been shown to be able to interact with artificial phospholipid bilayers, with the presence of DmsD appearing to decrease the level of membrane binding, with previous studies also suggesting lipid interactions by DmsD, TorA and OmpA signal peptides (Shanmugham et al., 2012, Shanmugham et al., 2006). It may be that the hydrophobic region of the signal peptide is involved in both interactions with the DmsD chaperone and initial interactions with the cytoplasmic membrane. It could be hypothesised that the binding interactions to spDmsA by both DmsD and the lipid bilayer is of a competitive nature, with

DmsD being forced to release the signal peptide upon spDmsA interactions with the membrane. DmsD was unable to recruit large amounts of spDmsA back from the membrane 99

(Shanmugham et al., 2012), likely due to reduced accessibility, and this could be the final step in passing of substrates from chaperone to Tat translocase.

Of course, as this work has involved altering the native sequence of the YnfE signal sequence, it could be argued that Tat transport of this substrate has been adversely affected in some way, and this is the real reason behind the disruption of selenate reductase activity observed in KM01 and KM02 mutants. For this reason, we attempted to raise antibody towards YnfE with the idea that fractionation and Western immunoblots could be performed to identify the localisation of spYnfE variants in comparison to the native protein. Unfortunately, the antibody produced did not appear to identify YnfE when produced at native levels in the cell. Even after attempts to affinity-purify the antibody, it was still impossible to identify differences by

Western analysis between wild-type S. Typhimurium strains and those with deletions in genes for YnfE and YnfF, and was therefore of no use for analysing YnfE localisation in cells. Because of this issue, we are unable to say for sure that the effects observed with spYnfE variants are because of disruptions to the interactions of the signal sequence to DmsD. However, one possible indication that these results are predominantly due to affected DmsD interactions

(and so incorrectly assembled enzyme) rather than inhibited Tat transport per se is that of

KM03, producing the S. Typhimurium spYnfE L33Q, variant still able to reduce selenate to selenium. This shows that the introduction of a single glutamine on the signal peptide is tolerated and selenate reductase is still transported. Whilst overall hydrophobicity of the signal sequence is important for transport (Cristóbal et al., 1999b), altering the hydrophobicity of the peptide by introducing a single additional glutamine, especially close to the polar c-region, is unlikely to cause such a change as to affect Tat transport.

It was also attempted to document transport activity of signal peptides in the absence of

DmsD binding. To do this an in vivo screening assay was designed to show if any of the amino acids highlighted in previous work were important in Tat translocation ability of substrates.

Signal peptides of Tat substrates YnfE, YnfF and DmsA were fused to chloramphenicol acetyl 100 transferase (CAT) on the plasmid pSU-PROM. CAT confers resistance to the antibiotic chloramphenicol when located in the cytoplasm of the cell, where it catalyses the acetyl-S-

CoA-dependent acetylation of chloramphenicol at the 3-hydroxyl group, rendering it inactive.

Therefore, with CAT fusions to Tat signal peptides, if transport is active cells will be sensitive to high levels of chloramphenicol, whereas if transport is disrupted cells will become resistant. In this way it would be possible to investigate whether disruptions to DmsD binding affected transport.

Cells were co-transformed with both pSU-PROM signal peptide vectors and pUNI-

PROM_DmsDst. It was then attempted to obtain the minimal inhibitory concentration of chloramphenicol with these cells before the assay could be used, as E. coli has an inherent low level of resistance. However, despite originally determining an approximate MIC of 150 µg/ml, in further tests using the same controls expressing native signal peptide fusions and DmsD it was not possible to obtain the same results, with cells growing on all concentrations of chloramphenicol tested. Because of this, the experiments had to be abandoned. However, if it were possible to obtain working negative controls then this reporter system would be useful in investigating the effects of signal peptide amino acid substitutions had on Tat transport activity. It could also be ideal for use in screening a dmsD mutant library to identify variants with disrupted transport.

In summary, this Chapter has shed new light on the involvement of the Tat signal sequence of

YnfE in DmsD interactions. The binding epitope for DmsD has been identified as within the hydrophobic stretch of the signal peptide, with at least amino acids Leu-24 and Ala-28 within this binding site also being important for overall physiological selenate reductase activity.

101

3 Tat-dependent selenate metabolism in S.

Typhimurium: the role of DmsD

102

3.1 Introduction

It has been widely reported that many enzymes transported out of the bacterial cell cytoplasm by the Tat system require incorporation of metal cofactors prior to translocation, and this biosynthetic process is monitored in many cases by cytoplasmic chaperone proteins in an event termed ‘Tat proofreading’ (Dubini & Sargent, 2003a, Sargent, 2007a). The cytoplasmic chaperone protein DmsD is a member of the TorD family of cytoplasmic chaperone proteins, and is encoded for in S. Typhimurium by the gene STM1495 as part of the ynfEFGHdmsD operon (Turner et al., 2004).

Figure 3.1: Three dimensional crystal structure of S. Typhimurium DmsD. The monomeric structure shows an all-α fold arrangement with 12 helices. The N- and C-terminal helices are separated by a loop region. Another feature to note is a protruding solvent-exposed N-terminal helix containing a number of hydrophobic residues (Qiu et al., 2008).

The 3D crystal structure of S. Typhimurium DmsD has been solved to a resolution of 1.38 Å

(Figure 3.1), and displays an all α-helical fold structure with 12 helices (Qiu et al., 2008). DmsD interactions have been studied using isothermal titration calorimetry, and unusually for a Tat chaperone protein, has been shown to interact with the signal peptides of at least three different Tat substrates - DmsA, YnfE and YnfF. Signal peptide fusions to MalE were found to bind to DmsD with varying affinities, with spYnfE interacting with an apparent Kd of 45 nM, 103 spYnfF with an apparent Kd of 10 nM and spDmsA with an apparent Kd of 104 nM (Guymer et al., 2009). It has also been suggested that DmsD may play a role in recruiting other proteins, such as those involved in molybdenum cofactor biosynthesis, to the Tat substrates in order to aid enzyme maturation (Li et al., 2010). In Chapter 2, the binding epitope for DmsD on the S.

Typhimurium YnfE signal peptide was determined. However, the regions of importance on

DmsD for binding to Tat signal peptides are yet to be fully described. Work has previously been done on investigating the residues of E. coli DmsD involved in interactions with the signal peptide of E. coli DmsA. A number of residues were inferred to be involved in signal recognition through mutagenesis as having a role in the binding of DmsD to spDmsA, appearing to form a number of so-called “hot-pockets” in the N-terminal region (Figure 3.2)

(Chan et al., 2008).

Figure 3.2: ‘Hot pocket’ for signal peptide interactions shown on 3D crystal structure of E. coli DmsD Crystal structure of DmsD as solved by Stevens et al. (2009). Indicated in yellow are residues highlighted to play a role in DmsA signal peptide binding from site-directed mutagenesis studies (Chan et al., 2008). 104

In terms of the wider DmsD/TorD family of peptide binding proteins, studying DmsD activity has advantages over studying TorD. The main advantage is that dmsD mutants, from both S.

Typhimurium and E. coli, are completely devoid of selenate reductase activity. In contrast, torD mutants retain at least 50% TMAO reductase activity. This enables a facile colourimetric assay to be used in order to follow DmsD function in vivo, which has the potential to be used as a powerful genetic screen.

3.2 Aims

The aim of this Chapter was to characterise the YnfE signal peptide binding site on S.

Typhimurium DmsD. In order to better understand the role of DmsD in YnfE biosynthesis, a dmsD random mutant library was used to identify DmsD variants that had negative effects on

YnfE interactions/proofreading. Purified DmsD protein was also utilised in biochemical studies for in vitro characterisation of the ssYnfE binding activity of this protein.

105

3.3 Results

3.3.1 Genetic analysis of the YnfE – DmsD interaction.

3.3.1.1 Isolating DmsD variants affecting selenate reductase activity

A S. Typhimurium dmsD mutant library was prepared using an error-prone PCR protocol in collaboration with Dr Holger Kneuper, a temporary postdoctoral researcher in the Sargent team. The library was prepared in the pUNI-PROM vector (Jack et al., 2004), which drives gene expression from the constitutive tat promoter. In Chapter 2, S. Typhimurium strains KM01

(spYnfE L24Q) and KM02 (spYnfE A28Q) were constructed. Both these strains lack selenate reductase activity under normal conditions, and were considered ideal candidates for the selection of second-site suppressor variants of DmsD that could rescue selenate reductase activity in the ssYnfE amino acid substituted strains.

First, KM01 (spYnfE L24Q) and KM02 (spYnfE A28Q) were transformed with pUNI-

PROM_DmsDst, and the empty vector pUNI-PROM, as controls. Following overnight growth with sodium selenate, KM01 (spYnfE L24Q) expressing either pUNI-PROM or pUNI-

PROM_DmsDst demonstrated a lack of active selenate reductase (Figure 3.3). Next, the KM01

(ssYnfE L24Q) strain was screened using the dmsD mutant library. This involved transformation of KM01 (spYnfE L24Q) with the mutant library and transformed cells being plated directly onto LB agar plates containing 10 mM sodium selenate. Plates were then incubated for 36 hours anaerobically, with colonies then being picked for further analysis according to colour.

For KM01, red colonies were desired as this would indicate an amino acid substitution on

DmsD able to overcome the L24Q substitution present on the YnfE signal peptide.

Unfortunately, following numerous attempts involving approximately 10,000 colonies, no selenate reductase-positive colonies were identified in this screen (Figure 3.4A). 106

In the case of KM02, when this strain (ssYnfE A28Q) was transformed with pUNI-

PROM_DmsDst only selenate reductase activity was surprisingly rescued and red selenium deposits were observed (Figure 3.3). This suggested that by simply expressing excess S.

Typhimurium dmsD using the tat promoter present on a multicopy vector was enough to overcome the original disruption to selenate reductase biosynthesis caused by the A28Q substitution within the YnfE signal sequence. This result prevented to screening the dmsD mutant library for second-site suppressors. Instead it was decided to adapt the screen to identify inactive mutants of dmsD.

Figure 3.3: An spYnfE A28Q phenotype can be rescued by excess DmsD. S. Typhimurium strains KM01 (ynfE L24Q) and KM02 (ynfE A28Q) were transformed with either empty pUNI-PROM or pUNI-PROM_DmsDst. A resultant colony was then grown overnight in microaerobic liquid LB culture containing 10 mM sodium selenate.

Since transformation of KM02 (spYnfE A28Q) with pUNI-PROM_DmsDst alone was enough to restart selenate reduction, the dmsD mutant library was screened instead for white colonies indicating DmsD activity was not present. Cells were transformed with the dmsD mutant library and directly plated onto LB agar plates containing sodium selenate. In this case, both red and white colonies could be observed on the plate after 36 hours of anaerobic growth.

Individual white colonies, along with a small number of red colonies to check observations were correct, were then patched onto fresh LB agar + selenate plates and again grown 107 anaerobically for 36 hours. Red and white colonies could then clearly be distinguished (Figure

3.4B).

108

Figure 3.4: Identification of amino acid substitutions in DmsD that disrupt selenate reductase activity. A – Screen of dmsD random mutant library in KM01 (ynfE L24Q). Cells were transformed with the pUNI- PROM_DmsDst mutant library and plated directly onto LB agar plates containing 10 mM sodium selenate. Positive control (1) was DIG103 pUNI-PROM and negative control (2) was KM01 pUNI-PROM_DmsDst. The presence or absence of selenate reductase activity can be observed through the lack or appearance of red cells. After 36 hours of anaerobic growth, plates were examined and colonies displaying the desired phenotype, for KM01 red cells, were patched onto LB agar + 10 mM sodium selenate plates to confirm the observed phenotype. B – Screen of DmsD random mutant library as in A in KM02 (ynfE A28Q). Positive control (1) was KM02 pUNI-PROM_DmsDst and negative control (2) was KM02 pUNI-PROM. C – DmsD variants observed as disrupting selenate reductase activity in KM02 were tested for expression and stability by Western analysis of whole cell samples. Arrows indicate DmsD. Antibodies: Anti-DmsD – 1:20,000. Anti-rabbit – 1:10,000. Plasmids expressing DmsD variants present at similar levels to native DmsD were mini-prepped and sequenced (*). D – Primary amino acid sequence of S. Typhimurium DmsD with amino acid residues highlighted that were substituted in DmsD variants found to have reduced selenate reductase activity.

109

With a negative screen such as this it is possible that a number of the errors introduced into dmsD in the library could have generated nonsense codons that affect expression levels, protein synthesis and protein stability. In order to filter out these types of mutation all of the white colonies identified were therefore tested by Western analysis in order to select the

DmsD variants that are being expressed properly (Figure 3.4C). By comparing the levels of

DmsD present in these samples to those in cells expressing native DmsD (Figure 3.4C), three clones were identified that appeared to be expressing stable DmsD in similar levels to the native pUNI-PROM_DmsDst. Vectors expressing these three DmsD variants were isolated from cells and analysed by DNA sequencing. It was found that in the three dmsD mutants sequenced, a total of seven codons had been mutated (Figure 3.4D). In one variant DmsD four amino acid residues were substituted, namely V16G, S96G, T103I and G171R. Another DmsD variant had just one substitution at position 91, where the tryptophan was substituted for an arginine. The final variant contained two amino acid residue substitutions – R94H and G100S.

3.3.1.2 Analysis of DmsD variants for effects on selenate reductase activity

Amino acids of S. Typhimurium DmsD identified in the library screen were investigated for their importance during selenate reductase biosynthesis. In order to do this, each of the seven amino acid residues identified in the screen was individually substituted for a glutamine.

Mutations were introduced onto both vector pT25-DmsD, from the two-hybrid system, and pUNI-PROM_DmsDst.

The seven pUNI-PROM_DmsDst mutants were tested for their importance in the production of functional selenate reductase by testing in both KM01 and KM02 strains. This was done through transformation of the cells individually with the seven mutant vectors before overnight growth in liquid LB cultures containing 10 mM sodium selenate. It was observed that substitutions R94Q, S96Q, T103Q and G171Q were individually not enough to disrupt selenate reductase activity alone (Figure 3.5). W91R was picked as an inactive variant from the original 110 screen, and this was confirmed here with selenate reductase activity again observed to be disrupted in a W91Q version. In addition, it is clear that a single V16Q or G100Q substitution in

DmsD is sufficient to inactivate the chaperone (Figure 3.5).

Figure 3.5: DmsD residues V16, W91 and G100 are required for selenate reductase assembly and activity. S. Typhimurium strain KM02 was transformed with pUNI-PROM_DmsDst single amino acid variants. Single colonies were inoculated into LB media containing 10 mM sodium selenate before being grown overnight in microaerobic conditions. Activity of selenate reductase is visible through production of red deposits in the cell culture.

3.3.1.3 Analysis of DmsD variants for interactions with spYnfE

For quantification of the ability of the identified DmsD variants to interact with the signal sequence of YnfE, the bacterial two-hybrid system was utilised. In this experiment, DmsD variants were encoded for by the plasmid pT25-DmsD, producing DmsD variant proteins covalently fused to the C-terminus of the T25 fragment of the adenylate cyclase. Interactions of these DmsD variants were tested against the spYnfE-T18 fusion. Interactions were quantified through use of the β-galactosidase assay (Figure 3.6). The results show that DmsD

R94Q, S96Q, G100Q and T103Q displayed β-galactosidase activity levels of between 150 and

200 % that of the native DmsD-spYnfE interaction, with W91Q and G171Q having closer to 100

% β-galactosidase activity levels. These are therefore considered to be unimpaired in signal peptide recognition. However, the V16Q experiment clearly differed in its β-galactosidase activity, with levels almost identical to that of the empty vector negative control (Figure 3.6).

This suggests that of these amino acid residues identified from the random mutant screen, only Valine 16 is actually important in DmsD interactions with the signal peptide of YnfE. 111

Figure 3.6: DmsD V16Q variant is unable to recognise the YnfE signal peptide. MG1655 ΔcyaA::Apra was co-transformed with vectors UT18-spYnfE and T25-DmsD plus variants and β- galactosidase was measured as an indicator of interactions. β-galactosidase activities are displayed relative to native spYnfE-DmsD activity (‘WT’). Cells containing both empty vectors were used as the negative control (‘UT18 + T25’). Data expressed as means (n = 3) ± SEM.

3.3.1.4 Alternative roles for DmsD in YnfE maturation

The genetic screens employed so far in this Chapter have identified three amino acid residues important for DmsD function. The first is V16. When substituted with glutamine, the resultant

V16Q variant DmsD is unable to bind the YnfE signal peptide and therefore cannot rescue the activity of the KM02 strain. The second and third are W91 and G100. When substituted with glutamine, the resultant W91Q and G100Q variant DmsD cannot rescue the activity of the

KM02 strain, but retains the ability to bind the YnfE signal peptide. Thus it may be that DmsD in fact has two different roles in selenate reductase biosynthesis. Other cytoplasmic Tat chaperones are known to be important for cofactor insertion as well as signal peptide interactions, for example TorD is thought to bind the signal peptide of TorA but simultaneously interact with the mature region of the protein (Dow et al., 2013). This second interaction is believed to aide cofactor loading, probably by holding the active site of TorA in a conformation 112 that promotes insertion (Dow et al., 2013). Potential additional interactions of DmsD with full- length YnfE, and its partner protein YnfG, were therefore investigated.

In order to characterise DmsD interactions vectors were constructed to express several different protein fusions to the UT18 fragment for bacterial two-hybrid analysis. A truncation of YnfE, lacking the signal peptide, was to be tested to identify the ability of DmsD to interact with YnfE in another region other than that already identified. In addition to this, a fusion of

UT18 to full length YnfE was generated, as previous work had used a construct encoding for just the signal peptide itself. A fusion with YnfG, an Fe-S subunit of selenate reductase that lacks a Tat signal peptide, was also constructed.

Figure 3.7: DmsD interacts preferentially with the twin-arginine signal peptide. Vectors T25-DmsD and UT18-spYnfE, YnfE Δsp, YnfE or YnfG were co-transformed into MG1655 ΔcyaA::Apra and β- galactosidase was measured as an indicator of interactions. β-galactosidase activities are displayed relative to spYnfE-DmsD activity (spYnfE). Cells containing both empty vectors were used as the negative control (UT18 + T25). Data expressed as means (n = 3) ± standard error of means (SEM).

Interactions of these fusion proteins with a DmsD-T25 fusion were tested using the bacterial two-hybrid system and β-galactosidase activity as an indicator (Figure 3.7). In comparison to the interaction between DmsD and spYnfE, truncated YnfE displayed activity levels of 13 %, 113 very similar to negative control levels and indicating there was no interaction with YnfE when the signal peptide is missing. Full length YnfE showed 31 % activity, whilst YnfG showed 35 %, both relatively low levels of interaction. From this result it would appear that DmsD does not interact to any great extent with the mature region of YnfE or YnfG. Of course, through the fusing of proteins of the size of YnfE and YnfG to UT18 fragment of adenylate cyclase, there is a chance that folding of these proteins could be negatively affected. This could lead to problems with stability that could not be tested for. Stability issues would also result in low β- galactosidase activity as observed in this experiment. Therefore we cannot conclude for definite that DmsD is not interacting with these proteins, rather that interactions cannot be detected in this particular experiment.

3.3.2 Biochemical investigations of the spYnfE interaction with DmsD

3.3.2.1 Overproduction and purification of DmsD

In order to purify S. Typhimurium DmsD a vector was constructed using pQE-80L as a backbone in order to allow for IPTG-inducible overproduction. Primers were designed to introduce a His- tag followed by a TEV cleavage site onto the N-terminus of DmsD, and this was cloned into pQE-80L. This allowed for purification of His-tagged DmsD (DmsDHis) followed by cleavage of the His-tag to produce native DmsD protein for use in in vitro experiments (Figure 3.8A).

114

Figure 3.8: Purification of recombinant DmsD. A – Construct design for pQE80-NHis TEV DmsD. Primers designed are indicated by green arrows. B – BL21 pLysS cells overexpressing DmsDhis were lysed, with the resultant soluble fraction being subjected to Ni-IMAC. Fractions were analysed through SDS-PAGE (12.5 % acrylamide gel), and those containing DmsDhis were pooled and incubated overnight in the presence of TEV protease in order to cleave the His-tag. Cleaved DmsD was then isolated by reverse Ni-IMAC. C – DmsD isolated from reverse Ni-IMAC was pooled, concentrated and subjected to further purification by SEC through a Superdex 75 10/300 GL column. Fractions were visualised by SDS-PAGE (12.5 % [w/v] acrylamide gel).

To isolate DmsD using this vector, the soluble cell fraction from cultures overproducing

DmsDHis was subjected to immobilised metal affinity chromatography (IMAC). Fractions from peaks observed during the initial IMAC elution profile were taken and analysed by SDS-PAGE

(Figure 3.8B). Through staining of the SDS-PAGE gel the presence of a strong band corresponding to the approximate size of monomeric DmsDHis was observed. Fractions appearing to contain DmsDHis were then pooled and mixed with purified TEV protease for overnight cleavage of the N-terminal His-tag. Cleaved DmsD was separated from uncleaved

DmsDHis by reverse IMAC, and fractions again analysed by SDS-PAGE (Figure 3.8C). Those containing the purest DmsD were pooled and concentrated, ready for further purification through size exclusion chromatography using a Superdex 75 column. The resulting protein was confirmed to be DmsD through Western analysis, before being concentrated and flash frozen to be stored until used. 115

3.3.2.2 Crosslinking analysis of purified DmsD and synthetic spYnfE

In order to probe YnfE signal peptide binding ability of DmsD, a chemical crosslinking approach was undertaken. It has been shown in Chapter 2 that DmsD binds the hydrophobic region of the YnfE signal peptide. However, less is known about the binding site on DmsD. This was investigated by a chemical crosslinking approach, using purified DmsD protein and a number of synthesised peptides of various sequence stretches of the YnfE signal peptide (Table 3.1).

Table 3.1 Details of synthetic peptides used in this study.

Name Amino acid sequence Length Molecular Use weight Peptide 1 SRRTLVKSAALGSLALAAGGVSLPFGMRK 29 a.a. 2.915 kDa crosslinking

Peptide 2 SLALAAGGVSLPFGMRK 17 a.a. 1.675 kDa crosslinking

Peptide 3 SRRTLVKSAALGSLALAAGGVSLPFGMR 28 a.a. 2.787 kDa crosslinking

Peptide 4 SLALAAGGVSLPFGMR 16 a.a. 1.547 kDa ITC, crystal trials

In a first experiment, purified DmsD and the synthetic peptide named peptide 1 were incubated together in solution, and three different chemical crosslinkers were tested for their ability to induce crosslinks to form. Disuccinimidyl suberate (DSS) is a chemical crosslinker that is able to covalently link lysine residues and other primary amine groups that are within 11.4 Å of each other. 1-Ethyl-3-[3-dimethylaminopropyl] carbodiimide hydrochloride (EDC) is a zero- length crosslinker capable of inducing crosslinks between carboxyl groups and primary amines.

The well-known chemical formaldehyde is a bifunctional crosslinker that can be used to form covalent bonds over distances of approximately 2.3 Å, and primarily reacts with lysines and tryptophans as well as primary and secondary amines at the N-terminus (Sutherland et al.,

2008). After induction of crosslinking, levels of crosslinked protein-peptide could then be analysed through SDS-PAGE and Western blotting.

116

Figure 3.9: DSS or formaldehyde, but not EDC, can crosslink a peptide ligand to DmsD Purified DmsD (2 mM) and synthetic peptide 1 (100 mM) were incubated together in a final volume of 100 µl prior to the addition of chemical crosslinkers to induce crosslinks between DmsD and synthetic peptide. Crosslinker concentrations were: 1 mM DSS, 1 mM EDC or 1% formaldehyde. Reactions were stopped after 30 minutes with 50 mM Tris. 10 µl samples were analysed by: (A) SDS-PAGE (17 % acrylamide gel) and (B) western blotting. Antibodies used were anti-DmsD (1:20,000) and anti-rabbit (1:10,000).

Analysis of DmsD samples exposed to chemical crosslinkers in the presence and absence of synthetic peptide 1 were compared to DmsD with no crosslinker present (Figure 3.9). Samples incubated with DSS had an additional two bands in the presence of peptide 1 compared to that without, one a strong band of higher molecular mass and the other a weaker band at still higher molecular mass (Figure 3.9). This indicated DSS was probably able to form a covalent crosslink between DmsD and the synthetic peptide. In comparison, samples with EDC did not show any additional band in the presence of peptide 1 (Figure 3.9). This suggests that EDC is not a suitable crosslinker for such experiments. In addition, it is recommended to include N- hydroxysuccinimide (NHS) or its water-soluble analogue sulfo-NHS in the reaction with EDC to improve efficiency of the reaction. This was not done in these experiments and may be the reason for a lack of crosslinking. Formaldehyde was able to induce a large amount of crosslinks

(Figure 3.9). However, it was notable here that in the sample of DmsD without peptide 1 formaldehyde formed covalent bonds between DmsD monomers, leading to the presence of a band at approximately 40 kDa, the correct size for DmsD dimer (Figure 3.9). Interestingly, upon the addition of peptide 1, the amount of DmsD dimer noticeably decreased and an additional 117 band appeared of the correct size for crosslinked DmsD-peptide (Figure 3.9). From these results, EDC was no longer used for these crosslinking experiments, whilst DSS and formaldehyde both appeared suitable.

Figure 3.10: SDS-PAGE and Western blot analysis of DmsD crosslinked by DSS with synthetic peptide 1 and 2. Purified DmsD (2 mM) and synthetic peptides 1 & 2 (100 mM) were incubated together in a final volume of 100 µl prior to the addition of 1 mM DSS, with the reaction being stopped after 30 minutes with 50 mM Tris. 10 µl samples were analysed by: (A) SDS-PAGE (17 % acrylamide gel) and (B) western blotting. Antibodies used were anti-DmsD (1:20,000) and anti-rabbit (1:10,000).

As it had now been shown that crosslinking could be induced using DSS and peptide 1, interactions of the DmsD protein with a smaller region of the signal peptide of YnfE was tested using synthetic peptide 2 (Figure 3.10). This peptide 2 was limited to only the hydrophobic region of the signal sequence, and contained just one non-native lysine at the C-terminal end with the intention for use in DSS crosslinking (Table 3.1). These synthetic peptides do not contain primary amines at the “N-termini”. Covalent links were introduced between DmsD and synthetic peptides with DSS, with samples then being evaluated through SDS-PAGE and

Western analysis (Figure 3.10). For peptide 2 a weak crosslinked band was observed above that of noncrosslinked DmsD (Figure 3.10), whilst for peptide 1 a very strong crosslinked band was present, along with a further band slightly larger again (Figure 3.10). In addition the presence of a band at approximately 40 kDa could be observed in the DmsD only sample, as 118 was observed in Figure 3.9. Interestingly, the dimer band appears to reduce in intensity upon the addition of peptide 1 and can barely be seen at all in the presence of peptide 2 (Figure

3.10). This suggests that DmsD is capable of interacting in a dimeric form that becomes disrupted upon spYnfE binding, perhaps hinting at a shared binding site on the chaperone protein.

Given peptide 1 contains two lysine residues, next the requirement of the additional non- native lysine residue present in these peptides was tested. A third synthetic peptide was generated of the full Tat motif and hydrophobic region of spYnfE, but without the lysine at the

C-terminus, therefore only containing one internal native lysine (Table 3.1). The ability of this peptide to form covalent links with DmsD will give information on how much the native and non-native lysines are involved in interactions and DSS crosslinking. Western blots showed crosslinked bands in both 1:25 and 1:100 ratio of DmsD protein to peptide 3. 1:1 ratio sample did not appear to have DmsD crosslinked (Figure 3.11). This indicates that peptide 3 is also able to form interactions with DmsD.

Figure 3.11: Ability of synthetic peptide 3 to form crosslinks with DmsD. DSS crosslinking was induced between purified DmsD and synthetic peptide 3 lacking the C-terminal lysine present in peptides 1 and 2. Varying millimolar ratios of DmsD protein to peptide as indicated were incubated for 1 hour prior to addition of 1 mM DSS. Reactions were stopped by the addition of 50 mM Tris, before 10 µl samples were analysed by SDS-PAGE (17 % acrylamide gel) and Western blot (Anti-DmsD – 1:20,000. Anti-rabbit – 1:10,000).

3.3.2.3 Investigating the location of the crosslinks between DmsD and ssYnfE

As it was observed that purified DmsD was able to reliably crosslink synthetic peptides using

DSS, it was attempted to analyse this crosslinking further through mass spectrometry. This was 119 hoped to give a more definitive picture of where the crosslinks are occurring between DmsD and the peptide. However, results only indicated crosslinks between the synthetic peptide itself (data not shown). Examination of the amino acid sequence of DmsD shows only two lysines in the entire protein (Figure 3.12A), and these do not fall in ideal areas for analysis with trypsin digest and mass spectrometry to be successful, with the fragments obtained after digest being too large for identification. Therefore a different approach to analyse where interactions were occurring was required.

It was decided that DmsD-peptide interactions would instead be investigated through the use of genetics – a bank of DmsD lysine variants was prepared. The reasoning behind this was that because DSS forms covalent links between free amines, and as there are only three potential crosslinking sites on DmsD (including the N-terminus), it should be fairly straightforward to identify which of these free amines is involved in crosslinking through the individual and double substitution of lysines for a glutamine. Any disruption in crosslinking caused by the loss of a lysine can be observed by SDS-PAGE and western as a lack of crosslinked sample bands.

Using the pQE80 NHis-TEV-DmsD vector as template, Lys-61 and Lys-200 were individually substituted by Gln residue along with a double Lys-Gln variant using Quikchange site-directed mutagenesis. Substitutions were confirmed by DNA sequencing. Proteins were then purified as before by IMAC, TEV cleavage and reverse IMAC, and SEC. DSS crosslinking was then performed using the purified variant proteins and native DmsD protein with synthetic peptide

2 (Figure 3.12).

120

Figure 3.12: Evidence that peptide ligands are crosslinking to the N-terminus of DmsD. A – DmsD (S. Typhimurium) primary amino acid sequence. The two lysines present in DmsD are highlighted. It was observed that they are located in different halves of the protein sequence. These were substituted for glutamines by site-directed mutagenesis, and purified by Ni IMAC and SEC. B – Expression tests with native DmsD and K61Q and K200Q variants. Overnight cultures of BL21 pLysS cells containing vectors for these DmsD variants were subcultured and grown at 37 oC for 1 hour. Samples were taken before addition of 1 mM IPTG (-) and after three hours growth in the presence of IPTG (+). These were mixed 1:1 with Laemmli buffer before being visualised by staining of SDS-PAGE gels (12.5 % acrylamide gel). C - Western blot analysis of purified DmsD Lys variants crosslinked with synthetic peptide 2. DmsD and peptide were incubated together in varying molar ratios as indicated (DmsD:peptide) before the addition of 1 mM DSS to induce crosslinking. The reaction was stopped after 30 minutes by the addition of 50 mM Tris. 10 µl samples were separated by SDS-PAGE (17 % acrylamide gel) and visualised by western blot (anti- DmsD – 1:20,000 & anti-rabbit – 1:10,000).

Western blots showed the presence of an additional band that corresponds to a crosslinked

DmsD-peptide complex in all protein samples incubated with synthetic peptide 2 (Figure

3.12C). This result indicates that neither a single nor a double Lys substitution was able to disrupt crosslinks between synthetic peptide 2 and DmsD. It was concluded that DSS crosslinks must be able to form between the synthetic peptide and the N-terminus of DmsD, which contains a free amine, as crosslinking was not completely disrupted even with the loss of both 121 lysines within DmsD (Figure 3.12C). To investigate this further, purified native DmsD was used that did not have the added histidine-tag removed, which should result in an additional 15 residues at the N-terminus. All other purification steps were performed as described before, but TEV-cleavage and reverse Nickel IMAC was not required. The resulting DmsDHis from SEC purification was checked and seen to be of a relatively high purity through SDS-PAGE (Figure

3.12A). Western analysis was used to confirm the presence of the His-tag (Figure 3.12B).

Figure 3.13: Analysis of purified DmsDhis interactions with synthetic ssYnfE peptides. A - DmsDhis was purified by Ni IMAC and SEC, with resultant fractions checked for purity by SDS-PAGE (12% acrylamide gel). B – The presence of the N-terminal His-tag was checked through Western blot of pooled fractions from SEC with anti-His (1:5,000) and anti-rabbit (1:10,000) antibodies. C - DmsDhis was incubated with synthetic peptides 1 and 2 at varying molar ratios (His-DmsD:peptide). 1 mM DSS was added to induce crosslinking and after 30 minutes the reaction was stopped by the addition of 50 mM Tris. 10 µl samples were analysed by SDS-PAGE (17 % acrylamide gel) and western blot (anti-DmsD – 1:20,000 & anti-rabbit – 1:10,000).

As a test to see if interactions could be disrupted to the N-terminal free amine, DmsD with His- tag still attached was also subjected to DSS crosslinking. DmsDHis was incubated with both synthetic peptides 1 and 2, with crosslinked samples then analysed by Western blot (Figure 122

3.13C). It appeared that both synthetic peptides were still forming covalent bonds with

DmsDHis, although crosslinked peptide 2-DmsDHis did appear at very much lower levels than that of peptide 2-DmsD (Figure 3.9, Figure 3.10)

3.3.2.4 DmsD can be co-purified in solution with synthetic spYnfE

The crystal structure of S. Typhimurium DmsD in its monomeric form was solved several years ago by Qiu et al. (2008). However the protein arrangement whilst in interactions with a signal peptide of one of its Tat substrate partners remains unknown. It was therefore attempted to obtain crystals of DmsD whilst in complex with a synthetic peptide of the hydrophobic region of spYnfE (residues 23-38), named synthetic peptide 4 (Table 3.1). For this experiment, purified native DmsD was incubated with synthetic peptide 4 prior to the final purification step of size exclusion chromatography, in order to attempt to co-elute protein and peptide together in a complex.

123

Figure 3.14: Elution profile for SEC of DmsD purified protein, with or without synthetic peptide 4. DmsD protein purified by IMAC was subjected to SEC through a Superdex 75 10/300 GL column, either alone or after several hours incubation with synthetic peptide 4 in a 1:1 molar ratio. Elution of protein was monitored by measurement of absorbance at 280 nm.

The soluble fraction of cells overexpressing DmsD was subjected to the initial IMAC purification followed by overnight TEV-cleavage and reverse IMAC as described previously. The resulting purified DmsD was then incubated with synthetic peptide in an approximate 1:1 molar ratio for three hours before the sample was concentrated and eluted through a Superdex 75 10/300

GL column. A single peak was observed during elution, similar to that observed for SEC purification of DmsD alone (Figure 3.14). Fractions were visualised by SDS-PAGE, with those appearing to contain pure DmsD being pooled. As synthetic peptide 4 is too small to be visualised using SDS-PAGE, its presence in solution with DmsD required a more sensitive analytic approach. This was performed using electrospray ionisation (ESI) mass spectrometry.

Samples from pooled fractions of DmsD alone and DmsD plus peptide 4 after SEC were 124 analysed (Figure 3.15). This confirmed the presence of peptide 4, as an additional mass peak of

1546.8 Daltons was seen in the DmsD plus peptide sample from SEC compared to DmsD alone, exactly corresponding to the mass of peptide 4. These data show that DmsD and synthetic peptide 4 are able to interact with each other strongly enough to co-elute when subjected to size-exclusion chromatography.

As DmsD and synthetic peptide 4 had been shown to have co-eluted during SEC, it was concluded that they had most likely formed a complex. It was decided that this complex should be subjected to crystal screens in an attempt to obtain crystals. As the crystal structure of

DmsD from S. Typhimurium has already been solved, it should be possible to obtain a crystal to elucidate the DmsD-spYnfE complex structure using the original crystallisation conditions as a starting point. The protein-peptide sample was concentrated down to approximately 18 mg/ml, with this then being entered into crystallisation trials using a number of commercially available crystallisation screens. A number of conditions were therefore taken further and trays screening around these conditions were manually set up. The original conditions from which a crystal was acquired and used to solve the DmsD structure by (Qiu et al., 2008) were also used and optimised around. Unfortunately these manual crystal screens did not yield any usable crystals, and as such it has not yet been possible so far to gain a clearer picture of the spYnfE-DmsD interface using this approach.

125

Figure 3.15: ESI mass spectrometry analysis of DmsD subjected to size exclusion chromatography (SEC) with and without synthetic peptide 4. The ability of synthetic peptide 4 to co-purify with DmsD was tested. DmsD alone or DmsD and synthetic peptide 4 (1:1 molar ratio) were eluted through a Superdex 75 10/300 column, with resultant protein being analysed through ESI mass spectrometry. A – Analysis of sample from DmsD alone. A single peak was identified with a mass of 23538 Daltons. B - Purified DmsD protein incubated with synthetic peptide 4 prior to SEC was also analysed by ESI and showed a profile with two peaks, having masses of 1546.8 Daltons and 23538.3 Daltons respectively. Inset boxes show base peak chromatograms for each sample, whilst main boxes show averaged mass spectra.

126

3.3.2.5 Does nucleotide binding affect peptide crosslinking?

Finally, previous experimental and bioinformatics analysis of DmsD and other TorD family chaperone members has hinted at a possible involvement of nucleotide in the activity of these chaperones. For example, it has been shown that TorD from E. coli is able to bind to and hydrolyze GTP in the presence of magnesium chloride (Guymer et al., 2010, Hatzixanthis et al.,

2005). In addition, although the protein sequence itself does not possess the entirety of the highly conserved guanine nucleotide-binding domain found in canonical GTPases, it does have some similarities and contains a potential G-4 guanine specificity motif (Guymer et al., 2010).

Surface analysis of Shewanella massilia TorD and S. Typhimurium DmsD revealed corresponding surfaces with similarities to a GTP binding surface from guanylyl cyclase from

Homo sapiens (Qiu et al., 2008). Many GTPases are thought to act as molecular switches through conformational changes induced through GTP/GDP binding, and it could be hypothesised that such GTPase activity may be involved in conformational changes required for Tat signal peptide binding, or vice versa. Therefore, the effects of the presence of GTP as well as several other nucleotides were tested on crosslinking of DmsD with synthetic peptide.

127

Figure 3.16: Analysis of crosslinking interactions of DmsD and peptide 1 when pre-incubated with various nucleotides. Purified DmsD (2 mM) and synthetic peptide 2 (100 mM) were incubated together with varying molar concentrations of the nucleotides ATP, GTP, ADP, XTP, GMP and cGMP with MgCl2 added to HEPES buffer, prior to the addition of 1 mM DSS and the reaction eventually being stopped with 50 mM Tris after 30 minutes. 10 µl samples were separated by SDS-PAGE (17 % acrylamide gel) before the gel was stained to reveal protein bands.

DmsD was incubated with synthetic peptide 1 in addition to various nucleotides. ATP, GTP,

ADP, XTP, GMP and cGMP were all added to the reaction mixture at final concentrations of 0.1 mM, 0.2 mM and 0.5 mM. MgCl2 was also included as a number of GTPases have been shown be magnesium-dependent (Zhang et al., 2000). DmsD alone was also incubated with each molar concentration of nucleotide to act as negative control. After several hours incubation the chemical crosslinker DSS was introduced to the reaction mixture in order to crosslink any closely interacting proteins. After ending the reaction by the addition of Tris samples were analysed through SDS-PAGE and Western blot. Upon staining, acrylamide gels did not appear to show any significant differences in the levels of peptide binding between samples containing DmsD and peptide 1 without any nucleotide and those with a nucleotide present

(Figure 3.16).

128

3.4 Discussion

3.4.1 New insight into the Tat proofreading function of DmsD using a genetic

screen

Through the utilisation of a random mutant DmsD library, it was possible to identify a number of amino acids on DmsD that had potential importance in interactions with YnfE. In S.

Typhimurium strain KM02 (ynfE A28Q) three DmsD variants with a total of seven amino acid substitutions were identified that had disrupted selenate reductase activity. When plotted on the crystal structure of DmsD from S. Typhimurium it could be observed that five of these residues, namely W91, R94, S96, G100 and T103, are located within the same region on the protein, with these five all also being exposed on the surface. These residues are nearby the loop region that is known to separate the N- and C-terminal regions of the protein. This loop region contains the conserved E(Q)PxDH motif that is found in all TorD family chaperone proteins and has been previously suggested to be involved in DmsA signal peptide binding

(Chan et al., 2008). Val-16 and Glu-171 are present in separate regions of the DmsD protein.

V16, the only amino acid found to be involved in signal peptide recognition, is found near a region of DmsD that has been extensively studied and suggested as the region involved in peptide binding (Figure 3.17) (Chan et al., 2008, Shanmugham et al., 2012).

129

Figure 3.17: Potential amino acid residues on DmsD involved in signal peptide binding. Amino acids have been highlighted in a number of studies as playing a role in signal peptide binding. Residues of interest are indicated in yellow. A – V16, identified in this work with a DmsD mutant screen as displaying disrupted selenate reductase activity in S. Typhimurium strain KM02. B – Residues of importance for spDmsA interactions as shown through random and site-directed mutagenesis (Chan et al., 2008). C – Residues found to display chemical shift perturbations in their 15N-HSQC TROSY spectra upon binding spDmsA (Stevens et al., 2013). These are highlighted on the crystal structure of S. Typhimurium DmsD (PDB 1S9U).

In a multiple alignment of a number of DmsD and TorD sequences, it was observed that all amino acids that were highlighted in this study apart from Gly-171 are conserved at least amongst DmsD, with a number being conserved or conservatively substituted across all sequences used (Figure 3.18). This high level of conservation does indeed suggest these amino

acids have some level of importance in DmsD function. V16 is conserved in DmsD proteins, as

is the nearby G18 residue, which was shown by NMR as displaying perturbations in the

presence of DmsA signal peptide (Shanmugham et al., 2012).

130

Figure 3.18: Multiple alignment of DmsD/TorD-like family chaperones showing possible signal peptide interacting residues. Aligned amino acid sequences of TorD from S. Typhimurium, E. coli and Shewanella, DmsD from S. Typhimurium and E. coli and predicted cytoplasmic chaperone protein STM0610 from S. Typhimurium. Amino acid residues highlighted through random mutant screen are indicated by arrows. V16, shown to be significant in this work, is indicated by a red arrow. Residues discussed by Stevens et al. (2013) are shown in blue, whilst those of significance from Chan et al. (2008) are shown in green.

Through individual DmsD variant testing, it was observed that Val-16, Trp-91 and Gly-100 were important for selenate reductase activity, with substitutions in these positions disrupting this activity. From this result it could be suggested that these three amino acids are involved in the proofreading activity of DmsD with YnfE, with substitutions disrupting the interaction and therefore preventing proofreading and formation of mature YnfE. Interestingly, a more specific look at interactions of DmsD variants with the signal peptide of YnfE through bacterial two- hybrid analysis revealed only DmsD V16Q to be obviously impaired in signal recognition. Val-16 is on a separate face from the majority of other residues highlighted through the mutant 131 screen, near to a region of the protein that has been previously highlighted as a putative binding site for the signal peptide of DmsA (Chan et al., 2008). DmsD Val-16 substitutions were also shown in Chan et al. (2008) to show a decreased interaction with the signal peptide.

Valine is a hydrophobic amino acid rarely found on the surface of proteins, and there appears to be a number of additional hydrophobic residues in the surrounding area, so it is feasible that it forms interactions with the hydrophobic stretch of the YnfE signal peptide important for binding. It is already known that the hydrophobic region of the signal peptide is important for binding, at least in the case of DmsA and its cognate substrate protein DmsD. It has been shown that for E. coli the hydrophobic region of spDmsA, and potentially the C- and H- regions in combination, are essential for interactions (Shanmugham et al., 2012, Winstone et al.,

2013a). This agrees with the idea of a hydrophobic pocket or groove being required for signal peptide binding via this hydrophobic stretch. In addition, negatively charged amino acids are also found nearby, potentially interacting with the positively charged amino acids of the Tat motif and n-region of the signal peptide. However, this part of the signal peptide does not appear to be essential for binding and so this is less likely to be important in primary interactions (Shanmugham et al., 2012, Winstone et al., 2013a).

Whilst Val-16 was unable to bind spYnfE in bacterial two-hybrid studies, Trp-91 and Gly-100 retained this ability despite displaying disrupted selenate reductase activity in KM02 (spYnfE

A28Q). It may be that these two residues are not involved in signal peptide binding.

Alternatively, from this result it could be proposed that these residues are required for signal peptide binding, hence the positive bacterial two-hybrid result, and are in fact involved in binding the A28 residue of spYnfE. This would explain the lack of selenate reductase activity observed in KM02 expressing these DmsD variants, as substitution of A28 could disrupt DmsD interactions and subsequently interfere with YnfE proofreading.

Recent work into the signal peptide binding site of DmsD has highlighted a particular region of the protein as having importance in interactions. Through analysis of NMR chemical shift 132 patterns of DmsD residues in the presence and absence of spDmsA, the area around helices 6 and 7 and including the loop between these helices, was suggested to form part of a hydrophobic cleft in DmsD that extends around towards the front of the protein (Stevens et al., 2013). This potential binding site was also thought to extend towards the front of the protein and perhaps also involve helices 3 and 4. Whilst not in one of these helices, residue

G18 was one of a number of amino acids shown to display a perturbation in the presence of spDmsA, indicating it also plays a role in peptide binding. Residues V16 and G18 may form one end of the hydrophobic binding cleft involved in signal peptide interactions, being relatively close to the hydrophobic pocket previously identified. These two residues are conserved between S. Typhimurium and E. coli DmsD, as are many of the other residues thought to play a role in signal peptide binding (Figure 3.18), and this conservation suggests an important role in function.

3.4.2 Chemical crosslinking as a tool to study the signal peptide binding site on

DmsD

A synthetic in vitro approach with purified DmsD protein was also used to investigate regions of importance for signal peptide binding on DmsD. Using synthetic peptides of differing lengths and sequences, the presence of covalent crosslinks formed by DSS was observed. Three different peptides were utilised during this work; a peptide covering the entire hydrophobic region containing just one Lys residue artificially added to the C-terminus of the peptide sequence (Peptide 1, Table 3.1), another covering both the twin-arginine motif and the hydrophobic stretch and containing two lysines – one native Lys present in the Tat motif plus the synthetically introduced Lys at the C-terminus (Peptide 2, Table 3.1), and a third peptide also covering both the twin-arginine motif and the hydrophobic stretch but without the addition of a Lys to the C-terminus, therefore containing just one lysine in the Tat motif

(Peptide 3, Table 3.1). It was observed that all three synthetic peptides were able to crosslink 133 with purified DmsD, with peptide 2, the longest and only peptide containing two lysine residues, appearing to be the most efficient at interacting with DmsD. Peptides 1 and 3 also appeared to be able to crosslink with purified DmsD but at a lower level. This suggests that both the hydrophobic region and the twin-arginine motif regions of spYnfE interact with DmsD, and whilst one of these regions is enough for an interaction to occur, interactions and DSS crosslinking can occur much more efficiently with the peptide covering both these regions.

It had been originally planned to induce crosslinks between DmsD and the synthetic peptides before identifying the crosslinked residues in more detail using tryptic digest and mass spectrometric analysis. Unfortunately, however, this was unsuccessful and no crosslinks in

DmsD-peptide samples could be detected other than peptide-with-peptide. DmsD only contains two lysines to which peptides could be crosslinked using DSS (Figure 3.19) (plus the primary amine at the N-terminal), and upon digest with trypsin which cleaves next to Lys and

Arg residues, the resulting fragments containing the crosslink would be too small to be detected by a mass spectrometer. An alternative genetic approach therefore was developed.

134

Figure 3.19: Locations of the two lysines of DmsD. A - Lysine residues 61 and 200 are shown in yellow on the 3D structure of S. Typhimurium DmsD. Both are present in loop regions of the structure. B - Aligned amino acid sequences of TorD from S. Typhimurium, E. coli and Shewanella, DmsD from S. Typhimurium and E. coli and predicted cytoplasmic chaperone protein 610 from S. Typhimurium, as before. Lysine residue locations in S. Typhimurium DmsD are indicated with red arrows.

It was decided to create genetic variants of DmsD where the lysines in the sequence were substituted, either individually or together for glutamines, with purified protein of these variants then being used in DSS crosslinking experiments as before. If crosslinking could be completely disrupted by substituting either one or both lysines of DmsD, this would reveal the 135 important lysine residue and point to the region involved in peptide interactions. To this end,

DSS crosslinking of peptide with all three DmsD variants was performed, however was not observed to be disrupted in any of the three samples. From this it could be suggested that the free amine at the N-terminus is able to interact with the peptide and this is where crosslinks are forming. Attempts to prevent this interaction by testing purified, nonprocessed DmsDHis were also unsuccessful. Of course, testing of the crosslinking ability of a nonprocessed His- tagged double lysine variant would be the key experiment to further confirm the DmsD N- terminus as the region of interaction.

Although crosslinking did not appear to be completely inhibited in any of the described conditions, it did appear to be reduced to some extent in some cases. Both DmsD K61Q and

DmsDHis show slightly lower intensity crosslinked bands with Western analysis when compared to native DmsD or other variants. One hypothesis that could be put forward from these results is that the peptide interacts with a region of DmsD that enables it to form crosslinks with both the N-terminal residue and with Lys-61. By removing the primary amine either present on the lysines or the N-terminus individually crosslinking would not be completely prevented, perhaps instead merely limiting the amount of crosslinking occurring. This would also fit with the apparent importance of the amino acid residue Val-16 for signal peptide binding, as Val-16 and the N-terminal region are both relatively near to Lys-61. With the chemical crosslinker DSS having a spacer arm of 11.4 Å, it may be close enough to form crosslinks with a peptide interacting with this region. Study of the amino acid residues present in the surrounding area indicates a number of hydrophobic residues are present around Lys-61, as well as at the N- terminus. This information reinforces this as the proposed binding site for the YnfE signal peptide, as hydrophobic interactions may be possible between this region and the hydrophobic region of spYnfE. From these results however, it is not possible to comment on the stoichiometry of the synthetic peptide binding to DmsD, and this would require further testing using the different synthetic peptides with DmsD variants. In addition, further work is 136 required on crystallisation of the DmsD-spYnfE complex purified by SEC, as this would divulge much greater detail about the location of these interactions.

Similar (unpublished) interaction studies were carried out by former PhD student Dr Jennifer

Dow in Professor Palmer’s group (Dundee), investigating interactions between the substrate chaperone pair TorA and TorD in E. coli. DSS crosslinking experiments were performed using purified TorD and TorAD complexes (with and without native TorA signal peptide) with synthetic TorA signal peptide containing a non-native lysine on the C-terminal end. Using mass spectrometry it was possible to identify three crosslinking sites on TorD, forming crosslinks with two sites on the synthetic peptide. Both the N-terminal primary amine and the C-terminal non-native lysine were able to crosslink to K105 on TorD, in addition to K98 being shown to crosslink in TorAΔspTorD samples (Figure 3.20). The occurrence of crosslinks between the non- native lysine on the synthetic peptide and K98 of TorD hints at interactions of amino acids near the C-terminus of the TorA signal peptide. Looking at this information together with previous

DmsD studies highlighting additional residues (Figure 3.17) it could be suggested that the C- terminal area of the DmsA signal peptide may interact with DmsD near these residues Leu-92 and Arg-107, with the hydrophobic and n-regions of the signal peptide likely to stretch across the DmsD molecule along the hydrophobic groove identified and past residues such as P86,

W87, P124 and D126 (Stevens et al., 2013). This would bring the peptide into close proximity to residue V16, and may explain its apparent importance in spYnfE binding.

137

Figure 3.20: Residues of DmsDSt equivalent to those identified in E. coli TorD crosslinking experiments. Crosslinking experiments have identified residues K98 and K105 of E. coli TorD as interacting with a synthetic TorA signal peptide. The equivalent amino acids in S. Typhimurium DmsD are L92 and R107, highlighted here in red.

3.4.3 Evidence for DmsD biochemical activities other than signal peptide binding

Whilst interactions with spYnfE were disrupted in DmsD V16Q, DmsD G100Q appeared to be able to interact with spYnfE at normal levels, in fact with the highest level of interaction of all

DmsD variants tested. In addition, W91Q displayed near normal levels of interaction, despite completely disrupting selenate reductase activity in KM02. Therefore it must be concluded that DmsD Val-16 is required for YnfE signal peptide interactions, whilst Trp-91 and Gly-100 is not, instead playing a role elsewhere in the biosynthetic pathway of selenate reductase. Whilst

DmsD residue Val-16 is implicated in YnfE signal peptide binding (as shown by bacterial two- hybrid analysis), the role of residues Trp-91 and Gly-100 must also be questioned. Through examination of chaperone protein sequence alignments, it can be seen that G100 is conserved in DmsD proteins but not in TorD. In addition, G100 is found at the end of a loop region of the protein that does not appear to exist in TorD, suggesting this region may be involved in some function not conserved in TorD (Figure 3.21). Meanwhile, W91, present just before this DmsD loop region, is conserved amongst DmsD proteins and conservatively substituted in all other proteins aligned (Figure 3.21). The apparent importance of these residues in selenate 138 reductase activity despite the lack of a clear role in peptide binding suggests another role in enzyme biosynthesis. There are a number of possibilities for additional functions of the DmsD chaperone in enzyme biosynthesis.

Figure 3.21: W91 and G100 residues are conserved amongst DmsD proteins. A – W91 and G100 residues shown in yellow were demonstrated in this work to play a role in YnfE activity outside of signal peptide binding during proofreading. Also highlighted in pale cyan is the loop region of which G100 is a part of. This loop is unique to DmsD proteins. B - Aligned amino acid sequences of TorD from S. Typhimurium, E. coli and Shewanella, DmsD from S. Typhimurium and E. coli and predicted cytoplasmic chaperone protein 610 from S. Typhimurium, as before. W91 and G100 residues identified in DmsDst is indicated with red arrow, whilst the loop region of which it is a part is outlined in purple. 139

In the past it has been demonstrated that in addition to showing interactions with a number of substrates of the Tat system, DmsD is in fact able to interact with a large range of general molecular chaperones involved in protein biosynthesis (Li et al., 2010). It is possible that Gly-

100 and this area of DmsD is involved in bringing more general chaperone proteins into close proximity with YnfE in order to aid protein maturation, or to guide the Tat substrate along a cascade of a number of molecular chaperones. There have also been suggestions that DmsD is involved in interactions with components of the twin-arginine translocase in the cytoplasmic membrane, and it may be that DmsD functions both as a proofreading chaperone and also to shuttle the Tat substrate towards the twin-arginine translocase itself (Papish et al., 2003).

Another way DmsD may be required for selenate reductase activity is perhaps through bringing subunits together. YnfG does not have a Tat signal sequence and so must be transported to the periplasm via Sec or act in a ‘hitch-hiking’ fashion and be transported through Tat with YnfE. Therefore it could be hypothesised that whilst interacting with YnfE,

DmsD also forms an interaction with YnfG in order to bring it into close contact with YnfE and aid this so-called ‘hitch-hiking’ across the cytoplasmic membrane. However, bacterial two- hybrid work showed that DmsD was not able to interact with YnfG (Figure 3.7). Of course, negative results from bacterial two-hybrid analysis can also occur through production of an unstable protein. Therefore it may be tentatively concluded that DmsD is not required for aiding formation of the complex that is selenate reductase through bringing its subunits together.

A third option could be that DmsD is required for correct assembly of the molybdenum cofactor in YnfE. It is known that other Tat cytoplasmic chaperones are important for this process and play an active part in cofactor insertion. For example the chaperone TorD, a cytoplasmic chaperone highly similar in amino acid sequence to DmsD, has been shown to be able to interact with the mature region of TorA when in an unfolded state, thought to promote 140 cofactor insertion (Pommier et al., 1998). TorD also demonstrates interactions with enzymes of molybdenum cofactor synthesis and the molybdenum cofactor itself (Genest et al., 2008).

Bacterial two-hybrid analysis apparently revealed that full length or the mature region only of

YnfE did not interact with DmsD (Figure 3.7), in agreement with previous studies (Winstone et al., 2013a). However, whilst DmsD may not be involved in active site interactions as TorD is with TorA, it could still be involved in cofactor biosynthesis. In addition to finding interactions of DmsD with general molecular chaperone proteins, Li et al. (2010) also observed interactions with protein components of the molybdopterin (MPT) biosynthetic pathway, for example

MoeA, MoeB and MobB. This suggested that DmsD may take its substrates to a ‘metabolon complex’ where proteins required for molybdenum cofactor biosynthesis are localised. This is a possibility for the Trp-91 and Gly-100 residues and the surrounding region of DmsD, and the potential role of DmsD in molybdenum cofactor assembly through recruitment of molybdopterin biosynthetic proteins would be an interesting route of inquiry. In summary,

DmsD may have important additional roles in Tat substrate maturation yet to be revealed.

141

4 Understanding the biosynthesis of E. coli

TorA and its relationship with the Tat

pathway

142

4.1 Introduction

One of the first cytoplasmic chaperone proteins to be reportedly involved in Tat proofreading was TorD from E. coli. This is a 22.5 kDa protein encoded by the torCAD operon that also encodes for the molybdoenzyme trimethylamine N oxide reductase (TorA) and the membrane bound pentahaem cytochrome TorC (Figure 4.1). Periplasmic TorA is the catalytic subunit of the enzyme possessing a Mo-bis-MGD cofactor, and receives electrons from TorC to reduce

TMAO to TMA.

143

Figure 4.1 Organisation of TMAO reductase in E. coli. A – TMAO reductase components TorA and TorC are encoded in an operon alongside the torD gene coding for the cognate chaperone protein of TorA. Expression of the torCAD operon is under the control of the torSTR and torI genes and is positively regulated by TMAO. The torSTR genes are proximal to the torCAD operon whilst torI is encoded elsewhere on the chromosome. Chromosomal locations as base numbers are indicated. B – Arrangement of TMAO reductase subunits at the cytoplasmic membrane. Membrane-bound TorC contains five haems and acts as the site for menaquinol (MQH2) oxidation, whilst periplasmic protein TorA transiently interacts with TorC and catalyses the reduction of TMAO to TMA. Crystal structures of TorA (1TMO) (Czjzek et al., 1998) and TorD (PDB 1N1C) (Tranier et al., 2003) from Shewanella massilia are shown, which share high sequence identity with the E. coli equivalents.

The 3D crystal structure of homodimeric TorD from Shewanella massilia was determined a number of years ago and follows the typically all-helical architecture as seen throughout the

TorD cytoplasmic chaperone family. Interestingly, when in dimeric form TorD exhibits a unique structural feature of extreme domain swapping between subunits, where the N-domain of one subunit interacts with the C-domain of the other (Tranier et al., 2003). This fold has not been observed in any other solved TorD-like chaperone structures to date and this level of domain 144 swapping, which is often restricted to a single helix or strand between proteins, is considered very unusual.

Early investigations demonstrated that cells devoid of TorD were affected in the stability and activity of TorA, with TMAO reduction decreasing two fold in a torD- strain, hinting at the importance of TorD for TMAO reductase activity (Pommier et al., 1998). E. coli TorD shares 24

% overall sequence identity with E. coli DmsD, and is thought to prevent premature export as well as stabilizing the apoprotein and facilitating cofactor insertion (Genest et al., 2009,

Sargent, 2007a).

Recent work showed binding of TorD to TorA appears to occur with a 1:1 stoichiometry (Dow et al., 2013). Moreover, Dow et al. (2013) demonstrated that that even a signal-less TorA could bind tightly to the TorD protein. Indeed, the low-resolution small angle X-ray scattering (SAXS) envelope generated from this signal-less complex appeared very similar in architecture to the

SAXS envelope of the signal-containing complex (Dow et al., 2013). These data suggest a single

TorD molecule binds to the TorA precursor at two different, but nearby, sites within the N- terminus. It also suggests that there are two distinct binding sites for different TorA epitopes on the surface of TorD. The work of Dow et al. (2013) concluded with some evidence that corroborated long-held views on the role of TorD, which was that the chaperone could maintain TorA in a conformation conducive to cofactor loading where “Domain IV” at the extreme C-terminus of TorA was displaced. In follow-up, unpublished, experiments a complex of TorD with truncated TorA (lacking Domain IV) was isolated (J.M. Dow, T. Palmer & F.

Sargent). SAXS analysis of this complex revealed that TorD was bound to truncated TorA exactly as before, thus corroborating the view that a single TorD molecule bind directly to the

N-terminus of TorA (Figure 4.2A,B).

145

Figure 4.2: The TorAD binding interaction. A – Ab initio dummy atom models of full length (grey) and ΔC-terminus (blue) TorAD complexes generated from experimental SAXS data, superimposed for direct comparison. Both complexes appear to display similar structures, with the slightly shorter appearance of the truncated complex due to its lacking domain IV. B – Superimposition of ab initio and rigid body models of ΔC-terminus TorAD complex. Rigid body modelling was created using PDB structures 1TMO (Czjzek et al., 1998) and 3CWO (Ramasamy & Clemons Jr, 2009), with green representing TorD and cyan representing TorA domains I-III. NMR data in the absence (C) and presence (D) of TorA signal peptide. A number of chemical shifts can be seen in the presence of signal peptide, highlighted in red. Work in (A) and (B) was performed by Dr Jennifer Dow, whilst (C) and (D) was performed by Dr Julian Maillard.

Additional unpublished work from this laboratory has attempted to understand the binding mechanism between TorD and the TorA signal peptide (J. Maillard, G. W. Vuister, & F.

Sargent). NMR experiments were attempted where 15N-labelled TorD was incubated with unlabelled synthetic TorA signal peptide in the spectrometer (Figure 4.2C,D). The peptide– 146 protein interaction was attempted to be followed through perturbation of the cross-peaks in the [1H-15N]–heteronuclear single quantum correlation (HSQC) spectrum (Figure 4.2C,D).

However, upon addition of the peptide, a general broadening of some of the resonances was observed, which is consistent with a complex in “intermediate exchange” i.e. a complex with neither very tight nor very loose binding, and so not straightforward to study by NMR.

Although quantitative measurements of binding affinities, or the determination of solution structures, cannot be made from this data, it can be concluded, however, that those peaks most readily affected are likely closely involved with the signal peptide binding site. Those include conserved residues such as His-125 and Asp-124, which were already implicated in

TorD function (Jack et al. 2004), and a number of new candidates including Glu-23 and Gly-122

(Figure 4.2C,D).

4.2 Aims

The aims of this Chapter were to elaborate on previous knowledge of the E. coli TorA-TorD and

TorD-TorD binding interactions in an attempt to gain further information on the TorD binding surface and its physiological function and the interrelationship between Tat substrates and the

Tat translocase. To assess the Tat proofreading activity of TorD, site-directed mutagenesis and biochemical assays were employed. To study TorD-TorD dimerization in live cells, high resolution microscopy techniques were attempted. Finally, a new approach employing super resolution microscopy was taken to explore the interplay between Tat components and Tat substrates.

147

4.3 Results

4.3.1 Molecular dissection of TorD activity

TorD is a 200 amino acid residue polypeptide that folds into an all-helical conformation dedicated to sequence-specific peptide binding. TorD binds a hydrophobic motif on the TorA twin-arginine signal peptide with high affinity (apparent Kd by ITC ~59 nM (Buchanan et al.,

2008). The location of the peptide binding site on E. coli TorD has not been described, and indeed the location of the peptide binding site of other TorD-like proteins remains a point of contention. For example, the peptide binding pockets derived for E. coli DmsD by mutagenesis

(Chan et al., 2008) and by NMR (Stevens et al., 2013) do not fully agree. Initial NMR work in the laboratory has identified a number of TorD residues implicated in peptide binding (Figure

4.2C,D). At least two of these residues (Asp-124, His-125) had been previously investigated

(Jack et al., 2004), and here the additional side chains identified in the NMR experiments were targeted. Those residues are Glu-23, Gly-112, Thr-115, and Glu-142. The glycine and threonine at positions 112 and 115, respectively, are of particular interest because they lie at the N- terminus of the unstructured ‘loop’ region implicated in domain swapping (Tranier et al.,

2003). These residues, together with the conserved Pro-122 residue within the loop that would not be visible by 15N-NMR and lies between Thr-115 and Asp-124/His-125 at the C-terminus of the loop, were chosen for mutagenesis.

The vector pUNI-PROM_TorD was chosen for these studies at it produces E. coli TorD constitutively and has been used previously in Tat proofreading assays (Jack et al., 2004). Site- directed mutagenesis was performed in order to substitute the amino acids in question with alanine (Figure 4.3A). Alanine is often used in this way due to the extremely low biochemical reactivity of its side chain without loss of secondary structure. Six side-chains were targeted, including Glu-23, Gln-31 (previously studied by Hatzixanthis et al. (2005)), Gly-112, Thr-115,

Pro-122, and Glu-142. In addition, a Gly-112-Ala / Pro-122-Ala double mutant was prepared 148 based on the hypothesis that these side chains may confer some flexibility to the hinge region.

Once mutagenesis was confirmed by DNA sequencing, the TorD proteins being produced were checked for stability through Western blotting. E. coli strain MC4100 was transformed with pUNI-PROM_TorD plus its derivatives encoding the variant proteins. Strains were grown anaerobically in LB medium (without TMAO to prevent transcription of the native torCAD operon) before whole cell samples were taken for Western blotting after separation by SDS-

PAGE. It was observed that all TorD variants appeared to be produced at levels similar to that of the native protein (Figure 4.3B), and so it could be concluded that introducing these point substitutions does not affect protein expression or stability.

Figure 4.3: Production of TorD variants. A – Primary amino acid sequence of TorD from E. coli. Amino acids targeted for substitution are shown in red. Secondary structure predictions are indicated by blue cylinders for α-helices and red cylinders for 310 (η)-helices (Tranier et al., 2003). B – Western blot analysis of pUNI-PROM_TorD variant stability. Plasmids were expressed in E. coli strain MC4100, with whole cell samples being separated by SDS-PAGE (12 % acrylamide gels) prior to visualisation by Western blot. Primary antibody – in-house anti-TorD rabbit serum, 1:20,000. Secondary antibody – anti-rabbit, 1:10,000.

149

4.3.1.1 Assembly of TMAO reductase by TorD

It was clear that TorD variants were synthesised and likely stable (Figure 4.3B). Next, the effects of amino acid substitutions on activity of TMAO reductase were tested. This assay is designed to show whether the TorD variants can still participate in cofactor loading and maturation of TorA. In this assay the oxidation of reduced benzyl viologen is coupled to the reduction of TMAO, allowing the amount of TMAO reduced in µmol per minute per mg of protein to be determined. The redox-dye benzyl viologen (BV) can be oxidised by TMAO reductase and acts as an indicator of enzymatic activity as it turns from blue/purple to colourless as it is oxidised. Using periplasmic samples from FTD100 (∆torD) cells producing the seven TorD variants, the TMAO reductase activity was calculated (Figure 4.4). Interestingly, in contrast to many published reports, in this case the ∆torD strain transformed with empty vector exhibited only 10% of the activity of a strain harbouring pUNI-PROM_TorD (Figure 4.4).

TMAO reductase activity in the periplasm of cells producing TorD variants Q31A, G112A and

T115A was similar to the cells with native TorD (Figure 4.4), thus the mutant phenotype was clearly complemented in trans by these plasmids. However, TorD E142A appeared to have a defect in TorA assembly since a two-fold decrease in activity levels compared to native TorD was observed (Figure 4.4). Moreover, TorD P122A had near 30 % native levels and the double variant TorD G112A/P122A had approximately 14 % of the activity observed in the native control. These results indicate that Pro-122 and Glu-142 are important for the ability of TorD to direct the maturation of TMAO reductase.

150

Figure 4.4: TMAO reductase activity directed by TorD variants. E. coli strain FTD100 (ΔtorD) containing the pUNI-PROM_TorD and variants was grown in anaerobic overnight cultures in LB supplemented with 0.5 % glycerol and 0.4 % TMAO, prior to isolation of the periplasmic fraction from cells. Using the redox-dye BV as an electron donor, TMAO reductase activity (µm/min/mg) could be determined through measurement of the colourimetric change at A600 as BV is oxidised. Results are displayed as a percentage of native TorD sample activity. Data expressed as means (n = 3) ± SEM.

4.3.1.2 Tat proofreading by TorD.

The TMAO reductase assay relies on the coordination of a number of biosynthetic processes including Tat proofreading (signal peptide binding), cofactor insertion, protein folding, and protein transport. In order to specifically assess the Tat proofreading abilities of these TorD variants, a proofreading assay was utilised that was previously developed by (Jack et al., 2004).

In this system, the TorA signal peptide is fused to the N-terminus of HybO, the small subunit of hydrogenase-2, replacing its own native Tat signal peptide. This results in a strain that has low levels of hydrogenase-2 activity, which can be subsequently boosted by the expression of excess torD. Because hydrogenase-2 does not contain a molybdenum cofactor, this then leaves hydrogenase activity dependent upon the Tat proofreading activity of TorD. For this work, E. coli strain RJ607-D was used, encoding the described ɸtorA::hybO chimera as well as containing a torD gene deletion (Jack et al., 2004). Hydrogenase activity can be measured in a similar way to TMAO reductase activity, although in this case BV acts as an artificial electron 151 acceptor and becomes reduced during the assay. Again through spectrometric measurement of the rate of colour change the activity of the TorA::HybO chimera can be determined.

Figure 4.5: Tat proofreading ability of TorD variants. pUNI-PROM_TorD variants were expressed in E. coli RJ607-D, a strain containing a deletion in torD and a TorA signal peptide fusion to the hybO gene at the native hyb locus on the chromosome generating a ɸtorA::hybO chimera. Therefore TorD proofreading of this chimeric protein is essential for production of functional hydrogenase-2. Enzymatic activity (µm/min/mg) of whole cell samples grown anaerobically in LB supplemented with 0.5 % [v/v] glycerol and 0.4 % [w/v] TMAO was determined through measuring the rate of reduction of BV over time, observed as a colour change measurable at A600. Results are displayed as a percentage of native TorD sample activity. Data expressed as means (n = 3) ± SEM.

E. coli strain RJ607-D was transformed with pUNI-PROM_TorD and derivatives encoding TorD variants, before overnight cultures were grown anaerobically in the presence of 0.5 % [v/v] glycerol and 0.4 % [w/v] TMAO. Whole cell samples were taken for use in the activity assay. It could be observed that RJ607-D encoding the empty vector pUNI-PROM had relatively low levels of hydrogenase activity, whilst upon introduction of native TorD hydrogenase activity could be restored (Figure 4.5). This demonstrates that pUNI-PROM_TorD can rescue hydrogenase-2 activity in this strain. TorD variants were able to complement the strain to varying degrees. TorD E23A, G112A and E142A all demonstrated near or above 100 % Tat 152 proofreading activity, whilst TorD G112A/P122A showed only slightly decreased levels at approximately 85 % native levels. TorD Q31A and T115A displayed activity levels of 66 % and

70 % respectively, whilst the largest decrease in activity could be seen in TorD P122A with 53 % native TorD levels. These data suggest that Pro-122 plays an important role in Tat proofreading and TorA assembly overall.

4.3.2 Imaging TorD-TorD interactions in live cells

Having studied the relationship between TorD and the TorA signal peptide, attention then turned to the relationship between TorD and itself. A number of biochemical and structural studies have suggested that TorD-family proteins can dimerize, or further oligomerize, sometimes with consequences for biological activity (Guymer et al., 2010, Sarfo et al., 2004,

Tranier et al., 2002). Here, the aim was to address whether dimerization occurred in living cells. The TorD protein was tagged with fluorescent proteins (and FRET pair) mCerulean or yellow fluorescent protein (YFP) on the N- and C-termini in varying combinations on the same plasmid backbone of pUNI-PROM (Figure 4.6). It was hoped that these fusion proteins would enable live cell imaging in E. coli cells in order to gain an idea of TorD interactions and possibly movement and subcellular location within the cell during the Tat translocation cycle. These plasmids were named pFLUOR_mCerN-TorD_YFPN-TorD, pFLUOR_mCerN-TorD_TorD-YFPC, pFLUOR_TorD-mCerC_ TorD-YFPC and pFLUOR_TorD-mCerC _YFPN-TorD, but will be referred to as pFLUOR _CC, _NC, _CC and _CN, respectively, for ease.

153

Figure 4.6: Diagram of fluorescent TorD fusions. Fusions of the fluorescent proteins mCer and YFP to either the N- of C-termini of TorD in various arrangements were encoded on the same plasmid. Fluorescent proteins were either fused to both N-termini (NN) or C-termini (CC) of the two TorD molecules, or one to the N- and one to the C-terminus (NC and CN). These plasmids were designed and produced in collaboration with Dr Sarah J. Coulthurst, a former post-doctoral researcher in the Sargent group.

Before cell imaging was carried out, it was necessary to test the activity of the TorD fusion proteins to identify any negative effects attachment to YFP and mCer may have, as it is possible such a fusion may cause mislocalisation or inactivation of the chaperone. In order to test the ability of these TorD-fluorophore fusions to function in TorA assembly, the activity of

TMAO reductase was assayed. Parental strain MC4100 and its derivative FTD100 (ΔtorD) containing the plasmids pUNI-PROM or pFLUOR_CN were grown overnight in anaerobic conditions supplemented with TMAO, before the periplasm of these cells was harvested for use in TMAO reductase assays. The amount of TMAO reduced in µmol per minute per mg of protein was determined through spectrometric measurement of the rate of colour change of the sample. Activity levels of each sample was determined and converted to percentages of the MC4100 periplasmic activity recorded.

154

Figure 4.7: TMAO reductase activity in the presence of fluorescently-tagged TorD. E. coli strains MC4100 and FTD100 (ΔtorD) expressing vectors pUNI-PROM or pFLUOR-CerC-torD_YFPN-torD were grown in anaerobic overnight cultures in LB supplemented with 0.5 % glycerol and 0.4 % TMAO, prior to isolation of the periplasmic fraction from cells. Using the redox-dye benzyl viologen as an electron donor, TMAO reductase activity (µm/min/mg) could be determined through measurement of the colourimetric change at Ab600. Results are displayed as a percentage of native TMAO reductase activity. Data expressed as means (n = 3) ± SEM.

From this assay, it was determined that the TMAO reductase activity for MC4100 not containing any plasmids was 1.48 µmol/min/mg, and this was used as the native TMAO reductase level to compare other samples to (Figure 4.7). MC4100 containing the empty plasmid pUNI-PROM had unusually increased levels of TMAO reductase levels compared to

MC4100 alone, approximately 230 % native levels. MC4100 pFLUOR-CN had similar levels to

MC4100 pUNI-PROM, of just under 200 % native levels. The E. coli strain FTD100 (ΔtorD) with no plasmids as well and both empty pUNI-PROM and pFLUOR_CN were also tested. Consistent with the data presented above (Figure 4.4) FTD100 had very little TMAO reductase activity, only 12 % that of MC4100. FTD100 containing pUNI-PROM was of a similar low level, also approximately 12 % that recorded for MC4100. In contrast, when pFLUOR_CN was expressed in FTD100 the plasmid was able to largely rescue the mutant phenotype, increasing TMAO reductase levels to approximately 80 % native levels. This indicated that at least one of the 155

TorD-fluorophore fusion proteins expressed from pFLUOR_CN was able to perform its TorA maturation function.

The localisation of TorD-fluorophore fusions was also investigated through the use of cell fractionation and Western immunoblotting. FTD100 (∆torD) containing the plasmid pFLUOR_CC was grown overnight before being fractionated into cytoplasmic, periplasmic and membrane fractions. These fractions plus a whole cell sample for comparison were analysed through the use of SDS-PAGE and Western immunoblot, in order to identify the cellular localisation of fluorescently tagged TorD. In this case, TorD could be observed below the 50 kDa marker, due to the presence of YFP or mCer fused to TorD making the overall protein size much larger than the 25 kDa TorD protein alone. The results of this Western analysis indicate the presence of TorD in whole cell, cytoplasm and membrane fractions, whilst the periplasm does not appear to contain any TorD (Figure 4.8). This demonstrates that the TorD-fluorophore fusion proteins tested are localising to the same areas of the cell as have been suggested for native TorD.

Figure 4.8: Cellular localisation of TorD-mCer and TorD-YFP fusions FTD100 (∆torD) cells expressing pFLUOR_CC were fractionated into cytoplasmic, membrane and periplasmic portions. Along with a whole cell sample (WC), these samples were separated by SDS-PAGE (12 % acrylamide) and visualised by Western blot. Anti-TorD serum was used at 1:20,000 with secondary antibody anti-rabbit at 1:10,000.

156

4.3.2.1 Confocal microscopy of mCer-TorD and YFP-TorD

In order to investigate localisation of TorD within the cell using a live cell imaging approach, E. coli MC4100 was transformed with pFLUOR plasmids encoding N- and C-terminal TorD fusions to both mCerulean and YFP. Cell samples from liquid culture were taken at exponential phase and imaged using a Leica SP5 SMD (Single molecule detection) confocal laser scanning microscope. Merged images of MC4100 containing pUNI-mCerN-TorD_YFPN-TorD (‘CC’) appear to show mCer and YFP fluorescence coinciding (Figure 4.9A). The two-dimensional histogram for this strain appeared to roughly follow the line of best fit in a linear fashion, indicating co-localisation of the two fluorophores. However, a small outlying region suggests there were some cells not demonstrating co-localisation. The average Pearson’s correlation coefficient for this plasmid was 0.838 (Figure 4.9B). For MC4100 pUNI-mCerN-TorD_YFPC-TorD

(‘CN’) strong evidence for co-localisation was obtained (Figure 4.9A). The average Pearson’s correlation coefficient for this sample was 0.898 (Figure 4.9B). The MC4100 pUNI-mCerC-

TorD_YFPC-TorD (‘NN’) histogram displayed a fairly linear correlation (Figure 4.9A). The average Pearson’s coefficient was 0.869 (Figure 4.9B). For cells with pUNI-mCerC-TorD_YFPN-

TorD (‘MC’) the histogram indicated a high level of linearity but with one large outlying branch

(Figure 4.9A). The average Pearson’s correlation coefficient was 0.857 (Figure 4.9B).

Unfortunately, whilst it was possible to obtain these images we struggled with photobleaching effects that manifest particularly with confocal microscopy, and so it was not possible to follow up these results with further live cell imaging or fluorescence lifetime correlation spectroscopy experiments as we had hoped.

157

Figure 4.9: Co-localisation of mCer-TorD and YFP-TorD fusions within E. coli. A – MC4100 cells expressing plasmids pFLUOR_NN, CC, CN and NC respectively were grown to an OD600 0.3 before being immobilised onto coverslips using CellTak. Shown are mCerulean (false-coloured green) and YFP (false- coloured red) channels, plus a merge of the two images. Yellow colour indicates co-localisation. Two-dimensional histograms of the images where locations of mCer and YFP fluorescence were plotted against each other were created using ImageJ software. Cells were imaged using a Leica SP5 SMD (Single molecule detection) confocal laser scanning microscope. B – Pearson’s correlation coefficients for TorD fusion proteins, indicating the level of co- localisation between mCer-tagged and YFP-tagged TorD proteins.

158

4.3.3 Super-resolution imaging of the E. coli cell

In recent years, a wide range of new microscopy techniques have been developed in an attempt to improve on the limitations of conventional optical microscopes. Collectively these techniques can be referred to as super-resolution microscopy. One such technique is termed three dimensional structured illumination microscopy, or 3D-SIM. 3D-SIM illuminates the sample with a known light pattern generated through the use of interfering multiple beams of light, in comparison to the single beam of light used in conventional microscopy. When this illumination pattern interferes with the sample of unknown structure it generates interference patterns known as Moiré patterns. By imaging numerous cross-sections of the sample and then processing and reconstructing the data, images with an axial resolution twice that of conventional microscopy can be obtained (Gustafsson, 2000).

4.3.3.1 Super-resolution imaging of chromosomally tagged TatA-YFP.

Through the award of a Euro-BioImaging Proof-of-Concept Studies grant we were able to use the OMX microscope located in the College of Life Sciences Microscopy Facility. With this we obtained high resolution images of the Tat translocase using E. coli strain JARV16 att tatA::YFP, where the TatA protein is tagged on the C-terminus with yellow fluorescent protein (YFP) and expressed at single copy from the attB locus. Cells were grown to an OD600 of 0.3 before being immobilised on coverslips using poly-L-lysine. The samples were then fixed with ice cold methanol and permeabilised with lysozyme. As it was discovered that YFP fluorescence was relatively weak, the cells were then incubated with GFP-booster (Chromotek), a GFP binding protein with a fluorescent dye attached that is excited into fluorescence at the same wavelength as GFP and YFP. This helps to enhance the fluorescent signal from YFP to enable imaging with the OMX. The coverslips were mounted onto slides using ProLong Gold antifade reagent (Life Technologies), designed to help prevent photobleaching during imaging as this was also a problem in the first instance. 159

Image stacks were taken using both conventional wide-field imaging (Figure 4.10A) and the 3D structured illumination imaging available on the OMX (Figure 4.10B) in order to demonstrate the increased resolution achievable. Images of these slides have revealed 2-3 bright spots in each cell indicating the Tat translocase pore was able to be observed. 3D-SIM images showed much clearer and more well-defined foci than those obtained using wide-field microscopy.

Figure 4.10: Fluorescence imaging of TatA in E. coli. Imaging of JARV16 att tatA::YFP cells, where TatA is C-terminally tagged with YFP on the chromosome. A – Two cells as imaged by conventional wide-field microscopy, and B – the same field of view imaged using 3D-SIM microscopy. Cell culture was grown to OD600 0.3 before being immobilised on coverslips and fixed and permeablized using methanol and lysozyme respectively. Cells were then incubated with GFP-booster to enhance the fluorescent signal. Coverslips were mounted using ProLong Gold anti-fade reagent in an additional attempt to preserve fluorescence. Image stacks were taken of a thickness of approximately 125 nm through the sample, using an OMX microscope.

4.3.3.2 Effects of substrate overexpression on visible TatA-YFP spots

As TatA foci could be observed in E. coli cells using an OMX microscope, it was decided to image cells where Tat substrates were being overexpressed. It was hoped that a difference in the size of foci may be observed between cells with wild-type levels of Tat substrates and those where a protein substrate is being overproduced. The Tat substrates selected were HiPIP

– high potential iron-sulphur protein, SufI – involved in cell division, and TorA – the catalytic subunit of TMAO reductase. These were overexpressed in JARV16 att tatA::YFP under high- 160

IPTG conditions using the pQE80L vector. Cells were then fixed and imaged as previously described using 3D-SIM.

Figure 4.11: The effects of Tat substrate overexpression on TatA-YFP foci. Images of JARV16 att tatA::YFP cells containing plasmids A – pQE80L; B – pQE80L-HiPIP; C – pQE80L-SufI; D – pQE80L-TorA. Cell cultures were grown to OD0.6 prior to induction of protein overexpression by the addition of IPTG. After growth in the presence of IPTG for approximately 180 minutes, cells were washed in sterile water and diluted to OD600 0.3 before being immobilised on coverslips and fixed and permeablilsed using methanol and lysozyme respectively. Cells were then incubated with GFP-booster to enhance the fluorescent signal. Coverslips were mounted using ProLong Gold antifade reagent in an additional attempt to preserve fluorescence. Image stacks were taken of a thickness of approximately 125 nm through the sample, using 3D-SIM with an OMX microscope.

It was observed that upon the induction of high levels of Tat substrates, there may be a slight increase in the number of TatA foci per cell. In the images shown, cells with both increased

HiPIP and TorA levels had six or seven TatA foci (Figure 4.11B & D), compared to the two or

three seen in cells only expressing empty pQE80L vector (Figure 4.11A). However, in order to

say for sure a more accurate way of analysing such data needs to be developed. In addition,

MG1655 cells expressing the empty vector pQE80L were smaller in appearance to those grown 161 without, and so there may be detrimental effects due to growth in the presence of IPTG.

MG1655 cells overexpressing SufI protein were abnormal in their appearance, displaying a single TatA spot of larger size than in all other samples.

4.3.3.3 Imaging TatA using fluorescent antibodies

It is known that the E. coli strain used in all previous work described to image TatA, JARV16 att tatA::YFP, has very low levels of Tat transport compared to wild-type cells where TatA is not fused to YFP (Leake et al., 2008). Therefore, any results obtained may not be applicable to wild-type E. coli. Because of this, we attempted to image TatA in wild-type MG1655 using fluorescent antibodies. MG1655 cells were treated similarly to JARV16 att tatA::YFP cells in previous experiments, where immobilised cells on coverslips were fixed using methanol before being permeabilised with lysozyme. They were then treated first by rabbit antiserum raised against TatA, before secondary incubation with the fluorescently labelled antibody DyLight 594 goat anti-rabbit (Abcam). This antibody generates red fluorescence and should highlight TatA in cells by binding to any anti-TatA antibody left on the sample after washing. Image stacks were then taken using 3D-SIM (Figure 4.12). Although TatA foci were not as clearly visible as in

JARV16 att tatA::YFP cells, a number of brighter foci could be observed. Fluorescence from the rest of the cell also appeared brighter in these samples. From this experiment it could be concluded that TatA is visible but perhaps to a lower extent than when TatA is chromosomally tagged with YFP. 162

Figure 4.12: Imaging TatA using fluorescent antibodies. MG1655 cells were grown to OD600 0.3 and immobilised on cover slips prior to methanol fixation and permeabilization by lysozyme. They were then incubated first by anti-TatA and secondly by DyLight 594 goat anti- rabbit to fluorescently highlight any TatA bound by primary antibody. Coverslips were mounted using ProLong Gold antifade reagent in an additional attempt to preserve fluorescence. Image stacks were taken of a thickness of approximately 125 nm through the sample, using 3D-SIM with an OMX microscope. 163

4.4 Discussion

4.4.1 TorD Pro-122 is important for TorD function

From the results of the TMAO reductase activity assay with TorD variants, it was clear that a number of the amino acids tested were more important in TMAO reductase assembly than others. Whilst TorD E23A, Q31A, G112A and T115A did not show any significant decrease in

TMAO reductase activities, P122A, E142A and double variant G112A/P122A all displayed activity levels of less than half that of that produced by native TorD (Figure 4.4). From this data, Proline 122 is the key novel finding as an important side-chain required for TorD activity.

Proline is an interesting amino acid (or more correctly imino acid), with its unique side chain forming a five point nitrogen-containing ring. Being relatively non-reactive due to this unusual side chain structure, it is less common to find this amino acid in protein binding domains.

However, it is often important in creating tight turns or kinks in protein structure. Therefore, it may be that TorD Pro-122 is required for the correct conformation of the TorD ‘hinge’ in order to enable binding to TorA. In the crystal structure of TorD from S. massilia (1N1C) the equivalent proline residue is on the edge of the hinge region in the TorD dimer, and it could be suggested that this proline is in fact important for the termination of the alpha-helix in order to enable formation of the hinge region. By substituting this amino acid for an alanine, it is clear that TorD no longer recognises TorA correctly, perhaps as the hinge region is not properly defined. Alternatively, whilst not being involved in a large amount of protein interactions, proline is known to be able to act as a hydrogen bond acceptor and so it is conceivable that it may be involved directly in TorA binding and this is the reason behind the decreased TMAO reductase activity levels in this TorD variant.

The proofreading assay using the Tat substrate HybO with the native signal peptide replaced for that of TorA enables separation of TorD interactions with the signal peptide from interactions with the mature region of the TorA protein, in order to only investigate the 164 proofreading aspect of TorAD interactions. When the TMAO reductase activity assay is compared to the results of the TorD proofreading assay, a number of differences can be seen.

Interestingly, although Glu-142 appears to be extremely important for the production of active

TMAO reductase (Figure 4.4), an alanine substitution does not seem to affect interactions with the signal peptide of TorA to any great extent, with E142A maintaining hydrogenase levels of nearly 100 % wild-type TorD levels (Figure 4.5). From this it can be construed that Glu-142 is possibly involved in TorD interactions with the mature region of TorA rather than the signal peptide. In comparison the residue Pro-122, as well as being implicated in the TMAO reductase assay, is also highlighted here as having a hydrogenase activity level half that of the parental strain when substituted for an alanine in the Tat proofreading assay. This suggests that this amino acid is important for binding to the signal peptide of TorA. Such a role would concur with results of published work investigating the binding region of the related E. coli DmsD protein, that highlighted not only the Pro-122 equivalent in DmsD, residue Pro-124, but also additional residues nearby such as Glu-123, Asp-126 and His-127, were important for signal recognition (Chan et al., 2008, Stevens et al., 2013). This certainly emphasizes the importance of this region of TorD-like chaperones in signal peptide binding. In addition to Pro-122, Gln-31 and Thr-115 have a lower Tat proofreading activity when substituted and so may be involved in signal peptide interactions too.

In conclusion, taken together with recently published observations of TorD and related proteins, this work provides further evidence of the two distinct functions of these types of chaperones. It is clear that signal peptide binding activity can be separated from a second binding event, in addition to unusual GTPase and cofactor-binding functions.

4.4.2 Imaging TorD and TatA in cells

It was shown through TMAO reductase activity assays that TorD tagged with both mCer or YFP were still able to function in proofreading and enable the production of mature TorA (Figure 165

4.7). Whilst problems with photobleaching prevented a large amount of work to be performed through the use of live cell confocal imaging of TorD fusions to mCer and YFP, it was possible to acquire images indicating the localisation of TorD within the cell (Figure 4.9). From these images it could be observed that plasmid-encoded TorD was uniformly present throughout the cytoplasm of the cells, with images of both mCer and YFP-tagged TorD clearly showing the typical rod shape of E. coli cells. In addition, when images of cells using both fluorescent channels were super-imposed onto one another, it was shown, unsurprisingly, that the fluorescent regions with TorD-mCer and TorD-YFP were co-localised. Through statistical analysis using Pearson’s correlation coefficients, it was shown that for all four TorD-fusion arrangements, the Pearson’s correlation coefficients were all between 0.8 and 0.9. Pearson’s correlation coefficients give an indication of the linear relationship between two sets of data, in this case TorD-mCer and TorD-YFP localisation within the cell. Values such as those found here indicate a high level of correlation, a suggestion that corresponds with the 2D histograms showing a high level of conformity along the lines of best fit. This result is unsurprising, as TorD is a soluble cytoplasmic protein that is most likely present diffused across the cell rather than localised at any particular point. At the outset, it had been hoped that TorD fusions could be produced from single copy at the native tor locus. However, in the course of this work it became clear that the fluorescence produced was not bright enough or stable enough for extensive study, even from the multi-copy plasmid-based constructs. Ideally, fluorescence lifetime correlation spectroscopy could have been employed to understand the TorD-TorD interaction during the translocation cycle (i.e. in the presence of induced substrate, for example). Unfortunately, it was extremely difficult to get past the problem of photobleaching in this work.

One unexpected result was obtained through examination of cellular localisation of fluorescently tagged TorD using Western analysis (Figure 4.8). TorD is a cytoplasmic protein, not normally associated with the inner membrane. It may be that this unusual observation was due to an effect of the fusion of TorD to YFP and mCer. This would explain the decreased 166

TMAO activity noted in cells relying upon these fusion TorD proteins for TMAO reductase maturation (Figure 4.7). However, any potential membrane association did not affect the entire fluorescent TorD population due to the presence of low level TMAO reductase activity, plus fluorescence microscopy indicated the presence of these proteins in the cytoplasm of cells.

In addition to imaging of the TorD chaperone, we also attempted to image the TatA protein, the subunit understood to form the pore of the Tat translocase. Using OMX 3D SIM we were able to demonstrate the power of the OMX microscope, with images taken using 3D SIM showing much higher resolution than conventional confocal imaging allows (Figure 4.10). This also demonstrated that it was possible to image chromosomally expressed TatA-YFP in the cell, with TatA appearing to form a small number of foci distributed around the surface of each cell.

This high resolution imaging was then used to investigate the effects of Tat substrate overexpression on these TatA foci. It was observed that there was perhaps a slight increase in

TatA foci in cells overexpressing HiPIP and TorA. This is in agreement with data from other studies where it has been shown that the overexpression of Tat substrates can cause an increase in the number of TatA foci seen in the cell (Alcock et al., 2013, Rose et al., 2013).

Unexpectedly, upon induction of SufI overexpression TatA formed one individual spot in the cell that was much larger than the foci seen in other cell samples. It may be that by overexpressing SufI, the Tat translocase becomes saturated and causes the TatA protein to accumulate together in such a way.

A published study using live cell imaging of E. coli TatA-YFP strains was able to show TatA foci forming in a transport-active cell background (Alcock et al., 2013). These foci appeared to be independently mobile and able to transit between high numbers of assembled foci and dispersed halos with few foci, with this distribution being affected by substrate availability to induce TatA assembly, and the rate of substrate removal by translocation in order to cause

TatA disassembly. The effects of substrate overexpression on TatA-YFP foci formation was also 167 investigated, with higher levels of Tat substrates clearly causing an increase in the numbers of

TatA-YFP foci compared to TatA-YFP in the diffuse halos. These foci were dependent upon the substrate possessing a functional Tat signal peptide, as well as the cell containing the TatBC complex. The clear observations of higher TatA-YFP foci in this study differed from results obtained using E. coli strain JARV16 att tatA::yfp, with only slight increases in foci being apparent. Such differences could be due to the different imaging methods used between this study and Alcock et al. (2013). Whereas in this published work live cells were imaged, we imaged fixed cells for our study. This could mean that due to the constant turnover of TatA complexes during translocation, by using fixed cells only a snapshot of the process is acquired, and perhaps the highly increased foci number being observed required a certain period of time.

Upon imaging of native TatA using fluorescent antibodies, individual TatA foci were observed similar to those seen in E. coli strain JARV16 att tatA::yfp, although the background fluorescence appeared to be greater than in cells with chromosomally tagged TatA. This demonstrated that the images obtained using a chromosomal TatA-YFP fusion did give an accurate representation of how TatA localises in wild-type E. coli cells. Clearly although transport through the Tat translocase is disrupted in JARV16 att tatA::yfp (Leake et al., 2008), this is likely not due to inhibition of the formation of the TatA pore and more probably a steric inhibition of Tat substrate binding or movement through the pore.

Photobleaching was also a large problem in this study, particularly when imaging wild-type

TatA using fluorescent antibodies, and this would require a better solution if this work were to be continued. In this work it was only possible to obtain images after cells had been fixed. This unfortunately meant that live imaging of the TatA protein during the cell cycle was not possible. It is possible that through the use of different more photostable fluorescent proteins or dyes this may allow for different techniques to be utilised such as a live cell imaging approach, which would be highly interesting as it may give an idea of the movement of TatA 168 within the cell throughout the Tat translocation cycle. This work represents a starting point for imaging of the Tat translocase using such high resolution microscopy.

169

5 Discovery of novel Tat-related phenotypes

using phenotype microarrays

170

5.1 Introduction

S. Typhimurium, a member of the Gamma branch of the Proteobacteria, is closely related to E. coli and shares approximately 71 % of its genes (McClelland et al., 2001). The Tat system of S.

Typhimurium is also thought to be similar to that of E. coli, with components TatA, TatB and

TatC of S. Typhimurium shown to be able to form functionally active heterologous structures when expressed in E. coli (Oates et al., 2003). The genetics of the S. Typhimurium Tat system are also identical to that found in E. coli (Figure 5.1).

Figure 5.1: Genetic organisation of the tat genes of S. Typhimurium LT2. The tat genes of S. Typhimurium are located in two clusters in the genome, encoding TatABCD (STM3973-6) and TatE (STEM0632). Genes of the Tat system are shown in green, whilst nearby genes are shown as white.

The ability to adapt to differing environments is highly important for survival of bacterial species. Similar to E. coli, S. Typhimurium also has a highly flexible respiratory chain that is modular in design, enabling it to make use of many substrates for respiration. S. Typhimurium is able to utilise all the potential electron acceptors and donors that E. coli can use, and in addition is also able to use tetrathionate and thiosulphate as terminal electron acceptors

(Clark & Barrett, 1987, Hensel et al., 1999). The central role of the Tat system in the biosynthesis of these types of enzymes (e.g. James et al., 2013), and the growing recognition that respiratory electron transport chains are central to the infection process (e.g. Winter et 171 al., 2010; Winter et al., 2013), has meant that the S. Typhimurium Tat system has come under the spotlight for infection studies. Indeed, during the course of this work it has been reported that S. Typhimurium tat mutants are impaired in their ability to both colonize and infect host organisms (Craig et al., 2013, Reynolds et al., 2011, Weatherspoon-Griffin et al., 2011).

In terms of Tat proofreading research, the S. Typhimurium system has also been a happy hunting ground, with the reporting of the first crystal structure of DmsD (Qiu et al., 2008).

However, the DmsD research field remains controversial in some areas, with some reports suggesting DmsD may interact directly with the Tat translocase (e.g. Papish et al., 2003;

Kostecki et al., 2010) or with general housekeeping chaperones in the cell (e.g. Li et al., 2010).

These works imply DmsD may have a more generalized role in Tat transport and protein folding than being a dedicated specific chaperone for DMSO and selenate reductases.

5.2 Aim

The aim of this Chapter was to characterise the broadest aerobic growth and aerobic respiratory phenotypes of a dmsD mutant strain of S. Typhimurium, alongside a number of other S. Typhimurium deletion strains, in order to better understand the general roles of these genes in the physiology of the bacterial cell.

172

5.3 Results

5.3.1 Metabolic abilities of S. Typhimurium strains were investigated using high-

throughput phenotype microarray analysis.

Through the use of genetic knockouts and specific assays, and a little insider knowledge, the specific functions of individual bacterial genes can often be determined. However, it can be of great advantage to perform an unbiased high-throughput screen in order to identify any unexpected effects of a specific gene deletion on phenotypes that could not have been predicted. In this case, three mutant strains of S. Typhimurium, namely DIG0610 (ΔSTM0610 – lacking a gene encoding a putative TorD-like chaperone protein for a predicted oxidoreductase); DIG100 (ΔtatABC); and DIG1495 (ΔdmsD), were investigated using such a high-throughput screen. The respiratory capability of these strains was compared using the

Biolog Phenotype MicroArray technology.

The Phenotype MicroArray system uses cell respiration with succinate as a universal reporter for cell viability, utilising the redox dye tetrazolium in order to demonstrate active respiration of the particular cell strains tested. Cells are incubated in 96 well plates, under aerobic conditions with shaking, where each well contains a defined cell media with a single carbon, nitrogen, phosphorus or sulphur source. If the cell phenotype is ‘positive’, cells are able to respire causing the dye to be reduced and initiating formation of a dark purple colour. If the phenotype is weakly positive or negative, respiration does not occur or occurs on a slower scale, leading to no or less colour formation. Incubation and monitoring of the phenotypic data is performed by the OmniLog system, enabling simultaneous monitoring of thousands of phenotypes over a particular time period to allow respiration activity to be observed. Through the use of this system, identification of any effects of gene deletions on a wide range of metabolic pathways can be visualised. 173

5.3.1.1 Aerobic metabolism with carbon sources

The first Biolog plates to be examined were those containing sole carbon sources (Figure 5.2).

These plates contained substrates such as glycerol, fumaric acid and dextrin, and totalled 190 carbon sources. 174

Figure 5.2: Phenotype microarray data for carbon sources. Respiratory rates for strains grown over 24 hours in 96 well plates of defined minimal media, with each well containing a single carbon source. Duplicate respiratory curves of each strain are shown. Key: LT2a – blue. DIG0610 (ΔSTM0610) – red. DIG100 (ΔtatABC) – green. DIG1495 (ΔdmsD) – yellow. Outlined in black is PM 1, well C10, containing maltose. Negative control wells were in position A01 of each plate, containing no carbon source. 175

Respiratory rates were recorded over 24 hours for the four strains tested, namely the parental strain LT2a, DIG0610 (ΔSTM0610), DIG100 (ΔtatABC) and DIG1495 (ΔdmsD), with any disparity between the wild-type LT2a and mutant strains being clearly observable in the consolidated data. Of these data (Figure 5.2), most wells appeared to show strains respiring similarly to parental control, with the majority showing either normal respiration of all strains or no respiration at all across all strains. Well number C10 of plate PM 1 (Figure 5.2) gave a more interesting result, however, where strain DIG100 (ΔtatABC – green trace in Figure 5.2) appeared to respire at a lower level than that of other mutant strains and the wild-type LT2a.

This well contains maltose as the sole carbon source.

Figure 5.3: The respiratory ability of S. Typhimurium strains in minimal media with maltose as the single carbon source. The respiratory curves of LT2a – blue, DIG0610 (ΔSTM0610) – red, DIG100 (ΔtatABC) – green, and DIG1495 (ΔdmsD) – yellow. Duplicate results for each strain are indicated by triangles and squares of the same colour. Open shapes indicate same strains in the negative control well of plate PM 1 (A01), whilst solid shapes refer to data of well C10. The metabolic activity as measured by respiration was recorded over 24 hour hours, in minimal media containing maltose as the sole carbon source.

176

The respiratory rates of well C10 containing maltose were analysed in isolation (Figure 5.3). It is clear that mutant strain DIG100, lacking a functional Tat system, is affected in its ability to respire in the presence of maltose. Respiratory activity appears to be much slower than wild- type, with it appearing to stabilise at approximately half the level of LT2a. In comparison, respiratory rates of the DIG1495 and DIG0610 strains were unaffected under these particular conditions (Figure 5.3).

Figure 5.4: Growth of a ΔtatABC mutant is defective in M9 media containing maltose. LT2a and DIG100 (ΔtatABC) were grown aerobically over 48 hours at 37 oC in M9 media supplemented with 0.5 % [w/v] maltose. Absorbance at OD600 was recorded at 20 minute intervals. Data expressed as means (n = 3) ± SEM.

As the results of the Biolog phenotype microarray provide details in terms of respiration rate and metabolism, the growth rates of LT2a and DIG100 with maltose a sole carbon source were next examined (Figure 5.4). Cells were incubated for 48 hours in M9 media supplemented with

0.5 % [w/v] maltose, with optical density OD600 being recorded every 20 minutes in order to obtain growth curves. It could be seen that the wild-type LT2a strain was able to grow faster than the tat mutant DIG100 across the entire 48 hour time period. DIG100 was unable to reach the same OD600 level as LT2a by the end of the 48 hours. 177

Figure 5.5: Location of maltose binding protein in S. Typhimurium strains. A – S. Typhimurium strains LT2a and DIG100 (ΔtatABC) were grown overnight in aerobic conditions with minimal M9 media supplemented with 0.5 % [w/v] maltose. Cells were fractionated into periplasmic (P), membrane (M) and cytoplasm (C) fractions, and these fractions plus a whole cell (WC) sample were separated by SDS-PAGE (12 % [w/v] acrylamide) before Western analysis. B – S. Typhimurium strains LT2a and DIG100 (ΔtatABC) were grown for 24 hours in aerobic conditions with minimal M9 media supplemented with 0.5 % [w/v] maltose. Cells were pelleted and the supernatant was removed and filtered through a 0.2 µm filter. Cell (WC) and supernatant (SN) samples were then separated by SDS-PAGE (12 % [w/v] acrylamide) before Western analysis. Antibodies for both Westerns were: Primary – anti-maltose binding protein, 1:20,000. Secondary – anti-mouse, 1:10,000.

Since a version of S. Typhimurium lacking a functional Tat system exhibits a growth and respiratory defect upon metabolism with maltose, it was then questioned as to how this defect could arise. Maltose uptake into the cell is governed by a very well-characterised transport system of the ‘ABC Superfamily’ (Bordignon et al., 2010). The maltose transporter consists of two transmembrane proteins (MalF and MalG) that together interact with a cytoplasmic ATP- binding protein (MalK), and a periplasmic substrate binding protein (MalE or MBP). Once in the periplasm, exogenous maltose binds to the maltose binding protein (often termed MBP, which is the product of the malE gene) and this triggers the complete system to actively import maltose in an ATP-dependent manner. MalE is a Sec-dependent periplasmic protein and it was considered that this was the most likely component to have been affected in some way by the tat mutation. The subcellular localisation of MBP was tested and LT2a and DIG100 cells were fractionated into individual cell components of periplasm, total membranes and cytoplasm, and these alongside a whole cell sample were visualised by Western immunoblotting (Figure

5.5A). Results for localisation of maltose-binding protein for both LT2a and DIG100 were the same, with it being found in whole cell, membrane and periplasmic fractions (Figure 5.5A). 178

It is known that tat mutants show defects in their cell envelope (Ize et al., 2003) and it was therefore subsequently considered a possibility that deletion of tatABC genes may result in a

“leaky” periplasm. If MBP was being continually lost from the periplasm, perhaps this could explain the growth phenotype of a tat mutant on maltose. To test this, whole cell and filtered culture supernatant samples were subjected to Western analysis. Results showed MBP to be present only within the cells, with the supernatant showing no protein (Figure 5.5B). This implies the cells are still intact and the deletion of the Tat system does not affect MBP localisation.

5.3.1.2 Metabolism of other substrates

Four of the Biolog phenotype microarray plates are designed to screen for use of nitrogen sources, encompassing 380 substrates (Figure 5.6). The four S. Typhimurium strains under investigation were screened using these plates to identify any metabolic deficiencies. Plate

PM 3-B contained nitrogen sources such as ammonia, nitrate and nitrite, whilst plates PM 6, 7 and 8 contained peptide nitrogen sources such as Ala-Ala, Ala-Arg and Leu-Ser. In these three plates there was an additional positive control well (A02) containing L-glutamine, a primary amine donor for key amino acid and nucleotide biosynthetic pathways (Figure 5.6).

179

Figure 5.6: Phenotypic microarray data for nitrogen sources. Respiratory rates for strains grown in defined minimal media with single specific nitrogen sources. Duplicate respiratory curves of each strain are shown. Key: LT2a – blue. DIG0610 (ΔSTM0610) – red. DIG100 (ΔtatABC) – green. DIG1495 (ΔdmsD) – yellow. Negative control wells were in position A01 of each plate, containing no nitrogen source. PM 6, 7 and 8 also had a positive control well (A02), containing L-glutamine. 180

Figure 5.6: Phenotypic microarray data for nitrogen sources (contd).

Analysis of the phenotype microarray data for nitrogen sources demonstrated that all strains were able to respire as normal using most nitrogen sources (Figure 5.6). Very few differences were observed across the range of nitrogen substrates. On occasion the DIG100 strain 181 appeared to display a slightly slower metabolic activity compared to wild-type, for example in plate PM 3-B, well A10. This contains the amino acid L-aspartic acid. In addition, DIG100 and

DIG1495 appear to demonstrate slightly decreased respiration levels compared to LT2a and

DIG0610 in PM 3-B A12, with L-glutamic acid. Also on this plate, well G2 containing xanthosine demonstrated a lower level of respiration for the DIG100 strain. In PM 6 all strains displayed the same phenotype, and either all demonstrated metabolic activity or all did not respire at all.

Respiration in PM 7 appeared normal for all except well B3, containing Lys-Thr, where DIG1495

(∆dmsD) showed respiratory activity whilst all others did not (Figure 5.6). Lastly, PM 8 also only had one well where a difference between strains was seen (Figure 5.6). This was in well F7, containing the amino acid D-Ala-D-Ala (D-alanyl-D-alanine), in the presence of which DIG100

(∆tat) was unable to respire.

Biolog phenotype microarray plate 4-A contains 95 sources of phosphorus and sulphur (Figure

5.7). In a number of these wells there appeared to be slight differences between strain respiration rates, however the duplicate samples shown did not always show the same patterns making it impossible to conclude on the actual effect from this data alone. One well with a clear difference in respiration rates was well D1, containing mannose-1-phosphate

(Figure 5.7). Interestingly strains LT2a, DIG100 and DIG1495 were unable to respire in this condition, whilst DIG0610, containing a deletion of a gene encoding unknown protein

STM0610 predicted to be a Tat chaperone, demonstrated strong respiration rates.

182

Figure 5.7: Phenotype microarray data for phosphorus and sulphur sources. Respiratory rates for strains grown in defined minimal media with single phosphorus or sulphur sources. Duplicate respiratory curves of each strain are shown. Key: LT2a – blue. DIG0610 (ΔSTM0610) – red. DIG100 (ΔtatABC) – green. DIG1495 (ΔdmsD) – yellow. Negative control wells were in position A01 of each plate, containing no phosphorus/sulphur source.

Biolog also market a phenotype microarray plate containing a number of different nutrient supplements. Testing was also performed using this plate with the four S. Typhimurium strains

(Figure 5.8). Interestingly with this plate, one of the supplements is L-glutamic acid (well A8), already highlighted in nitrogen source plate PM 3-B, well A12 (Figure 5.6). As with other plates, the ΔtatABC strain did show a slight decrease in respiration rates compared to the other strains in several wells.

It was attempted to investigate a number of these substrates such as L-glutamic acid further through the production of growth curves, however the strains tested did not appear to grow at all and there was no obvious increase in OD600 even over a 48 hour period. It is clear that, in these assays, active respiratory activity is not always linked to growth and proliferation. 183

Figure 5.8: Phenotype microarray data for respiration rates of S. Typhimurium strains on nutrient sources. Respiratory rates for strains grown in defined minimal media with single specific nutrient supplements. Duplicate respiratory curves of each strain are shown. Key: LT2a – blue. DIG0610 (ΔSTM0610) – red. DIG100 (ΔtatABC) – green. DIG1495 (ΔdmsD) – yellow. Negative control wells were in position A01 of each plate, with a positive control media in well A02.

5.4 Discussion

5.4.1 A requirement of a functional Tat translocase for maltose metabolism?

The S. Typhimurium strain DIG100, lacking a functional Tat system, appeared to have a defect in aerobic respiration when the sole carbon source provided was maltose (Figure 5.2). The maltose transport system of E. coli and S. Typhimurium is a member of the ATP-cassette binding superfamily of transporters. This membrane-associated transported consists of four subunits; two are transmembrane subunits MalF and MalG, plus two copies of the cytoplasmic

ATP-binding cassette (ABC) subunits MalK. Maltose-binding protein (MBP), or MalE, is involved in delivering maltose to the maltose transporter through interactions with two periplasmic 184 loops of the MalFG subunits, stimulating ATP hydrolysis (Daus et al., 2007, Davidson et al.,

1992). Maltose-binding protein is now a popular protein for use as a fusion tag for purification, due to its ability to form stable soluble protein fusions and to bind tightly to maltose-based resins (Kapust & Waugh, 1999, Lebendiker & Danieli, 2011).

The reasons behind the dependence upon the Tat system for maltose respiration remain unclear. From data presented here, the ΔtatABC deletion does not appear to disrupt cell integrity to such an extent that MBP is shed from the cell, and localisation of MBP which would in turn affect maltose uptake is potentially normal. Furthermore, the other components of the maltose transport complex are not substrates of the Tat system and so would not be expected to require the translocase for function. The maltose transporter system is known to be inhibited by the phosphorylatable protein enzyme IIA (EIIA) in conditions where a preferred carbon source such as glucose is available (Chen et al., 2013). However, in the conditions of the phenotype microarray maltose is the only carbon source and this so-called carbon catabolite repression system should not be working. Published work has shown that tatABCD tatE mutant E. coli is able to uptake maltose even in the absence of MBP, a phenomenon thought to arise through Δtat causing a loss of outer membrane integrity (Caldelari et al.,

2008). Moreover, the ΔtatABCD tatE strain used in this work was clearly able to utilise maltose as a carbon source. This at least provides some additional evidence that there is a link between

Tat transport and correctly functioning maltose metabolism. It should be noted, however, that the growth rate of an E. coli tat mutant in maltose media has not been studied here. In addition, well E10 of carbon plate PM1 contains maltotriose, a sugar consisting of three glucose units and therefore very similar to maltose. Therefore it could be expected that a similar respiratory phenotype would be observed with strains in this well as with that containing maltose. However, all strains appear to be able to respire normally under this condition, surprising given the maltose phenotype of the ΔtatABC deletion strain.

185

5.4.2 Further involvement of the Tat pathway in cell wall synthesis and

remodelling

Unusually, strain DIG1495 (∆dmsD), lacking the chaperone protein DmsD, was able to respire with Lys-Thr as the sole nitrogen source. The parental strain, and the other strains tested, were unable to do this. The presence of this peptide could potentially interfere with the structure of the cell wall. Possibly by inhibiting penicillin binding protein 3 (PBP3), which contains three conserved sequence motifs for activity, one of which being Lys-Thr-Gly-Thr which is thought to interact with substrates (Sauvage et al., 2014). PBP3 is important for catalysing crosslinking of the cell wall peptidoglycan during cell division and performs a transpeptidase activity that is reliant upon these three conserved motifs. However, it is difficult to understand how a strain lacking DmsD might be able to use Lys-Thr as a nitrogen source, while the parental strain does not. It is possible that Lys-Thr could be rapidly degraded in the parental strain, while it is not in the dmsD mutant. An alternative is that DmsD itself binds tightly to Lys-Thr thus sequestering the potential nitrogen source. In any case, the identification of a distinct phenotype for a

∆dmsD strain grown under aerobic, or microaerobic, conditions must be considered a tentative step forward in research into this protein.

Also on the theme of cell wall synthesis, in PM8 well F7, containing D-Ala-D-Ala, DIG100 (∆tat) displayed very low metabolic activity (Figure 5.6). This peptide is present in the cell wall, where

D-Ala-D-Ala makes up the peptide stem of each disaccharide tetrapeptide subunit (Vollmer et al., 2008a). In bacterial strains with vancomycin resistance, this may be changed to D-Ala-D- lactate, creating a modified peptidoglycan that binds vancomycin 1000-fold less than the original D-Ala-D-Ala peptidoglycan (Walsh et al., 1996). The Tat system is, as previously mentioned, important for peptidoglycan modelling through the transport of cell wall amidases

AmiA and AmiC that are involved in hydrolysis of peptidoglycan-side chain bonds during cell separation. The D-Ala-D-Ala part of the peptide stem are usually added as a dipeptide, synthesised by the enzyme Ddl (Bugg et al., 1991) and incorporated by the ligase MurF

(Duncan et al., 1990). The deficiency observed here with using D-Ala-D-Ala as a nitrogen source 186 may reflect the fact that the cell wall of a Tat mutant is deficient in D-Ala-D-Ala and as such this molecule is captured and used in the periplasm in an attempt to repair the cell envelope, rather than being taken into the cytoplasm to be used as a nitrogen source. It may be useful to perform some basic microscopic analysis of a tat mutant growing in the presence of D-Ala-D-

Ala, in case this has an effect on the cell-separation phenotype normally observed.

187

6 Final Conclusions and Future Perspectives

188

6.1 Final conclusions

The maturation and translocation steps of the twin-arginine translocation pathway are still being unravelled, with the mechanisms of ‘Tat proofreading’ involved in biosynthesis of the most complex substrates of the Tat system beginning to be understood in greater detail. A large family of proofreading chaperone proteins has been identified, with these chaperones appearing to play a number of vital roles in protein maturation prior to transport. In this study, two substrate-chaperone pairs have been investigated, namely YnfE and DmsD, and

TorA and TorD. We have been able to reveal a more specific model of interactions occurring during the Tat proofreading mechanism.

Glutamine scanning of the signal peptide of YnfE indicated that the hydrophobic stretch of the signal peptide was the region of importance for interactions with the DmsD chaperone protein. This is in line with other work that has identified the so-called h-region of the DmsA signal peptide as being important for DmsD interactions. Certainly the signal peptides of all three known substrates (YnfE, YnfF and DmsA) that interact with DmsD are conserved particularly in their h-regions, hinting at conservation in signal peptide recognition and Tat proofreading mechanism. Investigations into the binding site on DmsD have highlighted a hydrophobic groove on the DmsD surface that appears to be involved in signal peptide binding

(Chan et al., 2008, Stevens et al., 2013). The work presented here demonstrated that the

DmsD amino acid residue Val-16 was important in binding to the signal peptide of YnfE. This hydrophobic amino acid, whilst slightly outside of the groove suggested by other studies, still has the potential to interact with signal peptides if, as suggested in this work, the peptide binds to the chaperone in a specific orientation with the N-terminal of the peptide near the area containing this Val-16 residue. Given the fact that investigations into interactions of this substrate-chaperone pair from both the substrate and the chaperones perspective has shown hydrophobic residues as important suggest that hydrophobic interactions are at the heart of the binding interaction. Key questions remaining in the field are big ones. For example, what 189 are the chaperones actually doing when they are bound to the signal peptide? Do they simply supply some stearic hindrance, or is the location and mode of binding important? Is there some kind of molecular clock involved, timing binding and release of the chaperone, or is the hand-over of the signal peptide to the TatC protein driven purely by thermodynamics and binding affinities? Does the chaperone perform a biochemical activity when bound to a signal peptide, since variants can be isolated that are inactive for Tat proofreading in our in vivo assays, but seemingly unimpaired for peptide binding.

In this project we have additionally demonstrated that DmsD may play a number of roles in

YnfE maturation, rather than just binding the signal peptide until maturation is complete. As well as Val-16, residues Trp-91 and Gly-100 were highlighted in individual amino acid variants of DmsD as disrupting the activity of the selenate reductase enzyme despite there being no negative effects on signal peptide binding by these variant proteins. It is interesting to think of these chaperone proteins as involved in more than simple protein folding and prevention of premature transport, and certainly would be feasible. Both Trp-91 and Gly-100 are located around a loop region of DmsD that is not present in TorD chaperones, and it could be suggested that this loop region is involved in a number of possible functions, such cofactor assembly protein recruitment or general chaperone protein assembly.

In addition to revealing the interactions involved in ssYnfE-DmsD interactions, the importance of TorD in TorA maturation was also studied. A number of amino acids were shown to have a potential involvement in TorA binding, both to the signal peptide and separately the mature region of the protein. We have presented further evidence of two distinct roles of TorD, as has been suggested in previous studies. Through the use of super high-resolution microscopy it has been shown that Tat components and chaperone proteins are able to be observed in vivo, and this represents a first step in the utilisation of such imaging for studies of the Tat pathway.

190

6.2 Future Perspectives

In this study we have discussed the specific and essential roles of chaperone proteins during the maturation of Tat substrates. A major question that is still to be asked of these proteins is their importance in real terms in the cell. What are their roles during host symbiosis/infection?

6.2.1 How are these Tat substrate enzymes and their chaperones involved in

aiding cell survival, for example during host infection?

With the great respiratory flexibility demonstrated by bacteria such as E. coli and S.

Typhimurium, it should come as no surprise that a number of electron acceptors are utilised to the bacterial cells advantage during pathogenesis. Salmonella is one of the most common causes of food-borne infections and diarrhoea in developed countries (de Bruyn, 2000), and has been shown to be able to make use of a number of products of inflammation as terminal electron acceptors during infection in order to enhance competitive ability in the environment of the host.

- - - For example, the sulphur compounds tetrathionate (S4O6 ) and thiosulphate ( S-SO3 ) can be used as terminal electron acceptors during anaerobic respiration (Hensel et al., 1999, Hinsley

& Berks, 2002). It has long been known that S. Typhimurium is able to use tetrathionate in anaerobic respiration, with this fact being exploited for many decades in order to allow isolation of Salmonella serovars through the technique commonly termed tetrathionate enrichment, enabling Salmonella to outcompete other species of microbe by growth in broth containing tetrathionate (Winter et al., 2010). It now appears that S. Typhimurium makes use of the inflammatory response induced in the host during infection, where reactive oxygen species generated during this response is used to oxidise thiosulphate (produced in the gut from hydrogen sulphide generated by the microflora) and produce tetrathionate, a compound that can be reduced by tetrathionate reductase under anaerobic conditions enabling anaerobic respiration. Growth of S. Typhimurium is promoted as this terminal electron acceptor is 191 provided through the respiratory burst, thereby giving it an advantage in the competitive environment of the gut (Thiennimitr et al., 2012, Winter et al., 2010). The Tat-dependent enzyme tetrathionate reductase in Archaeglobus fulgidus has been shown to bind the TorD- family proofreading chaperone TtrD during maturation (Coulthurst et al., 2012), and whilst no chaperone has yet been identified in S. Typhimurium it is likely to require a similar protein chaperone due to its MGD cofactor. Tetrathionate reductase and its cognate chaperone is therefore an important tool in bacterial pathogenesis, at least in the case of Salmonella.

Early studies involving TtrA and TtrB were performed during the course of this study. As there is yet to be an identified TorD chaperone participating in Tat proofreading of TtrA or TtrB, we screened a number of known and predicted S. Typhimurium TorD chaperone proteins (namely

STM0610, STM4308, TorD, DmsD and YcdY) for their interactions against these two Tat substrates. From this bacterial two-hybrid study it was confirmed that none of these chaperones appeared to interact with TtrA or TtrB (Figure 6.1), and so any proteins that may be involved in Tat proofreading of these substrates remain elusive.

192

Figure 6.1: TorD family chaperone proteins did not display interactions with spTtrA or spTtrB. MG1655 ΔcyaA::Apra was co-transformed with vectors UT18-spTtrA or UT18-spTtrB and one of T25- STM0610/STM4308/TorDst/DmsDst/YcdYst. β-galactosidase was measured as an indicator of interactions. β- galactosidase activities are displayed relative to activity of the positive control (‘NarJ + spNarG’). Cells containing both empty vectors were used as the negative control (‘UT18 + T25’). Data expressed as means (n = 3) ± SEM.

In addition, studies into interactions of TtrB with the Tat component protein TatC were begun.

A number of amino acid residues of TatC were targeted for site directed mutagenesis in order to observe regions of importance for interaction with the signal peptide of TtrB. These residues were V3, L9, E15, K101, R105, E187 and R241, and were substituted for glutamines before interactions were tested using the bacterial two-hybrid system. Results suggested the N- terminal cytoplasmic region as playing a role in TatC interactions with spTtrB (Figure 6.2A), with both V3Q and E15Q displaying lower β-galactosidase activity (Figure 6.2B), in keeping with other published work highlighting the role of this region of TatC (Buchanan et al., 2002,

Kneuper et al., 2012, Zouflay et al., 2012). Interactions of TatC and signal peptides of Tat substrates are important for Tat translocation, and so understanding the binding interaction at the amino acid level is important. A greater understanding of this may allow for the possibility of fragment screening in order to identify compounds capable of disrupting binding, and reveal potential antimicrobial targets. 193

Figure 6.2: Specific interactions between spTtrB and TatC were analysed. A – TatCst structure, based on figure of E. coli TatC from Buchanan et al. (2002). Amino acid residues targeted for site directed mutagenesis are outlined in colour, where green indicates no subsequent disruption of interaction with spTtrB, and red indicates spTtrB interactions were disrupted. B - MG1655 ΔcyaA::Apra was co-transformed with vectors UT18-spTtrB and T25-TatC plus variants. β-galactosidase was measured as an indicator of interactions. β-galactosidase activities are displayed relative to activity of the positive control (‘NarJ + ssNarG’). Cells containing both empty vectors were used as the negative control (‘UT18 + T25’). Data expressed as means (n = 3) ± SEM.

Tetrathionate reductase is not the only Tat substrate that is thought to be involved in giving an advantage to bacteria during infection. Another by-product of the host inflammatory response is nitrate, the preferred electron acceptor utilised by bacteria such as S. Typhimurium. The

production of nitrate can be induced through S. Typhimurium expression of the secreted 194 protein SopE, an effector protein of the invasion-associated type III secretion system that activates a number of innate immune pathways and simultaneously leads to increased host- derived nitrate production from the activity of the inducible nitric oxide synthase (Lopez et al.,

2012). Three operons encode for nitrate reductases in S. Typhimurium, namely narGHJI, narZYWV and napFDAGHBC. Cytoplasmic-oriented nitrate reductase A, although not an active

Tat substrate, has been shown to harbour a remnant Tat signal peptide, and relies upon the

TorD-family chaperone protein NarJ for assembly (Turner et al., 2004). This puts another (Tat) substrate-chaperone pair in an important role in infection.

These examples of mechanisms by which S. Typhimurium is able to promote its own advantage during infection of a host all require the successful maturation and translocation of enzymes that are Tat substrates. Because of this, it can be said that the Tat proofreading activity of Tat chaperones of the TorD-like family are of great importance for Salmonella pathogenesis, and indeed a wide range of bacterial species, as the activity of many Tat enzymes requires successful Tat proofreading. A new direction of investigation could be to look at the roles of these cytoplasmic chaperone proteins during the process of infection. Given the dependence upon these chaperones for enzymatic activity, they could represent a new family of drug targets. Antimicrobial resistance is a major threat to human health, with many bacteria causing common infections in the community and in hospital settings now found to demonstrate high proportions of resistance (WHO, 2014). There is now an urgent need for the identification of new drug targets and new potential drugs.

195

Figure 6.3: Salmonella is able to utilise host inflammatory response during infection for competitive advantage in the intestine. Salmonella has been shown to be able to make use of products of the host inflammatory response during invasion in order to give itself a competitive advantage over the commensal microbiota. Production of reactive oxygen species (ROS) by neutrophils during inflammation enables both tetrathionate and nitrate respiration. ROS can react with nitric oxide, produced by nitric oxide synthase, to form peroxynitrate (ONOO-). This reactive nitrogen species - - quickly isomerises to nitrate (NO3 ) that can be reduced to nitrite (NO2 ) by Salmonella nitrate reductase enzymes, enabling nitrate respiration and energy production. In the same way, ROS can react with thiosulphate, produced from microbiotal hydrogen sulphide (H2S), to give tetrathionate that can then be reduced by Salmonella tetrathionate reductase. Again this allows Salmonella to carry out anaerobic respiration. Both pathways can lead to a growth bloom of Salmonella due to its competitive edge over other bacterial strains present in the microbiota.

As the signal peptide binding sites on TorD-like cytoplasmic chaperone proteins of the Tat

system are becoming more defined in their locations and composition through further study,

they represent the potential for development of new mechanisms of antibiotic targeting.

Targeting the proofreading function of these cytoplasmic chaperones would be an ideal

mechanism to reduce the invasion capabilities of virulent bacterial strains, as it has been

shown in the past that enzymatic activity of Tat substrates such as those mentioned is

dependent upon this interaction. The creation of deletion mutants and examination of their 196 invasion abilities through the use of virulence models could give an indication to their importance during infection. The use of high throughput drug screens may be useful in identifying compounds that interfere with the proofreading mechanism in some way, perhaps through disruption to binding of the signal peptide. Targeting these proofreading pathways and preventing enzyme maturation would rid Salmonella of its competitive edge in the host environment. This would be an interesting and highly relevant future topic of research.

6.2.2 Could Tat substrates and chaperones be utilised to the advantage of human

health?

The human intestinal microbiota is a highly complex community consisting of bacteria, viruses and fungi, thought to have coevolved with the host. The total number of commensal bacterial cells living on or in the host is thought to outnumber human cells by ten-fold (Turnbaugh et al.,

2007), with the gut microflora consisting of a large number of anaerobes plus some facultative anaerobes mainly residing in the large intestine (Gorbach, 1996). This complex ecosystem is extremely important for human health, and is thought to protect us against invading pathogens, contribute to normal immune functions within the body and aid energy extraction from our diets (Lozupone et al., 2012). The intestinal microflora are important for the production of enzymes that enable the breakdown of a number of vitamins, including vitamin

K, an essential cofactor in production of blood clotting factors such as prothrombin. In addition, other enzymes produced aid fermentation of indigestible carbohydrates such as pectins, cellulose and hemicellulose from plant material into short-chain fatty acids including acetate and butyrate (Gorbach, 1996). The protective abilities of the intestinal microflora against pathogen colonisation are also well documented, with these commensals occupying similar ecological niches to those required by pathogenic bacteria. It is thought that infections are prevented both through direct interactions such as alteration of the environmental conditions (such as pH) or the production of bacteriocins, proteinaceous toxins that inhibit 197 members of closely related bacterial species, and indirectly through the stimulation of host immunity (Kamada et al., 2013). Disruptions to the intestinal bacterial flora have been implicated in many diseases and conditions such as chronic kidney disease (Sabatino et al.,

2014), inflammatory bowel disease (Orel & Trop, 2014), obesity (Moran & Shanahan, 2014) and Type 2 diabetes mellitus (Puddu et al., 2014).

Activities of the human microflora can have influences on host health in many different ways.

Recent studies into the gut microbiota and cardiovascular disease (CVD) revealed links between dietary lipid intake, gut microflora and atherosclerosis (Wang et al., 2011). In this pathway, the compound trimethylamine N-oxide (TMAO) is produced as a result of the metabolism of the lipid phosphatidylcholine (PC) to choline, trimethylamine (TMA) and betaine by the gut microflora. TMA produced during this reaction is then absorbed into the circulation and becomes oxidised to TMAO by hepatic FMO3, a human enzyme that has been linked to the genetic disorder trimethylaminuria. Trimethyaminuria is a rare autosomal recessive condition which results in accumulation of TMA in the body, which is excreted in the sweat and urine causing a rotten fish smell (Ulman et al., 2014). It was observed that there was a high correlation between systemic levels of TMAO and coronary artery atherosclerosis and cardiac risk (Vinjé et al., 2014, Wang et al., 2011). Another published study demonstrated that consumption of red meat, high in the nutrient L-carnitine that contains a trimethylamine structure similar to that of choline, leads to the production of TMAO through L-carnitine metabolism by the gut microflora, and again increases atherosclerosis and CVD risks (Koeth et al., 2013, Vinjé et al., 2014). Increased plasma levels of TMAO show a positive association with the risk of major adverse cardiovascular events, with this TMAO increase being related to gut microbiota through its inhibition upon the application of antibiotics (Tang et al., 2013). The increased risk of CVD caused by high TMAO levels is thought to be due to its influence over the production of a pro-atherosclerotic phenotype amongst macrophages (Wang et al., 2011). 198

The importance of the gut microbiota for human health is clear. In recent years the idea has been touched upon that modulation of the intestinal flora may be a novel therapeutic approach for prevention or treatment of serious conditions such as heart disease (Jia et al.,

2008, Vinjé et al., 2014). Pre- and probiotics are already popular ways of attempting to influence the proportions of certain bacterial species within the intestine in the interests of health. It is possible that through control of the gut microbiota population or through specific drug targeting health issues linked to the intestinal microbiome could be prevented or treated.

In the case of high TMAO levels and the related CVD risks, it could be beneficial to be able to reduce the amount of TMAO present in the blood plasma. Certainly, it has been shown that certain dietary foods can have an effect on the levels of TMAO in the body (Miller et al., 2014).

Another potential way to control this is to catalyse the reduction of TMAO produced by the gut microbiota back to the compound TMA. E. coli and S. Typhimurium amongst other bacterial species are already capable of this reaction under anaerobic conditions, where they utilise

TMAO as a final electron acceptor during respiration. If it were possible to somehow harness this ability and increase levels of the enzyme TMAO reductase in the host it may be possible to decrease patient risks of atherosclerosis and CVD. Therefore it is essential to understand the entire pathway of enzyme maturation and production in order to investigate future uses of

TMAO reductase in cardiovascular disease treatment. In addition, there may be new roles to be identified for enzymes such as these in other conditions linked to the metabolism activities of the gut microbiota.

199

7 Materials and methods

200

7.1 Bacterial strains

Table 7.1: Bacterial strains used in this study. Bacterial Strain Genotype Antibiotic Reference Resistance - - - BL21(DE3) pLysS F ompT gal dcm lon hsdSB(rB mB ) Cml λ(DE3) pLysS(cmR) DH5α F_ φ80d ∆(lacZ)M15 recA1 endA1 None (Hanahan, - + gyrA96 thi-1 hsdR17 (rk mk ) supE44 1983) relA1 deoR ∆(lacZYA-argF)U169 DIG0610 Wild-type S. Typhimurium with None (Guymer et al., STM0610 deleted by lambda-red 2009) leaving 81bp scar DIG1495 Wild-type S. Typhimurium with None (Guymer et al., STM1495 (dmsD) deleted by lambda- 2009) red leaving 81bp scar DIG4308 Wild-type S. Typhimurium with None (Guymer et al., STM4308 deleted by lambda-red 2009) leaving 81bp scar DIG100 Wild-type S. Typhimurium with None (Guymer et al., tatABC genes deleted by lambda-red 2009) leaving 81bp scar DIG101 Wild-type S. Typhimurium with None (Guymer et al., STM1499 and STM1498 deleted by 2009) λRED leaving 81bp scar DIG102 S. Typhimurium strain LT2 with the None (Guymer et al., gene STM1499 deleted by λRED and 2009) replaced with an 81bp scar DIG103 Wild-type S. Typhimurium strain LT2 None (Guymer et al., with the gene STM1498 (ynfF) deleted 2009) by λRED and replaced with an 81bp scar DIG105 S. Typhiumrium strain LT2 with the None Dr David dmsA gene deleted by λRED and Guymer, replaced with an 81bp scar unpublished DIG107 S. Typhimurium DIG104 (ΔtorA) with None Dr David the dmsA gene deleted by λRED and Guymer, replaced with an 81bp scar unpublished FTD100 As MC4100, ΔtorD None (Jack et al., 2004) JARV16 att tatAYFP As B1LK0 (ΔtatB) (Bogsch et al., 1998), Kan but with tatA-YFP fusion in the lambda site KM01 As DIG103, with L24Q substitution in None This study spYnfE KM02 As DIG103, with A28Q substitution in None This study spYnfE KM03 As DIG103, with L33Q substitution in None This study spYnfE LT2a Wild-type S. Typhimurium strain. Attenuated due to altered rpoS allele. 201

MG1655 Wild-type E. coli strain. None (Blattner et F-, lambda-, rph-1 al., 1997) MG1655 ΔcyaA::Apra As MG1655, with PCR from Apra pMAK705DcyaA::Apra electroporated into host using λ red protocol - MC1061 F Δ(ara-leu)7697 [araD139]B/r Δ(codB- None (Casadaban & lacI)3 galK16 galE15 λ- e14- mcrA0 Cohen, 1980) relA1 rpsL150(strR) spoT1 mcrB1 hsdR2(r-m+) MC4100 F’ lacΔ169 araD139 rpsL150 thi None (Casadaban & flbB5301 deoC7 ptsF25 relA ara+ Cohen, 1980) MCBYFP As MC4100, but with tatB C- None terminally tagged with YFP. NCTC12023 Wild-type S. Typhimurium strain None RJ607-D As RJ607 (ssTorA::hybO ΔhybA) with Kan (Jack et al., torD gene replaced by Kan insert 2004) SL1344 Wild-type S. Typhimurium strain None

7.2 Media, additives and growth conditions Growth media and supplements used in this work are shown in Table 7.2 and Table 7.3. Strains were typically grown at 37 oC unless otherwise stated. Aerobic growth was conducted with agitation at 200 rpm, with cultures maintaining at least a 1:4 liquid to air ratio to ensure aerobiosis. Anaerobic growth was conducted by filling of containers to the top with medium and incubation without agitation. Growth on solid media was also performed at 37 oC unless otherwise stated, with anaerobic growth when required being conducted in sealed anaerobic jars with an Oxoid AnaeroGen sachet (Thermo Scientific).

Long term storage of strains was carried out at -80 oC by the addition of a final concentration of 25% glycerol to a stationary phase culture, with this being flash frozen with liquid nitrogen.

Table 7.2: Growth media used in this study. Media and their components were sterilised by autoclaving unless marked by as asterisk, in which case they were filter sterilised and added prior to media use.

Media Components Concentration Luria Bertani (LB) medium Tryptone 1.0% [w/v] Yeast extract 0.5% [w/v] NaCl 1.0% [w/v] 202

Low salt LB medium Tryptone 1.0% [w/v] Yeast extract 0.5% [w/v] NaCl 0.5% [w/v] LB-agar Tryptone 1.0% [w/v] Yeast extract 0.5% [w/v] NaCl 1.0% [w/v] Agar 1.5% [w/v] M9 medium M9 salts (10x) 10% [v/v] * (for Biolog PM growth tests) 1 M MgSO4 0.2% [v/v] * 1 M CaCl2 2% [v/v] 20% maltose 0.1-2% [v/v]

Table 7.3: Additives used in this study, with their respective stock and working concentrations. Supplement Stock Solution Final Concentration Dimethyl sulfoxide (DMSO) 20% 0.4% Glucose 20% [w/v] 0.5% Glycerol 50% [v/v] 0.5% Isopropyl β-D-1-thiogalactopyranoside 1 M 2mM Sodium fumarate 16% [w/v] 0.4%

Sodium selenate (Na2SeO4) 100 mM 10 mM Trimethylamine N-oxide (TMAO) 20% [w/v] 0.4%

Table 7.4: Antibiotics used in this study, with their stock and working concentrations. Antibiotic Stock Concentration Final Concentration Ampicillin (Amp) 125 mg ml-1 100 µg ml-1 Apramycin (Apra) 50 mg ml-1 50 µg ml-1 Chloramphenicol (Cm) 25 mg ml-1 25 µg ml-1 Kanamycin (Kan) 50 mg ml-1 50 µg ml-1

7.3 Buffers and solutions

Table 7.5: General buffers and solutions used in this study. Buffer/Solution Components Concentration APS Ammonium persulfate 10% [w/v] Crosslinking buffer HEPES 50 mM KCl 150 mM DNA loading buffer Bromophenol blue 0.25% [w/v] Xylene cyanol blue 0.25% [w/v] Sucrose 40% [w/v] Laemmli sample buffer (2x) Tris-HCl, pH 6.8 62.5 mM SDS 2% [w/v] Glycerol 26.3% [w/v] 203

Bromophenol blue 0.01% [w/v] β-mercaptoethanol 15% [v/v] M9 salts (10x stock) NaCl 5 g L-1 -1 Na2HPO4.7H2O 64 g L -1 KH2PO4 30 g L -1 NH4Cl 10 g L -1 ONPG ortho-Nitrophenyl-β-galactoside 4 mg ml -1 0.1 M Phosphate buffer NaH2PO4-2H2O 6.59 g L -1 Na2HPO4 8.19 g L

Phosphate Buffered Saline (PBS) KH2PO4 1 mM NaCl 155 mM 3 mM Na2HPO4-7H2O Ponceau S solution Ponceau S 0.1% [w/v] Acetic acid 5% [w/v] SDS running buffer Tris-HCl, pH 8.3 250 mM Glycine 1.92 M EDTA 1.0% [w/v] TBS/Tween (TBS-T) Tris-HCl, pH 8.0 20 mM NaCl 137 mM Tween 20 0.1% [v/v] TEV storage buffer Tris-HCl, pH 8 50 mM NaCl 25 mM Glycerol 10% EDTA 1 mM Dithiothreitol (DTT) 1 mM Tris Buffered Saline (TBS) Tris-HCl, pH 7.5 20 mM NaCl 137 mM Tris-Glycine transfer buffer Tris 25 mM Glycine 192 mM Methanol 10% [v/v] -1 Z-buffer Na2HPO4 8.52 g L -1 NaH2PO4.2H2O 6.24 g L KCl 0.75 g L-1 -1 MgSO4.7H2O 0.25 g L β-mercaptoethanol 0.7 ml L-1

Table 7.6: Buffers used during the purification of proteins by FPLC. Buffer Use Components Concentration Buffer A Nickel affinity Tris-HCl, pH 7.5 25 mM chromatography Imidazole 25 mM NaCl 250 mM Buffer B Nickel affinity Tris-HCl, pH 7.5 25 mM chromatography Imidazole 500 mM NaCl 250 mM Buffer C Size exclusion Tris-HCl, pH 7.5 25 mM chromatography NaCl 250 mM 204

Buffer D TEV purification – nickel Tris-HCl, pH 8 50 mM affinity chromatography NaCl 300 mM Glycerol 10% Buffer E TEV purification – nickel Tris-HCl, pH 8 50 mM affinity chromatography NaCl 300 mM Glycerol 10% Imidazole 500 mM Buffer F TEV purification – lysis Tris-HCl, pH 8 50 mM buffer NaCl 300 mM Glycerol 10% Imidazole 25 mM

MgCl2 2 mM

7.4 Genetic manipulations

7.4.1 Plasmid DNA preparation Plasmids were extracted from E. coli and Salmonella strains using the QIAprep Spin Miniprep kit (Qiagen). This technique is based on an alkaline lysis method modified by Birnboim & Doly

(1979). A 5 ml culture inoculated with a single colony was grown overnight prior to centrifugation at 10,000 rpm for 10 minutes. Cells were resuspended in an alkaline lysis buffer in order to release cell contents, with the lysate being treated with neutralisation buffer and cleared of precipitate. Plasmid DNA was then able to be isolated through adsorption onto a silica membrane under high-salt conditions. After washing to remove endonucleases and salts, plasmid DNA was eluted with dH2O.

205

Table 7.7: Plasmids used throughout this study. Plasmid Name Characteristics Resistance Reference pBluescript KS+ Cloning vector, with polylinker oriented in the KS orientation. Amp Stratagene pFLUOR_mCerN-TorD_YFPN-TorD Production of N-terminal mCer and YFP fusion to individual TorD Amp Dr Sarah Coulthurst, proteins unpublished pFLUOR_mCerN-TorD_TorD-YFPC Production of N-terminal mCer and C-terminal YFP fusion to individual Amp Dr Sarah Coulthurst, TorD proteins unpublished pFLUOR_TorD-mCerC_ TorD-YFPC Production of C-terminal mCer and YFP fusion to individual TorD Amp Dr Sarah Coulthurst, proteins unpublished pFLUOR_TorD-mCerC _YFPN-TorD Production of C-terminal mCer and C-terminal YFP fusion to individual Amp Dr Sarah Coulthurst, TorD proteins unpublished pQE70L-stmynfE Over-expression vector with lac operator and T5 promoter, with codons Amp expressing hexa-histidine tag at 3’ end after multiple cloning site. Allows IPTG-inducible over-expression of YnfE-his protein from S. Typhimurium pQE80L-Nhis-TEV-DmsD Over-expression vector with lac operator and T5 promoter, with codons Amp This study expressing hexa-histidine tag at 5’ end before multiple cloning site. DmsD from S. Typhimurium with N-terminal TEV cleavage site cloned in, allowing IPTG-inducible over-expression of his-TEV-DmsD pQE80L-Nhis-TEV-DmsD K61Q As pQE80L-Nhis-TEV-DmsD, with residue K61 substituted for Q Amp This study pQE80L-Nhis-TEV-DmsD K200Q As pQE80L-Nhis-TEV-DmsD, with residue K200 substituted for Q Amp This study pQE80L-Nhis-TEV-DmsD double As pQE80L-Nhis-TEV-DmsD, with all K residues substituted for Q Amp This study pQE80L-HiPIP HiPIP from E. coli cloned in, allowing IPTG-inducible over-expression of Amp HiPIP pQE80L-SufI SufI from E. coli cloned in, allowing IPTG-inducible over-expression of Amp SufI pQE80L-TorA TorA from E. coli cloned in, allowing IPTG-inducible over-expression of Amp TorA pMAK705 Derivative of pMAK700, suicide vector that allows E. coli chromosomal Cm (Hamilton et al., 1989) mutations pMAK-spYnfE pMAK705 containing chromosomal nucleotide sequence 500 bp Cm This study upstream and downstream of S. Typhimurium spYnfE 206

pMAK-spYnfE L24Q pMAK705 containing nucleotide sequence 500 bp upstream and Cm This study downstream of spYnfE with L24Q substitution pMAK-spYnfE A28Q pMAK705 containing nucleotide sequence 500 bp upstream and Cm This study downstream of spYnfE with A28Q substitution pMAK-spYnfE L33Q pMAK705 containing nucleotide sequence 500 bp upstream and Cm This study downstream of spYnfE with L33Q substitution pT25 Cloning vector for bacterial-two hybrid system. Encodes the T25 Cm (Karimova et al., 1998) fragment of CyaA (a.a. 1–224) from B. pertussis in frame with a multicloning site sequence, under the control of the lac UV5 promoter pT25-DmsD Encodes DmsD-T25 fusion for bacterial-two hybrid analysis Cm pT25-DmsD V16Q As pT25-DmsD, with residue V16 substituted for Q Cm This study pT25-DmsD W91Q As pT25-DmsD, with residue W91 of DmsD substituted for Q Cm This study pT25-DmsD S96Q As pT25-DmsD, with residue S96 of DmsD substituted for Q Cm This study pT25-DmsD G100Q As pT25-DmsD, with residue G100 of DmsD substituted for Q Cm This study pT25-DmsD T103Q As pT25-DmsD, with residue T103 of DmsD substituted for Q Cm This study pT25-DmsD G171Q As pT25-DmsD, with residue G171 of DmsD substituted for Q Cm This study pUNI-PROM pT7.5 with tat promoter cloned EcoRI-BamHI Amp (Jack et al., 2004) pUNI-PROM_DmsDst As pUNI-PROM, with dmsD from S. Typhimurium cloned BamHI-PstI Amp pUNI-PROM_DmsDst mutant library As pUNI-PROM_DmsDst, subjected to error-prone PCR, giving ~400,000 Amp Dr Holger Knueper, clones. unpublished pUNI-PROM_DmsDst V16Q As pUNI-PROM_DmsDst, with residue V16 of DmsD substituted for Q Amp This study pUNI-PROM_DmsDst W91Q As pUNI-PROM_DmsDst, with residue W91 of DmsD substituted for Q Amp This study pUNI-PROM_DmsDst S96Q As pUNI-PROM_DmsDst, with residue S96 of DmsD substituted for Q Amp This study pUNI-PROM_DmsDst G100Q As pUNI-PROM_DmsDst, with residue G100 of DmsD substituted for Q Amp This study pUNI-PROM_DmsDst T103Q As pUNI-PROM_DmsDst, with residue T103 of DmsD substituted for Q Amp This study pUNI-PROM_DmsDst G171Q As pUNI-PROM_DmsDst, with residue G171 of DmsD substituted for Q Amp This study pUNI-PROM_TorD As pUNI-PROM, with torD from E. coli Amp pUNI-PROM_TorD E23A As pUNI-PROM-TorD, with residue E23 substituted for A Amp pUNI-PROM_TorD Q31A As pUNI-PROM-TorD, with residue Q31 substituted for A Amp pUNI-PROM_TorD G112A As pUNI-PROM-TorD, with residue G112 substituted for A Amp pUNI-PROM_TorD T115A As pUNI-PROM-TorD, with residue T115 substituted for A Amp pUNI-PROM_TorD P122A As pUNI-PROM-TorD, with residue P122 substituted for A Amp 207

pUNI-PROM_TorD E142A As pUNI-PROM-TorD, with residue E142 substituted for A Amp pUNI-PROM_TorD G112A P122A As pUNI-PROM-TorD, with residues G112 & P122 substituted for A Amp pUT18 Derivative of pBluescript II KS (Stratagene) for bacterial-two hybrid Amp system. Encodes T18 fragment of CyaA (a.a. 225-399) from B. pertussis in frame with a multicloning site sequence from pBluescript II KS. pUT18-spYnfE Encodes T18-spYnfE (S. Typhimurium) fusion for bacterial-two hybrid Amp analysis pUT18-spYnfE (G3-A42)Q pUT18-spYnfE (S. Typhimurium) vectors with residues G3 to A42 Amp This study individually substituted for Q

pUT18-YnfEΔsp As pUT18, with YnfE (S. Typhimurium) missing N-terminal signal peptide Amp This study (a.a. 1-42) pUT18-YnfE As pUT18, encoding full length YnfE (a.a 1-813) (S. Typhimurium) Amp This study pUT18-YnfG As, pUT18, encoding YnfG (S. Typhimurium), an Fe-S subunit of selenate Amp This study reductase

Table 7.8: Oligonucleotides used throughout this study. All oligonucleotides were obtained from Sigma Aldrich.

Oligonucleotide Name Sequence (5’-3’) Use TorD E23A for CAGTTGTTCTCCCGTGCGCTGGACGATGAACAA For TorDec Quikchange TorD E23A rev TTGTTCATCGTCCAGCGCACGGGAGAACAACTG For TorDec Quikchange TorD Q31Q for GATGAACAACTGACGGCAATCGCCAGTGCGCAG For TorDec Quikchange TorD Q31A rev CTGCGCACTGGCGATTGCCGTCAGTTGTTCATC For TorDec Quikchange TorD G112A for TTGTTAGTTGAGGCAGCGATGGAAACCAGCGGC For TorDec Quikchange TorD G112A rev GCCGCTGGTTTCCATCGCTGCCTCAACTAACAA For TorDec Quikchange TorD T115A for GAGGCAGGGATGGAAGCCAGCGGCAATTTCAAC For TorDec Quikchange TorD T115A rev GTTGAAATTGCCGCTGGCTTCCATCCCTGCCTC For TorDec Quikchange 208

TorD P122A for GGCAATTTCAACGAAGCGGCAGATCATCTGATC For TorDec Quikchange TorD P122A rev GATCAGATGATCTGCCGCTTCGTTGAAATTGCC For TorDec Quikchange TorD E142A for CATTTTTCGCTGGGAGCGGGGACCGTTCCTGCG For TorDec Quikchange TorD E142A rev CGCAGGAACGGTCCCCGCTCCCAGCGAAAAATG For TorDec Quikchange YnfEst G3Q for GGAGGATCCATGCCCCAAGAAAAGCAACAAACAGGC For spYnfEst Quikchange YnfEst G3Q rev GCCTGTTTGTTGCTTTTCTTGGGGCATGGATCCTCC For spYnfEst Quikchange YnfEst E4Q for GTGATGCCCGGACAAAAGCAACAAACAGG For spYnfEst Quikchange YnfEst E4Q rev CCTGTTTGTTGCTTTTGTCCGGGCATCAC For spYnfEst Quikchange YnfEst K5Q for ATGCCCGGAGAAGACCAACAAACAGGCG For spYnfEst Quikchange YnfEst K5Q rev CGCCTGTTTGTTGGTCTTCTCCGGGCAT For spYnfEst Quikchange YnfEst Q6W for CCCGGAGAAAAGTGGCAAACAGGC For spYnfEst Quikchange YnfEst Q6W rev GCCTGTTTGCCACTTTTCTCCGGG For spYnfEst Quikchange YnfEst Q7W for CCATGCCCGGAGAAAAGCAATGGACAGGCGTTAGCCGCAGGACG For spYnfEst Quikchange YnfEst Q7W rev CGTCCTGCGGCTAACGCCTGTCCATTGCTTTTCTCCGGGCATGG For spYnfEst Quikchange YnfEst T8Q for GAAAAGCAACAACAAGGCGTTAGCCGC For spYnfEst Quikchange YnfEst T8Q rev GCGGCTAACGCCTTGTTGTTGCTTTTC For spYnfEst Quikchange YnfEst G9Q for GCAACAAACACAAGTTAGCCGCAGG For spYnfEst Quikchange YnfEst G9Q rev CCTGCGGCTAACTTGTGTTTGTTGC For spYnfEst Quikchange YnfEst A19Q for CGTTAGTGAAATCACAGGCGCTTGGCTCACTG For spYnfEst Quikchange YnfEst A19Q rev CGTTAGTGAAATCACAGGCGCTTGGCTCACTG For spYnfEst Quikchange YnfEst A20Q for GTGAAATCAGCGCAGCTTGGCTCACTGGCG For spYnfEst Quikchange YnfEst A20Q rev CGCCAGTGAGCCAAGCTGCGCTGATTTCAC For spYnfEst Quikchange YnfEst L21Q for GTGAAATCAGCGGCGCAAGGCTCACTGGCGCTG For spYnfEst Quikchange YnfEst L21Q rev CAGCGCCAGTGAGCCTTGCGCCGCTGATTTCAC For spYnfEst Quikchange YnfEst G22Q for GTGAAATCAGCGGCGCTTCAATCACTGGCGCTG For spYnfEst Quikchange YnfEst G22Q rev CAGCGCCAGTGATTGAAGCGCCGCTGATTTCAC For spYnfEst Quikchange YnfEst S23Q for GCGGCGCTTGGCCAACTGGCGCTGGCTGCC For spYnfEst Quikchange YnfEst S23Q rev GGCAGCCAGCGCCAGTTGGCCAAGCGCCGC For spYnfEst Quikchange YnfEst L24Q for GCGGCGCTTGGCTCACAGGCGCTGGCTGCC For spYnfEst Quikchange YnfEst L24Q rev GGCAGCCAGCGCCTGTGAGCCAAGCGCCGC For spYnfEst Quikchange YnfEst A25Q for GCGCTTGGCTCACTGCAGCTGGCTGCCGGAGGC For spYnfEst Quikchange YnfEst A25Q rev GCCTCCGGCAGCCAGCTGCAGTGAGCCAAGCGC For spYnfEst Quikchange 209

YnfEst L26Q for GCGCTTGGCTCACTGGCGCAGGCTGCCGGAGGC For spYnfEst Quikchange YnfEst L26Q rev GCCTCCGGCAGCCTGCGCCAGTGAGCCAAGCGC For spYnfEst Quikchange YnfEst A27Q for GGCTCACTGGCGCTGCAGGCCGGAGGCGTATC For spYnfEst Quikchange YnfEst A27Q rev GATACGCCTCCGGCCTGCAGCGCCAGTGAGCC For spYnfEst Quikchange YnfEst A28Q for CTGGCGCTGGCTCAGGGAGGCGTATCTTTACCG For spYnfEst Quikchange YnfEst A28Q rev CGGTAAAGATACGCCTCCCTGAGCCAGCGCCAG For spYnfEst Quikchange YnfEst G29Q for CTGGCGCTGGCTGCCCAAGGCGTATCTTTACCG For spYnfEst Quikchange YnfEst G29Q rev CGGTAAAGATACGCCTTGGGCAGCCAGCGCCAG For spYnfEst Quikchange YnfEst G30Q for GCGCTGGCTGCCGGACAGGTATCTTTACCGTTTGGG For spYnfEst Quikchange YnfEst G30Q rev CCCAAACGGTAAAGATACCTGTCCGGCAGCCAGCGC For spYnfEst Quikchange YnfEst V31Q for CTGGCTGCCGGAGGCCAATCTTTACCGTTTGGG For spYnfEst Quikchange YnfEst V31Q rev CCCAAACGGTAAAGATTGGCCTCCGGCAGCCAG For spYnfEst Quikchange YnfEst S32Q for GCTGCCGGAGGCGTACAGTTACCGTTTGGGATG For spYnfEst Quikchange YnfEst S32Q rev CATCCCAAACGGTAACTGTACGCCTCCGGCAGC For spYnfEst Quikchange YnfEst L33Q for GCCGGAGGCGTATCTCAACCGTTTGGGATGCGTAC For spYnfEst Quikchange YnfEst L33Q rev GTACGCATCCCAAACGGTTGAGATACGCCTCCGGC For spYnfEst Quikchange YnfEst P34Q for GGAGGCGTATCTTTACAGTTTGGGATGCGTACC For spYnfEst Quikchange YnfEst P34Q rev GGTACGCATCCCAAACTGTAAAGATACGCCTCC For spYnfEst Quikchange YnfEst F35Q for GGCGTATCTTTACCGCAGGGGATGCGTACCGCG For spYnfEst Quikchange YnfEst F35Q rev CGCGGTACGCATCCCCTGCGGTAAAGATACGCC For spYnfEst Quikchange YnfEst G36Q for GTATCTTTACCGTTTCAGATGCGTACCGCGGCC For spYnfEst Quikchange YnfEst G36Q rev GGCCGCGGTACGCATCTGAAACGGTAAAGATAC For spYnfEst Quikchange YnfEst M37Q for CTTTACCGTTTGGGCAGCGTACCGCGGCCGCG For spYnfEst Quikchange YnfEst M37Q rev CGCGGCCGCGGTACGCTGCCCAAACGGTAAAG For spYnfEst Quikchange YnfEst R38Q for CTTTACCGTTTGGGATGCAGACCGCGGCCGCG For spYnfEst Quikchange YnfEst R38Q rev CGCGGCCGCGGTCTGCATCCCAAACGGTAAAG For spYnfEst Quikchange YnfEst T39Q for GTTTGGGATGCGTCAGGCGGCCGCGGCGGTG For spYnfEst Quikchange YnfEst T39Q rev CACCGCCGCGGCCGCCTGACGCATCCCAAAC For spYnfEst Quikchange YnfEst A40Q for CCGTTTGGGATGCGTACCCAGGCCGCGGCGGTGCAG For spYnfEst Quikchange YnfEst A40Q rev CTGCACCGCCGCGGCCTGGGTACGCATCCCTTTCGG For spYnfEst Quikchange YnfEst A41Q for GGGATGCGTACCGCGCAGGCGGCGGTGCAGCAG For spYnfEst Quikchange YnfEst A41Q rev CTGCTGCACCGCCGCCTGCGCGGTACGCATCCC For spYnfEst Quikchange 210

YnfEst A42Q for GATGCGTACCGCGGCCCAGGCGGTGCAGCAGTCG For spYnfEst Quikchange YnfEst A42Q rev CGACTGCTGCACCGCCTGGGCCGCGGTACGCATC For spYnfEst Quikchange YnfEst sp for GCGCAAGCTTAGAACATCGTCATTATCACAG For cloning of 500 bp upstream of spYnfE. Contains HindIII restriction site. YnfEst sp rev GCGCGGATCCGTTGCCGTCGTTGGAACC For cloning of 500 bp upstream of spYnfE. Contains BamHI restriction site. YnfEst for GCGCACAATGGAGTGAGTGATGCCC For sequencing YnfE clones. YnfEst rev GCGCAGCCTGCTGCACCGCCGCGGC For sequencing YnfE clones. pQE80 TEV DmsDst for GCGCGGATCCGAAAACCTGTATTTTCAGGGCATGACCACTTTTTTACAACGTGATG For construction of His-TEV- DmsDst vector pQE80 DmsDst rev GCGCAAGCTTTTATTAACGGAATAACGGTTTTACAGCG For construction of His-TEV- DmsDst vector DmsDst V16Q for CAGTAACGGCGCGTCAGCTTGGCGCGTTG For DmsDst Quikchange DmsDst V16Q rev CAACGCGCCAAGCTGACGCGCCGTTACTG For DmsDst Quikchange DmsDst W91Q for CGTGGGGTTCCGTCCAGCTGGATCGTGAGAGCG For DmsDst Quikchange DmsDst W91Q rev CGCTCTCACGATCCAGCTGGACGGAACCCCACG For DmsDst Quikchange DmsDst R94Q for CCGTCTGGCTGGATCAGGAGAGCGTATTATTTGGC For DmsDst Quikchange DmsDst R94Q rev GCCAAATAATACGCTCTCCTGATCCAGCCAGACGG For DmsDst Quikchange DmsDst S96Q for GGCTGGATCGTGAGCAGGTATTATTTGGC For DmsDst Quikchange DmsDst S96Q rev GCCAAATAATACCTGCTCACGATCCAGCC For DmsDst Quikchange DmsDst G100Q for GAGCGTATTATTTCAGGATTCTACATTGGC For DmsDst Quikchange DmsDst G100Q rev GCCAATGTAGAATCCTGAAATAATACGCTC For DmsDst Quikchange DmsDst T103Q for GTATTATTTGGCGATTCTCAATTGGCGCTACGTC For DmsDst Quikchange DmsDst T103Q rev GACGTAGCGCCAATTGAGAATCGCCAAATAATAC For DmsDst Quikchange DmsDst G171Q for CGATCATGCCCAGCATCCGTTTTATC For DmsDst Quikchange DmsDst G171Q rev GATAAAACGGATGCTGGGCATGATCG For DmsDst Quikchange Full YnfE HindIII for GCGCAAGCTTGTGATGCCCGGAGAAAAGC For construction of vector containing full length YnfEst. Contains HindIII restriction site. Full YnfE EcoRI rev GCGCAATTCTTAGATTTTTTCGATTTCC For construction of vector 211

containing full length YnfEst Contains EcoRI restriction site. FullYnfG HindIII for GCGCAAGCTTATGACAACCCAGTATGG For construction of vector containing full length YnfGst. Contains HindIII restriction site. FullYnfG EcorRI rev GCGCAATTCTTATACCTCCTTCGGATTGGC For construction of vector containing full length YnfGst Contains EcoRI restriction site. Nosig YnfE EcoRI for GCGCGAATTGCGGTGCAGCAGGCTATGCGC For construction of vector containing YnfE lacking the N- terminal signal peptide (a.a. 1- 42). Contains EcoRI restriction site. Nosig YnfE HindIII rev GCGCAAGCTTTTAGATTTTTTCGATTTCC For construction of vector containing YnfE lacking the N- terminal signal peptide (a.a. 1- 42). Contains HindIII restriction site. DmsDst K61Q for CCTGTGGCGGCTATGTTTCAGACCCACAGCGAAGAGTCG For lysine Quikchange with pQE80L-Nhis-TEV-DmsDst DmsDst K61Q rev CGACTCTTCGCTGTGGGTCTGAAACATAGCCGCCACAGG For lysine Quikchange with pQE80L-Nhis-TEV-DmsDst DmsDst K200Q for CTCATTATTCCCGTCGCTGTACAACCGTTATTCCGTTAATAAAAGC For lysine Quikchange with pQE80L-Nhis-TEV-DmsDst DmsDst K200Q rev GCTTTTATTAACGGAATAACGGTTGTACAGCGACGGGAATAATGAG For lysine Quikchange with pQE80L-Nhis-TEV-DmsDst

212

7.4.2 Amplification of DNA by Polymerase Chain Reaction (PCR) The polymerase chain reaction (PCR) is a popular biochemical technique that allows the amplification of specific regions of DNA by several orders of magnitude through the utilisation of thermostable DNA polymerase enzymes and oligonucleotides. A typical PCR reaction uses oligonucleotide primers designed to be complementary to 5’ and 3’ ends of a thermally denatured single stranded DNA template. DNA elongation occurs in a 5’ to 3’ direction, enabling the extension of the 5’ and 3’ DNA strands towards one another in order to form a complete double strand of DNA.

PCR consists of three essential steps: denaturation, annealing and elongation. PCR reaction step temperatures and timings were modified depending on the size of DNA to be amplified and the specific PCR technique being utilised.

7.4.3 Site-directed mutagenesis by QuikchangeTM PCR Site-directed mutagenesis was performed based on the QuikchangeTM manual from

Stratagene. In order to introduce specific amino acid substitutions into a given sequence, complementary primers of approximately 30 base pairs in length were designed (Sigma). The substituted codon of interest was positioned in the centre of the sequence. DNA with the substituted codon was then amplified by PCR, before methylated template DNA present in the sample was digested by DpnI (New England Biolabs). DNA was then transformed into chemically competent E. coli strain DH5α and extracted by Miniprep.

7.4.4 Agarose gel electrophoresis DNA samples were analysed by agarose gel electrophoresis using 1 % w/v agarose gels prepared with TAE buffer and containing 0.001 % v/v Gel Red dye (Biotium). Loading dye was added to samples to create a visible running front, whilst DNA size markers (Stratagene) were 213 run alongside samples to allow an estimation of band sizes. Gels were run in TAE buffer at 100

V, with bands beings visualised in a Bio-Rad Gel-Doc imaging system.

7.4.5 DNA digestion and preparation for cloning DNA was incubated with 10 units of the restriction enzymes appropriate for sites to be cut, in a final reaction volume of 40 or 50 μl, depending on whether the DNA being digested was plasmid or PCR product. Also within this reaction volume was the required restriction enzyme buffer and DNA, usually either 1 μg of plasmid DNA or 43 μl of PCR product. Samples were incubated for at least 2 hours or overnight at 37 oC.

Digested vectors were dephosphorylated by treatment with alkaline phosphatase, in order to prevent self-religation. This involved the incubation of 40 μl digested vector with 5 μl of 10x alkaline phosphatase buffer and 5 μl alkaline phosphatase for 30 minutes at 37 oC. Samples were then analysed by agarose gel electrophoresis and visualised using UV light. Bands of the correct product size were excised and purified using a QIAquick Gel Extraction kit (Qiagen).

7.4.6 DNA ligation Digested DNA fragments with compatible ends were joined using T4 DNA ligase (Roche).

Samples were prepared to a final volume of 10 μl in ligation buffer, containing 1 μl of vector and 1 μl or 3 μl of PCR product to be inserted, as well as 1 μl of T4 DNA ligase. Reactions were incubated overnight at 18 oC before the entire reaction volume was used in transformations with chemically competent cells.

7.4.7 DNA sequencing Sequencing of DNA was performed in order to confirm plasmid sequences after cloning or

QuikchangeTM. DNA sequencing was performed by DNA Sequencing & Services (MRCPPU, 214

College of Life Sciences, University of Dundee, Scotland, www.dnaseq.co.uk) using Applied

Biosystems Big-Dye Ver 3.1 chemistry on an Applied Biosystems model 3730 automated capillary DNA sequencer. Chromatogram files obtained from sequencing reactions were analysed using BioEdit Sequence Alignment Editor (Copyright © 1997-2007 Tom Hall).

7.4.8 Preparation of competent cells and transformation with plasmid DNA For chemically competent cells, 5 ml liquid LB plus any required antibiotics was inoculated

1:100 from an overnight culture, and grown aerobically at 37 oC with shaking at 200 rpm until an OD600 pf 0.5 was achieved. Cells were then harvested by centrifugation at 3020x g for 10 minutes. Cells were resuspended in TSB medium and kept on ice for at least 15 minutes before use or freezing in liquid nitrogen and stored at -80 oC. Transformation was performed by incubating 100 μl of competent cells with 1 μl of plasmid for 15 minutes on ice. Cells were then heat-shocked at 42 oC for 90 seconds followed by a further 2 minute incubation on ice. 1 ml of

LB was then added and cells were grown at 37 oC for 1 hour, before cells were isolated by centrifugation at 16060x g for 1 minute. Cells were plated onto LB agar plates containing any required antibiotics and grown overnight at 37 oC.

For transformation with Salmonella, electrocompetent cells were prepared. 25 ml low salt LB plus required antibiotics was inoculated 1:100 with overnight culture. Cells were grown

o aerobically at 30 C to OD600 0.5. After cooling on ice for 40 minutes, cells were isolated by 10 minutes centrifugation at 4500 rpm, and then resuspended in 10 ml ice cold sterile 10 % glycerol. Cells were washed with 10 ml ice cold 10 % glycerol another two times before being resuspended in 0.5 ml ice cold 10 % glycerol. 50 μl aliquots were then taken and either used immediately or frozen in liquid nitrogen and stored at -80 oC. To transform, 1 μl plasmid was added to a 50 μl aliquot and incubated on ice for 10 minutes. Cells were placed into an ice-cold

Gene Pulser cuvette (Bio-Rad) and an electric pulse was passed through. 1 ml LB was mixed 215 with the sample and cells were recovered at 37 oC for 1 hour prior to spreading on LB agar plates with appropriate antibiotics overnight.

7.4.9 In-frame chromosomal integration using pMAK705 vector Gene replacement on the chromosome of Salmonella was achieved through the use of the suicide vector pMAK705 for homologous recombination (Hamilton et al., 1989). The gene region including 500 base pairs up and downstream of the region of interest was cloned into pMAK705 using DNA amplification, digestion and ligation techniques as described above, using pBluescript KS+ as a cloning intermediate. For integration of YnfE signal peptide sequence variants, pBluescript KS+ containing the appropriate gene region was subjected to

QuikchangeTM PCR prior to subcloning into pMAK705. Plasmids were transformed into a recA+ host strain using slightly different transformation conditions of 5 minutes heat shock at 37 oC and cell recovery at 30 oC for 1.5 hours. Cells were then grown on LB + chloramphenicol (Cm) plates overnight at 30 oC. All following steps were performed using selective LB containing chloramphenicol unless otherwise stated.

Overnight liquid cultures of transformed strains were grown aerobically at 30 oC. Serial dilutions at 1 x 10-7 (control), 1 x 10-5, 1 x 10-4 and 1 x 10-3 were plated onto LB + Cm, with the control plate then incubated at 30 oC and others at 44 oC overnight. Cells are only able to grow at 44 oC if plasmid integration has occurred. 10 ml liquid media was inoculated with four cell colonies per culture, and grown for 24 hours at 30 oC. A loopful of culture was used to inoculate fresh media before a further 24 hour incubation at 30 oC, with this step then being repeated once more. From this final culture plates were streaked to obtain single colonies. This process should lead to resolution of the co-integrated plasmid. To cure cells of the plasmid, a number of 10 ml LB culture without Cm were inoculated with one colony each before being grown at 44 oC for at least 16 hours. These cultures were streaked onto LB only plates and incubated at 44 oC overnight. Resulting individual colonies could then be patched onto both LB 216 only and LB + Cm plates to test sensitivity. Those that appeared Cm sensitive were analysed for desired chromosomal mutations by PCR and DNA sequencing.

7.5 Protein methods

7.5.1 Overproduction of proteins Proteins were overexpressed using pQE vector pQE80, encoding for an N-terminally His-tagged protein. Plasmids were transformed into the E. coli strains BL21 (DE3) pLysS or MC4100, before a single colony was used to inoculate a 5 ml overnight starter culture in LB plus appropriate antibiotics. 500 ml cultures were then inoculated using these starter cultures, at 1:1000 dilution. Cultures were grown aerobically in 2 L baffled flasks at 37 oC with 170 rpm shaking until they had reached OD600 0.6. Induction of recombinant protein expression was then initiated by addition of 2 mM isopropyl β-D thiogalatopyranoside (IPTG, Sigma). Cultures were then grown overnight at 18 oC before cells were harvested the next day by centrifugation.

7.5.2 Purification of hexa-histidine tagged proteins using immobilised metal ion affinity chromatography Cell pellets after induction of protein overexpression were resuspended in buffer A by homogenisation, in a volume to cell weight ratio of 10ml/g cells. In addition, protease inhibitor

(either EDTA-free complete protease inhibitor cocktail – Roche, or protease inhibitor cocktail set III, EDTA-Free - Calbiochem), lysozyme and DNase were added to the cell suspension. Cells were then lysed by three times passage at 15 000 psi through an Emulsiflex C3 high pressure homogenizer. Cell debris was then removed from the supernatant by centrifugation at 27143x g for 30 min at 4 oC. His-tagged proteins could then be isolated from the supernatant by immobilized metal ion affinity chromatography (IMAC) on an ÄKTATM FPLC system (GE

Healthcare). 217

Supernatant of lysed cells was filtered through a 0.45 µM membrane filter, before being loaded onto a 5 ml HisTrapTM HP column (GE Healthcare) pre-equilibrated with Buffer A. The column was then further washed with Buffer A to remove all unbound and non-specifically bound protein, usually approximately 5 column volumes. Bound protein was then eluted using an imidazole gradient of 25 mM to 500 mM (0 – 100 % Buffer B) over 5 column volumes and collected in 5 ml collection tubes. Fractions containing the protein of interest were identified by SDS-PAGE and were pooled and concentrated if required.

7.5.3 Purification of tobacco etch virus protease Tobacco etch virus (TEV) protease is a 27 kDa enzyme commonly used for the cleavage of affinity tags from recombinant proteins after purification. TEV protease recognises the general epitope E-X-X-Y-X-Q-G/S, where cleavage occurs between the Q and the G or S (Carrington &

Dougherty, 1988). Cell cultures expressing pET24b TEVhis were grown and overexpression induced as previously described. Cells were then resuspended in Buffer F plus EDTA-free complete protease inhibitor cocktail (Roche), lysozyme and DNase to a volume 5 times that of cell pellet mass. Lysis was performed, as before, by three times passage at 15 000 psi through an Emulsiflex C3 high pressure homogenizer followed by centrifugation at 27143x g for 30 min at 4 oC to remove cell debris. The supernatant was then filtered through a 0.45 µM membrane filter.

A 5 ml HisTrapTM HP column (GE Healthcare) was pre-equilibrated with Buffer D and 6 % Buffer

E prior to sample loading. The column was then washed with a further 5 column volumes before elution of protein using a Buffer E gradient of 0-60 % over 5 column volumes. Fractions were collected in 5 ml collection tubes, with EDTA being immediately added to a final concentration of 1 mM. SDS-PAGE was used to identify fractions containing TEV protease, with those fractions being pooled and dialysed overnight into Buffer G at 4 oC using 6.45 ml/cm 218 volume, 8 kDa molecular weight cut-off dialysis tubing (BioDesign). Protein concentration was then adjusted to 0.5 mg/ml prior to flash freezing and storage at -80 oC.

7.5.4 TEV cleavage of hexa-histidine tags For removal of the hexa-histidine tag, protein purified by IMAC as described above was mixed with TEVhis protease in a mg/ml ratio of approximately 1:10. This was then subjected to overnight dialysis at 4 oC using 6.45 ml/cm volume, 8 kDa molecular weight cut-off dialysis tubing (BioDesign) into Buffer A containing 1 mM DTT in order to remove unwanted imidazole from the sample. Cleaved DmsD protein was isolated from TEVhis protease and any uncleaved protein by reverse nickel chromatography with a 5 ml HisTrapTM HP column. The flow-through containing cleaved DmsD in Buffer A was collected in 5 ml collection tubes before the column was washed with Buffer B to remove any bound TEVhis and cleaved hexa-histidine tag.

Fractions were checked for the presence of the protein of interest by SDS-PAGE before being pooled and concentrated.

7.5.5 Purification of proteins using size-exclusion chromatography For further purification after IMAC, proteins can be separated according to size by size- exclusion chromatography (SEC). Samples were concentrated to 500 µl in a Vivaspin® 20 spin concentrator (Vivaproducts) with an appropriate molecular weight cut-off. SEC was performed using a Superdex 75 10/300 GL prepacked column with an ÄKTATM FPLC system (both GE

Healthcare). The column was first equilibrated with 1 ½ column volumes of degassed miliQ water, then 1 ½ column volumes of buffer C, prior to loading of the protein sample via a 500 µl loop. The column was then washed with a further 1 ½ column volumes of buffer C, in order to allow isocratic elution of proteins into 5 ml collection tubes. Fractions were analysed using

SDS-PAGE before being pooled and concentrated.

219

7.5.6 Protein concentration determination Concentrations of purified protein samples were determined through measurement of absorbance at 280 nm (A280) with an ND-1000 Nanodrop spectrophotometer (Thermo

Scientific). Whilst the relationship of absorbance to protein concentration is linear, in reality the degree of absorbance of a protein can be greatly affected by specific amino acid composition, particularly due to the presence of aromatic amino acids. Therefore, this method is only suitable for proteins of high purity with a known amino acid composition. A correction factor based on the molar extinction coefficient is used to calculate relative absorbance and therefore accurately predict protein concentration. Correction factors were calculated using the ExPASy ProtParam server (http://web.expasy.org/protparam/).

7.5.7 Purified protein crystallisation trials

Crystallisation trials were carried out at room temperature by the sitting-drop vapour-diffusion method in 96-well plates. Plates were set up using a Phoenix liquid handling system with purified DmsD protein (14 mg/ml) (Rigaku, Art Robbins Instruments) and the commercially available screens: PEG (Qiagen), The Classics suite (Qiagen), Hampton crystal screen (Hampton

Research), and Wizard (Emerald Biosystems). In addition conditions from Qiu et al. (2008) (20%

PEG400, 1 M ammonium sulfate, 10% glycerol, imidazole, pH 8.0) were tested. Drops were checked under microscope every day for the first week, then around once every week after this.

7.5.8 Polyclonal antibody production Purified proteins were concentrated and flash frozen before being sent to Eurogentec S.A.

(Belgium) for antibody to be produced, raised using the 28-day Speedy protocol with rabbits.

Antiserum received was aliquoted and stored at -80 oC prior to use.

220

7.5.9 Affinity purification of polyclonal antiserum Purified protein against which the antiserum was raised was ran on an SDS-PAGE gel and transferred onto nitrocellulose membrane by Western blot. The nitrocellulose membrane was incubated with ponceau stain for 1 min to reveal protein, which was marked prior to destaining of membrane with distilled water. The region of nitrocellulose membrane containing the protein was cut out and blocked in 1x PBS containing 5% milk for 1 hour, and subsequently was washed three times with 1x PBS. 200 µl antiserum and 800 µl 1x PBS was then used to incubate the membrane overnight at 4 oC with gentle shaking. The supernatant was then removed and the membrane was washed three times with 1x PBS, before 1 ml 100 mM glycine, pH 2.5, was added for 5 min with gentle shaking to elute the purified antibody bound to the membrane. This liquid was recovered and 100 µl 1 M Tris-HCl, pH 7.5 added to neutralise the pH and prevent antibody damage. The elution step was repeated for a second time, with first and second elutes then able to be tested in Western blots for antibody purity.

Purified antibody was used at a dilution of 1:200.

7.5.10 SDS-PAGE SDS polyacrylamide gel electrophoresis is used to separate proteins according to molecular weight under denaturing conditions. Tris-glycine gels were prepared with varying concentrations of polyacrylamide, according to experimental requirements, with a 0.75 mm thickness for use with the Mini-PROTEAN II system (Bio-Rad). Samples were prepared for PAGE by mixing 1:1 with Laemmli buffer. Tris-glycine gels were placed into a Bio-Rad gel electrophoresis tank before being submerged in SDS running buffer. Samples were loaded alongside Precision Plus ProteinTM All Blue standards (Bio-Rad) for allow approximation of protein molecular weights. The gel was run at 100 V until the dye front had passed through the stacking gel layer, before the voltage was increased to 200 V until the dye front or the appropriate marker had reached the end of the resolving gel. Gels could then be incubated in

Coomassie based InstantBlue stain (Expedon) for visualisation of proteins, or used further in 221 semi-dry western blotting. Stained gels were washed with water to destain prior to drying at

80 oC for 75 minutes in a Gel Dryer Model 583 (Bio-Rad).

7.5.11 Semi-dry western immunoblot For analysis by western immunoblot, samples were separated by SDS-PAGE as described above. Gels were soaked in Tris glycine transfer buffer before being placed on top of nitrocellulose membrane (Amersham Hybond-ECL, GE Healthcare), also pre-soaked in Tris glycine transfer buffer. These were stacked with 2 pieces of pre-soaked 3MM Whatmann paper on either side, and placed inside a TransBlot SD SemiDry Transfer Cell (Bio-Rad) for protein transfer. Transfer was performed for 30 minutes at a constant voltage of 10 V.

Nitrocellulose membrane was then blocked in 5 % skimmed milk TBS at room temperature for at least 1 hour, or at 4 oC overnight, with shaking.

The membrane was washed with TBS-T prior to a 1 hour incubation with primary antibody diluted in 20 ml TBS-T. A further 3 10 minute washes with TBS-T were then performed before the membrane was incubated for a further 1 hour with the secondary antibody, again diluted in 20 ml TBS-T. Working concentrations of antibody are listed below. Bands were visualised using enhanced chemiluminescent detection. This involved the use of horseradish peroxidase- conjugated secondary antibodies, where the conjugated horseradish peroxidase is able to oxidise luminol in the presence of peroxide, resulting in the emission of light proportional to protein quantities as the luminol decays. Bands were pictured using Konica Minolta medical film after the membrane was incubated with ImmobilonTM Western Chemiluminescent HRP

Substrate (Milipore), and developed using a SRX-101A medical film processor.

222

Table 7.9: Antibodies used in this work. General working concentrations used are shown.

Antibody Working Raised in Reference

- Primary concentration

Monoclonal penta-His 1:2,000 Mouse QIAGEN (Cat.# 34660)

Polyclonal DmsD antiserum 1:20,000 Rabbit

Polyclonal TorD antiserum 1:20,000 Rabbit

Polyclonal mCherry antiserum 1:20,000 Rabbit This study

Polyclonal TatA antiserum 1:500 Rabbit

Polyclonal YnfEst antiserum 1:20,000 Rabbit This study

Anti-MBP monoclonal 1:20,000 Mouse New England Biolabs

(Cat.# E8032S)

- Secondary

Anti-mouse IgG HRP conjugate 1:10.000 Goat Bio-Rad

Anti-rabbit IgG HRP conjugate 1:10,000 Goat Bio-Rad

DyLight 594 anti-rabbit 1:200 Goat

7.5.12 Chemical crosslinking of proteins For evaluation of interactions between DmsD and spYnfEst, purified DmsD prepared as described in sections 7.5.2, 7.5.4 and 7.5.5 was used in experiments involving chemical crosslinking. The process of crosslinking involves the chemical joining of two or more molecules through the addition of a covalent bond by a crosslinking reagent. Crosslinking reagents are molecules that consist of two or more reactive ends, usually on either end of a spacer arm of varying length, which can chemically attach to the proteins of interest via specific functional groups such as primary amines. In these experiments the crosslinkers disuccinimidyl suberate (DSS), formaldehyde and 1-Ethyl-3-[3-dimethylaminopropyl] carbodiimide hydrochloride (EDC) were used. 223

DSS is a non-cleavable, homobifunctional crosslinker that reacts with primary amines. The structure of DSS contains two N-hydroxysuccinimide (NHS) esters on either side of an 8 carbon spacer arm that is 11.4 Å in length (Figure 7.1). NHS esters are highly reactive specifically towards primary amines (-NH2) such as those found on lysine residues as well as the N- terminus of proteins. Providing there are no other primary amines in solution, specific crosslinking of the target molecules can occur between lysine residues within 11.4 Å of each other.

EDC, an amine-to-carboxyl crosslinker, is a zero-length crosslinker, meaning that it can cause direct conjugation of primary amines and carboxylates (-COOH) but does not become part of the final crosslink between molecules (Figure 7.1). Instead it causes the formation of an amine- reactive O-acylisourea intermediate through interactions with a carboxyl group. This intermediate can go on to interact with an amine and form an amine bond, with the release of an isourea by-product.

Formaldehyde is a chemical widely used for crosslinking and chemical fixation of proteins and cells (Figure 7.1).

Figure 7.1: Chemical structures of crosslinkers DSS, EDC and formaldehyde.

DmsD and synthetic peptide were co-incubated in HEPES crosslinking buffer in a reaction volume of 100 µl, on ice for 1 hour, prior to the addition of the chemical crosslinker of choice.

DSS was added to the reaction to a final concentration of 1 mM, EDC to 2 mM and formaldehyde to 1 % before subsequent 30 minute incubation on ice. Reactions were stopped by the addition of 10 mM Tris before being mixed in a 1:1 ratio with Laemli buffer. Samples 224 were then analysed through the use of SDS-PAGE (17 % acrylamide gels) and Western blot with

DmsD antiserum. The occurrence of crosslinks between DmsD and synthetic peptides could be observed by the appearance of a bandshift on gels.

7.6 Protein assays

7.6.1 BV-linked oxidoreductase assays Oxidation or reduction of substrates by oxidoreductase enzymes can be coupled to the in vitro reduction or oxidation of an artificial electron acceptor such as benzyl viologen (BV). BV changes from colourless when oxidised (BVOx) to blue-purple when reduced (BVRed), and therefore the level of oxidation or reduction can be quantified through measuring A600 using a

Lambda UV/Vis spectrophotometer (Perkin Elmer). Reactions can be initiated by addition of substrate or sample dependent upon the assay, and the conversion between BVOx and BVRed is then monitored. Oxidoreductase activity was calculated using the following equation:

× [ ] × BVOx [µmol min-1 mg-1] = × [ −1 ] × [ ]× [ ] ∆퐴600�푚푚푚 � 푉푐푐푐푐푐푐푐 푙 1000 −1 −1 2 휀600 푚푚 푐� 푚푝푝푝𝑝푝푝 푚푚 1 푐�

ΔA600 = absorption change at 600 nm per minute Vcuvette = volume of cuvette (0.0018 l) -1 -1 ε600 = extinction coefficient of oxidised BV at 600 nm (7.4 mM cm ) mprotein = mass of protein in periplasmic fraction

The extinction coefficient ε600 is multiplied by a factor of 2 due to the transfer of two electrons to or from BV for each substrate molecule.

7.6.1.1 TMAO reductase activity assay TMAO reductase activity was tested using the artificial electron donor BV. TMAO reductase can be reduced by BV, leading to a colour change from blue-purple to colourless. 225

Bacterial cells were grown overnight anaerobically in LB supplemented with 0.5% glycerol,

0.4% TMAO and appropriate antibiotics. Cells were isolated by centrifugation at 3000 x g for

10 minutes before washing in 10 ml 50 mM Tris-HCl, pH 7.5. Cells were again subjected to centrifugation at 3000 x g for 10 minutes and resuspended in 1 ml 50 mM Tris-HCl, 40% sucrose, pH 7.5. Periplasmic samples were isolated from cells, through removal of the outer membranes and peptidoglycan by 30 min incubation with 5 mM EDTA and 0.6 mg/ml lysozyme at 37 oC. Periplasmic protein concentration was determined through the use of the Lowry method (Lowry et al., 1951) with a DCTM protein assay kit (BioRad). A standard curve was generated using bovine serum albumin (BSA) for sample comparison. 10-100 µl periplasmic sample was used for determination depending upon protein concentration.

10 µl of 200 mM BV was added to a glass cuvette along with 10 µl of periplasmic sample. The cuvette was then filled to the top with 0.1 M phosphate buffer that had been deoxygenated using N2 before the cuvette was sealed with a perforated stopper. The cuvette was inverted gently to mix the sample after this step and every subsequent addition. Using a Hamilton syringe approximately 10 µl 1% dithionate (dissolved in 1 mM NaOH) was added to reduce the

BV and an intense purple colour was achieved. To initiate the reaction 10 µl 20% TMAO was added, with the colorimetric output being measured at A600. The specific periplasmic TMAO reductase activity was determined using the earlier described equation (Section 7.6.1). TMAO reductase activity is quantified as µmol of oxidised benzyl viologen min-1 mg-1 protein.

7.6.1.2 Hydrogenase-2 (Tat proofreading) activity assay Under anaerobic conditions, the E. coli hydrogenase-2 complex is able to catalyse the oxidation of molecular hydrogen. In this assay this oxidation reaction was coupled with the reduction of

BV, enabling activity to be calculated based on the colour change as BV was increasingly reduced. Strains were grown anaerobically overnight in LB supplemented with 0.5% glycerol,

0.4% TMAO and appropriate antibiotics, with cells then being isolated by centrifugation at at

3000 x g for 10 minutes. Whole cell samples were washed in 10 ml 50 mM Tris-HCl, pH7.5 per 226 gram of cell pellet, and taken for use in the activity assay. 100 µl 250 mM BV was added to a glass cuvette before it being filled to the top with 50 mM Tris-HCl, pH7.5, that had been deoxygenated with H2. The cuvette was stoppered and inverted to mix, before approximately

10 µl 1% dithionate was added. Once the mixture UV absorbance was steady, 10 µl cell sample was added. The colorimetric output was measured at A600 in order to detect BV reduction.

Cellular hydrogenase activity could then be determined using the earlier described equation

(Section 7.6.1), and is quantified as µmol of oxidised benzyl viologen min-1 g-1 cells.

7.6.2 Bacterial two-hybrid system and β-galactosidase assay The bacterial two-hybrid system is based on the reconstitution of an adenylate cyclase signal transduction pathway in E. coli, where the two complementary fragments, namely T18 and

T25, of adenylate cyclase from Bordetella pertussis are fused to two potentially interacting proteins (Karimova et al., 1998). If protein interactions occur, the two fragments are brought into close proximity and cAMP is produced, inducing downstream activation of catabolic genes. Chemically competent MG1655 ΔcyaA::Apra cells were transformed with vectors encoding T18 and T25 fusions to proteins of interest before plating onto MacConkey agar plates containing 1 % maltose as the carbon source. Plates were incubated for 2 days at 30 oC to allow time for colonies to grow. MacConkey contains the dye phenol red, and acts as an indicator for sugar fermentation through a colour change. MG1655 ΔcyaA::Apra cells are unable to ferment lactose or maltose under normal conditions, except in the case of an interaction between the two proteins of interest where expression of genes such as the lac or mal operons is enabled. A positive interaction can be then interpreted on MacConkey by the appearance of red colonies, while pink colonies indicate no interaction is occurring.

For quantification of interactions, a β-galactosidase assay was used. Colonies from MacConkey agar plates were used to inoculate 5 ml LB plus antibiotic cultures and grown at 30 oC overnight, 200 rpm shaking. These cultures were sub cultured and grown under the same 227 conditions until an approximate OD600 of 0.5 was reached. OD600 was recorded and 1 ml samples were taken for assay. 50 ml toluene was added to cell suspensions followed by 30 seconds of vortexing and incubation on ice for 15 minutes. 0.5 ml of the sample was then removed from below the toluene layer and transferred to a fresh tube, with 50 µl then added to

450 µl Z buffer. Following equilibration in a 28 oC waterbath for 10 minutes, 100 µl ONPG (4 mg/ml in Z buffer) was added and ‘start’ time noted. Samples were incubated at 28 oC until development of a yellow colour, with the reaction then being stopped by addition of 250 µl 1 M Na2CO3 and ‘stop’ time noted. A420 of samples was recorded. The specific β-galactosidase could then be calculated using the following formula:

× β-galactosidase activity (Miller units)= × × 3 퐴420 10 푡 푉 푂푂600 = time of reaction (min) = volume of cells (ml) 푡 푉

7.7 Cellular fractionation

7.7.1 Preparation of cytoplasmic/membrane/periplasmic fractions Cells were grown overnight in 500 ml volumes, before being subjected to centrifugation at

5,000 x g for 30 min at 4 oC. Cell pellets were washed with 10 ml 50 mM Tris-HCl, pH 7.5, and again isolated by centrifugation at 18,860 x g for 12 min at 4 oC. The pellet was resuspended in

10 ml g-1 wet weight of cells with 50 mM Tris-HCl pH 7.5 containing 40% sucrose. At this point a 100 µl sample was removed and mixed with 100 µl Laemmli sample buffer and then frozen with liquid nitrogen, for the whole cell sample. EDTA and lysozyme were added to the resuspended cells to a final concentration of 5 mM and 0.6 mg/ml respectively, before the cells being incubated at 37 oC for 30 min. Samples were then centrifuged for 20 min at 25,000 x g at 4 oC. A 100 µl sample of the upper level of the resultant supernatant was taken and flash frozen with Laemmli sample buffer as the periplasmic fraction. The rest of the supernatant was 228 discarded, and the pellet gently resuspended in 50 mM Tris-HCl in the same volume as the last step. Again a 100 µl sample was mixed with Laemmli buffer and frozen for the sphaeroplast fraction. A few flakes of DNase I (Sigma-Aldrich) were added to the cells before lysis by sonication. Cell debris could then be removed by centrifugation twice at 18,860 x g for 12 min at 4 oC. A small amount of the supernatant was then decanted. This encompassed the crude extract fraction, and an aliquot could be taken here and flash frozen with Laemmli buffer. A 1 ml sample was subjected to ultracentrifugation at 278,000 for 30 min at 4 oC. Immediately, 100

µl was removed from the top of the supernatant to give the cytoplasmic fraction; this was flash frozen as previously described. The remainder of the cytoplasm was removed and discarded by pipetting, with the membrane pellet then being resuspended in 800 µl 50 mM Tris-HCl pH7.5.

A 100 µl aliquot of this membrane suspension was also mixed with Laemmli sample buffer and flash frozen.

7.8 Growth curves

7.8.1 Set-up of 96-well plates Growth curves were performed to further analyse results of Biolog Phenotype Microarray screen. S. Typhimurium strains were grown overnight in minimal media or LB. Subcultures were then grown until they reached OD600 of approximately 0.6. 2 µl was then inoculated into

100 µl defined minimal media containing the required carbon source in 96-well plates. Plates

o were incubated for 24 or 48 hours at 37 C with gentle agitation, and A600 readings were taken every 20 minutes. Readings could then be plotted in excel to give growth curves.

229

7.9 Microscopy methods

7.9.1 Confocal microscopy

7.9.1.1 Slide preparation Coverslips for imaging were first prepared with Cell-TakTM (BD Biosciences) in order to allow bacterial cell immobilisation. 1.5 µl Cell-Tak per coverslip was mixed with 25 µl per coverslip of fresh 0.1 M NaHCO3 solution. Immediately, 25 µl of the mixture was spotted onto coverslips and incubated for at least 30 min at room temperature prior to use. Excess liquid was removed from the coverslip surface immediately prior to immobilisation of cells.

Bacterial cells for confocal microscopy were grown aerobically at 37 oC in LB and appropriate antibiotics, before being subcultured and grown until reaching OD600 0.3. Cells were then washed in 1x PBS prior to spotting of 25 µl onto surface of coverslip, over the adsorbed Cell-

Tak spot. Coverslips were covered with foil to protect from the light and incubated at room temperature for at least 30 min, with unbound cells being washed off with 1x PBS. For live cell imaging, coverslips were placed onto slides with a drop of 1x PBS between the two surfaces to hold them together by surface tension.

7.9.1.2 Imaging Cells were imaged using a Leica SP5 SMD (Single molecule detection) confocal laser scanning microscope at Heriot-Watt University, under the supervision of Professor Rory Duncan and Dr

Colin Rickman. Image reconstruction and generation of residual maps showing colocalisation was performed using ImageJ software.

7.9.2 OMX microscopy

7.9.2.1 Slide preparation Cell cultures were grown aerobically overnight at 37 oC in LB plus appropriate antibiotics. Cells were subcultured and grown to mid-logarithmic phase. Those cells in which Tat substrate expression was to be induced were grown for 2 ½ hours before IPTG was added to a final 230 concentration of 2 mM. Cultures were then allowed to grow for a further 2 ½ hours prior to slide preparation. Bacterial cells were washed and resuspended in mili-Q ultrapure water, before being diluted to OD600 0.3.

Number 1.5 coverslips were used for imaging with the OMX. Cells were mounted using poly- lysine. This required spotting of 5 µl poly-lysine onto the coverslip before allowing it to air dry.

Washed cells were then spotted onto the poly-lysine and also allowed to air dry. Fixation was performed using ice-cold 100% methanol, which was spotted onto the coverslip for 5 min before rehydration with three washes with TBS/Tween. Cells were permeabilised by incubation with 10 µl of lysozyme (10µg/ml) for 10 min, followed by washing with 1x PBS. A 1 hour incubation with GFP-booster (1:200 in PBS + 2% bovine serum albumin) (Chromotek) was performed to enhance fluorescence. Coverslips were washed thoroughly with 1x PBS, before being mounted onto glass slides using 10 µl Pro-Long Gold (Life Technologies). After overnight curing at room temperature, slides were sealed using clear nail varnish.

7.9.2.2 Imaging Cells were imaged using 3D structured illumination microscopy (3D SIM) with the OMX microscope in the College of Life Sciences at the University of Dundee, with the assistance of

OMX Scientific Officer Dr Markus Posch. Images were reconstructed and analysed using

SoftWorx software.

231

8 Appendix

232

8.1 Biolog plates

Maps of respiratory sources in individual Biolog plates. Information obtained from http://www.biolog.com/products/?product=Phenotype%20MicroArrays%20for%20Microbial%

20Cells&view=Product%20Literature.

233

234

235

236

9 Bibliography 237

Agar, J.N., C. Krebs, J. Frazzon, B.H. Huynh, D.R. Dean & M.K. Johnson, (2000) IscU as a Scaffold for Iron−Sulfur Cluster Biosynthesis: Sequential Assembly of [2Fe-2S] and [4Fe-4S] Clusters in IscU†. Biochemistry 39: 7856-7862. Alami, M., I. Lüke, S. Deitermann, G. Eisner, H.-G. Koch, J. Brunner & M. Müller, (2003) Differential interactions between a twin-arginine signal peptide and its translocase in Escherichia coli. Mol. Cell 12: 937-946. Alcock, F., M.A.B. Baker, N.P. Greene, T. Palmer, M.I. Wallace & B.C. Berks, (2013) Live cell imaging shows reversible assembly of the TatA component of the twin-arginine protein transport system. Proc. Natl. Acad. Sci. U. S. A. 110: E3650-E3659. Alvarez, A.F., C. Rodriguez & D. Georgellis, (2013) Ubiquinone and menaquinone elctron carriers represent the yin and yang in the redox regulation of the ArcB sensor kinase. J. Bacteriol. 195: 3054-3061. Anderson, L.A., E. McNairn, T. Leubke, R.N. Pau & D.H. Boxer, (2000) ModE-Dependent Molybdate Regulation of the Molybdenum Cofactor Operon moa in Escherichia coli. J. Bacteriol. 182: 7035-7043. Ansaldi, M., C. Bordi, M. Lepelletier & V. Méjean, (1999) TorC apocytochrome negatively autoregulates the trimethylamine N-oxide (TMAO) reductase operon in Escherichia coli. Mol. Microbiol. 33: 284-295. Ansaldi, M., C. Castelli-Jourlin, M. Lepelletier, L. Théraulaz & V. Méjean, (2001) Rapid dephosphorylation of the TorR response regulator by the TorS unorthodox sensor in Escherichia coli. J. Bacteriol. 183: 2691-2695. Ansaldi, M., L. Théraulaz & V. Méjean, (2004) TorI, a response regulator inhibitor of phage origin in Escherichia coli. Proc. Natl. Acad. Sci. U. S. A. 101: 9423-9428. Arkowitz, R.A., J.C. Joly & W. Wickner, (1993) Translocation can drive the unfolding of a preprotein domain. EMBO J. 12: 243-253. Ayala-Castro, C., A. Saini & F.W. Outten, (2008) Fe-S cluster assembly pathways in bacteria. Microbiol. Mol. Biol. Rev. 72: 112-125. Bageshwar, U.K. & S.M. Musser, (2007) Two electrical potential-dependent steps are required for transport by the Esherichia coli Tat machinery. J. Cell Biol. 179: 27-99. Baraquet, C., L. Théraulaz, M. Guiral, D. Lafitte, V. Méjean & C. Castelli-Jourlin, (2006) TorT, a member of a new periplasmic binding protein family, triggers induction of the Tor respiratory system upon trimethylamine N-oxide electron-acceptor binding in Escherichia coli. J. Biol. Chem. 281: 38189-38199. Barkocy-Gallagher, G.A. & P.J. Bassfor Jr., (1992) Synthesis of precursor maltose-binding protein with proline in the +1 position of the cleavage site interferes with the activity of Escherichia coli signal peptidase I in vivo. J. Biol. Chem. 267: 1231-1238. Barnett, J.P., R.T. Eijlander, O.P. Kuipers & C. Robinson, (2008) A minimal Tat system from a Gram-positive organism: a bifunctional TatA subunit participates in discrete TatAC and TatA complexes. J. Biol. Chem. 283: 2534-2542. Barrett, E.L. & H.S. Kwan, (1985) Bacterial reduction of trimethylamine oxide. Annu. Rev. Microbiol. 39: 131-149. Baud, C., S. Karamanou, G. Sianidis, E. Vrontou, A.S. Politou & A. Economou, (2002) Allosteric communication between signal peptides and the SecA protein DEAD motor ATPase domain. J. Biol. Chem. 277: 13724-13731. Bearson, S.M.D., J.A. Albrecht & R.P. Gunsalus, (2002) Oxygen and nitrate-dependent regulation of dmsABC operon expression in Escherichia coli: sites for Fnr and NarL protein interactions. BMC Microbiol. 2. Behrendt, J., K. Standar, U. Lindenstrauß & T. Brüser, (2004) Topological studies on the twin- arginine translocase component TatC. FEMS Microbiol. Lett. 234: 303-308. Bekker, M., S. Alexeeva, W. Laan, G. Sawers, J.T. de Mattos & K.J. Hellingwerf, (2010) The ArcBA two-component system of Escherichia coli is regulated by the redox state of both the ubiquinone and the menaquinone pool. J. Bacteriol. 192: 746-754. 238

Bendtsen, J.D., H. Nielsen, D. Widdick, T. Palmer & S. Brunak, (2005) Prediction of twin-arginine signal peptides. BMC Bioinformatics 6: 167-174. Berg, B.v.d., W.M. Clemons, I. Collinson, Y. Modis, E. Hartmann, S.C. Harrison & T.A. Rapoport, (2004) X-ray structure of a protein-conducting channel. Nature 427: 36-44. Berks, B.C., (1996) A common export pathway for proteins binding complex redox cofactors? Mol. Microbiol. 22: 393-404. Berks, B.C., S.M. Lea & P.J. Stansfeld, (2014) Structural biology of Tat protein transport. Curr. Opin. Struct. Biol. 27: 32-37. Berks, B.C., T. Palmer & F. Sargent, (2005) Protein targeting by the bacterial twin-arginine translocation (Tat) pathway. Curr. Opin. Microbiol. 8: 174-181. Berks, B.C., D.J. Richardson, A. Reilly, A.C. Willis & S.J. Ferguson, (1995) The napEDABC gene cluster encoding the periplasmic nitrate reductase system of Thiosphaera pantotropha. Biochem. J. 309: 983-992. Berks, B.C., F. Sargent, E. De Leeuw, A.P. Hinsley, N.R. Stanley, R.L. Jack, G. Buchanan & T. Palmer, (2000) A novel protein transport system involved in the biogenesis of bacterial electron transfer chains. Biochim. Biophys. Acta 1459: 325-330. Berry, R.M., (1993) Torque and switching in the bacterial flagellar motor. An electrostatic model. Biophys. J. 64: 961-973. Bertero, M.G., R.A. Rothery, N. Boroumand, M. Palek, F. Blasco, N. Ginet, J.H. Weiner & N.C.J. Strynadka, (2005) Structural and biochemical characterization of a quinol binding site of Escherichia coli nitrate reductase A. J. Biol. Chem. 280: 14836-14843. Bieker-Brady, K. & T.J. Silhavy, (1992) Suppressor analysis suggests a multistep, cyclic mechanism for protein secretion in Escherichia coli. EMBO J. 11: 3165-3174. Bieker, K.L. & T.J. Silhavy, (1990) PrlA (SecY) and PrlG (SecE) interact directly and function sequentially during protein translocation in E. coli. Cell 61: 833-842. Bilous, P.T. & J.H. Weiner, (1985a) Demethyl sulfoxide reductase activity by anaerobically grown Escherichia coli HB101. J. Bacteriol. 162: 1151-1155. Bilous, P.T. & J.H. Weiner, (1985b) Dimethyl sulfoxide reductase activity by anaerobically grown Escherichia coli HB101. J. Bacteriol. 162: 1151-1155. Bilous, P.T. & J.H. Weiner, (1988) Molecular cloning and expression of the Escherichia coli dimethyl sulfoxide reductase operon. J. Bacteriol. 170: 1511-1518. Birnboim, H.C. & J. Doly, (1979) A rapid alkaline extraction procedure for screening recombinant plasmid DNA. Nucleic Acids Res. 7: 1513-1523. Blasco, F., J.-P. Dos Santos, A. Magalon, C. Frixon, B. Guigliarelli, C.-L. Santini & G. Giordano, (1998) NarJ is a specific chaperone required for molybdenum cofactor assembly in nitrate reductase A of Escherichia coli. Mol. Microbiol. 28: 435-447. Blasco, F., C. Iobbi, J. Ratouchniak, V. Bonnefoy & M. Chippaux, (1990) Nitrate reductases of Escherichia coli: Sequence of the second nitrate reductase and comparison with that encoded by the narGHJI operon. Molec. Gen. Genet. 222: 104-111. Blattner, F.R., G. Plunkett, C.A. Bloch, N.T. Perna, V. Burland, M. Riley, J. Collado-Vides, J.D. Glasner, C.K. Rode, G.F. Mayhew, J. Gregor, N.W. Davis, H.A. Kirkpatrick, M.A. Goeden, D.J. Rose, B. Mau & Y. Shao, (1997) The Complete Genome Sequence of Escherichia coli K-12. Science 277: 1453-1462. Blaudeck, N., P. Kreutzenbeck, R. Freudl & G.A. Sprenger, (2003) Genetic analysis of pathway specificity during posttranslational protein translocation across the Escherichia coli plasma membrane. J. Bacteriol. 185: 2811-2819. Bogsch, E., S. Brink & C. Robinson, (1997) Pathway specificity for a ΔpH-dependent precursor thylakoid lumen protein is governed by a 'Sec-avoidance' motif in the transfer peptide and a 'Sec-incompatible' mature protein. EMBO J. 16: 3851-3859. Bogsch, E.G., F. Sargent, N.R. Stanley, B.C. Berks, C. Robinson & T. Palmer, (1998) An Essential Component of a Novel Bacterial Protein Export System with Homologues in Plastids and Mitochondria. J. Biol. Chem. 273: 18003-18006. 239

Bolhuis, A., J.E. Mathers, J.D. Thomas, C.M. Barrett & C. Robinson, (2001) TatB and TatC form a functional and structural unit of the twin-arginine translocase from Escherichia coli. J. Biol. Chem. 276: 20213-20219. Bonnefoy, V. & J.A. DeMoss, (1994) Nitrate reductases in Escherichia coli. Antonie Van Leeuwenhoek 66: 47-56. Bonomi, F., S. Iametti, A. Morleo, D. Ta & L.E. Vickery, (2008) Studies on the Mechanism of Catalysis of Iron−Sulfur Cluster Transfer from IscU[2Fe2S] by HscA/HscB Chaperones†. Biochemistry 47: 12795-12801. Bordignon, E., M. Grote & E. Schneider, (2010) The maltose ATP-binding cassette transporter in the 21st century – towards a structural dynamic perspective on its mode of action. Mol. Microbiol. 77: 1354-1366. Bowman, L., T. Palmer & F. Sargent, (2013) A regulatory domain controls the transport activity of a twin-arginine signal peptide. FEBS Lett. 587: 3365-3370. Brondijk, T.H.C., A. Nilavongse, N. Filenko, D.J. Richardson & J.A. Cole, (2004) NapGH components of the periplasmic nitrate reductase of Escherichia coli K-12: location, topology and physiological roles in quinol oxidation and redox balancing. Biochem. J. 379: 47-55. Brundage, L., J.P. Hendrick, E. Schiebel, A.J.M. Driessen & W. Wickner, (1990) The purified E. coli integral membrane protein SecY E is sufficient for reconstitution of SecA- dependent precursor protein translocation. Cell 62: 649-657. Buchanan, G., E.d. Leeuw, N.R. Stanley, M. Wexler, B.C. Berks, F. Sargent & T. Palmer, (2002) Funtional complexity of the twin-arginine translocase TatC component revealed by site-directed mutagenesis. Mol. Microbiol. 43: 1457-1470. Buchanan, G., J. Maillard, S.B. Nabuurs, D.J. Richardson, T. Palmer & F. Sargent, (2008) Features of a twin-arginine signal peptide required for recognition by a Tat proofreading chaperone. FEBS Lett. 582: 3979-3984. Bugg, T.D.H., G.D. Wright, S. Dutka-Malen, M. Arthur, P. Courvalin & C.T. Walsh, (1991) Molecular basis for vancomycin resistance in Enterococcus faecium BM4147: biosynthesis of a depsipeptide peptidoglycan precursor by vancomycin resistance proteins VanH and VanA. Biochemistry 30: 10408-10415. Butler, C.S., C.M. Debieux, E.J. Dridge, P. Splatt & M. Wright, (2012) Biomineralization of selenium by the selenate-respiring bacterium Thauera selenatis. Biochem. Soc. Trans. 40: 1239-1243. Caldelari, I., T. Palmer & F. Sargent, (2008) Escherichia coli tat mutant strains are able to transport maltose in the absence of an active malE gene. Arch. Microbiol. 189: 597- 604. Carrington, J.C. & W.G. Dougherty, (1988) A viral cleavage site cassette: identification of amino acid sequences required for tobacco etch virus polyprotein processing. Proceedings of the National Academy of Sciences 85: 3391-3395. Casadaban, M., J. & S. Cohen, N. , (1980) Analysis of gene control signals by DNA fusion and cloning inEscherichia coli. J. Mol. Biol. 138: 179-207. Chan, C.S., E.F. Haney, H.J. Vogel & R.J. Turner, (2011) Towards understanding the Tat translocation mechanism through structural and biophysical studies of the amphipathic region of TatA from Escherichia coli. Biochim. Biophys. Acta 1808: 2289- 2296. Chan, C.S., T.M.L. Winstone, L. Chang, C.M. Stevens, M.L. Workentine, H. Li, Y. Wei, M.J. Ondrechen, M. Paetzel & R.J. Turner, (2008) Identification of residues in DmsD for twin-arginine leader peptide binding, defined through random and bioinformatics- directed mutagenesis. Biochemistry 47: 2749-2759. Chan, C.S., M.R. Zlomislic, D.P. Tieleman & R.J. Turner, (2007) The TatA subunit of Escherichia coli twin-arginine translocase has an N-in topology. Biochemistry 46: 7396-7404. 240

Chan, M.K., S. Mukund, A. Kletzin, M.W.W. Adams & D.C. Rees, (1995) Structure of a hyperthermophilic tungstopterin enzyme, aldehyde ferredoxin oxidoreductase. Science 267: 1463-1469. Chanal, A., C.L. Santini & L.F. Wu, (1998) Potential receptor function of three homologous components, TatA, TatB and TatE, of the twin-arginine signal sequence-dependent metalloenzyme translocation pathway in Escherichia coli. Mol. Microbiol. 30: 674-676. Chandramouli, K. & M.K. Johnson, (2006) HscA and HscB stimulate [2Fe-2S] cluster transfer from IscU to apo-ferredoxin in an ATP-dependent reaction. Biochemistry (Wash.) 45: 11087-11095. Chang, C.-K., K.-H. Teng, S.-W. Lin, T.-H. Chang & P.-H. Liang, (2013) Control Activity of Yeast Geranylgeranyl Diphosphate Synthase from Dimer Interface through H‑Bonds and Hydrophobic Interaction. Biochemistry 52: 2783-2792. Chang, L., L.I.C. Wei, J.P. Audia, R.A. Morton & H.E. Schellhorn, (1999) Expression of the Escherichia coli NRZ nitrate reductase is highly growth phase dependent and is controlled by RpoS, the alternative vegetative sigma factor. Mol. Microbiol. 34: 756- 766. Chen, M., J.C. Samuelson, F. Jiang, M. Muller, A. Kuhn & R.E. Dalbey, (2002) Direct interaction of YidC with the Sec-independent Pf3 coat protein during its membrane protein insertion. J. Biol. Chem. 277: 7670-7675. Chen, S., M.L. Oldham, A.L. Davidson & J. Chen, (2013) Carbon catabolite repression of the maltose transporter revealed by X-ray crystallography. Nature 499: 364-368. Chitra, N.R., A. Brown, I. & A. Rutenberg, D. , (2014) Protein translocation without specific quality control in a computational model of the Tat system. Phys. Biol. 11: 056005. Clark, M.A. & E.L. Barrett, (1987) The phs gene and hydrogen sulfide production by Salmonella typhimurium. J. Bacteriol. 169: 2391-2397. Cline, K. & H. Mori, (2001) Thylakoid ΔpH-dependent precursor proteins bind to a cpTatC- Hcf106 complex before Tha4-dependent transport. J. Cell Biol. 154: 719-729. Compan, I. & D. Touati, (1994) Anaerobic activation of arcA transcription in Escherichia coli: roles of Fnr and ArcA. Mol. Microbiol. 11: 955-964. Connolly, T. & R. Gilmore, (1989) The signal recognition particle receptor mediates the GTP- dependent displacement of SRP from the signal sequence of the nascent polypeptide. Cell 57: 599-610. Constantinidou, C., J.L. Hobman, G. Lesley, M.D. Patel, C.W. Penn, J.A. Cole & T.W. Overton, (2006) A reassessment of the FNR regulon and transcriptomic analysis of the effects of nitrate, nitrite, NarXL, and NarQP as Escherichia coli K12 adapts from aerobic to anaerobic growth. J. Biol. Chem. 281: 4802-4815. Coulthurst, S.J., A. Dawson, W.N. Hunter & F. Sargent, (2012) Conserved signal peptide recognition systems across the prokaryotic domains. Biochemistry 51: 1678-1686. Crack, J.C., A.J. Jervis, A. Gaskell, G.F. White, J. Green, A.J. Thomson & N.E. Le Brun, (2008) Signal perception by FNR: the role of the iron-sulfur cluster. Biochem. Soc. Trans. 36: 1144-1148. Crack, J.C., M.R. Stapleton, J. Green, A.J. Thomson & N.E. Le Brun, (2014) Influence of association state and DNA binding on the O2-reactivity of [4Fe-4S] fumarate and nitrate reduction (FNR) regulator. Biochem. J. 463: 83-92. Craig, M., A.Y. Sadik, Y.A. Golubeva, A. Tidhar & J.M. Slauch, (2013) Twin-arginine translocation system (tat) mutants of Salmonella are attenuated due to envelope defects, not respiratory defects. Mol. Microbiol.: n/a-n/a. Crane, J.M., C. Mao, A.A. Lilly, V.F. Smith, Y. Suo, W.L. Hubbell & L.L. Randall, (2005) Mapping of the Docking of SecA onto the Chaperone SecB by Site-directed Spin Labeling: Insight into the Mechanism of Ligand Transfer During Protein Export. J. Mol. Biol. 353: 295- 307. Cristóbal, S., J.-W. de Gier, H. Neielsen & G. von Heijne, (1999a) Competition between Sec- and TAT-dependent protein translocation in Escherichia coli. EMBO J. 18: 2982-2990. 241

Cristóbal, S., J.-W. de Gier, H. Nielsen & G. von Heijne, (1999b) Competition between Sec- and TAT-dependent protein in Escherichia coli. EMBO J. 18: 2982-2990. Cunningham, K. & W. Wickner, (1989) Specific recognition of the leader region of precursor proteins is required for the activation of translocation ATPase of Escherichia coli. Proc. Natl. Acad. Sci. U. S. A. 86: 8630-8634. Czjzek, M., J.-P. Dos Santos, J. Pommier, G. Giordano, V. Méjean & R. Haser, (1998) Crystal structure of oxidized trimethylamine N-oxide reductase from Shewanella massilia at 2.5 å resolution. J. Mol. Biol. 284: 435-447. Dabney-Smith, C., H. Mori & K. Cline, (2006) Oligomers of Tha4 organize at the thylakoid Tat translocase during protein transport. J. Biol. Chem. 281: 5476-5483. Dalbey, R.E. & A. Kuhn, (2012) Protein Traffic in Gram-negative bacteria – how exported and secreted proteins find their way. FEMS Microbiol. Rev. 36: 1023-1045. Danese, P.N. & T.J. Silhavy, (1998) Targeting and assembly of peirplasmic and outer-membrane proteins in Escherichia coli. Annu. Rev. Genet. 32: 59-94. Daniels, J.N., M.M. Wuebbens, K.V. Rajagopalan & H. Schindelin, (2007) Crystal Structure of a Molybdopterin Synthase−Precursor Z Complex: Insight into Its Sulfur Transfer Mechanism and Its Role in Molybdenum Cofactor Deficiency†,‡. Biochemistry 47: 615- 626. Daus, M.L., S. Berendt, S. Wuttge & E. Schneider, (2007) Maltose binding protein (MalE) interacts with periplasmic loops P2 and P1 respectively of the MalFG subunits of the maltose ATP binding cassette transporter (MalFGK2) from Escherichia coli/Salmonella during the transport cycle. Mol. Microbiol. 66: 1107-1122. Davidson, A.L., H.A. Shuman & H. Nikaido, (1992) Mechanism of maltose transport in Escherichia coli: transmembrane signaling by periplasmic binding proteins. Proc. Natl. Acad. Sci. U. S. A. 89: 2360-2364. de Bruyn, G., (2000) Infectious disease: diarrhea. West. J. Med. 176: 409-412. De Gier, J.-W.L., Q.A. Valent, G. Von Heijne & J. Luirink, (1997) The E. coli SRP: preferences of a targeting factor. FEBS Lett. 408: 1-4. De Lacey, A.L., V.M. Fernández, M. Rousset & R. Cammack, (2007) Activation and Inactivation of Hydrogenase Function and the Catalytic Cycle: Spectroelectrochemical Studies. Chem. Rev. 107: 4304-4330. De Leeuw, E., I. Porcelli, F. Sargent, T. Palmer & B.C. Berks, (2001) Membrane interactions and self-association of the TatA and TatB components of the twin-arginine translocation pathway. FEBS Lett. 506: 143-148. Dekker, C., B.d. Kruijff & P. Gros, (2003) Crystal structure of SecB from Escherichia coli. J. Struct. Biol. 144: 313-319. DeLisa, M.P., P. Lee, T. Palmer & G. Georgiou, (2004) Phage shock protein PspA of Escherichia coli relieves saturation of protein export via the Tat pathway. J. Bacteriol. 186: 366- 373. DeLisa, M.P., D. Tullman & G. Georgiou, (2003) 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. Dilks, K., R.W. Rose, E. Hartmann & M. Pohlschröder, (2003) Prokaryotic utilization of the Twin- Arginine Translocation pathway: a genomic survey. J. Bacteriol. 185: 1478-1483. Ding, H. & R.J. Clark, (2004) Characterization of iron binding in IscA, an ancient iron-sulphur cluster assembly protein. Biochem. J. 379: 433-440. Ding, H., R.J. Clark & B. Ding, (2004) IscA mediates iron delivery for assembly of iron-sulfur clusters in IscU under the limited accessible free iron conditions. J. Biol. Chem. 279: 37499-37504. Dow, J.M., F. Gabel, F. Sargent & T. Palmer, (2013) Characterization of a pre-export enzyme- chaperone complex on the twin-arginine transport pathway. Biochem. J. 452: 57-66. Dowhan, W., (1997) Molecular basis for membrane phospholipid diversity: Why Are There So Many Lipids? Annu. Rev. Biochem. 66: 199-232. 242

Dridge, E.J., C.A. Watts, B.J.N. Jepson, K. Line, J.M. Santini, D.J. Richardson & C.S. Butler, (2007) Investigation of the redox centres of periplasmic selenate reductase from Thauera selenatis by EPR spectroscopy. Biochem. J. 408: 19-28. Driessen, A.J.M., (1992) Precursor protein translocation by the Escherichia coli translocase is directed by the protonmotive force. EMBO J. 11: 847-853. du Plessis, D.J.F., N. Nouwen & A.J.M. Driessen, (2011) The Sec translocase. Biochim. Biophys. Acta 1808: 851-865. Dubini, A. & F. Sargent, (2003a) Assembly of Tat-dependent [NiFe] hydrogenases: identification of precursor-binding accessory proteins. FEBS Lett. 549: 141-146. Dubini, A. & F. Sargent, (2003b) Assembly of Tat-dependent [NiFe] hydrogenases:identification of precursor-binding accessory proteins. FEBS Lett. 549: 141-146. Duncan, K., J. Van Heijenoort & C.T. Walsh, (1990) Purification and characterization of the D- alanyl-D-alanine-adding enzyme from Escherichia coli. Biochemistry 29: 2379-2386. Economou, A., J.A. Pogliano, J. Beckwith, D.B. Oliver & W. Wickner, (1995) SecA membrane cycling at SecYEG is driven by distinct ATP binding and hydrolysis events and is regulated by SecD and SecF. Cell 83: 1171-1181. Economou, A. & W. Wickner, (1994) SecA promotes preprotein translocation by undergoing ATP-driven cycles of membrane insertion and deinsertion. Cell 78: 835-843. Facey, S.J., S.A. Neugebauer, S. Krauss & A. Kuhn, (2007) The Mechanosensitive Channel Protein MscL Is Targeted by the SRP to The Novel YidC Membrane Insertion Pathway of Escherichia coli. J. Mol. Biol. 365: 995-1004. Feilmeier, B.J., G. Iseminger, D. Schroeder, H. Webber & G.J. Phillips, (2000) Green fluorescent protein functions as a reporter for protein localization in Escherichia coli. J. Bacteriol. 182: 4068-4076. Fekkes, P., C. van der Does & A.J.M. Driessen, (1997) The molecular chaperone SecB is released from the carboxy-terminus of SecA during initiation of precursor protein translocation. EMBO J. 16: 6105-6113. Fisher, A.C. & M.P. DeLisa, (2004) A little help from my friends: quality control of presecretory proteins in bacteria. J. Bacteriol. 186: 7467-7473. Fontecave, M. & S. Ollagnier-de-Choudens, (2008) Iron–sulfur cluster biosynthesis in bacteria: Mechanisms of cluster assembly and transfer. Arch. Biochem. Biophys. 474: 226-237. Freudl, R., (2013) Leaving home ain't easy: protein export systems in Gram-positive bacteria. Res. Microbiol. 164: 664-674. Fröbel, J., P. Rose, F. Lausberg, A.-S. Blummel, R. Freudl & M. Müller, (2012) Transmembrane insertion of twin-arginine signal peptides is driven by TatC and regulated by TatB. Nature Communications. Genest, O., V. Méjean & C. Iobbi-Nivol, (2009) Multiple roles of TorD-like chaperones in the biogenesis of molybdoenzymes. FEMS Microbiol. Lett. 297: 1-9. Genest, O., M. Neumann, F. Seduk, W. Stöcklein, V. Méjean, S. Leimkühler & C. Iobbi-Nivol, (2008) Dedicated Metallochaperone Connects Apoenzyme and Molybdenum Cofactor Biosynthesis Components. J. Biol. Chem. 283: 21433-21440. Gérard, F. & K. Cline, (2006) Efficient Twin Arginine Translocation (Tat) pathway transport of a precursor protein covalently anchored to its initial cpTatC binding site. J. Biol. Chem. 281: 6130-6135. Gerrard, T.L., J.N. Telford & H.H. Williams, (1974) Detection of selenium deposits in Escherichia coli by electron microscopy. J. Bacteriol. 119: 1057-1060. Gohlke, U., L. Pullan, C.A. McDevitt, I. Porcelli, E. de Leeuw, T. Palmer, H.R. Saibil & B.C. Berks, (2005) 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. Gon, S., M.-T. Giudici-Orticoni & V. Méjean, (2001) Electron transfer and binding of the c-type cytochrome TorC to the trimethylamine N-oxide reductase in Escherichia coli. J. Biol. Chem. 276: 11545-11551. 243

Gon, S., J.-C. Patte, V. Méjean & C. Iobbi-Nivol, (2000) The torYZ (yecK bisZ) operon encodes a third respiratory trimethylamine N-Oxide reductase in Escherichia coli. J. Bacteriol. 182: 5779-5786. Gorbach, S.L., (1996) Microbiology of the gastrointestinal tract. In: Medical Microbiology. 4th edition. S. Baron (ed). Galveston (TX): University of Texas Medical Branch at Galveston, pp. Gouffi, K., F. Gerard, C.-L. Santini & L.-F. Wu, (2004) Dual topology of the Escherichia coli TatA protein. J. Biol. Chem. 279: 11608-11615. Grahl, S., J. Maillard, C.A.E.M. Spronk, G.W. Vuister & F. Sargent, (2012) Overlapping transport and chaperone-binding functions within a bacterial twin-arginine signal peptide. Mol. Microbiol. 83: 1254-1267. Graubner, W., A. Schierhorn & T. Brüser, (2007) DnaK plays a pivotal role in Tat targeting of CueO and functions beside SlyD as a general Tat signal binding chaperone. J. Biol. Chem. 282: 7116-7124. Greene, N.P., I. Porcelli, G. Buchanan, M.G. Hicks, S.M. Schermann, T. Palmer & B.C. Berks, (2007) Cysteine scanning mutagenesis and disulfide mapping studies of the TatA component of the bacterial twin arginine translocase. J. Biol. Chem. 282: 23937-23945. Gustafsson, M.G.L., (2000) Surpassing the lateral resolution limit by a factor of two using structured illumination microscopy. J. Microsc. 198: 82-87. Gutzke, G., B. Fischer, R.R. Mendel & G. Schwarz, (2001) Thiocarboxylation of molybdopterin synthase provides evidence for the mechanism of dithiolene formation in metal- binding pterins. J. Biol. Chem. 276: 36268-36274. Guymer, D., J. Maillard, M.F. Agacan, C.A. Brearly & F. Sargent, (2010) Intrinsic GTPase activity of a bacterial twin-arginine translocation proofreading chaperone induced by domain swapping. FEBS J. 277. Guymer, D., J. Maillard & F. Sargent, (2009) A genetic analysis of in vivo selenate reduction by Salmonella enterica serovar Typhimurium LT2 and Escherichia coli K12. Arch. Microbiol. 191: 519-528. Hall, J.A., A.K. Ganesan, J. Chen & H. Nikaido, (1997) Two modes of ligand binding in maltose- binding protein of Escherichia coli. J. Biol. Chem. 272: 17615-17622. Hamilton, C.M., M. Aldea, B.K. Washburn, P. Babltzke & S.R. Kushner, (1989) New method for generating deletions and gene replacements in Escherichia coli. J. Bacteriol. 171: 4617- 4622. Hanahan, D., (1983) Studies on transformation of Escherichia coli with plasmids. J. Mol. Biol. 166: 557-580. Hänzelmann, P., H.L. Hernández, C. Menzel, R. García-Serres, B.H. Huynh, M.K. Johnson, R.R. Mendel & H. Schindlen, (2004) Characterization of MOCS1A, an oxygen-sensitive iron- sulfur protein involved in human molybdenum cofactor biosynthesis. J. Biol. Chem. 279: 34721-34732. Hardy, S.J.S. & L.L. Randall, (1991) A kinetic partitioning model of selective binding of nonnative proteins by the bacterial chaperone SecB. Science 251: 439-443. Hartl, F.-U., S. Lecker, E. Schiebel, J.P. Hendrick & W. Wickner, (1990) The binding cascade of SecB to SecA to SecY E mediates preprotein targeting to the E. coli plasma membrane. Cell 63: 269-279. Hatzixanthis, K., T.A. Clarke, A. Oubrie, D.J. Richardson & R.J. Turner, (2005) Signal peptide- chaperone interactions on the twin-arginine protein transport pathway. Proc. Natl. Acad. Sci. U. S. A. 102: 8460-8465. Hensel, M., A.P. Hinsley, T. Nikolaus, G. Sawers & B.C. Berks, (1999) The genetic basis of tetrathionate respiration in Salmonella typhimurium. Mol. Microbiol. 32: 275-287. Hicks, M.G., E. de Leeuw, I. Porcelli, G. Buchanan, B.C. Berks & T. Palmer, (2003) The Escherichia coli twin-arginine translocase: conserved residues of TatA and TatB family components involved in protein transport. FEBS Lett. 539: 61-67. Hille, R., (1996) The mononuclear molybdenum enzymes. Chem. Rev. 96: 2757-2816. 244

Hille, R., (2002) Molybdenum and tungsten in biology. Trends Biochem. Sci. 27: 360-367. Hinsley, A.P. & B.C. Berks, (2002) Specificity of respiratory pathways involved in the reduction of sulfur compounds by Salmonella enterica. Microbiology 148: 3631-3638. Hinsley, A.P., N.R. Stanley, T. Palmer & B.C. Berks, (2001) A naturally occurring bacterial Tat signal peptide lacking one of the 'invariant' arginine residues of the consensus targeting motif. FEBS Lett. 497: 45-49. Hiraoka, S., H. Matsuzaki & I. Shibuya, (1993) Active increase in cardiolipin synthesis in the stationary growth phase and its physiological significance in Escherichia coli. FEBS Lett. 336: 221-224. Hu, Y., E. Zhao, H. Li, B. Xia & C. Jin, (2010) Solution NMR Structure of the TatA Component of the Twin-Arginine Protein Transport System from Gram-Positive Bacterium Bacillus subtilis. J. Am. Chem. Soc. 132: 15942-15944. Ilbert, M., V. Méjean, M.-T. Giudici-Orticoni, J.-P. Samama & C. Iobbi-Nivol, (2003) Involvement of a Mate Chaperone (TorD) in the Maturation Pathway of Molybdoenzyme TorA. J. Biol. Chem. 278: 28787-28792. Ilbert, M., V. Méjean & C. Iobbi-Nivol, (2004) Functional and structural analysis of members of the TorD family, a large chaperone family dedicated to molybdoproteins. Microbiology 150. Iobbi-Nivol, C., T. Palmer, P. Whitty, E. McNairn & D.H. Boxer, (1995) The mob locus of Escherichia coli K12 required for molybdenum cofactor biosynthesis is expressed at very low levels. Microbiology 141: 1663-1671. Iobbi, C., C.-L. Santini, V. Bonnefoy & G. Giordano, (1987) Biochemical and immunological evidence for a second nitrate reductase in Escherichia coli K12. Eur. J. Biochem. 168: 451-459. Iuchi, S., V. Chepuri, H.-A. Fu, R.B. Gennis & E.C.C. Lin, (1990) Requirement for terminal cytochromes in generation of the aerobic signal for the arc regulatory system in Escherichia coli: study utilizing deletions and lac fusions of cyo and cyd. J. Bacteriol. 172: 6020-6025. Izard, J.W. & D.A. Kendall, (1994) Signal peptides: exquisitely designed transport promoters. Mol. Microbiol. 13: 765-773. Ize, B., S.J. Coulthurst, K. Hatzixanthis, I. Caldelari, G. Buchanan, E.C. Barclay, D.J. Richardson, T. Palmer & F. Sargent, (2009a) Remnant signal peptides on non-exported enzymes: implications for the evolution of prokaryotic respiratory chains. Microbiology 155: 3992-4004. Ize, B., S.J. Coulthurst, K. Hatzixanthis, I. Caldelari, G. Buchanan, E.C. Barclay, D.J. Richardson, T. Palmer & F. Sargent, (2009b) Remnant signal peptides on non-exported enzymes: implications for the evolution of prokaryotic respiratory chains. Microbiology 155: 3992-4004. Ize, B., I. Porcelli, S. Lucchini, J.C. Hinton, B.C. Berks & T. Palmer, (2004) Novel phenotypes of Escherichia coli tat mutants revealed by global gene expression and phenotypic analysis. J. Biol. Chem. 279: 47543-47554. Ize, B., N.R. Stanley, G. Buchanan & T. Palmer, (2003) Role of the Escherichia coli Tat pathway in outer membrane integrity. Mol. Microbiol. 48: 1183-1193. Jack, R.L., G. Buchanan, A. Dunbini, K. Hatzixanthis, T. Palmer & F. Sargent, (2004) Coordinating assembly and export of complex bacterial proteins. EMBO (European Molecular Biology Organization) Journal 23: 3962-3972. Jack, R.L., F. Sargent, B.C. Berks, G. Sawers & T. Palmer, (2001) Constitutive expression of Escherichia coli tat genes indicates an important role for the Twin-Arginine Translocase during aerobic and anaerobic growth. J. Bacteriol. 183: 1801-1804. James, M.J., S.J. Coulthurst, T. Palmer & F. Sargent, (2013) Signal peptide etiquette during assembly of a complex respiratory enzyme. Mol. Microbiol. 90: 400-414. 245

Jepson, B.J.N., S. Mohan, T.A. Clarke, A.J. Gates, J.A. Cole, C.S. Butler, J.N. Butt, A.M. Hemmings & D.J. Richardson, (2007) Spectropotentiometric and structural analysis of the periplasmic nitrate reductase from Escherichia coli. J. Biol. Chem. 282: 6425-6437. Jia, W., H. Li, L. Zhao & J.K. Nicholson, (2008) Gut microbiota: a potential new territory for drug targeting. Nat. Rev. Drug Discov. 7: 123-129. Johnson, M.E. & K.V. Rajagopalan, (1987) Involvement of chlA, E, M and N loci in Escherichia coli molybdopterin biosynthesis. J. Bacteriol. 169: 117-125. Johnson, T.J. & L.K. Nolan, (2009) Pathogenomics of the virulence plasmids of Escherichia coli. Microbiol. Mol. Biol. Rev. 73: 750-774. Joly, J.C. & W. Wickner, (1993) The SecA and SecY subunits of translocase are the nearest neighbors of a translocating preprotein, shielding it from phospholipids. EMBO J. 12: 255-263. Jong, W.S.P., C.M. ten Hagen-Jongman, P. Genevaux, J. Brunner, B. Oudega & J. Luirink, (2004) Trigger factor interacts with the signal peptide of nascent Tat substrates but does not play a critical role in Tat-mediated export. Eur. J. Biochem. 271: 4779-4787. Jormakka, M., D. Richardson, B. Byrne & S. Iwata, (2004) Architecture of NarGH reveals a structural classification of Mo-bisMGD enzymes. Structure 12: 95-104. Joshi, M.S., J.L. Johnson & K.V. Rajagopalan, (1996) Molybdenum cofactor biosynthesis in Escherichia coli mod and mog mutants. J. Bacteriol. 178: 4310-4312. Jourlin, C., A. Bengrine, M. Chippaux & V. Méjean, (1996a) An unorthodox sensor protein (TorS) mediates the induction of the tor structural genes in response to trimethylamine N-oxide in Escherichia coli. Mol. Microbiol. 20: 1297-1306. Jourlin, C., G. Simon, J. Pommier, M. Chippaux & V. Méjean, (1996b) The periplasmic TorT protein is required for trimethylamine N-oxide reductase gene induction in Escherichia coli. J. Bacteriol. Virol. 178: 1219-1223. Kamada, N., G.Y. Chen, N. Inohara & G. Nunez, (2013) Control of pathogens and pathobionts by the gut microbiota. Nat. Immunol. 14: 685-690. Kapust, R.B. & D.S. Waugh, (1999) Escherichia coli maltose-binding protein is uncommonly effective at promoting the solubility of polypeptides to which it is fused. Protein Sci. 8: 1668-1674. Karimova, G., J. Pidoux, A. Ullmann & D. Ladnant, (1998) A bacterial two-hybrid system based on a reconstituted signal transduction pathway. Proc. Natl. Acad. Sci. U. S. A. 95: 5752- 5756. Kebir, M.O. & D.A. Kendall, (2002) SecA Specificity for Different Signal Peptides. Biochemistry 41: 5573-5580. Kirillova, O., M. Chruszcz, I.A. Shumilin, T. Skarina, E. Gorodichtchenskaia, M. Cymborowski, A. Savchenko, A. Edwards & W. Minor, (2007) An extremely SAD case: structure of a putative redox-enzyme maturation protein from Archaeoglobus fulgidus at 3.4 A resolution. Acta Crystallographica Section D 63: 348-354. Kisker, C., H. Schindlen & D.C. Rees, (1997) Molybdenum-cofactor-containing enzymes: structure and mechanism. Annu. Rev. Biochem. 66: 233-267. Klein, M.J., S.L. Grage, C. Muhle-Goll, J. Bürck, S. Afonin & A.S. Ulrich, (2012) Structure analysis of the membrane-bound PhoD signal peptide of the Tat translocase shows an N- terminal amphiphilic helix. Biochim. Biophys. Acta 1818: 3025-3031. Kneuper, H., B. Maldonado, F. Jäger, M. Krehenbrink, G. Buchanan, R. Keller, M. Müller, B.C. Berks & T. Palmer, (2012) Molecular dissection of TatC defines critical regions essential for protein transport and a TatB-TatC contact site. Mol. Microbiol. 85: 954-961. Koch, S., M.J. Fritsch, G. Buchanan & T. Palmer, (2012) The Escherichia coli TatA and TatB proteins have an N-out C-in topology in intact cells. J. Biol. Chem. Koeth, R.A., Z. Wang, B.S. Levison, J.A. Buffa, E. Org, B.T. Sheehy, E.B. Britt, X. Fu, Y. Wu, L. Li, J.D. Smith, J.A. DiDonato, J. Chen, H. Li, G.D. Wu, J.D. Lewis, M. Warrier, J.M. Brown, R.M. Krauss, W.H.W. Tang, F.D. Bushman, A.J. Lusis & S.L. Hazen, (2013) Intestinal 246

microbiota metabolism of L-carnitine, a nutrient in red meat, promotes atherosclerosis. Nat. Med. 19: 576-585. Kojima, S., K. Imada, M. Sakuma, Y. Sudo, C. Kojima, T. Minamino, M. Homma & K. Namba, (2009) Stator assembly and activation mechanism of the flagellar motor by the periplasmic region of MotB. Mol. Microbiol. 73: 710-718. Kostecki, J.S., H. Li, R.J. Turner & M.P. DeLisa, (2010) Visualizing interactions along the Escherichia coli twin arginine translocation pathway using protein fragment complementation. PLoS One 5: e9225. Kwan, H.S. & E.L. Barrett, (1983) Roles for menaquinone and the two trimethylamine oxide (TMAO) reductases in TMAO respiration in Salmonella typhimurium: Mu d(Apr lac) insertion mutations in men and tor. J. Bacteriol. 155: 1147-1155. Lanciano, P., A. Vergnes, S. Grimaldi, B. Guigliarelli & A. Magalon, (2007) Biogenesis of a respiratory complex is orchestrated by a single accessory protein. J. Biol. Chem. 282: 17468-17474. Leake, M.C., N.P. Greene, R.M. Godun, T. Granjon, G. Buchanan, S. Chen, R.M. Berry, T. Palmer & B.C. Berks, (2008) Variable stoiciometry of the TatA component of the twin-arginine protein transport system observed by in vivo single-molecule imaging. Proc. Natl. Acad. Sci. U. S. A. 105: 15376-15381. Lebendiker, M. & T. Danieli, (2011) Purification of Proteins Fused to Maltose-Binding Protein. In: Protein Chromatography. D. Walls & S.T. Loughran (eds). Humana Press, pp. 281- 293. Lee, P.A., G. Buchanan, N.R. Stanley, B.C. Berks & T. Palmer, (2002) Truncation analysis of TatA and TatB defines the minimal functional units required for protein translocation. J. Bacteriol. 184: 5871-5879. Lee, P.A., G.L. Orriss, G. Buchanan, N.P. Greene, P.J. Bond, C. Punginelli, R.L. Jack, M.S.P. Sansom, B.C. Berks & T. Palmer, (2006a) Cysteine-scanning mutagenesis and disulphide mapping studies of the conserved domain of the Twin-arginine Translocase TatB component. J. Biol. Chem. 281: 34072-34085. Lee, P.A., D. Tullman-Ercek & G. Georgiou, (2006b) The bacterial Twin-arginine Translocation pathway. Annu. Rev. Microbiol. 60: 373-395. Leimkühler, S., (2007) The Biosynthesis of the Molybdenum Cofactor and Its Incorporation into Molybdoenzymes. In: The Periplasm. American Society of Microbiology, pp. Leimkühler, S., M.M. Wuebbens & K.V. Rajagopalan, (2001) Characterization of Escherichia coli MoeB and its involvement in the activation of molybdopterin synthase for the biosynthesis of the molybdenum cofactor. J. Biol. Chem. 276: 34695-34701. Leimkühler, S., M.M. Wuebbens & K.V. Rajagopalan, (2011) The history of the discovery of the molybdenum cofactor and novel aspects of its biosynthesis in bacteria. Coord. Chem. Rev. 255: 1129-1144. Lewis, R.J., D.J. Scott, J.A. Brannigan, J.C. Ladds, M.A. Cervin, G.B. Spiegelman, J.G. Hoggett, I. Barák & A.J. Wilkinson, (2002) Dimer formation and transcription activation in the sporulation response regulator Spo0A. J. Mol. Biol. 316: 235-245. Li, H., L. Chang, J.M. Howell & R.J. Turner, (2010) DmsD, a Tat system specific chaperone, interacts with other general chaperones and proteins involved in the molybdenum cofactor biosynthesis. Biochim. Biophys. Acta 1804: 1301-1309. Lill, R., K. Cunningham, L.A. Brundage, K. Ito, D. Olicer & W. Wickner, (1989) SecA hydrolyzes ATP and is an essential of the protein translocation ATPase of Escherichia coli. EMBO J. 8: 961-966. Lin, A.V. & V. Stewart, (2010) Functional roles foe the GerE-family carboxyl-terminal domains of nitrate response regulators NarL and NarP of Escherichia coli K-12. Microbiology 156: 2933-2943. Lin, E.C.C. & S. Iuchi, (1991) Regulation of gene expression in fementative and respiratory systems in Escherichia coli and related bacteria. Annu. Rev. Genet. 25: 3361-3387. 247

Liu, M.T.W., M.M. Wuebbens, K.V. Rajagopalan & H. Schindelin, (2000) Crystal structure of the gephyrin-related molybdenum cofactor biosynthesis protein MogA from Escherichia coli. J. Biol. Chem. 275: 1814-1822. Liu, X. & P. De Wulf, (2004) Probing the ArcA-P modulon of Escherichia coli by whole genome transcriptional analysis and sequence recognition profiling. J. Biol. Chem. 279: 12588- 12597. Lopez, C.A., S.E. Winter, F. Rivera-Chávez, M.N. Xavier, V. Poon, S.-P. Nuccio, R.M. Tsolis & A.J. Bäumler, (2012) Phage-mediated acquisition of a Type III secreted effector protein boosts growth of Salmonella by nitrate respiration. mBio 3: e00143-00112. Loschi, L., S.J. Brokx, T.L. Hills, G. Zhang, M.G. Bertero, A.L. Lovering, J.H. Weiner & N.C.J. Strynadka, (2004) Structural and biochemical identification of a novel bacterial oxidoreductase. J. Biol. Chem. 279: 80391-50400. Lowry, O.H., N.J. Rosebrough, A.L. Farr & R.J. Randall, (1951) Protein measurement with the Folin phenol reagent. J. Biol. Dyn. 193: 265-275. Lozupone, C.A., J.I. Stombaugh, J.I. Gordon, J.K. Jansson & R. Knight, (2012) Diversity, stability and resilience of the human gut microbiota. Nature 489: 220-230. Lubitz, S.P. & J.H. Weiner, (2003) The Escherichia coli ynfEFGHI operon encodes polypeptides which are paralogues of dimethyl sulfoxide reductase (DmsABC). Arch. Biochem. Biophys. 418: 205-216. Lüke, I., G. Butland, K. Moore, G. Buchanan, V. Lyall, S. Fairhurst, J. Greenblatt, A. Emili, T. Palmer & F. Sargent, (2008) Biosynthesis of the respiratory formate dehydrogenases from Escherichia coli: characterization of the FdhE protein. Arch. Microbiol. 190: 685- 696. Lüke, I., J.I. Handford, T. Palmer & F. Sargent, (2009) Proteolytic processing of Escherichia coli twin-arginine signal peptides by LepB. Arch. Microbiol. 191: 919-925. Madueño, F., J.A. Napier & J.C. Gray, (1993) Newly imported rieske iron-sulfur protein associates with both Cpn60 and Hsp70 in the chloroplast stroma. The Plant Cell 5: 1865-1876. Magalon, A., J.G. Fedor, A. Walburger & J.H. Weiner, (2011) Molybdenum enzymes in bacteria and their maturation. Coord. Chem. Rev. 255: 1159-1178. Magalon, A., R.A. Rothery, G. Giordano, F. Blasco & J.H. Weiner, (1997) Characterization by electron paramagnetic resonance of the role of the Escherichia coli nitrate reductase (NarGHI) iron-sulfur clusters in electron transfer to nitrate and identification of a semiquinone radical intermediate. J. Bacteriol. 16: 5037-5045. Maier, R.J., (2005) Use of molecular hydrogen as an energy substrate by human pathogenic bacteria. Biochem. Soc. Trans. 33: 83-85. Maillard, J., C.A.E.M. Spronk, G. Buchanan, V. Lyall, D.J. Richardson, T. Palmer, G.W. Vuister & F. Sargent, (2007) Structural diversity in twin-arginine signal peptide-binding proteins. Proc. Natl. Acad. Sci. U. S. A. 104: 15641-15646. Maldonado, B., H. Kneuper, G. Buchanan, K. Hatzixanthis, F. Sargent, B.C. Berks & T. Palmer, (2011) Characterisation of the membrane-extrinsic domain of the TatB component of the twin arginine protein translocase. FEBS Lett. 585: 478-484. Maloney, P.C., E.R. Kashket & T.H. Wilson, (1974) A protonmotive force drives ATP synthesis. Proc. Natl. Acad. Sci. U. S. A. 71: 3896-3900. Mandrand-Berthelot, M.-A., G. Couchoux-Luthaud, C.-L. Santini & G. Giordano, (1988 ) Mutants of Escherichia coli specifically deficient in respiratory formate dehydrogenase activity. J. Gen. Microbiol. 134: 3129-3139. Mangels, D., J. Mathers, A. Bolhuis & C. Robinson, (2005) The core TatABC complex of the twin- arginine translocase in Escherichia coli: TatC drives assembly whereas TatA is essential for stability. J. Mol. Biol. 345: 415-423. Maris, A.E., M. Kaczor-Grzeskowiak, Z. Ma, M.L. Kopka, R.P. Gunsalus & R.E. Dickerson, (2005) Primary and Secondary Modes of DNA Recognition by the NarL Two-Component Response Regulator†,‡. Biochemistry 44: 14538-14552. 248

Maris, A.E., M.R. Sawaya, M. Kaczor-Grzeskowiak, M.R. Jarvis, S.M.D. Bearson, M.L. Kopka, I. Schröder, R.P. Gunsalus & R.E. Dickerson, (2002) Dimerization allows DNA target site recognition by the NarL response regulator. Nat. Struct. Biol. 7: 771-778. Matos, C.F.R.O., C. Robinson & A. Di Cola, (2008) The Tat system proofreads FeS protein substrates and directly initiates the disposal of rejected molecules. EMBO J. 27: 2055- 2063. Maupin-Furlow, J.A., J.K. Rosentel, J.H. Lee, U. Deppenmeier, R.P. Gunsalus & K.T. Shanmugam, (1995) Genetic analysis of the modABCD (molybdate transport) operon of Escherichia coli. J. Bacteriol. 177: 4851-4856. McClelland, M., K.E. Sanderson, J. Spieth, S.W. Clifton, P. Latreille, L. Courtney, S. Porwollik, J. Ali, M. Dante, F. Du, S. Hou, D. Layman, S. Leonard, C. Nguyen, K. Scott, A. Holmes, N. Grewal, E. Mulvaney, E. Ryan, H. Sun, L. Florea, W. Miller, T. Stoneking, M. Nhan, R. Waterston & R.K. Wilson, (2001) Complete genome sequence of Salmonella enterica serovar Typhimurium LT2. Nature 413: 852-856. McDevitt, C.A., P. Hugenholtz, G.R. Hanson & A.G. McEwan, (2002) Molecular analysis of dimethyl sulphide dehydrogenase from Rhodovulum sulfidophilum: its place in the dimethyl sulphoxide reductase family of microbial molybdopterin-containing enzymes. Mol. Microbiol. 44: 1575-1587. McEwan, A.G., J.P. Ridge, C.A. McDevitt & P. Hugenholtz, (2002) The DMSO reductase family of microbial molybdenum enzymes; molecular properties and role in the dissimilatory reductase of toxic elements. Geomicrobiol. J. 19: 3-21. McLuskey, K., J.A. Harrison, A.W. Schüttelkopf, D.H. Boxer & W.N. Hunter, (2003) Insight into the role of Escherichia coli MobB in molybdenum cofactor biosynthesis based on the high resolution crystal structure. J. Biol. Chem. 278: 23706-23713. McNicholas, P.M., R.C. Chiang & R.P. Gunsalus, (1998) Anaerobic regulation of the Escherichia coli dmsABC operon requires the molybdate-responsive regulator ModE. Mol. Microbiol. 27: 197-208. McNicholas, P.M., S.A. Rech & R.P. Gunsalus, (1997) Characterization of the ModE DNA-binding sites in the control regions of modABCD and moaABCDE of Escherichia coli. Mol. Microbiol. 23: 515-524. Méjean, V., C. Lobbi-Nivol, M. Lepelletier, G. Giordano, M. Chippaux & M.-C. Pascal, (1994) TMAO anaerobic respiration in Escherichia coli: involvement of the tor operon. Mol. Microbiol. 11: 1169-1179. Mihara, H. & N. Esaki, (2002) Bacterial cysteine desulfurases: their function and mechanisms. Appl. Microbiol. Biotechnol. 60: 12-23. Miller, C.A., K.D. Corbin, K.-A. da Costa, S. Zhang, X. Zhao, J.A. Galanko, T. Blevins, B.J. Bennett, A. O'Connor & S.H. Zeisel, (2014) Effect of egg ingestion on trimethylamine-N-oxide production in humans: a randomized, controlled, dose-response study. Am. J. Clin. Nutr. 100: 778-786. Molik, S., I. Karnauchov, C. Weidlich, R.G. Herrmann & R.B. Klosgen, (2001) The Rieske Fe/S protein of the cytochrome b6/f complex in chloroplasts. J. Biol. Chem. 279: 42761- 42766. Moore, L.J. & P.J. Kiley, (2001) Characterization of the dimerization domain in the FNR transcription factor. J. Biol. Chem. 276: 45744-45750. Moran, C.P. & F. Shanahan, (2014) Gut microbiota and obesity: Role in aetiology and potential therapeutic target. Best Practice & Research Clinical Gastroenterology 28: 585-597. Mori, H. & K. Cline, (2001) Post-translational protein translocastion into thylakoids by the Sec and ΔpH-dependent pathways. Biochemica et Biophysical Acta 1541: 80-90. Natale, P., T. Brüser & A.J.M. Driessen, (2008) Sec- and Tat-mediated protein secretion across the bacterial cytoplasmic membrane—Distinct translocases and mechanisms. Biochim. Biophys. Acta 1778: 1735-1756. 249

Neumann, M. & S. Leimkühler, (2008) Heavy metal ions inhibit molybdoenzyme activity by binding to the dithiolene moiety of molybdopterin in Escherichia coli. FEBS J. 275: 5678-5689. Neumann, M., F. Seduk, C. Iobbi-Nivol & S. Leimkühler, (2011) Molybdopterin dinucleotide biosynthesis in Escherichia coli: Identification of amino acid residues of molybdopterin dinucleotide transferases that determine specificity for binding of guanine or cytosine nucleotides. J. Biol. Chem. 286: 1400-1408. Nishiyama, K.-i., M. Hanada & H. Tokuda, (1994) Disruption of the gene encoding p12 (SecG) reveals the direct involvement and important function of SecG in the protein translocation of Escherichia coli at low temperature. EMBO J. 13: 3272-3277. Noriega, C.E., H.-Y. Lin, L.-L. Chen, S.B. Williams & V. Stewart, (2010) Asymmetric cross- regulation between the nitrate-responsive NarX–NarL and NarQ–NarP two-component regulatory systems from Escherichia coli K-12. Mol. Microbiol. 75: 394-412. Nouwen, N. & A.J.M. Driessen, (2002) SecDFyajC forms a heterotetrameric complex with YidC. Mol. Microbiol. 44: 1397-1405. Oates, J., C.M.L. Barrett, J.P. Barnett, K.G. Byrne, A. Bolhuis & C. Robinson, (2005) The Escherichia coli twin-arginine translocation apparatus incorporates a distinct form of TatABC complex, spectrum of modular TatA complexes and minor TatAB complex. J. Mol. Biol. 346: 295-305. Oates, J., J. Mathers, D. Mangels, W. Kühlbrandt, C. Robinson & K. Model, (2003) Consensus Structural Features of Purified Bacterial TatABC Complexes. J. Mol. Biol. 330: 277-286. Oliver, D.C. & M. Paetzel, (2008) Crystal structure of the major periplasmic domain of the bacterial membrane protein assembly facilitator YidC. J. Biol. Chem. 283: 5208-5216. Olmo-Mira, M.F., M. Gavira, D.J. Richardson, F. Castillo, C. Moreno-Vivián & M.D. Roldán, (2004) NapF is a cytoplasmic iron-sulfur protein required for Fe-S cluster assembly in the periplasmic nitrate reductase. J. Biol. Chem. 279: 49727-49735. Orel, R. & T.K. Trop, (2014) Intestinal microbiota, probiotics and prebiotics in inflammatory bowel disease. World J. Gastroenterol. 20: 11505-11524. Oresnik, I.J., C.L. Ladner & R.J. Turner, (2001) Identification of a twin-arginine leader-binding protein. Mol. Microbiol. 40: 323-331. Orriss, G.L., M.J. Tarry, B. Ize, F. Sargent, S.M. Lea, T. Palmer & B.C. Berks, (2007) TatBC, TatB, and TatC form structurally autonomous units within the twin arginine protein transport system of Escherichia coli. FEBS Lett. 581: 4091-4097. Palmer, T. & B.C. Berks, (2012) The twin-arginine translocation (Tat) protein export pathway. Nature Reviews Microbiology 10: 483-496. Palmer, T., C.-L. Santini, C. Iobbi-Nivol, D.J. Eaves, D.H. Boxer & G. Giordano, (1996) Involvement of the narJ and mob gene products in distinct steps in the biosynthesis of the molybdoenzyme nitrate reductase in Escherichia coli. Mol. Microbiol. 20: 875-884. Palmer, T., F. Sargent & B.C. Berks, (2005) Export of complex cofactor-containing proteins by the bacterial Tat pathway. Trends Microbiol. 13: 175-180. Panahandeh, S., C. Maurer, M. Moser, M.P. DeLisa & M. Müller, (2008) Following the path of a twin-arginine precursor along the TatABC translocase of E. coli. J. Biol. Chem. 283: 33267-33275. Papanikou, E., S. Karamanou, C. Baud, M. Frank, G. Sianidis, D. Keramisanou, C.G. Kalodimos, A. Kuhn & A. Economou, (2005) Identification of the preprotein binding domain of SecA. J. Biol. Chem. 280: 43209-43217. Papish, A.L., C.L. Ladner & R.J. Turner, (2003) The Twin-arginine leader-binding protein, DmsD, interacts with the TatB and TatC subunits of the Escherichia coli Twin-arginine Translocase. J. Biol. Chem. 278: 82501-82506. Park, A.K., J.H. Moon, K.S. Lee & Y.M. Chi, (2012) Crystal structure of receiver domain of putative NarL family response regulator spr1814 from Streptococcus pneumoniae in the absence and presence of the phosphoryl analog beryllofluoride. Biochem. Biophys. Res. Commun. 421: 403-407. 250

Park, A.K., J.H. Moon, J.S. Oh, K.S. Lee & Y.M. Chi, (2013a) Crystal structure of the response regulator spr1814 from Streptococcus pneumoniae reveals unique interdomain contacts among NarL family proteins. Biochem. Biophys. Res. Commun. 434: 65-69. Park, D.M., M.S. Akhtar, A.Z. Ansari, R. Landick & P.J. Kiley, (2013b) The Bacterial Response Regulator ArcA Uses a Diverse Binding Site Architecture to Regulate Carbon Oxidation Globally. PLoS Genet. 9: e1003839. Pérez-Rodríguez, R., A.C. Fisher, J.D. Perlmutter, M.G. Hicks, A. Chanal, C.-L. Santini, L.-F. Wu, T. Palmer & M.P. DeLisa, (2007) An essential role for the DnaK molecular chaperone in stabilizing over-expressed substrate proteins of the bacterial twin-arginine translocation pathway. J. Mol. Biol. 367: 715-730. Peters, J.W., G.J. Schut, E.S. Boyd, D.W. Mulder, E.M. Shepard, J.B. Broderick, P.W. King & M.W.W. Adams, (2014) [FeFe]- and [NiFe]-hydrogenase diversity, mechanism, and maturation. Biochim. Biophys. Acta. Pilak, O., B. Mamat, S. Vogt, C.H. Hagemeier, R.K. Thauer, S. Shima, C. Vonrhein, E. Warkentin & U. Ermler, (2006) The Crystal Structure of the Apoenzyme of the Iron–Sulphur Cluster-free Hydrogenase. J. Mol. Biol. 358: 798-809. Pinske, C. & G. Sawers, (2012) Delivery of iron-sulfur clusters to the hydrogen-oxidizing [NiFe]- hydrogenases in Eshcerichia coli requires the A-type carrier proteins ErpA and IscA. PLoS One 7: e31755. Pinske, C. & R.G. Sawers, (2011) A-type carrier protein ErpA is essential for formation of an active formate-nitrate respiratory pathway in Escherichia coli K-12. J. Bacteriol. 194: 346-353. Pitterle, D.M. & K.V. Rajagopalan, (1989) Two proteins encoded at the chlA locus constitute the converting factor of Escherichia coli chlA1. J. Bacteriol. 171: 3373-3378. Pohlschröder, M., E. Hartmann, N.J. Hand, K. Dilks & A. Haddad, (2005) Diversity and evolution of protein translocation. Annu. Rev. Microbiol. 59: 91-111. Pommier, J., V. Méjean, G. Giordano & C. Iobbi-Nivol, (1998) TorD, a cytoplasmic chaperone that interacts with the unfolded trimethylamine N-oxide reductase enzyme (TorA) in Escherichia coli. J. Biol. Chem. 273: 16615-16620. Potter, L.C. & J.A. Cole, (1999) Essential roles for the products of the napABCD genes, but not napFGH, in periplasmic nitrate reduction by Escherichia coli K-12. Biochem. J. 344. Potter, L.C., P. Millington, L. Griffiths, G.H. Thomas & J.A. Cole, (1999) Competition between Escherichia coli strains expressing either a periplasmic or a membrane-bound nitrate reductase: does Nap confer a selective advantage during nitrate-limited growth? Biochem. J. 344: 77-84. Prinz, A., C. Behrens, T.A. Rapoport, E. Hartmann & K.-U. Kalies, (2000) Evolutionarily conserved binding of ribosomes to the translocation channel via the large ribosomal RNA. EMBO J. 19: 1900-1906. Puddu, A., R. Sanguineti, F. Montecucco & G.L. Viviani, (2014) Evidence for the gut microbiota short-chain fatty acids as key pathophysiological molecules improving diabetes. Mediators Inflamm.: 162021. Pugsley, A.P., (1993) The complete general secretory pathway in Gram-negative bacteria. Microbiol. Rev. 57: 50-108. Punginelli, C., B. Maldonado, S. Grahl, R. Jack, M. Alami, J. Schröder, B.C. Berks & T. Palmer, (2007) Cysteine scanning mutagenesis and topological mapping of the Escherichia coli twin-arginine translocase TatC component. J. Bacteriol. 189: 5482-5494. Qiu, Y., R. Zhang, T.A. Binkowski, V. Tereshko, A. Joachimiak & A. Kossiakoff, (2008) The 1.38 Å crystal structure of DmsD protein from Salmonella typhimurium, a proofreading chaperone on the Tat pathway. Proteins: Structure, Function, and Bioinformatics 71: 525-533. Raetz, C.R.H. & C. Whitfield, (2002) Lipopolysaccharide endotoxins. Annu. Rev. Biochem. 71: 635-700. 251

Ramasamy, S.K. & W.M. Clemons Jr, (2009) Structure of the twin-arginine signal-binding protein DmsD from Escherichia coli. Acta Crystallogr. Sect. F Struct. Biol. Cryst. Commun. F65: 746-750. Ramos, S., S. Schuldiner & H.R. Kaback, (1976) The electrochemical gradient of protons and its relationship to active transport in Escherichia coli membrane vesicles. Proc. Natl. Acad. Sci. U. S. A. 73: 1892-1896. Randall, L.L., S.J.S. Hardy, T.B. Topping, V.F. Smith, J.E. Bruce & R.D. Smith, (1998a) The interaction between the chaperone SecB and its ligands: Evidence for multiple subsites for binding. Protein Sci. 7: 2384-2390. Randall, L.L., T.B. Topping, D. Suciu & S.J.S. Hardy, (1998b) Calorimetric analyses of the interaction between SecB and its ligands. Protein Sci. 7: 1195-1200. Ray, N., J. Oates, R.J. Turner & C. Robinson, (2003) DmsD is required for the biogenesis of DMSO reductase in Escherichia coli but not for the interaction of the DmsA signal peptide with the Tat apparatus. FEBS Lett. 534: 156-160. Raymond, J.A. & G.E. Plopper, (2002) A bacterial TMAO transporter. Comparative Biochemistry and Physiology Part B: Biochemistry and Molecular Biology 133: 29-34. Redelberger, D., F. Seduk, O. Genest, V. Méjean, S. Leimkühler & C. Iobbi-Nivol, (2011) The YcdY protein of Escherichia coli, an atypical member of the TorD chaperone family. J. Bacteriol. Epub. Reynolds, M.M., L. Bogomolnaya, J. Guo, L. Aldrich, D. Bokhari, C.A. Santiviago, M. McClelland & H. Andrews-Polymenis, (2011) Abrogation of the Twin Arginine Transport System in Salmonella enterica Serovar Typhimurium Leads to Colonization Defects during Infection. PLoS One 6: e15800. Richardson, D.J., (2000) Bacterial respiration: a flexible process for a changing environment. Microbiology 146: 551-571. Richter, S., U. Lindenstrauß, C. Lücke, R. Bayliss & T. Brüser, (2007) Functional Tat transport of unstructured small, hydrophilic proteins. J. Biol. Chem. 282: 33257-33264. Robinson, C., C.F.R.O. Matos, D. Beck, C. Ren, J. Lawrence, N. Vasisht & S. Mendel, (2011) Transport and proofreading of proteins by the twin-arginine translocation (Tat) system in bacteria. Biochim. Biophys. Acta 1808: 876-884. Rocco, M.A., D. Waraho-Zhmayev & M.P. DeLisa, (2012) Twin-arginine translocase mutations that suppress folding quality control and permit export of misfolded substrate proteins. Proceedings of the National Academy of Sciences 109: 13392-13397. Rodrigue, A., N. Batia, M. Müller, O. Fayet, R. Böhm, M.-A. Mandrand-Berthelot & L.-F. Wu, (1996) Involvement of the GroE chaperonins in the nickel-dependent anaerobic biosynthesis of NiFe-hydrogenases of Escherichia coli. J. Bacteriol. 178: 4453-4460. Rodrigue, A., A. Chanal, K. Beck, M. Müller & L.-F. Wu, (1999) Co-translation of a periplasmic enzyme complex by a hitchhiker mechanism through the bacterial Tat pathway. J. Biol. Chem. 274: 13223-13228. Rodriguez, F., S.L. Rouse, C.E. Tait, J. Harmer, A. De Riso, C.R. Timmel, M.S.P. Samsom, B.C. Berks & J.R. Schnell, (2013) Structural model for the protein-translocating element of the twin-arginine transport system. Proc. Natl. Acad. Sci. U. S. A. 110: E1092-E1101. Roldán, M.D., H.J. Sears, M.R. Cheesman, S.J. Ferguson, A.J. Thomson, B.C. Berks & D.J. Richardson, (1998) Spectroscopic characterization of a novel multiheme c-type cytochrome widely implicated in bacterial electron transport. J. Biol. Chem. 273: 28785-28790. Rollauer, S.E., M.J. Tarry, J.E. Graham, M. Jääskeläinen, F. Jäger, S. Johnson, M. Krehenbrink, S.- M. Liu, M.J. Lukey, J. Marcoux, M.A. McDowell, F. Rodriguez, P. Roversi, P.J. Stansfeld, C.V. Robinson, M.S.P. Sansom, T. Palmer, M. Högbom, B.C. Berks & S.M. Lea, (2012) Structure of the TatC core of the twin-arginine protein transport system. Nature (London) 492: 210-212. Romao, M.J., (2009) Molybdenum and tungsten enzymes: a crystallographic and mechanistic overview. Dalton Transactions: 4053-4068. 252

Rose, P., J. Fröbel, P.L. Graumann & M. Müller, (2013) Substrate-Dependent Assembly of the Tat Translocase as Observed in Live Escherichia coli Cells. PLoS One 8: e69488. Rose, R.W., T. Brüser, J.C. Kissinger & M. Pohlschröder, (2002) Adaptation of protein secretion to extremely high-salt conditions by extensive use of the twin-arginine translocation pathway. Mol. Microbiol. 45: 934-950. Rothery, R.A., F. Blasco, A. Magalon & J.H. Weiner, (2001) The diheme cytochrome b subunit (NarI) of Escherichia coli nitrate reductase A (NarGHI): structure, function and interaction with quinols. J. Mol. Microbiol. Biotechnol. 3: 273-283. Rothery, R.A., J.L.S. Grant, J.L. Johnson, K.V. Rajagopalan & J.H. Weiner, (1995) Association of molybdopterin guanine dinucleotide with Escherichia coli dimethyl sulfoxide reductase: effect of tungstate and a mobmutation. J. Bacteriol. 177: 2057-2063. Rothery, R.A., A. Magalon, G. Giordano, B. Guigliarelli, F. Blasco & J.H. Weiner, (1998) The molybdenum cofactor of Escherichia coli nitrate reductase A (NarGHI): Effect of a mobAB mutation and interactions with [Fe-S] clusters. J. Biol. Chem. 273: 7462-7469. Rothery, R.A. & J.H. Weiner, (1993) Topological characterization of Escherichia coli DMSO reductase by electron paramagnetic resonance spectroscopy of an engineered [3Fe- 4S] cluster. Biochemistry 32: 5855-5861. Rudiger, S., A. Buchberger & B. Bukau, (1997) Interaction of Hsp70 chaperones with substrates. Nat. Struct. Biol. 4: 342-349. Ruiz, N., D. Kahne & T.J. Silhavy, (2009) Transport of lipopolysaccharide across the cell envelope: the long road of discovery. Nat Rev Micro 7: 677-683. Rusch, S.L. & D.A. Kendall, (2007) Interactions That Drive Sec-Dependent Bacterial Protein Transport. Biochemistry 46: 9665-9673. Sabatino, A., G. Regolisti, I. Brusasco, A. Cabassi, S. Morabito & E. Fiaccadori, (2014) Alterations of intestinal barrier and microbiota in chronic kidney disease. Nephrol. Dial. Transplant.: 1-10. Sachelaru, I., N.A. Petriman, R. Kudva, P. Kuhn, T. Welte, B. Knapp, F. Drepper, B. Warscheid & H.-G. Koch, (2013) YidC occupies the lateral gate of the SecYEG translocon and is sequentially displaced by a nascent membrane protein. J. Biol. Chem. 288: 16295- 16307. Sambasivarao, D., D.G. Scraba, C. Trieber & J.H. Weiner, (1990) Organization of dimethyl sulfoxide reductase in the plasma membrane of Escherichia coli. J. Bacteriol. 172: 5938-5948. Sambasivarao, D., R.J. Turner, J.L. Simala-Grant, G. Shaw, J. Hu & J.H. Weiner, (2000) Multiple roles for the twin-arginine leader sequence of dimethyl sulfoxide reductase of Escherichia coli. J. Biol. Chem. 275: 22526-22531. Sambasivarao, D. & J.H. Weiner, (1991) Dimethyl sulfoxide reductase of Escherichia coli: an investigation of function and assembly by use of in vivo complementation. J. Bacteriol. 173: 5935-5943. San Miguel, M., R. Marrington, P.M. Rodger, A. Rodger & C. Robinson, (2003) An Escherichia coli twin-arginine signal peptide switches between helical and unstructured conformations depending on the hydrophobicity of the environment. Eur. J. Biochem. 270: 3345-3352. Santini, C.-L., A. Bernadac, M. Zhang, A. Chanal, B. Ize, C. Blanco & L.-F. Wu, (2001) Translocation of jellyfish green fluorescent protein via the Tat System of Escherichia coli and change of its periplasmic localization in response to osmotic up-shock. J. Biol. Chem. 276: 8159-8164. Santini, C.-L., B. Ize, A. Chanal, M. Müller, G. Giordano & L.-F. Wu, (1998) A novel Sec- independent periplasmic protein translocation pathway in Escherichia coli. EMBO J. 17: 101-112. Sarfo, K.J., T.L. Winstone, A.L. Papish, J.M. Howell, H. Kadir, H.J. Vogel & R.J. Turner, (2004) Folding forms of Escherichia coli DmsD, a twin-arginine leader binding protein. Biochem. Biophys. Res. Commun. 315: 397-403. 253

Sargent, F., (2007a) Constructing the wonders of the bacterial world: biosynthesis of complex enzymes. Microbiology 153: 633-651. Sargent, F., (2007b) The twin-arginine transport system: moving folded proteins across membranes. Biochem. Soc. Trans. 35: 835-847. Sargent, F., B.C. Berks & T. Palmer, (2006) Pathfinders and trailblazers: a prokaryotic targeting system for transport of folded proteins. FEMS Microbiol. Lett. 254: 198-207. Sargent, F., E.G. Bogsch, N.R. Stanley, M. Wexler, C. Robinson, B.C. Berks & T. Palmer, (1998) Overlapping functions of components of a bacterial Sec-independent protein export pathway. EMBO (European Molecular Biology Organization) Journal 17: 3640-3650. Sargent, F., N.R. Stanley, B.C. Berks & T. Palmer, (1999) Sec-independent protein translocation in Escherichia coli. Journal of Biological Control 274: 36073-36082. Sasaki, S., S.-i. Matsuyama & S. Mizushima, (1990) In vitro kinetic analysis of the role of the positive charge at the amino-terminal region of signal peptides in translocation of secretory protein across the cytoplasmic membrane in Escherichia coli. J. Biol. Chem. 265: 4358-4363. Sauvage, E., A. Derouaux, C. Fraipont, M. Joris, R. Herman, M. Rocaboy, M. Schloesser, J. Dumas, F. Kerff, M. Nguyen-Distèche & P. Charlier, (2014) Crystal Structure of Penicillin-Binding Protein 3 (PBP3) from Escherichia coli. PLoS One 9: e98042. Schrag, J.D., W. Huang, J. Sivaraman, C. Smith, J. Plamondon, R. Larocque, A. Matte & M. Cygler, (2001) The crystal structure of Escherichia coli MoeA, a protein from the molybdopterin synthesis pathway. J. Mol. Biol. 310: 419-431. Schröder, I., S. Rech, T. Krafft & J.M. Macy, (1997) Purification and characterisation of the selenate reductase from Thauera selenatis. J. Biol. Chem. 272: 23765-23768. Schwarz, C.J., J.L. Giel, T. Patschkowski, C. Luther, F.J. Ruzicka, H. Beinert & P.J. Kiley, (2001) IscR, an Fe-S cluster-containing transcription factor, represses expression of Escherichia coli genes encoding Fe-S cluster assembly proteins. Proc. Natl. Acad. Sci. U. S. A. 98: 14895-14900. Schwarz, G., (2005) Molybdenum cofactor biosynthesis and deficiency. Cellular and Molecular Life Sciences CMLS 62: 2792-2810. Scotti, P.A., M.L. Urbanus, J. Brunner, J.-W.L. de Gier, G. von Heijne, C. van der Does, A.J.M. Driessen, B. Oudega & J. Luirink, (2000) YidC, the Escherichia coli homologue of mitochondrial Oxa1p, is a component of the Sec translocase. EMBO J. 19: 542-549. Self, W.T., A.M. Grunden, A. Hasona & K.T. Shanmugan, (1999) Transcriptional regulation of molybdoenzyme synthesis in Escherichia coli in response to molybdenum: ModE- molybdate, a repressor of the modABCD (molybdate transport) operon is a secondary transcriptional activator for the hyc and nar operons. Microbiology 145: 41-55. Shalel-Levanon, S., K.-Y. San & G.N. Bennett, (2005) Effect of oxygen on the Escherichia coli ArcA and FNR regulation systems and metabolic responses. Biotechnol. Bioeng. 89: 556-564. Shanmugham, A., A. Bakayan, P. Völler, J. Grosveld, H. Lill & Y.J.M. Bollen, (2012) The Hydrophobic Core of Twin-Arginine Signal Sequences Orchestrates Specific Binding to Tat-Pathway Related Chaperones. PLoS One 7: e34159. Shanmugham, A., H.W.W.F. Sang, Y.J.M. Bollen & H. Lill, (2006) Membrane binding of twin arginine preproteins as an early step in translocation. Biochemistry 45: 2243-2249. Sharma, P., S. Stagge, M. Bekker, K. Bettenbrock & K.J. Hellingwerf, (2013) Kinase Activity of ArcB from Escherichia coli Is Subject to Regulation by Both Ubiquinone and Demethylmenaquinone. PLoS One 8: e75412. Sharma, P., M.J. Teixeira de Mattos, K.J. Hellingwerf & M. Bekker, (2012) On the function of the various quinone species in Escherichia coli. FEBS J. 279: 3364-3373. Shima, S., O. Pilak, S. Vogt, M. Schick, M.S. Stagni, W. Meyer-Klaucke, E. Warkentin, R.K. Thauer & U. Ermler, (2008) The crystal structure of [Fe]-hydrogenase reveals the geometry of the active site. Sci. Cult. 321: 572-575. 254

Shima, S. & R.K. Thauer, (2007) A third type of hydrogenase catalyzing H2 activation. The Chemical Record 7: 37-46. Simala-Grant, J.L. & J.H. Weiner, (1998) Modulation of the substrate specificity of Escherichia coli dimethylsulfoxide reductase. Eur. J. Biochem. 251: 510-515. Simon, G., C. Jourlin, M. Ansaldi, M.-C. Pascal, M. Chippaux & V. Méjean, (1995) Binding of the TorR regulator to cis-acting direct repeats activates tor operon expression. Mol. Microbiol. 17: 971-980. Simon, G., V. Méjean, C. Jourlin, M. Chippaux & M.-C. Pascal, (1994) The torR gene of Escherichia coli encodes a response regulator protein involved in the expression of the trimethylamine N-oxide reductase genes. J. Bacteriol. 176: 5601-5606. Smith, A.D., G.N.L. Jameson, P.C. Dos Santos, J.N. Agar, S. Naik, C. Krebs, J. Frazzon, D.R. Dean, B.H. Huynh & M.K. Johnson, (2005) NifS-Mediated Assembly of [4Fe−4S] Clusters in the N- and C-Terminal Domains of the NifU Scaffold Protein†. Biochemistry 44: 12955- 12969. Sodergren, E.J. & J.A. DeMoss, (1988) narI region of the Escherichia coli nitrate reductase (nar) operone contains two genes. J. Bacteriol. 170: 1721-1729. Sodergren, E.J., P.-Y. Hsu & J.A. DeMoss, (1988) Roles of the narJ and narI gene products in the expression of nitrate reductase in Escherichia coli. J. Biol. Chem. 263: 16156-16162. Sofia, H.J., G. Chen, B.G. Hetzier, J.F. Reyes-Spindola & N.E. Miller, (2001) Radical SAM, a novel protein superfamily linking unresolved steps in familiar biosynthetic pathways with radical mechanisms: functional characterization using new analysis and information visualization methods. Nucleic Acids Res. 29: 1097-1106. Sparacino-Watkins, C., J.F. Stolz & P. Basu, (2014) Nitrate and periplasmic nitrate reductases. Chem. Soc. Rev. 43: 676-706. Stanley, N.R., K. Findlay, B.C. Berks & T. Palmer, (2001) Escherichia coli strains blocked in Tat- dependent protein export exhibit pleiotropic defects in the cell envelope. J. Bacteriol. 183: 139-144. Stanley, N.R., T. Palmer & B.C. Berks, (2000) The twin arginine consensus motif of Tat signal peptides is involved in Sec-independent protein targeting in Escherichia coli. J. Biol. Chem. 275: 11591-11596. Stanley, N.R., F. Sargent, G. Buchanan, J. Shi, V. Stewart, T. Palmer & B.C. Berks, (2002) Behaviour of topological marker proteins targeted to the Tat protein transport pathway. Mol. Microbiol. 43: 1005-1021. Stevens, C., M. Okon, L. McIntosh & M. Paetzel, (2013) 1H, 13C and 15N resonance assignments and peptide binding site chemical shift perturbation mapping for the Escherichia coli redox enzyme chaperone DmsD. Biomolecular NMR Assignments 7: 193-197. Stevens, C.M., T.M.L. Winstone, R.J. Turner & M. Paetzel, (2009) Structural analysis of a monomeric form of the twin-arginine leader peptide binding chaperone Escherichia coli DmsD. J. Mol. Biol. 389: 124-133. Stewart, V. & P.J. Bledsoe, (2003) Synthetic lac operator substitutions for studying the nitrate- and nitrate-responsive NarX-NarL and NarQ-NarP two-component regulatory systems of Escherichia coli K-12. J. Bacteriol. 185: 2104-2111. Stewart, V., Y. Lu & A.J. Darwin, (2002) Periplasmic Nitrate Reductase (NapABC Enzyme) Supports Anaerobic Respiration by Escherichia coli K-12. J. Bacteriol. 184: 1314-1323. Stolz, J.F., P. Basu, J.M. Santini & R.S. Oremland, (2006) Arsenic and selenium in microbial metabolism. Annu. Rev. Microbiol. 60: 107-130. Storf, S., F. Pfieffer, K. Dilks, Z.Q. Chen, S. Imam & M. Pohlschröder, (2010) Mutational and bioinformatic analysis of haloarchaeal lipobox-containing proteins. Archaea 2010. Sutherland, B.W., J. Toews & J. Kast, (2008) Utility of formaldehyde cross-linking and mass spectrometry in the study of protein–protein interactions. J. Mass Spectrom. 43: 699- 715. 255

Tang, W.H.W., Z. Wang, B.S. Levison, R.A. Koeth, E.B. Britt, X. Fu, Y. Wu & S.L. Hazen, (2013) Intestinal Microbial Metabolism of Phosphatidylcholine and Cardiovascular Risk. New Engl. J. Med. 368: 1575-1584. Tarry, M., S.J.R. Arends, P. Roversi, E. Piette, F. Sargent, B.C. Berks, D.S. Weiss & S.M. Lea, (2009a) The Escherichia coli cell division protein and model Tat substrate SufI (FtsP) localizes to the septal ring and has a multicopper oxidase-like structure. J. Mol. Biol. 386: 504-519. Tarry, M.J., E. Schafer, S. Chen, G. Buchanan, N.P. Greene, S.M. Lea, T. Palmer, H.R. Saibil & B.C. Berks, (2009b) Structural analysis of substrate binding by the TatBC component of the twin-arginine protein transport system. Proc. Natl. Acad. Sci. U. S. A. 106: 13284- 13289. Thauer, R.K., A.-K. Kaster, M. Goenrich, M. Schick, T. Hiromoto & S. Shima, (2010) Hydrogenases from Methanogenic Archaea, Nickel, a Novel Cofactor, and H2 Storage. Annu. Rev. Biochem. 79: 507-536. Thiennimitr, P., S.E. Winter & A.J. Bäumler, (2012) Salmonella, the host and its microbiota. Curr. Opin. Microbiol. 15: 108-114. Thomas, G., L.C. Potter & J.A. Cole, (1999) The periplasmic nitrate reductase from Escherichia coli: a heterodimeric molybdoprotein with a double-arginine signal sequence and an unusual leader peptide cleavage site. FEMS Microbiol. Lett. 174: 167-171. Thomas, J.D., R.A. Daniel, J. Errington & C. Robinson, (2001) Export of active green fluorescent protein to the periplasm by the twin-arginine translocase (Tat) pathway in Escherichia coli. Mol. Microbiol. 39: 47-53. Tottey, S., K.J. Waldron, S.J. Firbank, B. Reale, C. Bessant, K. Sato, T.R. Cheek, J. Gray, M.J. Banfield, C. Dennison & N.J. Robinson, (2008) Protein-folding location can regulate manganese-binding versus copper- or zinc-binding. Nature 455: 1138-1142. Tranier, S., C. Iobbi-Nivol, C. Birck, M. Ilbert, I. Mortier-Barrière, V. Méjean & J.-P. Samama, (2003) A Novel Protein Fold and Extreme Domain Swapping in the Dimeric TorD Chaperone from Shewanella massilia. Structure 11: 165-174. Tranier, S., I. Mortier-Barrière, M. Ilbert, C. Birck, C. Iobbi-Nivol, V. Méjean & J.-P. Samama, (2002) Characterization and multiple molecular forms of TorD from Shewanella massilia, the putative chaperone of the molybdoenzyme TorA. Protein Sci. 11: 2148- 2157. Truglio, J.J., K. Theis, S. Leimkühler, R. Rappa, K.V. Rajagopalan & C. Kisker, Crystal Structures of the Active and Alloxanthine-Inhibited Forms of Xanthine Dehydrogenase from Rhodobacter capsulatus. Structure 10: 115-125. Trun, N.J., J. Stader, A. Lupas, C. Kumamoto & T.J. Silhavy, (1988) Two cellular components, PrlA and SecB, that recognize different sequence determinants are required for efficient protein export. J. Bacteriol. 170: 5928-5930. Tsui, H.-C.T., G. Zhao, G. Feng, H.-C.E. Leung & M.E. Winkler, (1994) The mutL repair gene of Escherichia coli K-12 forms a superoperon with a gene encoding a new cell-wall amidase. Mol. Microbiol. 11: 189-202. Tullman-Ercek, D., M.P. DeLisa, Y. Kawarasaki, P. Iranpour, B. Ribnicky, T. Palmer & G. Georgiou, (2007) Export pathway selectivity of Escherichia coli twin arginine translocation signal peptides. J. Biol. Chem. 282: 8309-8316. Turnbaugh, P.J., R.E. Ley, M. Hamady, C.M. Fraser-Liggett, R. Knight & J.I. Gordon, (2007) The Human Microbiome Project. Nature 449: 804-810. Turner, R.J., A.L. Papish & F. Sargent, (2004) Sequence analysis of bacterial redox enzyme maturation proteins (REMPs). Can. J. Microbiol. 50: 225-238. Turner, R.J., J.H. Weiner & D.E. Taylor, (1998) Selenium metabolism in Escherichia coli. BioMetals 11: 223-227. Ullers, R.S., E.N.G. Houben, A. Raine, C.M. ten Hagen-Hongman, M. Ehrenberg, J. Brunner, B. Oudega, N. Harms & J. Luirink, (2003) Interplay of signal recognition particle and 256

trigger factor at L23 near the nascent chain exit site on the Escherichia coli ribosome. J. Cell Biol. 161: 679-684. Ulman, C.A., J.J. Trevino, M. Miller & R.K. Gandhi, (2014) Fish odor syndrome: a case report of trimethylaminuria. Dermatol. Online J. 20. Unden, G. & J. Bongaerts, (1997) Alternative respiratory pathways of Escherichia coli: energetics and transcriptional regulation in response to electron acceptors. Biochim. Biophys. Acta 1320: 217-234. Unden, G. & J.R. Guest, (1985) Isolation and characterization of the Fnr protein, the transcriptional regulator of anaerobic electron transport in Escherichia coli. Eur. J. Biochem. 146: 193-199. Valent, Q.A., P.A. Scotti, S. High, J.-W.L. di Gier, G. von Heijne, G. Lentzen, W. Wintermeyer, B. Oudega & J. Luirink, (1998) The Escherichia coli SRP and SecB targeting pathways converge at the translocon. EMBO J. 17: 2504-2512. van den Berg, B., W.M. Clemons Jr, I. Collinson, Y. Modis, E. Hartmann, S.C. Harrison & T.A. Rapoport, (2004) X-ray structure of a protein-conducting channel. Nature 427: 36-44. van der Laan, M., N.P. Nouwen & A.J.M. Driessen, (2005) YidC – an evolutionary conserved device for the assembly of energy-transducing membrane protein complexes. Curr. Opin. Microbiol. 8: 182-187. van der Laan, M., M.L. Urbanus, C.M.t. Hagen-Jongman, N. Nouwen, B. Oudega, N. Harms, A.J.M. Driessen & J. Luirink, (2003) A conserved function of YidC in the biogenesis of respiratory chain complexes. Proc. Natl. Acad. Sci. U. S. A. 100: 5801-5806. Vergnes, A., J. Pommier, R. Toci, F. Blasco, G. Giordano & A. Magalon, (2006) NarJ chaperone binds on two distinct sites of the aponitrate reductase of Escherichia coli to coordinate molybdenum cofactor insertion and assembly. J. Biol. Chem. 281: 2170-2176. Vignais, P.M. & B. Billoud, (2007) Occurrence, Classification, and Biological Function of Hydrogenases: An Overview. Chem. Rev. 107: 4206-4272. Vignais, P.M., B. Billoud & J. Meyer, (2001) Classification and phylogeny of hydrogenases. FEMS Microbiol. Rev. 25: 455-501. Vignais, P.M. & A. Colbeau, (2004) Molecular biology of microbial hydrogenases. Curr. Issues Mol. Biol. 6: 159-188. Vinjé, S., E. Stroes, M. Nieuwdorp & S.L. Hazen, (2014) The gut microbiome as novel cardio- metabolic target: the time has come! Eur. Heart J. 35: 883-887. Vollmer, W., D. Blanot & M.A. De Pedro, (2008a) Peptidoglycan structure and architecture. FEMS Microbiol. Rev. 32: 149-167. Vollmer, W., B. Joris, P. Charlier & S. Foster, (2008b) Bacterial peptidoglycan (murein) hydrolases. FEMS Microbiol. Rev. 32: 259-286. Walsh, C.T., S.L. Fisher, I.S. Park, M. Prahalad & Z. Wu, (1996) Bacterial resistance to vancomycin: five genes and one missing hydrogen bond tell the story. Chem. Biol. 3: 21-28. Walther, T.H., S.L. Grage, N. Roth & A.S. Ulrich, (2010) Membrane Alignment of the Pore- Forming Component TatAd of the Twin-Arginine Translocase from Bacillus subtilis Resolved by Solid-State NMR Spectroscopy. J. Am. Chem. Soc. 132: 15945-15956. Wang, P. & R.E. Dalbey, (2011) Inserting membrane proteins: The YidC/Oxa1/Alb3 machinery in bacteria, mitochondria, and chloroplasts. Biochim. Biophys. Acta 1808: 866-875. Wang, Z., E. Klipfell, B.J. Bennett, R. Koeth, B.S. Levison, B. DuGar, A.E. Feldstein, E.B. Britt, X. Fu, Y.-M. Chung, Y. Wu, P. Schauer, J.D. Smith, H. Allayee, W.H.W. Tang, J.A. DiDonato, A.J. Lusis & S.L. Hazen, (2011) Gut flora metabolism of phosphatidylcholine promotes cardiovascular disease. Nature 472: 57-63. Weatherspoon-Griffin, N., G. Zhao, W. Kong, Y. Kong, Morigen, H. Andrews-Polymenis, M. McClelland & Y. Shi, (2011) The CpxR/CpxA Two-component System Up-regulates Two Tat-dependent Peptidoglycan Amidases to Confer Bacterial Resistance to Antimicrobial Peptide. J. Biol. Chem. 286: 5529-5539. 257

Weiner, J.H., P.T. Bilous, G.M. Shaw, S.P. Lubitz, L. Frost, G.H. Thomas, J.A. Cole & R.J. Turner, (1998) A Novel and Ubiquitous System for Membrane Targeting and Secretion of Cofactor-Containing Proteins. Cell 93: 93-101. Weiner, J.H., D.P. MacIsaac, R.E. Bishop & P.T. Bilous, (1988) Purification and properties of Escherichia coli dimethyl sulfoxide reductase, an iron-sulfur molybdoenzyme with broad substrate specificity. J. Bacteriol. 170: 1505-1510. Weiner, J.H., G. Shaw, R.J. Turner & C. Trieber, (1993a) The topology of the anchor subunit of dimethyl sulfoxide reductase of Escherichia coli. J. Biol. Chem. 268: 3238-3244. Weiner, J.H., G. Shaw, R.J. Turner & C.A. Trieber, (1993b) The topology of the anchor subunit of dimethyl sulfoxide reductase of Escherichia coli. J. Biol. Chem. 268: 3238-3244. Wexler, M., F. Sargent, R.L. Jack, N.R. Stanley, E.G. Bogsch, C. Robinson, B.C. Berks & T. Palmer, (2000) TatD is a cytoplasmic protein with DNase activity. J. Biol. Chem. 275: 16717- 16722. Whitaker, N., U.K. Bageshwar & S.M. Musser, (2012) Kinetics of precursor interactions with the bacterial Tat translocase detected by real-time FRET. J. Biol. Chem. WHO, (2014) Antimicrobial resistance: global report on surveillance 2014. World Health Organisation. Wickles, S., A. Singharoy, J. Andreani, S. Seemayer, L. Bischoff, O. Berninghausen, J. Soeding, K. Schulten, E.O. van der Sluis & R. Beckmann, (2014) A structural model of the active ribosome-bound membrane protein insertase YidC. eLife 3: e03035. Winstone, T.L., V.A. Tran & R.J. Turner, (2013a) The hydrophobic region of the DmsA twin- arginine leader peptide determines specificity with chaperone DmsD. Biochemistry (Wash.) 52: 7532-7541. Winstone, T.L., M.L. Workentine, K.J. Sarfo, A.J. Binding, B.D. Haslam & R.J. Turner, (2006) Physical nature of signal peptide binding to DmsD. Arch. Biochem. Biophys. 455: 89-97. Winstone, T.M.L., V.A. Tran & R.J. Turner, (2013b) The hydrophobic region of the DmsA twin- arginine leader peptide determines specificity with chaperone DmsD. Biochemistry (Wash.) 52: 7532-7541. Winter, S.E., P. Thiennimitr, M.G. Winter, B.P. Butler, D.L. Huseby, R.W. Crawford, J.M. Russell, C.L. Bevins, L.G. Adams, R.M. Tsolis, J.R. Roth & A.J. Bäumler, (2010) Gut inflammation provides a respiratory electron acceptor for Salmonella. Nature 467: 426-429. Winter, S.E., M.G. Winter, M.N. Xavier, P. Thiennimitr, V. Poon, A.M. Keestra, R.C. Laughlin, G. Gomez, J. Wu, S.D. Lawhon, I.E. Popova, S.J. Parikh, L.G. Adams, R.M. Tsolis, V.J. Stewart & A.J. Bäumler, (2013) Host-derived nitrate boosts growth of E. coli in the inflamed gut. Science 339: 708-711. Xiang, S., J. Nichols, K.V. Rajagopalan & H. Schindelin, (2001) The Crystal Structure of Escherichia coli MoeA and Its Relationship to the Multifunctional Protein Gephyrin. Structure 9: 299-310. Xiong, Y., C.-L. Santini, B. Kan, J. Xu, A. Filloux & L.-F. Wu, (2007) Expression level of heterologous tat genes is crucial for in vivo reconstitution of a functional Tat translocase in Escherichia coli. Biochimie 89: 676-685. Xu, M., S.J.W. Busby & D.F. Browning, (2009) Activation and repression at the Escherichia coli ynfEFGHI operon promoter. J. Bacteriol. 191: 3172-3176. Xu, Z., J.D. Knafels & K. Yoshino, (2000) Crystal structure of the bacterial protein export chaperone SecB. Nat. Struct. Biol. 7: 1172-1177. Yahr, T.L. & W.T. Wickner, (2001) Functional reconstitution of bacterial Tat translocation in vitro. EMBO J. 20: 2472-2479. Yamamoto, K., K. Hirao, T. Oshima, H. Aiba, R. Utsumi & A. Ishihama, (2005) Functional characterisation in vitro of all two-component signal transduction systems from Escherichia coli. J. Biol. Chem. 280: 1448-1456. Yen, M.-R., Y.-H. Tseng, E. Nguyen, L.-F. Wu & M. Saier, (2002) Sequence and phylogenetic analyses of the twin-arginine targeting (Tat) protein export system. Arch. Microbiol. 177: 441-450. 258

Young, J.C., V.R. Agashe, K. Siegers & F.U. Hartl, (2004) Pathways of chaperone-mediated protein folding in the cytosol. Nat. Rev. Mol. Cell Biol. 5: 781-791. Young, L., R.L. Jernigan & D.G. Covell, (1994) A role for surface hydrophpobicity in protein- protein recognition. Protein Sci. 3: 717-729. Yuvaniyama, P., J.N. Agar, V.L. Cash, M.K. Johnson & D.R. Dean, (2000) NifS-directed assembly of a transient [2Fe-2S] cluster within the NifU protein. Proc. Natl. Acad. Sci. U. S. A. 97: 599-604. Zakian, S., D. Lafitte, A. Vergnes, C. Pimentel, C. Sebban-Kreuzer, R. Toci, J.-B. Claude, F. Guerlesquin & A. Magalon, (2010) Basis of recognition between the NarJ chaperone and the N-terminus of the NarG subunit from Escherichia coli nitrate reductase. FEBS J. 277: 1886-1895. Zhang, B., J.C. Crack, S. Subramanian, J. Green, A.J. Thomson, N.E. Le Brun & M.K. Johnson, (2012) Reversible cycling between cysteine persulfide-ligated [2Fe-2S] and cysteine- ligated [4Fe-4S] clusters in the FNR regulatory protein. Proceedings of the National Academy of Sciences 109: 15734-15739. Zhang, B., Y. Zhang, Z.-X. Wang & Y. Zheng, (2000) The role of Mg2+ cofactor in the guanine nucleotide exchange and GTP-hydrolysis reactions of Rho family GTP-binding proteins. J. Biol. Chem. 275: 25299-25307. Zhang, J.H., G. Xiao, R.P. Gunsalus & W.L. Hubbell, (2003) Phosphorylation Triggers Domain Separation in the DNA Binding Response Regulator NarL. Biochemistry 42: 2552-2559. Zhang, W., A. Urban, H. Mihara, S. Leimkühler, T. Kurlhara & N. Esaki, (2010) IscS functions as a primary sulfur-donating enzyme by interacting specifically with MoeB and MoaD in the biosynthesis of molybdopterin in Escherichia coli. J. Biol. Chem. 285: 2302-2308. Zhang, X., N. Ge & T.A. Keiderling, (2007) Electrostatic and hydrophobic interactions governing the interaction and binding of β-lactoglobulin to membranes. Biochemistry 46: 5252- 5260. Zhang, X., S. Kung & S.-o. Shan, (2008) Demonstration of A Multi-step Mechanism for Assembly of the SRP-SRP Receptor Complex: Implications for the Catalytic Role of SRP RNA. J. Mol. Biol. 381: 581-593. Zhang, X., C. Schaffitzel, N. Ban & S.-o. Shan, (2009) Multiple conformational switches in a GTPase complex control co-translational protein targeting. Proc. Natl. Acad. Sci. U. S. A. 106: 1754-1759. Zhang, Y., Y. Hu, H. Li & C. Jin, (2014a) Structural basis for TatA oligomerization: an NMR study of Escherichia coli TatA dimeric structure. PLoS One 9: e103157-e103157. Zhang, Y., L. Wang, Y. Hu & C. Jin, (2014b) Solution structure of the TatB component of the twin-arginine translocation system. Biochim. Biophys. Acta 1838: 1881-1888. Zhou, Y., T. Ueda & M. Müller, (2014) Signal recognition particle and SecA cooperate during export of secretory proteins with highly hydrophobic signal sequences. PLoS One 9: e92994. Zhu, X., X. Zhao, W.F. Burkholder, A. Gragerov, C.M. Ogata, M.E. Gottesman & W.A. Hendrickson, (1996) Structural analysis of substrate binding by the molecular chaperone DnaK. Science 272: 1606-1614. Zouflay, S., J. Fröbel, P. Rose, T. Flecken, C. Maurer, M. Moser & M. Müller, (2012) Mapping precursor-binding site on TatC subunit of twin-arginine-specific protein translocase by site-specific photo cross-linking. J. Biol. Chem. 287: 13430-13441.