The Rava-Viaa Chaperone-Like System Targets Specific Respiratory Complexes in Escherichia Coli

The RavA-ViaA Chaperone-like System Targets Specific Respiratory Complexes in Escherichia coli

Keith Wong

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy Department of Biochemistry University of Toronto

The RavA-ViaA Chaperone-like System Targets Specific Respiratory Complexes in Escherichia coli

Keith Wong

Doctor of Philosophy

Department of Biochemistry University of Toronto

2014 Abstract

MoxR AAA+ ATPases are widespread throughout bacteria and archaea. They perform chaperone-like functions in the maturation of specific target proteins, or aid in cofactor insertion.

In Escherichia coli, the MoxR ATPase, RavA has been characterized extensively. Its X-ray structure has been solved. Importantly, RavA was found to interact and modulate the activity of the inducible lysine decarboxylase (LdcI), an important acid stress response protein.

Under aerobic condition, RavA and ViaA are expressed in early stationary-phase at low levels. Both proteins are predominantly localized to the cytoplasm, but ViaA was also found in the cell membrane. High-throughput genetic studies indicated that RavA and ViaA are linked to

Fe-S cluster assembly and specific respiratory pathways. Systematic analysis of mutant strains indicated that RavA-ViaA sensitizes cells to sublethal concentrations of aminoglycosides.

Furthermore, this phenotype was dependent on RavA’s ATPase activity, and on the presence of specific subunits of NADH:ubiquinone oxidoreductase I (Nuo Complex). Importantly, both

RavA and ViaA were found to physically interact with specific Nuo subunits.

Under anaerobic condition, the expression of ravA and viaA was revealed to be dependent on Fnr, and at least two sites for Fnr binding were identified in the ravAviaA promoter region by ii electromobility shift assay. Furthermore, ViaA was found to physically interact with FrdA, the flavin-containing subunit of the anaerobic fumarate reductase (Frd complex). Both RavA and the

Fe-S-containing subunit of the Frd complex, FrdB, appeared to modulate this interaction.

Importantly, both RavA and ViaA were found to be necessary for optimal activity of the Frd complex. This mirrors similar findings for other MoxR proteins and their respective targets.

Together, the interaction with specific Nuo and Frd subunits suggests that RavA and

ViaA are likely to modulate the anaerobic respiration of fumarate.

(286 words)

iii

Acknowledgments

This project will not have come to fruition without the guidance and input of my supervisor, Professor Walid Houry, who have been exceptional in his role as the Principle Investigator of the project and an inspirational mentor.

I would also like to give special thanks to the other two members of my Supervisory Committee, Professors John Brumell and Gil Privé, for their innovative ideas and solutions brought forth from their own disciplines that greatly enriched my knowledge and research experience.

In several experiments, both the acquisition of materials that are difficult to synthesize in- house and the use of specialized experimental protocols had been instrumental. Especially, I express my gratitude to Professor Andrew Emili (CCBR, Toronto), Professor Mohan Babu (University of Regina, SK) and Dr. Christopher Grahm (University of Regina, SK) for their invaluable contribution throughout our collaboration in analyzing the genetic interactions of ravA and viaA genes using the high-throughput E. coli Synthetic Genetic Arrays (eSGA), presented in Chapter 2 of this thesis, as well as Professor Sarath Janga (Indiana University-Purdue University, Indiana, US) for the co-expression profiling of ravA and viaA, presented in Chapter 3. Also, I would like to thank Professor Joel Weiner (University of Alberta, AB) and Professor Jan Willem de Gier (Stockholm University, Sweden), who generously supplied several important primary antibodies for Western blotting.

In addition, I am very grateful to Professor Tomoko Yamamoto (Chiba University, Japan) and Professor Akiko Takaya (Chiba University, Japan) for their efforts in our collaborative analysis on RavA-ViaA functionality in Salmonella typhimurium, and Dr. Yoshiharu Sato (Chiba University, Japan) for his aid in the construction and computational analysis of the ViaA structural model during his stay here in Toronto. Especially, I greatly appreciate their extraordinary hospitality during my stay in Sapporo, Japan for the IUMS Conference in 2011.

To describe this research project as challenging is perhaps an understatement, and it would have been that much more difficult without all the people who have lent considerable academic and emotional support throughout my time at the Houry Lab. In particular, I would like to express my deepest appreciation to Dr. Jamie Snider, the pioneer of the RavA project, for his iv guidance in developing my experimental techniques and scientific thinking, and for being a great friend. I would also give special mentions to Mr. Romain Favier and Ms. Shirin Shahsavand for their contributions to the project during their stay at the lab as project students.

I also want to thank Dr. Yoshito Kakihara for indulging me in our daily exchanges on many interesting subjects, from science, politics, hobbies to modern Japanese visual culture, as well as Ms. Elisa Leung, our research technician, and Dr. Jennifer Huen, for their continuous support in scientific matters and for supplying the lab with delicious, homemade culinary treats. Last but not least, I would like to express my deepest appreciation to Dr. Angela Yu, Dr. Usheer Kanjee, Dr. Majida El-bakkouri, Dr. Philip Wong, Mr. Andre Pow, Mr. Kai-yin Liu, Mr. Deji Ologbenla, Ms. Nardin Nano, Mr. Liang Zhao, Mr. Kamran Rizzolo, Mr. Vaibhav Bhandari, and all members of the Houry Lab, past and present, for being an important part of my experience as a graduate student, which I will treasure always.

Sincerely,

Keith Wong

December, 2013

Table of Contents

Acknowledgments ...... iv

Table of Contents ...... vi

List of Tables ...... x

List of Figures ...... xi

Chapter 1 General Introduction on AAA+ ATPases & Bacterial Respiration ...... 1

1.1 Preface ...... 2

1.2 Conserved Structural & Functional Features of the AAA+ Domain ...... 2

1.2.1 The Walker A and Walker B Motifs of the AAA+ Domain ...... 3

1.2.2 Other Important Functional Motifs of the AAA+ Domain ...... 4

1.3 Classification of AAA+ ATPases within the Hierarchy of P-loop NTPases ...... 5

1.3.1 KG and ASCE: The Two Major Groups of P-loop NTPases ...... 9

1.3.2 AAA+ ATPases and their Classification ...... 10

1.4 Functional & Structural Characteristics of the MoxR Family of AAA+ ATPases ...... 12

1.4.1 General Introduction of the MoxR Proteins ...... 12

1.4.2 Recent Functional Characterization of MoxR Proteins ...... 14

1.4.3 Structural Characterization of MoxR Proteins ...... 21

1.4.4 Summary on the Functional and Structural Characteristics of the MoxR AAA+ ATPases ...... 31

1.5 Introduction to the Aerobic and Anaerobic Respiration of E. coli ...... 32

1.5.1 The Respiratory Electron Transport Chains of E. coli ...... 32

1.5.2 Biophysical and Functional Characteristics of the NADH:ubiquinone Oxidoreductase I and the Fumarate Reductase from E. coli ...... 44

1.5.3 Summary on the Aerobic and Anaerobic Respiration in E. coli ...... 50

Chapter 2 The MoxR ATPase RavA and its Cofactor ViaA Interact with the NADH:Ubiquinone Oxidoreductase I in E. coli ...... 51

2.1 Chapter Summary ...... 52

2.2 Introduction ...... 52

2.3 Materials and Methods ...... 54

2.3.1 Bacterial strains and plasmids used ...... 54

2.3.2 Quantification of RavA and ViaA levels in cells ...... 57

2.3.3 Subcellular localization of RavA and ViaA ...... 58

2.3.4 Microarray experiments and data analysis ...... 58

2.3.5 E. coli Synthetic Genetic Array (eSGA) analysis ...... 59

2.3.6 Growth of E. coli MG1655 in cultures containing sublethal dosages of different antibiotics ...... 60

2.3.7 Analysis of intracellular oxidative stress by DHR fluorescence ...... 60

2.3.8 Suppression mutation analysis to identify direct functional targets of RavA and ViaA ...... 61

2.3.9 Identifying the physical interactors of RavA and ViaA by immunoprecipitation ...... 61

2.4 Results ...... 62

2.4.1 Expression and localization of RavA and ViaA ...... 62

2.4.2 The function of RavA and ViaA is linked to Fe-S cluster assembly and specific respiratory pathways ...... 64

2.4.3 RavA and ViaA sensitize E. coli to aminoglycosides ...... 70

2.4.4 The RavA-ViaA phenotype is abolished by reduced glutathione and 2,2'-dipyridyl ...... 72

2.4.5 RavA-ViaA targets specific Nuo subunits and other respiratory proteins in sensitizing E. coli to kanamycin ...... 75

2.4.6 RavA and ViaA interact with specific Nuo subunits ...... 78

2.5 Discussion ...... 81

Chapter 3 The MoxR AAA+ ATPase RavA and its VWA Cofactor ViaA Interacts and Modulates the Activity of the Fumarate Reductase Complex during Anaerobiosis in E. coli ...... 83

3.1 Chapter Summary ...... 84 vii

3.2 Introduction ...... 84

3.3 Materials and Methods ...... 86

3.3.1 Bacterial strains and plasmids used ...... 86

3.3.2 Co-expression profiling of ravA and viaA in E. coli ...... 89

3.3.3 Expression analysis of RavA and ViaA in E. coli under aerobic and anaerobic conditions ...... 89

3.3.4 Electromobility shift assay ...... 90

3.3.5 Immunoprecipitation by SPA-tagged bait proteins ...... 91

3.3.6 Western Blotting ...... 91

3.3.7 Fumarate reductase activity assay ...... 92

3.4 Results ...... 93

3.4.1 ravA and viaA display similar co-expression profiles as those of the Fnr-inducible genes ...... 93

3.4.2 Fnr enhances the expression of RavA and ViaA during anaerobiosis ...... 95

3.4.3 Identification of potential Fnr-binding sites in the native promoter region of ravAviaA ...... 97

3.4.4 ViaA physically interacts with FrdA in anaerobically growing cells ...... 100

3.4.5 RavA-ViaA bind free FrdA ...... 102

3.4.6 RavA modulates the binding of ViaA to FrdA in an ATP-dependent manner 102

3.4.7 RavA-ViaA stimulate the activity of the Frd complex in E. coli ...... 104

3.5 Discussion ...... 106

Chapter 4 General Conclusion & Future Directions ...... 107

4.1 General Conclusion of Thesis ...... 108

4.1.1 Interaction of RavA-ViaA with the target subunits occurs prior to the assembly of the full complex ...... 108

4.1.2 RavA-ViaA facilitates the assembly of the full complex ...... 109

4.2 Future Directions ...... 111

viii

4.2.1 In-depth characterization of the interactions between RavA-ViaA and the subunits of the Nuo and Frd respiratory complexes ...... 111

4.2.2 Re-examination of the RavA-ViaA-induced sensitization to aminoglycosides independent of ROS ...... 114

4.2.3 Investigation of a Potential Role of RavA-ViaA in Stress-induced Mutagenesis ...... 115

4.3 Closing Remarks ...... 116

References ...... 117

Appendix A ...... 133

Appendix B ...... 140

Appendix C ...... 143

Appendix D ...... 149

List of Tables

TABLE 1.1. Recent Functional Characterization of MoxR Proteins ...... 14

TABLE 2.1. List of bacterial strains and plasmids used in this study ...... 55

TABLE 2.2. Suppression mutation analysis for the RavA-ViaA overexpression- induced sensitization to kanamycin in E. coli MG1655 ...... 77

TABLE 3.1. List of bacterial strains and plasmids used ...... 87

TABLE 3.2. List of primers used for generating the DNA substrates used in EMSA ...... 90

List of Figures

Figure 1.1. The canonical AAA+ Domain and its nucleotide-binding pocket ...... 3

Figure 1.2. Hexameric structures of AAA+ ATPases from various species ...... 6

Figure 1.3. Classification hierarchy of P-loop NTPases ...... 7

Figure 1.4. Structures of the Catalytic Cores of KG and ASCE NTPases, and representatives for the different hierarchical groups most relevant to RavA ...... 8

Figure 1.5. Structures of AAA+ domains of Helix-2 Insert Clade proteins ...... 12

Figure 1.6. Gene organization of different MoxR-containing gene clusters ...... 16

Figure 1.7. Nucleotide binding site in MoxR proteins ...... 23

Figure 1.8. Multiple sequence alignment of the AAA+ domain of RavA, CHU_153, and BchI ...... 26

Figure 1.9. The RavA and BchI hexameric assembly ...... 28

Figure 1.10. Structure of RavA from E. coli ...... 28

Figure 1.11. Structure of the RavA-LdcI complex ...... 30

Figure 1.12. Schematic representation of primary dehydrogenases, terminal reductases and specific transporters involved in the aerobic and anaerobic respiratory pathways of E. coli ...... 34

Figure 1.13. Chemical structures of the major quinones utilized in E. coli respiratory processes ...... 39

Figure 1.14. Schematic representation of signaling and regulatory pathways for the induction of respiratory proteins in E. coli in response to external stimuli ...... 43

Figure 1.15. Structures of the NADH:ubiquinone oxidoreductase I (NuoA-N) and fumarate reductase (FrdA-D), and schematic representation of their biochemical activities ...... 46

Figure 2.1. Expression and localization of RavA and ViaA in E. coli MG1655 ...... 63

Figure 2.2. Expression levels of RavA and ViaA in various strain backgrounds used in this study ...... 65

Figure 2.3. Genomic organization of genes relevant to Fe-S clusters assembly or bacterial respiration showing statistically significant changes in the microarray experiments ...... 67

Figure 2.4. Schematic representation of genes showing significant changes in transcript levels as a result of the deletion or overexpression of RavA/ViaA ...... 67

Figure 2.5. Genetic interactions between ravA/viaA and genes functionally relevant to Fe-S clusters assembly and bacterial respiration ...... 69

Figure 2.6. Growth profiles of cells in the presence of sublethal concentrations of aminoglycosides ...... 71

Figure 2.7. Effects of glutathione and 2,2'-dipyridyl on the growth profiles of cells in the presence of sublethal concentrations of kanamycin ...... 74

Figure 2.8. Growth profiles of selected single-gene knockouts ...... 76

Figure 2.9. Physical interactions between RavA and ViaA with specific subunits of the Nuo complex under different growth conditions ...... 80

Figure 2.10. Immunoprecipitation experiments on WT DY330 and strains expressing SPA-tagged NuoF or NuoCD ...... 80

Figure 3.1. Co-expression profiles of ravA and viaA ...... 94

xii

Figure 3.2. Expression analysis of RavA and ViaA in E. coli MG1655 WT, Δfnr::kanR and ΔrpoS::kanR by Western blotting ...... 96

Figure 3.3. Genomic DNA sequence in E. coli K-12 MG1655 corresponding to the ravAviaA promoter region ...... 98

Figure 3.4. Electromobility shift assays (EMSA) on the ravAviaA promoter region

and its variants using the mutant transcriptional regulator FnrD154A ...... 99

Figure 3.5. Physical interactors of RavA and ViaA by SPA-tag immunoprecipitation ...... 101

Figure 3.6. RavA and FrdB modulate the interaction between ViaA and FrdA-SPA ...... 103

Figure 3.7. RavA-ViaA enhance fumarate reductase activity in anaerobically grown E. coli MG1655 ...... 105

Figure 4.1. Hypothetical Role of RavA-ViaA as Molecular Chaperone for FrdA in the Assembly and Maturation of the Fumarate Reductase Complex ...... 110

xiii 1

Chapter 1 General Introduction on AAA+ ATPases & Bacterial Respiration

Section 1.4 is a reprint with slight modifications, with the authors’ permission, of the following published review: Wong, K. S. and Houry, W. A. (2012) Novel structural and functional insights into the MoxR family of AAA+ ATPases. J Struct Biol. 179(2): 211-21. [PMID: 22491058]

1 General Introduction on AAA+ ATPases & Bacterial Respiration 1.1 Preface

The primary objective in my research is the functional characterization of the AAA+ ATPase RavA, a member of the MoxR family, and the associated VWA protein, ViaA, in Escherichia coli. To facilitate presentation and discussion of the experimental data, a general introduction on AAA+ ATPases is provided to cover the many key concepts that are important in this field of research. These include the structural and functional features that are characteristic of the AAA+ ATPases, as well as the current classification system used for these proteins. Next, the MoxR family will be discussed in detail from both a functional and structural perspective. Importantly, RavA was found to physically and functionally interact with specific subunits of the respiratory complexes NADH:ubiquinone oxidoreductase I (NuoA-N) and fumarate reductase (FrdA-D). As such, an introduction on the respiratory processes in E. coli is also included in this chapter, with a focus on the structural and functional aspects of the major anaerobic electron transport chain utilized by the cell in the absence of nitrate, which is made up of the Nuo and Frd respiratory complexes.

1.2 Conserved Structural & Functional Features of the AAA+ Domain

AAA+ ATPases constitute a large and functionally diverse class of proteins that spreads across all kingdoms of life. “AAA” stands for “ATPases Associated with Diverse Cellular Activities”, and the “+” sign signifies an expansion from the original AAA classification to include additional members into the class (Neuwald et al., 1999; Hanson & Whiteheart, 2005; Snider & Houry, 2008; Snider et al., 2008). The definitive characteristic of AAA+ ATPases is the presence of at least one AAA+ domain. This domain is essential for AAA+ ATPases in harnessing the energy released during hydrolysis of nucleotides, usually ATP, for their diverse biomolecular activities (Ogura & Wilkinson, 2001; Snider & Houry, 2008; Snider et al., 2008). Structurally, the basic AAA+ domain consists of two subdomains – the αβα subdomain and the all-α subdomain (Fig. 1.1A).

Figure 1.1. The canonical AAA+ Domain and its nucleotide-binding pocket

(A) X-ray structure of the AAA+ Domain in RFC1 (Replication Factor C) from Saccharomyces cerevisiae (PDB ID 1SXJ). The αβα subdomain is coloured in light blue, and the all-α subdomain is coloured in light orange. Important functional motifs are coloured as indicated by the legend below panel (B). The bound ATPγS is shown in stick format and coloured in black. The catalytic Mg2+ ion is shown as a yellow sphere.

(B) The nucleotide-binding pocket of the AAA+ Domain in RFC1. Conserved / key residues from each functional motif are shown in stick format in their corresponding colour as shown in the legend below. As before, the bound ATPγS is shown in stick format and coloured in black, and the catalytic Mg2+ ion is shown as a yellow sphere.

______

1.2.1 The Walker A and Walker B Motifs of the AAA+ Domain

The αβα subdomain consists of a core β sheet sandwiched between two rows of α helices, and serves as the catalytic core of the AAA+ domain (Fig. 1.1A). It contains many sequence motifs that are highly conserved and crucial for nucleotide hydrolysis. Among the most important

functional motifs is the Walker A motif (Fig. 1.1A; coloured in red). It has the amino acid sequence GxxxxGK[T/S], where x can be any amino acid residue (Hanson & Whiteheart, 2005). The conserved lysine residue (K359 in S. cerevisiae RFC1; shown in Fig. 1.1B) is crucial for the binding of nucleotides (Walker et al., 1982; Saraste et al., 1990). Another important sequence motif within the αβα subdomain is known as the Walker B motif (ΨΨΨΨDE, where Ψ can be any hydrophobic residue) (Hanson & Whiteheart, 2005). Both the conserved aspartate and glutamate residues (D424 and E425, respectively, in S. cerevisiae RFC1; shown in Fig. 1.1B) are important for the hydrolysis of bound nucleotides. The aspartate residue coordinates the Mg2+ cation required for the process, while the glutamate residue is hypothesized to activate water for the hydrolytic reaction (Iyer et al., 2004).

1.2.2 Other Important Functional Motifs of the AAA+ Domain

Aside from Walker A and Walker B, the AAA+ domain also contains other conserved motifs that are important of its function. These include Box II (Fig. 1.1A; coloured in magenta), which is located within the first α-helix of the AAA+ domain, upstream of the Walker A motif in sequence (Snider & Houry, 2008). Box II is hypothesized to be involved in adenine recognition based on its proximity to the adenine ring of a bound ATP, although its amino acid residues are not as highly conserved as the Walker motifs (Neuwald et al., 1999). Sensor 1 is located after the Walker B motif. Physically, it is mapped to the β-strand adjacent to the Walker B strand within the catalytic core (Fig. 1.1A; coloured in green). This motif has a conserved asparagine residue (N456 in S. cerevisiae RFC1; shown in Fig. 1.1B), or in some cases, theronine, serine or histidine, that either directly interacts with the γ-phosphate of the bound nucleotide or indirectly via a water molecule. As its name implies, Sensor 1 acts as a sensor for nucleotide binding and/or hydrolysis (Guenther et al., 1997; Karata et al., 1999).

Downstream of Sensor 1 is another conserved motif known as the Arginine Finger (Arg Finger) (Fig. 1.1A; coloured in blue). The Arg Finger interacts with the γ-phosphate of the nucleotide bound to the AAA+ domain of the neighbouring subunit, and has been shown to be crucial for nucleotide hydrolysis and the communication between subunits, although it is not required for nucleotide binding (Karata et al., 1999; Rombel et al., 1999; Davey et al., 2003; Johnson & O'Donnell, 2003). The last major motif that follows the Arg Finger is Sensor 2, which is mapped to the all-α subdomain (Fig. 1.1A; coloured in purple). Sensor 2 has a highly

conserved arginine residue (R516 in S. cerevisiae RFC1; shown in Fig. 1.1B), which has been attributed with diverse functions, ranging from nucleotide binding, hydrolysis, sensing, inter- subunit interaction, to coordinating the movement between the αβα and all-α subdomains (Snider & Houry, 2008). The spatial position and orientation of these motifs defines the nucleotide-binding pocket within the AAA+ domain (Fig. 1.1B). Importantly, the occurrence of both Box II and the Arg Finger on the opposite side of the other functional motifs requires the interaction between neighbouring subunits to occur in a particular manner to provide a complete nucleotide-binding pocket (Fig. 1.1B). This in turn results in all AAA+ ATPases adopting a ring- like oligomeric structure, with the hexameric state being the most common, (Fig. 1.2) and is considered to be the biological unit of AAA+ ATPases in vivo (Ogura & Wilkinson, 2001; Hanson & Whiteheart, 2005; Snider & Houry, 2008).

Next, the current classification system adopted for AAA+ ATPases will be discussed. Emphasis is placed on the families, groups and clades that are the most relevant to the MoxR family of AAA+ ATPases.

1.3 Classification of AAA+ ATPases within the Hierarchy of P- loop NTPases

The AAA+ ATPases constitute one of six major classes of ATPases within the larger hierarchy of P-loop NTPases (Fig. 1.3). P-loop, short for phosphate-binding loop, refers to a flexible region downstream of the first α-helix within a catalytic core that is composed of a parallel β-sheet sandwiched between two rows of α-helices (Fig. 1.4A and B). This is where the highly conserved Walker A motif is located (Fig. 1.4; in red). The Walker B motif, which is also critical for nucleotide hydrolysis, is also found within the catalytic core of all P-loop NTPases (Fig. 1.4; in orange).

Figure 1.2. Hexameric structures of AAA+ ATPases from various species

X-ray structures of the hexameric form of (A) HslU of the HslUV protease complex from Escherichia coli K-12 (PDB ID 1DO0), (B) TIP49b DNA helicase from Homo sapiens (PDB ID 3UK6), and (C) LTag (Large T-antigen helicase) from Simian Virus 40 (SV40) (PDB ID 1SVM). Individual subunits are coloured as shown. Bound nucleotides (ATP or ADP), if present, are shown as sticks.

______

Figure 1.3. Classification hierarchy of P-loop NTPases

Shown are the major groups, clades, families, etc. that branch off from the route for RavA (shown as gray boxes) at each level of the classification hierarchy of P-loop NTPases. The STAND Group is shown with a dotted box due to uncertainty in its definition as a stand-alone group. For the PACTT Group, any proteins that do not belong to either the HCL or the Helix-2 Insert Clades are categorized as “Other”.

______

Figure 1.4. Structures of the Catalytic Cores of KG and ASCE NTPases, and representatives for the different hierarchical groups most relevant to RavA

(A)-(B) Catalytic core structures of (A) KG (Kinase-GTPase) Group, represented here with the core domain of Ffh from Thermus aquaticus (PDB ID 2C03), and (B) ASCE (Additional Strand Catalytic E)

Group, represented here with the αβα subdomain of RFC1 from S. cerevisiae (PDB ID 1SXJ). The Walker A motif (P-loop) and the Walker B motif are coloured in red and orange, respectively. The core β- strand preceding the Walker A motif is coloured in magenta. The core β-strand used to distinguish KG and ASCE NTPases is coloured in green (see main text for details). The other core β-strands are coloured in blue for visual clarification.

(C)-(E) AAA+ domains of (C) RFC1 from S. cerevisiae as an example of the canonical AAA+ domain, (D) ClpA-D2 (i.e. 2nd AAA+ domain of ClpA) from E. coli K-12 (PDB ID 1R6B) to represent the PACTT Group, and (E) PspF from E. coli K-12 (PDB ID 2C9C) to represent the Helix-2 Insert Clade. As in Fig. 1.1, the αβα subdomain is coloured in light blue, and the all-α subdomain is coloured in light orange. The Walker A and Walker B motifs are coloured in red and orange, respectively. Pre-sensor 1 β-hairpins are coloured in purple. Helix-2 inserts are coloured in yellow. Dotted lines indicate flexible regions that were not resolved in the X-ray structure.

______

1.3.1 KG and ASCE: The Two Major Groups of P-loop NTPases

P-loop NTPases are subdivided into defined groups, classes and clades arranged in a hierarchy based on unique sequence and structural features (Iyer et al., 2004; Ammelburg et al., 2006; Snider & Houry, 2008; Snider et al., 2008) (Fig. 1.3). There are two major groups of P-loop NTPases – the Kinase-GTPase (KG) group and the Additional Strand Catalytic E (ASCE) group. The primary structural feature that sets apart ASCE and KG proteins lies in the organization of their catalytic core. For proteins in the KG group, the β-strand that precedes the P-loop (Fig. 1.4A; β-strand in magenta, and the P-loop in red) is adjacent to the β-strand that carries the Walker B motif (Fig. 1.4A; the Walker B motif in orange). On the other hand, the catalytic core of ASCE proteins is structurally different in that the β-strand preceding the P-loop (Fig. 1.4B; β- strand in magenta, and the P-loop in red) is separated from the Walker B strand (Fig. 1.4B; in orange) by a third, parallel β-strand (Fig. 1.4B; in green). The equivalent strand within the KG catalytic core is positioned on the opposite side of the P-loop-preceding strand (Fig. 1.4A; in green). A second structural feature that distinguishes ASCE proteins from KG proteins is the presence of a highly conserved and catalytically important glutamate residue immediately downstream of the conserved aspartate residue in the Walker B motif (Fig. 1.4B; in orange).

The ASCE proteins are subdivided into six classes – KAP ATPases, RecA/F1 ATPases, ABC ATPases, SF1/2 ATPases, VirD/PilT ATPases and AAA+ ATPases (Snider & Houry, 2008; Snider et al., 2008). Our discussion will now focus on the AAA+ ATPases.

1.3.2 AAA+ ATPases and their Classification

An important structural characteristic of the AAA+ ATPases is the presence of the globular all-α subdomain (Fig. 1.4C; in yellow), which together with the αβα subdomain (i.e. the ASCE catalytic core) (Fig. 1.4C; in blue) makes up the AAA+ domain. As the name implies, the all-α subdomain consists primarily or entirely of α-helices. In a typical AAA+ domain, which is represented here by RFC1 from Saccharomyces cerevisiae (Fig. 1.4C), the all-α subdomain is located “on top” of αβα subdomain when viewed as shown. This particular spatial arrangement of the subdomains necessitates the formation of the ATP-binding pocket (Fig. 1.1 and 1.2).

Based on structural variations in the AAA+ domain, the AAA+ ATPases can be further subdivided into at least four groups – the Extended AAA Group, the HEC Group, the ExeA Group, and the PACTT Group (Fig. 1.3). Among them, members of the PACTT (Protease, Chelatase, Transcriptional Activators and Transport) Group are characterized by the presence of a β-hairpin that precedes the core β-strand carrying the Sensor 1 motif (Fig. 1.4D; coloured in purple), known as the pre-sensor 1 β-hairpin (Snider & Houry, 2008; Snider et al., 2008).

The PACTT Group consists of two defined clades – the HCL Clade and the Helix-2 Insert Clade – as well as a collection of other proteins that do not belong to either clade (Fig. 1.3). Of the two clades, the Helix-2 Insert Clade is characterized by the presence of an insertion that disrupts the second α-helix of the AAA+ domain (Fig. 1.4E; coloured in yellow), known as the helix-2 insert (H2-insert). Based on known structures of Helix-2 Insert Clade members, the H2-insert typically adopts a simple structure, although variations exist among different members of the clade. These range from a simple flexible loop in the AAA+ domains of PspF (Fig. 1.5A) (Rappas et al., 2006) and RavA (Fig. 1.5B) (El Bakkouri et al., 2010) from E. coli, a β-hairpin in CHU_0153 from Cytophaga hutchinsonii (Fig. 1.5C) (PDB ID 2R44, Joint Center for Structural Genomics) and MCM from Sulfolobus solfataricus (Fig. 1.5E) (Brewster et al., 2008), to the more complex strand-loop-helix-loop-strand motif in BchI from Rhodobacter capsulatus (Fig. 1.5D) (Fodje et al., 2001).

Figure 1.5. Structures of AAA+ domains of Helix-2 Insert Clade proteins

Shown are the X-ray structures of the AAA+ domain of (A) PspF from E. coli K-12 (PDB ID 2C9C), (B) RavA from E. coli K-12 (PDB ID 3NBX), (C) CHU_153 (YP_676785.1) from C. hutchinsonii ATCC 33406 (PDB ID 2R44), (D) BchI from R. capsulatus (PDB ID 1G8P), and (E) MCM from Sulfolobus solfataricus (PDB ID 3F9V). The αβα subdomains are coloured in light blue and the all-α subdomains in light orange. Linker regions of RavA, CHU_153, BchI and MCM are coloured in green. The Walker A and Walker B motifs are highlighted in red and orange, respectively. As in Fig. 1.4, Pre-sensor 1 β-hairpins are coloured in purple, and Helix-2 inserts (H2-inserts) are coloured in yellow. The α1-β2-β-hairpin unique to BchI (Snider et al., 2008) is highlighted in pink. α1 helices in RavA, CHU_153 and MCM are indicated with black arrows as shown. Dotted lines indicate unstructured regions that are not resolved in the X-ray structures of the proteins. Both the front view (with nucleotide binding site facing the reader), shown on the left in each panel, and the back view, shown on the right, of the AAA+ domains are illustrated for all five proteins. For simplicity, the bound nucleotides in RavA and HslU, as well as, the I-domain of HslU, have been omitted.

______

The Helix-2 Insert Clade is comprised of eight families – MCM, McrB/Unc-53, Midasin, Chelatase, σ54 Activator, YifB, ComM, and MoxR (Fig. 1.3) (Snider & Houry, 2008). In the following section, the MoxR family will be discussed in detail with respect to their functional and biophysical properties.

1.4 Functional & Structural Characteristics of the MoxR Family of AAA+ ATPases

The following section aims to provide an overview of the MoxR family of AAA+ ATPases from both its functional and structural perspectives, using the most recent experimental data collected in the functional characterization of several MoxR family members from multiple species, as well the structural and biophysical characterization of RavA from Escherichia coli and CHU_0153, from Cytophaga hutchinsonii.

1.4.1 General Introduction of the MoxR Proteins

MoxR proteins constitute a family of AAA+ ATPases that is widespread among bacteria and archaea. In a previous bioinformatic analysis on the amino acid sequences of 596 complete AAA+ domains of MoxR proteins, we classified MoxR proteins into 7 major subfamilies: MoxR Proper (MRP), TM0930, RavA, CbbQ/GvpN/NorQ (CGN), APE2220, PA2707, and YehL (Snider & Houry, 2006). A largely common feature of MoxR proteins is their co-occurrence with other proteins that carry the metal-binding von Willebrand factor A (VWA) domain. In most cases, these proteins are encoded in the genome immediately downstream of the gene for the

respective AAA+ protein (Snider & Houry, 2006). Although the VWA domain is better characterized in eukaryotes than in prokaryotes, proteins having this domain are generally involved in mediating protein-protein interactions (Whittaker & Hynes, 2002; Springer, 2006). The key feature of the VWA domain is the presence of the non-contiguous MIDAS (Metal Ion- Dependent Adhesion Site) motif, which provides the binding site of a single divalent cation (usually Mg2+) and is crucial for the domain’s function (Whittaker & Hynes, 2002; Springer, 2006).

Despite the diversity and widespread occurrence of MoxR proteins, their functional characterization remains relatively limited to genetic studies, primarily, of members of the MRP and CGN subfamilies [see reference (Snider & Houry, 2006)]. Based on these studies, MoxR proteins are thought to have a chaperone-like activity, which is important for the maturation or assembly of specific protein complexes that are involved in dedicated metabolic pathways. MoxR proteins are also thought to be involved in cofactor insertion into specific proteins. For example, in Paracoccus denitrificans, MoxR (MRP subfamily) is important for the maturation of methanol dehydrogenase (MDH), although it has no effect on the biosynthesis of MDH itself, its cofactor pyrroloquinoline quinone (PQQ), or the associated electron acceptor cytochrome c (van Spanning et al., 1991). These findings also apply to Methylobacterium extorquens MoxR, in which case the protein is implicated in the insertion of Ca2+ into MDH during the enzyme’s maturation process (Toyama et al., 1998). Similarly, NirQ and NorQ (CGN subfamily) have been shown experimentally to be important for the activation of the nitric oxide reductase enzyme involved in denitrification, but without any effects on the expression of the denitrification proteins (Jungst & Zumft, 1992; Arai et al., 1999). This has been illustrated for NirQ/NorQ in Pseudomonas aeruginosa, Pseudomonas stutzeri, Paracoccus denitrificans, and Rhodobacter sphaeroides 2.4.3 (Jungst & Zumft, 1992; de Boer et al., 1996; Bartnikas et al., 1997; Arai et al., 1999).

In the following sections, our discussion will focus on the most recent experimental characterization of known and newly identified members of the MoxR family. Importantly, the recently solved crystal structures of the Escherichia coli RavA and Cytophaga hutchinsonii CHU_0153, which belong to the RavA and MRP subfamilies, respectively, provide detailed structural insights into MoxR proteins that are important in identifying distinctive features that are potentially unique to MoxR family proteins.

1.4.2 Recent Functional Characterization of MoxR Proteins

In the following section, our discussion will focus on the most recent functional characterization of both known and newly identified MoxR proteins. The newly characterized functions of these proteins are summarized in Table 1.1.

TABLE 1.1. Recent Functional Characterization of MoxR Proteins

MoxR Subfamily Protein Organism Classification Cellular/Molecular Functions Reference

RL3499 Rhizobium MRP • Cell envelope development (Vanderlinde et leguminosarum al., 2011) Biovar viciae • Cell morphology • Stress tolerance • Bacterium-host symbiosis

FTL_0200 Francisella MRP • Acid & oxidative stress (Dieppedale et al., tularensis resistance 2011) • Bacterial pathogenesis

RavA Escherichia coli RavA • Prevents inhibition of the (El Bakkouri et K-12 inducible lysine decarboxylase al., 2010) LdcI • Role in acid stress & stringent responses

p618 Acidianus two- RavA • DNA binding (Scheele et al., tailed virus 2011) • Possible role in extracellular viral tail development

CoxD Oligotropha APE2220 • Partial unfolding of apo-CO (Pelzmann et al., carboxidovorans dehydrogenase 2009) OM5 • CO dehydrogenase maturation

1.4.2.1 Roles of MRP in Stress Tolerance, Cell Development, and Bacterial Pathogenesis

1.4.2.1.1 RL3499 from Rhizobium leguminosarum

RL3499 encodes a newly identified MoxR protein of the MRP subfamily in Rhizobium leguminosarum Biovar viciae (Vanderlinde et al., 2011), a soil-dwelling Gram-negative

bacterium that forms a symbiotic relationship with leguminous plants (Newton, 2000; Oldroyd & Downie, 2008). Using a general screen utilizing 1750 individual transposants, Vanderlinde and coworkers (Vanderlinde et al., 2011) identified RL3499, located in the gene cluster RL3499 – RL3502, as being essential for R. leguminosarum growth in rich medium. The gene cluster has the expected organization of MRP subfamily members (Snider & Houry, 2006) (Fig. 1.6A): RL3499 is followed immediately downstream by RL3500, which encodes a protein that contains a domain of unknown function, DUF58; RL3500 is in turn followed downstream by RL3501, which encodes a 937-residues-long protein with a predicted N-terminal double-transmembrane domain, a VWA domain in the middle, and a C-terminal Class-I Glutamine amidotransferase (GATase) domain; RL3502 is a putative transmembrane protein. The gene cluster RL3499 – RL3502 was found to be important for growth of R. leguminosarum in the presence of reagents such as SDS, erythromycin (a hydrophobic antibiotic), and polymyxin B, which weakens the cell envelope (Vanderlinde et al., 2011). The gene cluster was also found to be important for growth at alkaline pH (Vanderlinde et al., 2011), a condition that has been reported to reduce the rate of peptidoglycan cross-linking in Bacillus sp. C-125 (Aono & Sanada, 1994). RL3499 – RL3502 was required for growth in rich or minimal media in the presence of glycine or peptides, which are known to increase cell envelope permeability (Hammes et al., 1973; Li et al., 2009). Accordingly, Ca2+ and Mg2+ that stabilize cell envelope integrity (Vaara, 1992; Dominguez, 2004) can reverse the growth defects of mutants lacking RL3499 – RL3502 to various degrees, depending on the exact growth conditions (Vanderlinde et al., 2011). Importantly, RL3499 – RL3502 was found to be crucial for maintaining normal cell morphology, as well as, symbiosis with the host pea plant Pisum sativum (Vanderlinde et al., 2011). These phenotypes thus highlight a major role of RL3499 in maintaining cell envelope integrity, which in turn is important for the symbiosis between R. leguminosarum and the pea plant (Table 1.1). However, the protein target of RL3499 activity is not known.

Figure 1.6. Gene organization of different MoxR-containing gene clusters

Shown are schematics for the MoxR-containing gene clusters found in (A) Rhizobium leguminosarum Biovar viciae, (B) Francisella tularensis, (C) Escherichia coli K-12, (D) Acidianus two-tailed virus (ATV), and (E) Oligotropha carboxidovorans OM5. Gene clusters containing MoxR proteins from the same subfamily are grouped together as shown. All genes encoding for the different MoxR proteins, VWA proteins, and other proteins that are characteristic of the different MoxR subfamilies are coloured and their key domains/motifs are identified. All other genes are coloured in gray. For each gene cluster, genes are oriented relative to their respective MoxR-encoding gene.

______

1.4.2.1.2 FTL_0200 from Francisella tularensis

Another recently characterized MRP protein is FTL_0200 from Francisella tularensis (Dieppedale et al., 2011), a Gram-negative pathogen that causes tularemia in humans and many animal species (Santic et al., 2010). The FTL_0200 gene is located the furthest upstream in the gene cluster FTL_0200 – FTL_0206 (Dieppedale et al., 2011), which has the expected gene organization for MRP subfamily members (Snider & Houry, 2006) (Fig. 1.6B). FTL_0200 is

followed immediately downstream by FTL_0201, which contains sequential motifs for both VWA and DUF58 domains (Dieppedale et al., 2011). Two additional VWA proteins are encoded by FTL_0203 and FTL_0204, respectively, with FTL_0204 also carrying motifs for the Tetratricopeptide Repeat (TPR) domain. FTL_0205, which encodes a second TPR protein, follows immediately downstream. The last gene in this cluster is FTL_0206 (Dieppedale et al., 2011), which encodes a homologue of an oxygen tolerance-associated protein BatD from the obligate anaerobe Bacteroides fragilis (Tang et al., 1999). Promoter region analysis and gene expression data suggest that the expression of the FTL_0200 – FTL_0206 gene cluster is likely inducible by the heat shock-associated transcriptional regulator σ32 (Dieppedale et al., 2011).

Although insertional mutations to FTL_0200 (MRP gene), FTL_0205 (TPR gene) and FTL_0206 show no growth defects in liquid media, all three mutants show significantly impaired ability to infect THP1 (human acute monocytic leukemia cell line) and J774 (murine macrophage cell line) cells compared to wild type (Dieppedale et al., 2011). The three mutants also exhibit reduced infectivity in the mouse model (Dieppedale et al., 2011). Further analysis using a FTL_0200 deletion mutant shows that the MoxR protein is important for tolerance to oxidative stress, acid stress, heat stress, and the presence of protein denaturants such as ethanol and SDS (Dieppedale et al., 2011). The combined data thus strongly suggest a role for FTL_0200 in multiple stress tolerance pathways, as well as, in Francisella pathogenesis (Table 1.1).

Taken together, both RL3499 and FTL_0200 provide the first examples that MRP proteins are also associated with a variety of important cellular functions other than their traditional role in metabolism. These range from stress tolerance, cell morphology and development, bacterium-host interaction, to pathogenesis.

1.4.2.2 Roles of RavA in Acid Stress, Stringent Response, and Viral Tail Formation

1.4.2.2.1 RavA from Escherichia coli

RavA from Escherichia coli K-12 is the first member of the RavA subfamily that has been extensively characterized (Snider et al., 2006; El Bakkouri et al., 2010). Its corresponding VWA protein is known as ViaA (Snider et al., 2006; Snider & Houry, 2006). Like other members of the RavA subfamily, the ravA gene is encoded immediately upstream of the viaA gene in the same operon (Snider et al., 2006; Snider & Houry, 2006) (Fig. 1.6C), which is inducible by the

alternative transcriptional regulator σS in aerobically grown cells during stationary phase (Snider et al., 2006). In the presence of ATP or other nucleotides, RavA forms a hexamer via its AAA+ domain (see below) (Snider et al., 2006; El Bakkouri et al., 2010). The ATPase activity of RavA is optimal at neutral pH and is comparable to other well-characterized AAA+ ATPases such as HslU, NtrC, and Lon protease (Snider et al., 2006). The interaction of RavA with ViaA has been observed in vitro, which increases RavA’s ATPase activity by about 2-fold (Snider et al., 2006). Importantly, RavA interacts strongly and specifically with the inducible lysine decarboxylase LdcI (Snider et al., 2006; El Bakkouri et al., 2010), which is an important acid stress response enzyme (Kanjee et al., 2011). This interaction, which requires the LARA (LdcI associating domain of RavA) domain in RavA (El Bakkouri et al., 2010), results in a large cage-like structure consisting of two LdcI decamers and up to five RavA hexamers (Snider et al., 2006) (further discussed below). The RavA-LdcI interaction results in a 1.4-fold increase in RavA’s ATPase activity (Snider et al., 2006).

Although the function of E. coli RavA is not well established, RavA has been shown to de-inhibit LdcI activity by preventing the binding of a potent LdcI inhibitor, the stringent response alarmone, guanosine 3′, 5′-bis(diphosphate) (ppGpp) (Kanjee et al., 2011), both in vitro and in vivo (El Bakkouri et al., 2010) (Table 1.1). The alarmone ppGpp is a bacterial signaling molecule that mediates the bacterial stringent response, which is elicited as a consequence of amino acid starvation or other nutritional stresses resulting in changes to the transcriptional profile of cells; this causes a switch in growth phase from exponential to stationary phase (Cashel et al., 1996; Nystrom, 2004). Under these conditions, there is an overall down-regulation of processes involved in cell proliferation including DNA replication, rRNA and tRNA transcription, production of ribosomes, cell membrane synthesis, and general biomolecular synthesis, and a concomitant up-regulation of genes involved in stationary phase, stress survival, amino acid biosynthesis, and nutritional scavenging. The stringent response effects are primarily mediated by guanosine tetraphosphate ppGpp and guanosine pentaphosphate pppGpp (guanosine 3′-diphosphate, 5′-triphosphate), collectively known as (p)ppGpp. The E. coli RNA polymerase (RNAP) is a main target of (p)ppGpp binding, which results in the ‘reprogramming’ of the polymerase and change in the transcriptional profile of the cell. However, there are other protein targets of (p)ppGpp including LdcI.

Thus, the RavA-LdcI interaction has been proposed to be a regulatory mechanism by which the cell can switch LdcI activity between the LdcI-mediated acid stress response, which consumes lysine in the process, and the conservation of amino acids during stringent response (El Bakkouri et al., 2010; Kanjee et al., 2011). It should be noted that the exact role of ViaA in this process is not yet clear.

1.4.2.2.2 p618 from the crenarchaeal Acidianus two-tailed virus (ATV)

A novel member of the RavA subfamily that was recently characterized is p618 from the crenarchaeal Acidianus two-tailed virus (ATV) (Scheele et al., 2011). Its primary sequence shares a 21% identity and 42% similarity with E. coli RavA. Structurally, the hexameric form of the Walker B mutant of p618 shows a close resemblance to that of wild type E. coli RavA in the presence of nucleotides. Biochemically, the Km value for the ATPase activity of p618 (~0.55 mM) (Scheele et al., 2011) is very similar to that of E. coli RavA (~0.79 mM) (Snider et al., 2006). However, unlike E. coli RavA, the ATPase activity of p618 is optimal at high temperature, due to the fact that ATV’s host, the hyperthermophile Acidianus convivator, dwells in an acidic environment with temperatures ranging from 85°C to 93°C (Scheele et al., 2011).

The equivalent of the ViaA protein for p618 is p892. Unlike for other members of the RavA subfamily, the p892-encoding gene, ATV_gp61, does not immediately follow downstream of the p618-encoding gene, ATV_gp66. Instead, they are separated by four genes: ATV_gp62, ATV_gp63, ATV_gp64 and ATV_gp65 (Fig. 1.6D) (Prangishvili et al., 2006). Importantly, this is the first report of a RavA-containing gene cluster not having the typical gene organization of the RavA subfamily. Strong interaction between p618 and p892 has been observed in vitro, which requires the AAA+ domain of p618 and is enhanced by the presence of ATP and Mg2+ (Scheele et al., 2011). Similar to E. coli ViaA, the presence of p892 also increases the ATPase activity of p618 (Scheele et al., 2011). Unlike E. coli ViaA, however, p892 forms tetramers and hexamers in vitro and binds linear, double-stranded DNA in a non-specific manner in vitro (Scheele et al., 2011) – a property not observed in its E. coli counterpart [(Snider et al., 2006) and KW & WAH, unpublished data]. Interestingly, while p618 alone does not show any affinity for DNA, its interaction with p892 completely inhibits the VWA protein from binding to DNA (Scheele et al., 2011). Aside from p892, three additional ATV proteins, p387 (ATV_gp63, Fig. 1D), p653, and p800, were found to interact physically with p618 (Scheele et al., 2011). p387

has 61.5% sequence identity to p653, and they both exhibit affinity for linear, double-stranded DNA in a non-specific manner (Scheele et al., 2011). On the other hand, p800 is known to form filaments spontaneously via self-aggregation and has been implicated in viral tail development (Prangishvili et al., 2006). Taken together, the interactions of p618 with DNA-binding proteins and a putative viral tail structural protein suggests a possible role of this RavA subfamily member protein in the viral tail formation of ATV (Scheele et al., 2011) (Table 1.1). Given that the two viral tails are formed after the ATV virion has extruded from the host cell, the ATP required for p618 activity is proposed to be available from a pool of encapsulated cellular ATP within the ATV virion that it acquires from the host cell during the extrusion process (Scheele et al., 2011).

1.4.2.3 Newly Identified Role of APE2220 in Protein Maturation

1.4.2.3.1 CoxD from Oligotropha carboxidovorans

The Gram-negative chemolithoautotroph (an organism that obtains energy from inorganic

compounds and carbon from CO2) Oligotropha carboxidovorans OM5 requires the carbon monoxide (CO) dehydrogenase complex for utilizing CO as a sole source of energy. The complex exists as a dimer-of-trimers, with each trimer consisting of the subunits CoxL, CoxM and CoxS (Dobbek et al., 1999). The catalytic site for the oxidation of CO, located in the CoxL

subunit of the complex, contains a unique [CuSMoO2] cluster (Meyer et al., 2000). The proper assembly of this cluster into CoxL requires the AAA+ ATPase CoxD, which belongs to the APE2220 subfamily (Pelzmann et al., 2009). Notably, this is the first APE2220 protein to be assigned a definitive biological function (Snider & Houry, 2006). CoxD is encoded in the gene cluster coxDEFG (Fig. 1.6E) located downstream of coxMSL that encodes the CO dehydrogenase subunits (Pelzmann et al., 2009). The coxE gene, which immediately follows coxD (Fig. 1.6E), is predicted to encode a VWA protein (Fuhrmann et al., 2003) and is likely to be the corresponding VWA protein partner for CoxD.

CoxD can utilize both ATP and GTP (Pelzmann et al., 2009), and was found to be essential for the activity of CO dehydrogenase, as disruption of coxD inhibits O. carboxidovorans from utilizing CO as a sole energy source although there is no effect on the expression of CO dehydrogenase subunits (Pelzmann et al., 2009). CO dehydrogenase purified from the coxD insertional mutant was found to be inactive due to the lack of a properly

assembled [CuSMoO2] cluster, while the rest of the electron transfer relay, consisting of two [2Fe-2S] clusters found in the CoxS and a redox-active FAD in CoxM (Meyer et al., 2000), remained intact (Pelzmann et al., 2009). Importantly, the enzymatic activity of apo-CO dehydrogenase isolated from the coxD mutant can be restored by a step-wise reconstitution of the

[CuSMoO2] cluster in vitro, which first involves the resulfuration of the MoO3 moiety found in the apo-enzyme, followed by the introduction of copper (Pelzmann et al., 2009). Based on these data, the role of CoxD has been proposed to be for the partial unfolding of CoxL to allow for the

step-wise assembly of the [CuSMoO2] cluster (Table 1.1), although no in vitro data is available for verifying this claim due to the reported difficulty in isolating soluble CoxD in its functional form (Pelzmann et al., 2009). Given the apparent association of CoxD with the inner membrane of O. carboxidovorans and the fact that CO dehydrogenase is cytoplasmic, the CoxD-dependent assembly of the [CuSMoO2] cluster is proposed to occur on the cytoplasmic side of inner membrane (Pelzmann et al., 2009).

1.4.3 Structural Characterization of MoxR Proteins

The primary focus of this section will be placed on the recently solved crystal structures of the E. coli RavA (PDB ID 3NBX) (El Bakkouri et al., 2010) and the MoxR (MRP) AAA+ ATPase, CHU_0153, from Cytophaga hutchinsonii (PDB ID 2R44, Joint Centre for Structural Genomics), both of which possess structural features that are potentially unique to MoxR family proteins in general. We will first discuss the structural aspects of the AAA+ domain of the MoxR proteins and then provide a detailed discussion of the structure of the E. coli RavA-LdcI complex.

1.4.3.1 The Unique Subdomain Arrangement in the AAA+ Domain of MoxR Proteins

Examination of the AAA+ domains of E. coli RavA (Fig. 1.5B), C. hutchinsonii CHU_0153 (Fig. 1.5C), the closely related magnesium chelatase AAA+ subunit BchI from Rhodobacter capsulatus (Fig. 1.5D), and the AAA+ domain of S. solfataricus MCM (Fig. 1.5E) revealed important structural characteristics that are in common. All four proteins belong to the Helix-2 Insert Clade, which includes the MoxR family, the Chelatase family and the MCM family (Iyer et al., 2004). The most striking structural characteristic is the unique spatial arrangement of the αβα and all-α subdomains of the AAA+ domain, compared to the typical configuration found in other AAA+ proteins. Using the AAA+ domain of E. coli PspF as an example of an AAA+

ATPase with the canonical arrangement of the two subdomains (Fig. 1.5A), it can be noted that while the all-α subdomain is usually positioned on ‘top’ of the αβα subdomain, the all-α subdomain in RavA, CHU_0153, BchI and MCM is rearranged to the ‘side’ of the αβα subdomain. This unique spatial arrangement is made possible by the presence of a linker region (Fig. 1.5B-E; coloured in green) that spans along the ‘back’ of the αβα subdomain (back views in Fig. 1.5B-E). The linker region has been previously shown to be a unique characteristic among many AAA+ ATPases belonging to the Helix-2 Insert Clade (Iyer et al., 2004).

For the remainder of this section, we will focus our discussion on the E. coli RavA (Fig. 1.5B), C. hutchinsonii CHU_0153 (Fig. 1.5C), and the closely related R. capsulatus BchI (Fig. 1.5D), given that the AAA+ domains of these three proteins share a greater degree of structural similarity with each other compared to the AAA+ domain of S. solfataricus MCM (Fig. 1.5E).

The X-ray structures of E. coli RavA, C. hutchinsonii CHU_0153, and R. capsulatus BchI are those of monomers and not hexamers. However, 3D electron microscopy (EM) reconstructions of the hexamers are available for RavA (El Bakkouri et al., 2010) and BchI hexamers (Lundqvist et al., 2010). By fitting the monomer X-ray structure into the hexamer EM maps, it can be deduced that despite the unusual subdomain arrangement, the ATP-binding pocket at each dimer interface in RavA and BchI (and presumably CHU_0153) is expected to be similar to the one found in a canonical AAA+ domain such as that of E. coli HslU, whose X-ray structure was solved as a hexamer (Bochtler et al., 2000; Wang et al., 2001). As shown in Fig. 1.7, the five essential functional motifs (Walker A, Walker B, Sensor 1, Arg finger, and Sensor 2) are all conserved in their spatial coordination relative to the nucleotide-binding pocket. The only alteration to the pocket as a result of the different subdomain arrangement is in Sensor 2. Specifically, with the canonical subdomain arrangement of HslU, Sensor 2 is contributed by the same monomeric subunit as Walker A, Walker B and Sensor 1 (Fig. 1.7C), while in the alternative subdomain arrangement of RavA and BchI, it originates from the adjacent monomeric subunit instead (i.e. the same subunit that contributes the Arginine finger) (Fig. 1.7A and B). The functional significance of having an alternative subdomain arrangement in RavA, CHU_0153, and BchI is not known and warrants further investigation. It should be noted that, in BchI, the arginine side chain of Sensor 2 and the Arg finger both point away from the nucleotide-binding

23 pocket in the X-ray structure (Fig. 1.7B), which might reflect a different nucleotide state or may simply be an artifact of crystallization (Fodje et al., 2001).

Figure 1.7. Nucleotide binding site in MoxR proteins

Shown are two adjacent AAA+ domains with the respective nucleotide-binding site between the two subunits for (A) E. coli RavA, (B) R. capsulatus BchI, and (C) E. coli HslU (PDB ID 1DO0). One subunit is coloured dark gray, while the other subunit is coloured white. The important functional motifs of the nucleotide binding sites have their key amino acid residues displayed in stick format, and are coloured as follows: red = Walker A; blue = Walker B; green = Sensor 1; orange = Arg finger; magenta = Sensor 2. The key residues highlighted are as given in the chart at the lower-right corner. All bound nucleotides are shown in stick format and coloured in brown. BchI has no bound nucleotide. For simplicity, only the AAA+ domains are shown and other domains have been removed. The insets show the orientation of the two subunits (shown in ribbon format) relative to the full-length hexameric complex (with the accessible surfaces shown).

______

1.4.3.2 Structural Features Unique to the AAA+ Domain of MoxR Proteins

The overall structural layouts of the AAA+ domains of RavA and CHU_0153 highly resemble that of BchI (Fig. 1.5B-D). Nevertheless, a closer examination reveals key differences that distinguish RavA and CHU_0153 as members of a different protein family from BchI. For example, the linker region in BchI is largely helical in structure with two minor kinks at both ends, which gives it a slightly S-shaped conformation (Fig. 1.5D). However, in both RavA and CHU_0153, only a small helical segment is present in the C-terminal half of the linker, while the N-terminal half forms flexible loops (Fig. 1.5B and C). The orientation of the helical segment in the linker is also different. In BchI, the helical segment runs along the back of the αβα subdomain from top to the bottom before reaching the all-α subdomain (Fig. 1.5D), while in RavA and CHU_0153, it lies at a small angle to the base of the αβα subdomain, with the C- terminal end pointing downwards leading into the all-α subdomain (Fig. 1.5B and C). The primary reason for this difference is due to physical constraint, as both RavA and CHU_0153 have a long α1 helix at the N-terminus (back view in Fig. 1.5B and C) that occupies the same space as the helical linker region in BchI (back view in Fig. 1.5D). Note that the length of the linker region for RavA (33 residues), CHU_0153 (29 residues), and BchI (32 residues) are very similar (Fig. 1.8). Based on multiple sequence alignment of the three AAA+ domains (Fig. 1.8), it is interesting to note that although their respective linker regions are well aligned in the C- terminal half, where all three are helical in structure, the N-terminal half is not. This is likely to contribute to differences in tertiary structure.

Another structural difference between RavA/CHU_0153 and BchI lies in the H2-insert and its immediate surroundings. As observed in RavA and CHU_0153, the H2-insert adopts a simple unstructured loop (Fig. 1.5B) or β-hairpin motif (Fig. 1.5C), respectively. However in BchI, the H2-insert adopts a strand-loop-helix-loop-strand motif (Fig. 1.5D). There is also an α1- β2-β-hairpin in the vicinity (Fig. 1.5D, highlighted in pink and Fig. 1.8, highlighted in gray), which is unique to magnesium chelatases (Snider et al., 2008) and is absent in both RavA and CHU_0153. Thus, both RavA and CHU_0153 appear to have simpler H2-insert motifs compared to BchI.

Figure 1.8. Multiple sequence alignment of the AAA+ domain of RavA, CHU_153, and BchI

The multiple sequence alignment was generated using MUSCLE (Edgar, 2004). All information on secondary structures is derived from the X-ray structures of the three proteins. Residues coloured in red are found in α-helices, and those in blue are found in β-strands. The linker region is highlighted in yellow. The α1-β2-β-hairpin unique to magnesium chelatases (Snider et al., 2008) is highlighted in gray. Important sequence motifs of the AAA+ domain, the Helix-2 insert (H-2 insert), and the Pre-sensor 1 β- hairpin (PS1-βH) are all shown as indicated. In the alignment, ‘*’ denotes identical amino acid residues, ‘:’ denotes highly similar amino acid residues and ‘.’ denotes similar amino acid.

______

A third structural difference lies in the oligomeric state of the proteins as deduced from fitting the X-ray structures of the monomers into the EM density maps of the hexamers in the ADP-bound state (El Bakkouri et al., 2010; Lundqvist et al., 2010). While RavA adopts the typical hexameric state with a general six-fold symmetry like many other AAA+ ATPases (Fig. 1.9, top panel), BchI shows a distinct trimer-of-dimers arrangement with a three-fold symmetry (Fig. 1.9, bottom panel). In addition, the orientation of the AAA+ domains is different between the two proteins. For RavA, the back of the AAA+ domain faces the top of the hexamer and tilts away from the central pore, whereas for BchI, the back of the AAA+ domain faces the side of the hexamer with one monomer in each of the three dimeric subunits tilting at an angle relative to the other monomer (as indicated by their respective helical linker regions, coloured in green in Fig. 1.9, bottom panel). As a result, the RavA hexamer has a more compact central core, while the BchI hexamer shows a large pore in its center (Fig. 1.9). Hypothetically, the same inherent structural characteristics may be extended to other members of the respective families.

1.4.3.3 The Structure of the E. coli RavA-LdcI Complex

As mentioned earlier, the X-ray crystal structure of the RavA protomer was recently determined by our group (Fig. 1.10A and B) (El Bakkouri et al., 2010) and shows that the protein has three domains: an N-terminal AAA+ domain and a small, unique, β-rich domain that we named the LARA domain which interacts with LdcI; these domains are separated by a discontinuous triple helical bundle domain. A negative-stain electron-microscopy (EM) reconstruction of the RavA hexamer shows that the protein has a hexameric, flower-like arrangement (Fig. 1.9, top panel and Fig. 1.10C) (Snider et al., 2006; El Bakkouri et al., 2010).

Figure 1.9. The RavA and BchI hexameric assembly

Shown are the top, side, and bottom views of the RavA hexamer (top row) and BchI hexamer (bottom row). For the RavA hexamer, two consecutive subunits are shown in red and blue. For BchI, which forms a trimer-of-dimers, two subunits that belong to the same dimer are highlighted in red and blue. The linker regions are highlighted in green. Both RavA and BchI hexamers were obtained by modeling the X-ray structure of the monomer onto the respective EM structure of the hexamer in the ADP-bound state (El Bakkouri et al., 2010; Lundqvist et al., 2010).

______

Figure 1.10. Structure of RavA from E. coli

(A) A schematic of the RavA domain organization. Residues defining the boundaries of individual domains and the two ends of the polypeptide are as shown.

(B) X-ray structure of RavA protomer. αβα subdomain is shown in brown, all-α subdomain is shown in wheat, the linker between the two subdomains is shown in green, triple helical bundle domain is shown in

blue, the LARA domain is shown in dark blue, and bound ADP is shown in violet. The α-helices and β- strands are labeled sequentially except for βa and βb of the Pre-Sensor 1 β-hairpin insertion. Residues 88–97 and 438–441 were not observed in the X-ray structure and are indicated by a dashed line.

(C) Top and side views of the EM 3D reconstruction of the RavA-ADP hexamer. An atomic model of RavA hexamer was generated from the X-ray structure of the RavA protomer by docking into the EM envelope of the hexamer and comparison with the X-ray structure of the HslU hexamer (PDB ID code 1DO0).

Figures B and C are from reference (El Bakkouri et al., 2010).

______

LdcI is a 715 amino acids-long (Meng & Bennett, 1992) pyridoxal phosphate (PLP)- dependent decarboxylase (Sabo et al., 1974) that catalyzes the decarboxylation of L-lysine to

form the polyamine cadaverine in a reaction that consumes a proton and generates CO2, which diffuses out of the cell. Cadaverine is removed from the cell by the antiporter CadB, which at the same time imports lysine into the cell. We solved the structure of LdcI (Fig. 1.11A) and showed that it has three domains: Wing, Core, and C-terminal domain (Alexopoulos et al., 2008; Kanjee et al., 2011). Five LdcI dimers associate to form a decamer with distinct pentameric symmetry (Fig. 1.11A). In the course of solving the crystal structure of the enzyme, we made the serendipitous finding of the binding of the stringent response alarmone ppGpp at a specific interface between neighboring dimers in the LdcI decamer (Fig. 1.11A). Hence, there are ten ppGpp molecules per LdcI decamer. This led us to identify a novel regulatory function for ppGpp in controlling LdcI activity during the acid and stringent stress responses. We found that ppGpp dramatically inhibits LdcI activity.

The reconstitution of the RavA-LdcI complex using purified proteins results in a large, cage-like structure in which five RavA hexamers bridge two LdcI decamers as observed by negative stain electron microscopy (EM) (El Bakkouri et al., 2010) (Fig. 1.11B). We expect such a complex to be highly abundant in the cell under conditions of acid stress and anaerobiosis (Snider et al., 2006; Kanjee et al., 2011). The complex is about 3 MDa and is as big as the ribosome in size. At this stage it is not clear what is the cellular function of this complex. However, biochemical and cell biological studies showed that RavA can prevent the binding of ppGpp to LdcI to allow the enzyme to remain active even in the presence of the alarmone (El Bakkouri et al., 2010). More importantly, the LARA domain of RavA seems to bind at a site in LdcI close to the ppGpp-binding site and, hence, competes with ppGpp for binding to the decarboxylase. We showed that the LARA domain is required for the RavA-LdcI interaction, as

well as, for the RavA-RavA interaction in the RavA-LdcI cage structure (Fig. 1.11B and C) (El Bakkouri et al., 2010). As shown schematically in Fig. 1.11C, two legs (triple helical domain + LARA domain) from RavA bind the upper LdcI decamer: one leg binds an LdcI subunit in the upper pentamer and the second leg binds a corresponding subunit in the lower pentamer. The interactions seem to be mainly mediated by the LARA domain. The same set of interactions is found with an LdcI dimer at the bottom of the complex. The two remaining legs of RavA are interacting with a neighboring RavA leg on the left and on the right (Fig. 1.11C). The RavA- RavA leg-leg interactions seem to involve the triple-helical domain, as well as the LARA domain. Hence, the unique structure of RavA makes all these interactions possible.

Figure 1.11. Structure of the RavA-LdcI complex

(A) A cartoon representation of the LdcI decamer is shown beneath a transparent van der Waals’ surface. For the structure on the left, neighbouring monomers in a dimer are colour matched except for those of the top-most pair that are coloured according to domains. The dimension of one of the five side pores is indicated. For the structure on the right, the top ring of an LdcI decamer is shown as a cartoon representation, where each monomer has a different colour and the bottom ring is shown in gray as a van der Waals’ surface. The PLP in each active site and the ppGpp molecules are shown as van der Waals’ spheres. Oxygen atoms are in red, nitrogen atoms are in blue, and phosphate atoms are in cyan. The dimension of the central pore is indicated.

(B) Fit of the RavA hexameric model and LdcI decamer into the EM envelope of the RavA-LdcI complex viewed from the side (Left) and the top (Right). One LdcI dimer is coloured in red (the upper monomer) and green (the lower monomer). ppGpp bound to LdcI is drawn as blue spheres. For clarity, PLP is not shown.

Figure A is from reference (Kanjee et al., 2011), while figures B and C are from reference (El Bakkouri et al., 2010).

______

1.4.4 Summary on the Functional and Structural Characteristics of the MoxR AAA+ ATPases

The recent structural and functional characterization of known and newly identified MoxR proteins has provided valuable insight into this family of AAA+ ATPases that has remained largely poorly characterized. The majority of the examples discussed here have uncovered previously unknown roles of MoxR proteins in multiple stress response pathways (Table 1.1). The exact type of stress response associated depends largely on the organisms in which the proteins are found. For example, RL3499 from R. leguminosarum is important for stress tolerance in the cell envelope, which in turn is important for establishing optimal symbiosis between the bacterium and the host pea plant (Vanderlinde et al., 2011). FTL_0200 is associated with stress tolerance against the host’s immune responses for the pathogenic F. tularensis (Dieppedale et al., 2011). RavA is associated with both acid stress tolerance and the stringent response, both of which are experienced during the life cycle of E. coli (El Bakkouri et al., 2010; Kanjee et al., 2011). An equally intriguing revelation is the link of MoxR proteins to cellular morphology and development, as seen in the requirement of RL3499 for maintaining normal cell morphology in R. leguminosarum (Vanderlinde et al., 2011), and the implication of p618 in extracellular viral tail development of the crenarchaeal Acidianus two-tailed virus (ATV) (Scheele et al., 2011). Importantly, the functional diversity displayed among MoxR proteins is likely reflected in the nature of their physical interactors. This is supported by the interaction

between RavA and the acid stress-associated LdcI in E. coli (Snider et al., 2006; El Bakkouri et al., 2010), as well as, the interaction between p618 and the p800 associated with viral tail formation in ATV (Scheele et al., 2011). Thus, the identification and characterization of the physical interactors of MoxR proteins should be considered as equally important in providing a comprehensive view into the nature of MoxR proteins themselves.

From the structural perspective, the X-ray crystal structures of E. coli RavA and C. hutchinsonii CHU_0153 enable the identification of structural features that are potentially characteristic of the MoxR family proteins. These include: (a) an atypical spatial arrangement of the αβα and all-α subdomains of the AAA+ domain that is also found in the closely related BchI from R. capsulatus; (b) the presence of a linker region that, unlike the extended helical structure observed in BchI, takes on a flexible-loop conformation in its N-terminal part and a short helix in the C-terminal part; (c) the presence of an extended α1 helix that occupies the same space as the helical linker region in BchI; (d) a simple structural motif for the helix-2 insert; and (e) a hexameric state resembling the typical configuration of many other AAA+ ATPases instead of the trimer-of-dimers configuration of BchI. Nevertheless, to further generalize these structural features requires obtaining more structures of MoxR AAA+ proteins from the different subfamilies.

1.5 Introduction to the Aerobic and Anaerobic Respiration of E. coli

This following section aims to provide general background information on the protein machinery and biochemical processes involved in both the aerobic and anaerobic respiration of E. coli, with a focus on the respiratory complexes NADH:ubiquinone oxidoreductase I (Nuo) and the fumarate reductase (Frd), as well as the associated respiratory electron transport chain that utilizes NADH as the electron donor and fumarate as the electron acceptor. Specific subunits of the Nuo and Frd complexes were found to interact with RavA and ViaA. Experimental details of these interactions will be presented in Chapters 2 and 3 of this thesis.

1.5.1 The Respiratory Electron Transport Chains of E. coli

The bacterial respiratory electron transport machinery is highly versatile that allows the cell to fully optimize its respiratory processes to best fit its physiology, life style and the specific

environment of its natural habitat (Unden & Bongaerts, 1997; Richardson, 2000; Richardson et al., 2001; Bueno et al., 2012; Radzi Noor & Soulimane, 2012; Roger et al., 2012). In E. coli, a facultative anaerobe (i.e. it grows in both aerobic and anaerobic conditions, but prefers the former), there are multiple respiratory electron transport chains consisting of various combinations of primary dehydrogenases and terminal reductases. The multiplicity and diversity in its respiratory machinery allows E. coli to utilize different combinations of electron donors

(e.g. NADH, succinate, H2, etc.) and terminal electron acceptors (e.g. O2, nitrate, fumarate, etc.) as required under different growth conditions (Gennis & Stewart, 1996; Unden & Bongaerts, 1997; Cecchini et al., 2002; Price & Driessen, 2010). Fig. 1.12 provides a general representation of the various respiratory complexes that are utilized by E. coli under different growth conditions, as well as dedicated transporters for the movement of the metabolites required and byproducts produced from their activities.

The basic architecture of an E. coli respiratory electron transport chain consists of a primary dehydrogenase and a terminal reductase that are coupled via lipid-soluble quinones within the inner membrane (Gennis & Stewart, 1996; Unden & Bongaerts, 1997), through which

electrons are transferred from an electron donor (e.g. NADH) to an electron acceptor (e.g. O2) via the same basic process. First, electrons are extracted from the electron donor via its oxidation by the primary dehydrogenase. Next, the extracted electrons are shuttled through the dehydrogenase and transferred to a quinone molecule within the inner membrane, reducing it to a quinol. The quinol in turn interacts with the terminal reductase, which extracts electrons from it and re-oxidizing it back into quinone. At the same time, the extracted electrons are shuttled through the reductase and are transferred to the electron acceptor via its reduction (Gennis & Stewart, 1996; Unden & Bongaerts, 1997; Soballe & Poole, 1999). The energy released from the electron transfer from donor to acceptor generally results in the migration of protons from the cytoplasm to the periplasm for maintaining the proton gradient that is necessary for ATP biosynthesis (Gennis & Stewart, 1996; Unden & Bongaerts, 1997; Senior et al., 2002), although the number of protons translocated across the membrane per electron transferred (i.e. H+/e- ratio) and the exact mechanism involved vary greatly, depending on the respiratory complexes involved (Gennis & Stewart, 1996; Unden & Bongaerts, 1997) (Fig. 1.12).

Figure 1.12. Schematic representation of primary dehydrogenases, terminal reductases and specific transporters involved in the aerobic and anaerobic respiratory pathways of E. coli

Proteins coloured in white are primarily involved in aerobic respiration, while those in dark gray are primarily involved in anaerobic respiration. Proteins coloured in light gray (e.g. NuoA-N and TorA2C) participate in both aerobic and anaerobic respiration. Except for NuoA-N, all aerobic primary dehydrogenases are shown on the left, their anaerobic counterparts are shown on the right, and all terminal reductases are shown at the bottom of the diagram. The transporters are shown at the top-right corner. The biochemical activities catalyzed by each protein / protein complex are as shown. All information on topology, subunit composition and stoichiometry of the protein complexes shown are obtained from public online databases such as EcoCyc (Keseler et al., 2005) and UniProt (Magrane & Consortium, 2011). Q represents quinones in general without distinguishing them as ubiquinone (UQ), menaquinone (MK) or demethylmenaquinone (DMK) (see main text for more details). NADH / NAD+ = nicotinamide adenine dinucleotide (NADH = reduced form; NAD+ = oxidized form); Gly-3-P = sn-glycerol- 3-phosphate; DHAP = dihydroxyacetone phosphate; TMAO = trimethylamine N-oxide; TMA = - trimethylamine; HCO2 = formate.

______

1.5.1.1 Primary Dehydrogenases in E. coli

The E. coli genome encodes a diverse arsenal of primary dehydrogenases that vary greatly in primary sequence, structure and composition (Gennis & Stewart, 1996; Unden & Bongaerts, 1997), which allow the cell to utilize a variety of electron donors across different growth conditions (Fig. 1.12). Function wise, each primary dehydrogenase is dedicated to the oxidation of a specific electron donor or a highly similar variant, although isozymes (i.e. different enzymes that perform the same molecular function) with very different structures and compositions but share common electron donors are also present. For example, both the 14-subunit NADH:ubiquinone oxidoreductase I (NuoA-N) and the single-unit NADH:ubiquinone oxidoreductase II (Ndh) are dedicated to the oxidation of NADH (Matsushita et al., 1987; Hayashi et al., 1989; Calhoun & Gennis, 1993; Leif et al., 1995). Similarly, both the aerobic glycerol-3-phosphate dehydrogenase (GlpD) and its anaerobic counterpart (GlpABC) catalyze the oxidation of glycerol-3-phosphate (Schryvers & Weiner, 1981; Schweizer & Larson, 1987).

With few exceptions, the primary dehydrogenases in E. coli are involved almost exclusively in either aerobic or anaerobic respiration (Fig. 1.12). For aerobic dehydrogenases, the majority have single subunits consisting of a cytoplasmic domain and one or more small hydrophobic regions for anchoring to the inner membrane (e.g. Ndh, GlpD and PoxB in Fig. 1.12) (Gennis & Stewart, 1996; Unden & Bongaerts, 1997). Other dehydrogenases, such as the aforementioned Nuo complex (NuoA-N) and the tetrameric succinate dehydrogenase (SdhA-D), are composed of multiple subunits, such that the solvent-exposed cytoplasmic subunits are anchored to the inner membrane via direct interaction with integral membrane subunits (Fig. 1.12) (Yankovskaya et al., 2003; Baradaran et al., 2013). A key feature of all aerobic

dehydrogenases is that the primary active site, where electrons are extracted from the electron donor, is located within specific subunits that sit on the cytoplasmic side of the inner membrane (e.g. NuoF and SdhA; Fig. 1.12) (Unden & Bongaerts, 1997; Yankovskaya et al., 2003; Baradaran et al., 2013). Another important feature of the aerobic dehydrogenases is that with the exception of the Nuo complex, they do not translocate protons across the inner membrane (e.g. Ndh, GlpD, PoxB, SdhA-D; Fig. 1.12) and thus do not contribute to the maintenance / conservation of the proton gradient (Gennis & Stewart, 1996; Unden & Bongaerts, 1997). Hypothetically, the higher energy yields of aerobic electron transport chains minimizes the cell’s need to conserve energy, in exchange for the maximization of growth rates, turnover rates of metabolites and other parameters that are better suited for growth and survival under aerobic conditions (Gennis & Stewart, 1996; Unden & Bongaerts, 1997).

The primary dehydrogenases for anaerobic respiration are mostly composed of multiple subunits (e.g. HyaABC, HybABCO, GlpA-C; Fig. 1.12) (Schryvers & Weiner, 1981; Gennis & Stewart, 1996; Unden & Bongaerts, 1997), with the primary active site located within a specific solvent-exposed subunit (e.g. HyaB, HybC, GlpA; Fig. 1.12) (Schryvers & Weiner, 1981; Cole et al., 1988; Sargent et al., 1998; Vignais & Colbeau, 2004). In contrast to aerobic dehydrogenases, the subunits carrying the active site in most of the anaerobic dehydrogenases are located on the periplasmic side of the inner membrane (Fig. 1.12) (Unden & Bongaerts, 1997; Sargent et al., 1998; Vignais & Colbeau, 2004). Importantly, the oxidation of electron donors by these dehydrogenases results in the production of protons that are deposited into the periplasm (Fig. 1.12) (Gennis & Stewart, 1996; Unden & Bongaerts, 1997). At the same time, protons are removed from the cytoplasmic side by these dehydrogenases in the reduction of quinone to quinol (Fig. 1.12) (Gennis & Stewart, 1996; Unden & Bongaerts, 1997). Together, these two events result in a displacement of protons from the cytoplasm to the periplasm without their actual movement across the inner membrane, which nonetheless contributes to the conservation of the proton gradient. Given that the energy yields of anaerobic transport chains are typically much smaller compared to aerobic ones (Gennis & Stewart, 1996; Unden & Bongaerts, 1997), the maintenance of the proton gradient by the anaerobic dehydrogenases is likely to be more important for optimal cell growth under anaerobic conditions (Gennis & Stewart, 1996; Unden & Bongaerts, 1997).

1.5.1.2 Terminal Reductases in E. coli

Like primary dehydrogeneases, there are multiple terminal reductases encoded in the E. coli genome, which allows the cell to utilize various electron acceptors across different growth conditions (Fig. 1.12) (Gennis & Stewart, 1996; Unden & Bongaerts, 1997). These terminal reductases also show a high degree of variation in primary sequence, structure, composition and location of their primary active sites with respect to the inner membrane (Fig. 1.12) (Gennis & Stewart, 1996; Unden & Bongaerts, 1997). Similarly, the terminal reductases in E. coli are also dedicated to either aerobic or anaerobic respiration, with the trimethylamine N-oxide reductase

(TorA2C) being a known exception that participates in both (Ansaldi et al., 2007).

Under aerobic conditions, the cytochrome bo terminal oxidase (CyoA-D) and the

cytochrome bd-I terminal oxidase (CydABX) utilize molecular oxygen (O2) as the terminal electron acceptor. Both reductases are composed of multiple integral membrane subunits (Fig. 1.12). Given most of the aerobic primary dehydrogenases lack the ability to translocate protons across the inner membrane, the maintenance of the proton gradient is largely dependent on the activity of the two cytochrome oxidases. While both CyoA-D and CydABX release protons into the periplasm as they re-oxidize quinol to replenish the quinone pool in the inner membrane, CyoA-D also function as a proton pump that facilitates the migration of protons from the cytoplasm to the periplasm (Fig. 1.12) (Puustinen et al., 1991).

When oxygen is limited or unavailable, the cell switches to other terminal reductases such as nitrate reductase A (NarGHI), fumarate reductase (FrdA-D) and others that utilize alternative electron acceptors for energy metabolism (Fig. 1.12) (Gennis & Stewart, 1996; Unden & Bongaerts, 1997). Unlike the cytochrome oxidases, these alternative terminal reductases consist of both cytoplasmic and integral membrane subunits, with the primary active site located within a specific cytoplasmic subunit (e.g. NarG, FrdA; Fig. 1.12). Others such as the aforementioned TorA2C (Fig. 1.12) or the dimethyl sulfoxide reductase (DmsABC; not shown) adopt the opposite topology, and their primary active sites are positioned in the periplasm. In terms of function, the anaerobic terminal reductases differ from the cytochrome oxidases as they are not the sole contributors in the maintenance of the proton gradient during anaerobic respiration. For example, the utilization of fumarate as the electron acceptor by FrdA-D actually results in the release of protons back into the cytoplasm during the re-oxidation of quinol (Fig.

1.12) (Cecchini et al., 2002). In this case, the translocation of protons is left to the primary reductase NuoA-N (Tran et al., 1996).

1.5.1.3 Respiratory Quinones in E. coli

Quinones are lipid-soluble organic molecules within the inner membrane of E. coli that shuttle electrons from primary dehydrogenases to terminal reductases during respiration. There are three major types of quinones that are associated with the respiratory process in E. coli – ubiquinone (UQ), menaquinone (MK) and demethyl-menaquinone (DMK) (Soballe & Poole, 1999). As shown in Fig. 1.13, the basic structure of these quinones consists of a derivative of 1,4- benzoquinone (as in UQ; Fig. 1.13A) or 1,4-naphthoquinone (as in MK and DMK; Fig. 1.13B and C) that carries a side chain consisting of isoprenoid units (Soballe & Poole, 1999). In E. coli, the majority of UQ, MK and DMK all have 8 isoprenoid units in their side chains (Fig. 1.13A- C), although small amounts of these quinones carrying side chains of varying lengths are also present (Soballe & Poole, 1999). Despite their structural variations, all three types of quinones are reduced step-wise: first into their respective semiquinone radicals (Fig. 1.13D-F), then into their respective quinols (Fig. 1.13G-I) by primary dehydrogenase. Upon interaction with terminal reductases, the quinols are oxidized back into semiquinones and then into quinones (Song & Buettner, 2010).

The choice of quinone is largely determined by the type of electron acceptor that is available in a given growth environment. In general, UQ is primarily used during aerobic respiration, whereas MK and DMK are used for anaerobic respiration of fumarate or trimethylamine N-oxide (TMAO) (Gennis & Stewart, 1996). For the anaerobic respiration of nitrate, either UQ or MK can be utilized to carry electrons from the primary dehydrogenase to the terminal reductase (Wissenbach et al., 1990; Wissenbach et al., 1992). The preference for UQ in aerobic respiration and MK / DMK in anaerobic respiration is attributed to the higher

midpoint potential (or standard redox potential) of UQ (E’m = +113 mV) than either MK (E’m = -

74 mV) or DMK (E’m = +36 mV) (Unden & Bongaerts, 1997). As such, UQ is more suitable for

reducing the more electro-positive oxygen (E’m = +818 mV), while MK or DMK is used for

reducing anaerobic electron acceptors such as fumarate (E’m = +30 mV) and TMAO (E’m = +130 mV), which are far less electro-positive (Unden & Bongaerts, 1997). For nitrate (E’m = +433

mV), its midpoint potential is lower than O2 but higher than fumarate or TMAO, thus both UQ

39 and MK are suitable (Gennis & Stewart, 1996; Unden & Bongaerts, 1997; Soballe & Poole, 1999). Accordingly, the intracellular concentration of UQ in aerobically grown cells is 4-5 times higher than MK and DMK combined. In contrast, UQ concentration drops to approximately one- third of MK and DMK combined in anaerobically grown cells (Wissenbach et al., 1990; Wissenbach et al., 1992). Although the factors that determine the composition of the quinone pool remain unclear, there is experimental evidence in support of a post-transcriptional regulatory mechanism for quinone biosynthesis that is O2-dependent (Shestopalov et al., 1997).

Figure 1.13. Chemical structures of the major quinones utilized in E. coli respiratory processes

Shown are the chemical structures of (A) ubiquinone-8 (UQ-8), (B) menaquinone-8 (MK-8), (C) demethylmenaquinone-8 (DMK-8), their respective semiquinone radicals (D) semiubiquinone-8 (UQ•--8), (E) semimenaquinone-8 (MK•--8), (F) semidemethylquinone-8 (DMK•--8), and reduced form: (G) ubiquinol-8 (UQH2-8), (H) menaquinol-8 (MKH2-8) and (I) demethylmenaquinol (DMKH2-8). The number 8 indicates there are 8 isoprenoid units in the hydrophobic side chain. All forward reduction reactions (i.e. a quinone to a semiquinone, and a semiquinone to a quinol) are mediated by a dehydrogenase, while the reversed oxidation reactions are mediated by a terminal reductase. Dot (•) represents the unpaired electron of a radical.

______

1.5.1.4 Transporters for Respiratory Metabolites and Byproducts in E. coli

During respiration, both the influx of electron donors and acceptors into the cytoplasm and efflux of the reaction’s byproducts out to the periplasm are necessary to sustain the activity of the respiratory enzymes. In E. coli, there are various specialized transport proteins that facilitate the movement of their respective respiratory substrates and/or byproducts across the inner membrane. Several representative examples are given in Fig. 1.12.

The simplest transport of respiratory metabolites occurs via simple diffusion without any

transporters or protein channels. For example, O2, CO2 and other gaseous compounds can readily diffuse across the inner membrane (Lodish et al., 2000). Small ions can permeate the inner

membrane via specific ion channels (e.g. AmtB3 is utilized for the bi-directional movement of + NH4 across the inner membrane (Soupene et al., 1998)).

For the movement of larger organic molecules, specific transport proteins that require energy input are utilized. Among them, the secondary active transporters (i.e. the energy required for activity is derived by coupling the movement of substrate with the movement of another molecule that has high electrochemical potential energy across the membrane) are the most common. In the examples that are illustrated in Fig. 1.12, the C4-dicarboxylate:H+ symporter DctA facilitates the import of succinate, fumarate, L-malate and D-malate by coupling their influx to the proton gradient, such that two protons are co-imported for every substrate molecule (Gutowski & Rosenberg, 1975; Lo, 1977). Other secondary active transporters couple the influx of respiratory substrates with the efflux of byproducts. As shown in Fig. 1.12, the import of - nitrate (NO3 ), the electron acceptor for the anaerobic nitrate reductase (NarGHI), is coupled to - the export of nitrite (NO2 ) via the nitrate:nitrite antiporter (NarK) (Zheng et al., 2013).

Similarly, the influx of fumarate, substrate for the anaerobic fumarate reductase (FrdA-D), is coupled to the efflux of succinate via the C4-dicarboxylate transporter DcuB (Engel et al., 1994).

In addition to secondary active transporters, E. coli also utilizes primary active transporters (i.e. the energy required for activity is derived from ATP or other energy carriers) for a few respiratory substrates. For example, the import of sn-glycerol-3-phosphate (Gly-3-P), which is used as the electron donor by both the aerobic and anaerobic glycerol-3-phosphate dehydrogenases (GlpD and GlpABC, respectively), is facilitated by the ABC (ATP-binding cassette) transporter UgpABCE. ABC transporters are ATPases that constitute one of the six major classes of the ASCE ATPases (Fig. 1.3; see Section 1.2 for details) (Snider & Houry, 2008; Snider et al., 2008). Unlike the secondary active transporters discussed thus far, UgpABCE is composed of four subunits and adopts a structural layout that is typical of ABC transporters (Davidson & Chen, 2004). The periplasmic UgpB is responsible of binding Gly-3-P for import, while the integral membrane subunits UgpA and UgpE constitute the trans-membrane channel for the movement of substrate and the cytoplasmic UgpC provides the site for ATP hydrolysis (Fig. 1.12) (Brzoska et al., 1994).

1.5.1.5 Regulation of the Expression of Respiratory Proteins in E. coli

Theoretically, any combination of a primary dehydrogenase and a terminal reductase can form a respiratory electron transport chain in E. coli as long as the sequence of redox reactions satisfies all thermodynamic requirements (Gennis & Stewart, 1996; Unden & Bongaerts, 1997). Nevertheless, the actual number of possible combinations is far less due to the coordinated expression of specific sets of respiratory proteins that is regulated via a complex network of transcriptional regulators (Gennis & Stewart, 1996; Unden & Bongaerts, 1997). In general, the coupling of Ndh and CyoA-D serves as the primary respiratory chain under aerobic condition, while the coupling of NuoA-N and FrdA-D is the dominant one under anaerobic condition in the absence of nitrate (Gennis & Stewart, 1996; Unden & Bongaerts, 1997).

In E. coli, the expression of respiratory proteins is regulated in response to the type of electron acceptors available, such that the cell prioritizes the use of electron acceptors with high energy yield over those with low energy yield (Gennis & Stewart, 1996; Unden & Bongaerts, 1997). As such, aerobic respiratory chains, which have the highest energy yields, are given top priority, followed by those that utilize nitrate; anaerobic respiratory chains that utilize fumarate,

DMSO or TMAO have the lowest energy yields, and therefore are given the lowest priority (Gennis & Stewart, 1996; Unden & Bongaerts, 1997). The prioritization of electron acceptors is

achieved via primary transcriptional regulators such as the O2-sensing Fnr (Park & Gunsalus, 1995; Kang et al., 2005) and the ArcBA two-component signal transduction system (Bekker et al., 2010; Alvarez et al., 2013) (Fig. 1.14A and B), as well as the nitrate-sensing NarQL and NarXP two-component signal transduction systems (Fig. 1.14C) (Stewart, 1993; Stewart, 1994). Other transcriptional regulators such as AppY, IHF, H-NS, Fis and StpA offer additional control over the expression of specific respiratory proteins in response to different environmental factors (e.g. pH) as well as the growth rate and growth phase of the cell (Rowbury, 1997; King & Przybyla, 1999; Wolf et al., 2006; Bradley et al., 2007; Chib & Mahadevan, 2012).

Under aerobic condition, the cell prioritizes the use of aerobic respiratory chains by suppressing the induction of anaerobic respiratory genes, via inhibition of their respective transcriptional regulators that are O2-responsive (Gennis & Stewart, 1996; Unden & Bongaerts, 1997). As shown in Fig. 1.14A, Fnr is inactivated in the presence of oxygen due to oxidation of its iron-sulfur (Fe-S) cluster from [4Fe-4S] to [2Fe-2S] (Unden et al., 2002; Zhang et al., 2012). This process destabilizes the dimeric state of Fnr that is necessary for the activation of genes that are under its control (e.g. frdABCD) (Lazazzera et al., 1993; Salmon et al., 2003; Kang et al., 2005). The presence of oxygen also induces a shift of the quinone pool to the oxidized state that in turn prevents the autophosphorylation of the membrane-bound sensory kinase ArcB and subsequent transphosphorylation of the cytoplasmic response regulator ArcA, thereby preventing the activation of genes inducible by ArcBA (e.g. hyaA-F) (Fig. 1.14B) (Malpica et al., 2006; Bekker et al., 2010; Alvarez et al., 2013).

Figure 1.14. Schematic representation of signaling and regulatory pathways for the induction of respiratory proteins in E. coli in response to external stimuli

(A) Deactivation of the transcriptional regulator Fnr by molecular oxygen via oxidation of its [4Fe-4S] cluster into [2Fe-2S], and the subsequent dissociation of the Fnr dimer into monomers, which in turn prevents the induction of the Fnr regulon.

(B) Inhibition of the ArcBA two-component signaling pathway by molecular oxygen. A shift in the quinone pool to the oxidized form (QH2 → Q; quinol → quinone) inhibits the autophosphorylation of ArcB and subsequent activation of ArcA (see main text for details), which prevents the induction of the ArcA regulon. Pi = inorganic phosphate.

(C) Activation of the NarXL and NarQP two-component signaling pathways in the presence of exogenous nitrate. The autophosphorylation of NarX or NarQ activates NarL or NarP by transphosphorylation, which in turn induces the NarL and NarP regulons. At the same time, respiratory genes involved in the anaerobic respiration of H2 (hya, hyb), fumarate (frd), DMSO (dms), TMAO (tor) and other alternative electron acceptors are suppressed (see main text for more details). Pi = inorganic phosphate.

______

In the absence of oxygen, the nitrate-sensing regulatory system imposes additional control over the expression of anaerobic respiratory proteins, as the cell prioritizes the use of nitrate over fumarate and other alternative electron acceptors (Gennis & Stewart, 1996; Unden & Bongaerts, 1997). It consists of the paralogous membrane-bound sensory kinases NarQ and NarX, and the cytoplasmic response regulators NarL and NarP (Fig. 1.14C) (Stewart, 1994). Upon sensing the presence of nitrate in the environment, both NarQ and NarX are autophosphorylated (Fig. 1.14C) (Rabin & Stewart, 1993). The phosphoryl group is then transferred to either NarL or NarP, which leads to the induction of their respective regulons (Fig. 1.14C) (Rabin & Stewart, 1993; Stewart, 1993). Importantly, the activity of the NarQ/X and NarL/P inhibits the expression of alternative anaerobic respiratory proteins that are not part of the nitrate respiratory chains (Fig. 1.14C). These include HyaABC (hydrogenase 1), HybABCO

(hydrogenase 2), FrdA-D (fumarate reductase), TorA2C (TMAO reductase) and DmsABC (DMSO reductase) (Fig. 1.11) (Stewart, 1993; Stewart, 1994). Furthermore, the expression of transport proteins associated with the flux of metabolites utilized in non-nitrate respiratory chains, such as the C4-dicarboxylate transporters DcuA and DcuB (Fig. 1.12), is also inhibited (Stewart, 1993; Stewart, 1994).

1.5.2 Biophysical and Functional Characteristics of the NADH:ubiquinone Oxidoreductase I and the Fumarate Reductase from E. coli

In E. coli, the respiratory electron transport chain consists of the NADH:ubiquinone Oxidoreductase I (NuoA-N) and the fumarate reductase (FrdA-D) serves as the major anaerobic respiratory pathway in the absence of nitrate (Gennis & Stewart, 1996; Unden & Bongaerts, 1997). This section aims to provide more background details on the two respiratory complexes, as both NuoA-N and FrdA-D are highly relevant in our discussion of RavA and ViaA in chapters 2 and 3.

1.5.2.1 NADH:ubiquinone Oxidoreductase I (NuoA-N) from E. coli

The NADH:ubiquinone oxidoreductase I (NuoA-N) is considered one of the most sophisticated multi-subunit enzymes in E. coli. It is a membrane-bound protein complex composed of 13 distinct subunits – 6 cytoplasmic subunits and 7 integral membrane subunits (Fig. 1.15A) (Efremov et al., 2010; Baradaran et al., 2013). X-ray structures of both the E. coli membrane subunits at 3.0 Å (Efremov & Sazanov, 2011b) and the complete enzyme of the homologue from Thermus thermophilus at 3.3 Å (Baradaran et al., 2013) have been solved. The latter has been used to infer structural and functional information from E. coli K-12 (Fig. 1.15A) (Efremov & Sazanov, 2012).

The Nuo complex can be physically separated into 3 fragments: the soluble fragment consists of the subunits NuoE, NuoF and NuoG; the amphipathic connecting fragment consists of NuoB, the fused NuoCD and NuoI; and the membrane fragment composed of NuoA, NuoH, NuoJ, NuoK, NuoL, NuoM and NuoN (Fig. 1.15A) (Leif et al., 1995). Together the soluble and connecting fragments of the Nuo complex contain all of the cofactors – one FMN (flavin mononucleotide) and nine iron-sulfur (Fe-S) clusters – that are necessary for the transfer of electrons from the electron donor NADH to a quinone (Fig. 1.15B) (Euro et al., 2008; Verkhovskaya et al., 2008). The integral membrane subunits consist of trans-membrane (TM) α- helices (Fig. 1.15A). Among them, NuoL, NuoM and NuoN all adopt highly similar structures that resemble antiporters (Fig. 1.15A), with each possessing two half-channels – one opens to the cytoplasm and the other opens to the periplasm (Fig. 1.15B; indicated by the first and second halves of the H+ flow paths) – lined with conserved polar residues that facilitate proton translocation (Efremov & Sazanov, 2011b; Baradaran et al., 2013). In addition, NuoL also has a long α-helical segment, known as Helix HL, which runs along the cytosol-membrane interface adjacent to both NuoM and NuoN. The long Helix HL ends with a second α-helix, which is buried in the inner membrane and interacts with adjacent TM helices from NuoN (Fig. 1.15A) (Efremov & Sazanov, 2011b; Baradaran et al., 2013). On the periplasmic side, a linear arrangement of alternating β-hairpins and C-terminal α-helices from NuoL, NuoM and NuoN, known as the βH element (not shown in Fig. 1.15A) is positioned at the periplasm-membrane interface in a similar manner as Helix HL (Efremov & Sazanov, 2011b; Baradaran et al., 2013).

Figure 1.15. Structures of the NADH:ubiquinone oxidoreductase I (NuoA-N) and fumarate reductase (FrdA-D), and schematic representation of their biochemical activities

(A) The X-ray structure of the complete NADH:ubiquinone oxidoreductase I (Complex I) from Thermus thermophilus (PDB ID 4HEA) is used to represent its close homologue in E. coli K-12, whose X-ray structure of the membrane fragment is solved (PDB ID 3RKO). The assignment of subunits shown here is based on the structural layout of E. coli NuoA-N (Efremov & Sazanov, 2012). The cytoplasmic subunits Nqo15 and Nqo16 unique to T. thermophilus are omitted for clarity. The bound FMN and Fe-S clusters are shown in stick format.

(B) Schematic representation of the biochemical activity of NuoA-N. Movement of electrons extracted from NADH along the chain of FMN and Fe-S clusters to the quinone-binding site (Q) is shown with small red arrows. The long, narrow quinone-binding pocket is represented with the green dotted line. The cytoplasmic half of the forth proton channel (i.e. the E-channel) is represented with the thick magenta line. All other half-channels are represented with the flow paths of the individual protons. The longitudinal migration of protons as within each proton channel from the cytoplasmic to the periplasmic half-channel is indicated with the smaller horizontal black arrows along the thick blue line across the membrane subunits. The movements of quinone (Q) and quinol (QH2) into and out of the quinone-binding pocket are as shown.

(C) X-ray structure of the complete fumarate reductase (FrdA-D) from E. coli K-12 (PDB ID 1LOV). The bound FAD, Fe-S clusters and menaquinones are shown in stick format.

(D) Schematic representation of the biochemical activity of FrdA-D. As in (C), the movement of electrons extracted from quinol along the Fe-S clusters down to FAD to reduce fumarate is shown with small red arrows. The dotted arrows represent uncertainty pertaining to the exact role of the distal quinone-binding site (QD) in the electron transport chain (see main text for details).

For simplicity, all species of quinones are represented as Q, and quinols as QH2 in (B) and (D).

______

The other membrane subunits – NuoA, NuoH, NuoJ and NuoK – interact strongly with their respective neighbours via extensive protein-protein interfaces that are buried away from the inner membrane, as well as numerous hydrogen bonds and salt bridges (Efremov & Sazanov, 2011b). The tetrad forms the fourth proton channel of the Nuo complex, with a specialized “E- channel” consists of mostly Glu residues (Fig. 1.15B; represented schematically with the thick magenta line) inside subunit NuoH serving as the cytoplasmic half-channel for proton translocation (Fig. 1.15B; indicated by the first half of the H+ flow path). The periplasmic half- channel is produced by NuoA, NuoJ and NuoK (Fig. 1.15B; indicated by the second half of the H+ flow path) (Baradaran et al., 2013). Importantly, a quinone-binding pocket is identified at the interface between the membrane subunits NuoA, NuoH, and the cytoplasmic subunits NuoB and NuoCD (Fig. 1.15A; schematically represented in Fig. 1.15B) (Baradaran et al., 2013). It consists of a long, narrow hydrophilic crevice approximately 30 Å in length, which extends from the inner membrane into the cytoplasmic side of the complex (Baradaran et al., 2013). The spatial restriction of pocket forces the bound quinone molecule to adopt an extended conformation, with the quinone head group deep inside and the end of the poly-isoprenoid tail at the pocket’s opening (Baradaran et al., 2013).

As a proton pump, the Nuo complex couples the electron transport from NADH to the bound quinone to the translocation of 4 protons across the inner membrane from the cytoplasm

into the periplasm via a complex, coordinated set of local molecular rearrangements of its subunits. Upon the binding of NADH at the primary active site located within the cytoplasmic NuoF (Fig. 14A and B), the non-covalently bound FMN (Braun et al., 1998) oxidizes the electron donor into NAD+, and transfer the two electrons extracted down a column of strategically arranged Fe-S clusters that are carried within the subunits NuoF (N3 cluster), NuoG (clusters N1b, N4 and N5), NuoI (clusters N6a and N6b) and NuoB (N2 cluster) in the following sequence: N3→N1b→N4→N5→N6a→N6b→N2 (Fig. 1.15B; indicated with small red arrows) (Euro et al., 2008; Verkhovskaya et al., 2008). Neither the N1a cluster in NuoE nor the N7 cluster in NuoG participates in the electron transport (Fig. 1.15B) (Euro et al., 2008; Verkhovskaya et al., 2008).

Subsequently, the electrons are transferred from the N2 cluster to the head group of a bound quinone that is approximately 12 Å away (Verkhovskaya et al., 2008; Baradaran et al., 2013). The energy generated in the transfer of electrons results in a defined series of local molecular rearrangements that cascades throughout the entire complex (Baradaran et al., 2013). Specifically, prior to the reduction of the bound quinone, the antiporter-like NuoL, NuoM, NuoN and the NuoAHJK tetrad all adopt a similar conformation, such that the cytoplasmic half-channel opens up to allow the trapping of protons (Fig. 1.15B). After that, the reduction of the bound quinone induces a conformational change that spreads from NuoAHJK to NuoL, mediated by the interaction between neighboring subunits as well as the Helix HL and βH element (Baradaran et al., 2013). As a result, the cytoplasmic half-channels are shuttered, while the periplasmic half- channels open up (Baradaran et al., 2013). At the same time, the trapped protons are moved longitudinally within each subunit via the interactions among conserved and strategically positioned charged and polar residues (Fig. 1.15B; indicated with horizontal arrows along the blue line). The protons are then released into periplasm upon reaching the base of the opened periplasmic half-channel (Baradaran et al., 2013). The entire cycle then repeats itself upon the binding of a fresh quinone.

1.5.2.2 Fumarate Reductase (FrdA-D) from E. coli

The fumarate reductase (FrdA-D) complements the Nuo complex as the terminal reductase in the major anaerobic respiratory chain in E. coli (Gennis & Stewart, 1996; Unden & Bongaerts, 1997). It is composed of a soluble fragment containing two cytoplasmic subunits (FrdA and

FrdB), and a membrane fragment with two integral membrane subuits (FrdC and FrdD) (Fig. 1.15C), as illustrated by its X-ray structure solved at 2.7 Å (Iverson et al., 2002). Similar to the Nuo complex, the soluble fragment of the Frd complex contains all of the redox cofactors that are required for electron transport. These include a covalently bound FAD (flavin adenine dinucleotide) within the primary active site in FrdA, and three Fe-S clusters (one [2Fe-2S], one [4Fe-4S] and one [3Fe-4S]) in FrdB, all of which are aligned in a single column to facilitate electron flow (Fig. 1.15C) (Iverson et al., 2002). The membrane fragment of the Frd complex is composed of the subunits FrdC and FrdD, both of which are made up of three TM helices (Fig. 1.15C). Thus, the soluble fragment is anchored to the inner membrane via the membrane fragment, with FrdB in direct contact with the two membrane subunits while FrdA interacts solely with FrdB (Fig. 1.15C). Based on the bound menaquinone (MK) molecules detected in the X-ray structure, two quinone-binding sites were identified at the interface between FrdC and

FrdD (Fig. 1.15C) – one in proximity to FrdB, known as QP (P stands for proximal), and the other on the opposite end of the membrane fragment, known as QD (D stands for distal) (Fig. 1.15D) (Cecchini et al., 2002; Iverson et al., 2002). Based on the relatively large distance

between QP and QD (~25 Å) and the known structures of homologues in other species, QP is

believed to be the primary quinone-binding site for electron transport, while QD may have a structural role for the membrane anchor (Fig. 1.15D) (Cecchini et al., 2002).

During anaerobic respiration on fumarate, the Frd complex facilitates the transfer of electrons from the pool of menaquinol (MKH2) in the inner membrane to the terminal electron

acceptor, fumarate (Fig. 1.15D) (Cecchini et al., 2002). Upon binding at QP, MKH2 is oxidized to replenish the pool of MK, while the two electrons extracted are transferred to the [3Fe-4S] cluster in FrdB (Fig. 1.15D) (Cecchini et al., 2002). Importantly, unlike other terminal reductases discussed previously (Fig. 1.12), the protons released during the oxidation of MKH2 are released back into the cytoplasm instead of the periplasm (Fig. 1.15D) (Cecchini et al., 2002). Next, the electrons are passed to the [4Fe-4S] cluster and then to the [2Fe-2S] cluster, after which they are transferred to the FAD cofactor within the primary active site of FrdA (Fig. 1.15D) (Cecchini et al., 2002). The binding of fumarate at the substrate-binding site of FrdA, which is next to the primary active site, allows the transfer of electrons from FAD to fumarate, during with the electron acceptor is reduced to succinate (Fig. 1.15D) (Cecchini et al., 2002).

1.5.3 Summary on the Aerobic and Anaerobic Respiration in E. coli

In summary, the E. coli respiratory electron transport chains are comprised of one primary dehydrogenase and one terminal reductase that are coupled via lipid-soluble quinones in the inner membrane. The various dehydrogenase-reductase combinations allow the cell to be highly versatile in utilizing different electron donors / acceptors for energy across different growth conditions. Most of these respiratory proteins are dedicated to either aerobic or anaerobic respiration, and their expression is regulated via a complex network of transcriptional regulators that respond to changes in oxygen level and availability of specific electron acceptors, such as nitrate. Similarly, ubiquinone is used primarily for aerobic respiration, whereas menaquinone and demethylmenauqinone are used almost exclusively under anaerobic condition.

The use of different electron acceptors is prioritized based on energy yield. Thus, the high-yield aerobic respiratory chains are preferred over the anaerobic ones that are of lower- yields. Among the anaerobic respiratory chains, those involved in nitrate reduction are preferred over those utilizing alternative electron acceptors. In general, Ndh coupled with CyoA-D via ubiquinone is the dominant aerobic respiratory chain, while NuoA-N coupled with FrdA-D via menaquinone serves as the major respiratory chain under anaerobic condition in the absence of nitrate. Both the structure and molecular function of NuoA-N and FrdA-D have been characterized extensively, and the mechanisms underlying the electron transport across the two complexes have been described in detail.

In the next two chapters, I will present my research on the biological function of the MoxR AAA+ ATPase RavA and the associated VWA protein ViaA in E. coli. Experimental evidence will be presented with respect to the newly identified interactions between RavA-ViaA and specific subunits of both the Nuo and Frd respiratory complexes. The potential regulatory role of RavA-ViaA in the activity of the Frd complex will also be discussed in detail.

Chapter 2 The MoxR ATPase RavA and its Cofactor ViaA Interact with the NADH:Ubiquinone Oxidoreductase I in E. coli

Data attribution: I performed the majority of the experiments presented in this chapter. The microarray experiments and data analysis were performed by Dr. Jamie D. Snider (Department of Biochemistry, University of Toronto, Toronto, ON). The eSGA section was contributed by Dr. Chris Graham and Prof. Mohan Babu (Department of Biochemistry, Research and Innovation Centre, University of Regina, Regina, SK).

Publication details: This chapter is a reprint with slight modifications, with the authors’ permission, from the following PLoS One article:

Wong, K. S. et al. (2013) The MoxR ATPase RavA and its Cofactor ViaA Interact with the NADH:Ubiquinone Oxidoreductase I in Escherichia coli. PLoS One. 9(1): e88529. [PMID: 24454883]

2 The MoxR ATPase RavA and its Cofactor ViaA Interact with the NADH:Ubiquinone Oxidoreductase I in Escherichia coli 2.1 Chapter Summary

MoxR ATPases are widespread throughout bacteria and archaea. The experimental evidence to date suggests that these proteins have chaperone-like roles in facilitating the maturation of dedicated protein complexes that are functionally diverse. In Escherichia coli, the MoxR ATPase RavA and its putative cofactor ViaA are found to exist in early stationary-phase cells at 37 °C at low levels of about 350 and 90 molecules per cell, respectively. Both proteins are predominantly localized to the cytoplasm, but ViaA was also unexpectedly found to localize to the cell membrane. Whole genome microarrays and synthetic lethality studies both indicated that RavA- ViaA are genetically linked to Fe-S cluster assembly and specific respiratory pathways. Systematic analysis of mutant strains of ravA and viaA indicated that RavA-ViaA sensitizes cells to sublethal concentrations of aminoglycosides. Furthermore, this effect was dependent on RavA’s ATPase activity, and on the presence of specific subunits of NADH:ubiquinone oxidoreductase I (Nuo Complex, or Complex I). Importantly, both RavA and ViaA were found to physically interact with specific Nuo subunits. We propose that RavA-ViaA facilitate the maturation of the Nuo complex.

2.2 Introduction

The MoxR family of AAA+ ATPases is widespread across different bacterial and archaeal species (Snider & Houry, 2006; Wong & Houry, 2012). Based on sequence similarity and local genetic structure, MoxR proteins are subdivided into seven subfamilies: MRP (MoxR Proper), APE0892, RavA, CGN (CbbQ / GvpN / NorQ), APE2220, PA2707, and YehL (Snider & Houry, 2006). The exact roles of MoxR proteins in vivo are unclear, although the experimental evidence collected to date suggests that they have chaperone-like functions and are involved in the maturation and activation of specific protein complexes. For example, MoxR of the MRP subfamily in Paracoccus denitrificans and Methylobacterium extorquens is important for the activation of methanol dehydrogenase (MDH) (van Spanning et al., 1991; Toyama et al., 1998). NirQ/NorQ, which belong to the CGN subfamily, are necessary for the activity of nitric oxide

reductase in Pseudomonas stutzeri (Jungst & Zumft, 1992), Pseudomonas aeruginosa (Arai et al., 1999), Paracoccus denitrificans (de Boer et al., 1996), and Rhodobacter sphaeroides 2.4.3 (Bartnikas et al., 1997). In the chemolithoautotrophic eubacterium Oligotropha carboxidovorans OM5, CoxD, a member of the APE2220 subfamily, is required for the assembly of the

[CuSMoO2] cluster in the carbon-monoxide (CO) dehydrogenase, which enables the bacteria to utilize CO as a sole carbon source (Pelzmann et al., 2009).

MoxR proteins also have important roles in other biological processes. For example, in Rhizobium leguminosarum, RL3499 of the MRP subfamily is optimally expressed in stationary phase cells and is important for both membrane integrity and cell morphology (Vanderlinde et al., 2011). In the crenarachaeal Acidianus two-tailed virus (ATV), p618 of the RavA subfamily interacts with p892, which forms filamentous structures and is believed to play a role in the extracellular, host-independent formation of viral tails (Scheele et al., 2011). In Francisella tularensis, the MRP protein, FTL_0200, has been implicated in multiple stress tolerance pathways and was shown to be important for infection (Dieppedale et al., 2011; Dieppedale et al., 2013).

Generally, MoxR proteins co-occur with at least one cofactor that carries a von Willebrand factor A (VWA) domain. The genes encoding these proteins are usually in close proximity within the genome (Snider & Houry, 2006). The VWA domain contains a metal- binding motif, known as the MIDAS (Metal Ion-dependant Adhesion Site) motif. This motif binds a single divalent metal cation, usually Mg2+, and is often involved in mediating protein- protein interactions (Whittaker & Hynes, 2002). While eukaryotic VWA proteins have been characterized extensively, the cellular function of prokaryotic VWA proteins remains poorly understood. Current experimental evidence for these proteins suggests diverse functions, including surface adhesion, fibrinogen binding, metal insertion into protoporphyrin IX, and pathogenesis (Kachlany et al., 2000; Katerov et al., 2000; Willows, 2003; Konto-Ghiorghi et al., 2009).

Two MoxR proteins are encoded in the genome of Escherichia coli K-12 MG1655: RavA (Regulatory ATPase variant A) of the RavA subfamily, and YehL of the YehL subfamily. We have characterized RavA extensively using various biochemical and biophysical methods. RavA co-occurs with the VWA protein ViaA (VWA interacting with AAA+ ATPase), and the genes

encoding these proteins form an operon (Snider et al., 2006). Under aerobic conditions, the co- expression of RavA and ViaA is primarily dependent on the stationary phase sigma factor σS (RpoS) (Snider et al., 2006). RavA interacts physically with ViaA, which results in the enhancement of RavA ATPase activity (Snider et al., 2006). Typical of AAA+ ATPases, RavA forms a hexamer via its AAA+ module (Snider et al., 2006; El Bakkouri et al., 2010) as observed based on the X-ray crystal structure we solved for RavA protomer and the 3D electron microscopy reconstruction of the protein hexamer (El Bakkouri et al., 2010). We also found that RavA interacts strongly with the inducible lysine decarboxylase LdcI (or CadA), forming a large cage-like complex (Snider et al., 2006; El Bakkouri et al., 2010). LdcI is an important acid stress response protein in E. coli (Park et al., 1996; Kanjee et al., 2011).

Despite the detailed biochemical and biophysical characterization described above, the cellular function of RavA in vivo remains elusive. Association of RavA with LdcI suggests a potential role for the AAA+ ATPase in bacterial acid stress response. Recently, we discovered that LdcI binds the alarmone ppGpp, the primary activator of the stringent response (Jain et al., 2006), and that the binding inhibits LdcI activity (Kanjee et al., 2011). Furthermore, RavA was found to antagonize the effect of ppGpp inhibition on LdcI (El Bakkouri et al., 2010). While RavA and, indirectly, ViaA might function to modulate the activity of LdcI, we suspect that the system must have other roles in the cell.

To identify other cellular roles for the RavA-ViaA chaperone-like system, we carried out genome wide genetic interaction and microarray analyses, phenotypic screens, and physical interaction studies. These experiments demonstrated that both RavA and ViaA interact with specific subunits of the highly conserved NADH:ubiquinone oxidoreductase I complex (i.e. Nuo complex, or Complex I), particularly with NuoA and NuoF under aerobic conditions, and with the fused NuoCD under anaerobic conditions. To our knowledge, this is the first report of an interaction between the Nuo complex and a member of the MoxR AAA+ ATPases.

2.3 Materials and Methods

2.3.1 Bacterial strains and plasmids used

All bacterial strains used are listed in Table 2.1 with the exception of the 30 BW25113 single- gene knockouts (KO) used in our suppression mutation analysis (see below). Wild type (WT) E.

55 coli K-12 MG1655 was obtained from ATCC (catalog number 700926). The corresponding single KO mutants for ravA (ΔravA::cat) and viaA (ΔviaA::cat) were generated by transducing the required chloramphenicol resistance KO cassettes (cat) from the original DY330 strains to MG1655 via P1 phage (Sternberg & Maurer, 1991) as previously described (Snider et al., 2006). A double KO mutant for ravA and viaA (ΔravAviaA::cat) was also generated in the same manner. The required cat KO cassette was generated by PCR using the primers RKO_forward (5'- agaaacgtctatactcgcaatttacgcagaacttttgacgaaagggtgtaggctggagctgcttc-3') and VKO_reverse (5'- gcgagagcgtcccttctctgctgtaataatttatcgccgccagcgcatatgaatatcctccttag-3'), and the pKD3 template plasmid as described (Snider et al., 2006). The cat KO cassettes in ΔravA::cat, ΔviaA::cat and ΔravAviaA::cat was later removed using the pCP20 plasmid that expresses the FLP recombinase (Datsenko & Wanner, 2000) to obtain ΔravA, ΔviaA and ΔravAviaA, respectively, with no markers. The generated strains were verified by sequencing. Only KOs without markers (clean KOs) were used in the subsequent experiments with the exception of the microarray experiments.

TABLE 2.1. List of bacterial strains and plasmids used in this study

Bacterial Strains Genotype Reference

MG1655 F-, rph-1, λ- (Guyer et al., 1981) MG1655 ΔravA::cat* MG1655, ΔravA::cat (Snider et al., 2006) MG1655 ΔviaA::cat MG1655, ΔviaA::cat (Snider et al., 2006) MG1655 ΔravA MG1655, ΔravA This study MG1655 ΔviaA MG1655, ΔviaA This study MG1655 ΔravAviaA MG1655, ΔravAviaA This study

DY330 W3110, ΔlacU169, gal490, λcI857, Δ(cro-bioA) (Yu et al., 2000) DY330 ΔravA::cat* DY330, ΔravA::cat (Snider et al., 2006) DY330 ΔviaA::cat DY330, ΔviaA::cat (Snider et al., 2006) DY330 ΔravA-viaA::cat DY330, ΔravAviaA::cat This study DY330 nuoA-SPA::kanR DY330, nuoA-SPA::kanR (Butland et al., 2005) DY330 nuoB-SPA::kanR DY330, nuoB-SPA::kanR (Butland et al., 2005) DY330 nuoCD-SPA::kanR DY330, nuoCD-SPA::kanR (Butland et al., 2005) DY330 nuoE-SPA::kanR DY330, nuoE-SPA::kanR (Butland et al., 2005) DY330 nuoF-SPA::kanR DY330, nuoF-SPA::kanR (Butland et al., 2005) DY330 nuoG-SPA::kanR DY330, nuoG-SPA::kanR (Butland et al., 2005) DY330 sdhA-SPA::kanR DY330, sdhA-SPA::kanR (Butland et al., 2005) DY330 sdhB-SPA::kanR DY330, sdhB-SPA::kanR (Butland et al., 2005) DY330 cyoB-SPA::kanR DY330, cyoB-SPA::kanR (Butland et al., 2005)

DY330 cyoC-SPA::kanR DY330, cyoC-SPA::kanR (Butland et al., 2005) DY330 nuoA- DY330, nuoA-SPA::kanR, ΔviaA::cat This study SPA::kanR ΔviaA::cat DY330 nuoCD-SPA::kanR DY330, nuoCD-SPA::kanR, ΔviaA::cat This study ΔviaA::cat DY330 nuoF-SPA::kanR DY330, nuoF-SPA::kanR, ΔviaA::cat This study ΔviaA::cat

Hfr Cavalli (Hfr C) Hfr(PO2A), relA1, spoT1, metB1, rrnB-2, mcrB1, (Bachmann, 1972) creC510 Hfr C ΔravA::cat* Hfr C, ΔravA::cat This study Hfr C ΔviaA::cat Hfr C, ΔviaA::cat This study Hfr C ΔravA-viaA::cat Hfr C, ΔravAviaA::cat This study

Plasmids Description Reference

p11 Cloning vector derived from pET15b(+) (Zhang et al., 2001) pR p11-ravAp-ravA, for overexpression of RavA regulated This study by the native ravA promoter

pRK52Q p11-ravAp-ravA(K52Q), for overexpression of RavA This study Walker A mutant regulated by the native ravA promoter pRV p11-ravAp-ravAviaA, for RavA and ViaA overexpression This study regulated by the native ravA promoter

pRK52QV p11-ravAp-ravA(K52Q)viaA, for overexpression of RavA This study Walker A mutant and wild-type ViaA regulated by the native ravA promoter cat = chloramphenicol acetyltransferase gene; confers resistance to chloramphenicol. kanR = kanamycin resistance gene *ViaA expression is increased in ΔravA::cat compared to WT (see Fig. 2.2).

For the customized E. coli synthetic genetic arrays (eSGA) (Butland et al., 2008), ΔravA::cat, ΔviaA::cat and ΔravAviaA::cat, were generated by transducing the cat KO cassettes from MG1655 into the Hfr C background via P1 bacteriophage as described (Sternberg & Maurer, 1991). For immunoprecipitation, DY330 strains expressing endogenous proteins fused with a C-terminal SPA (Sequential Peptide Affinity) tag for NuoA, NuoB, NuoCD, NuoE, NuoF, NuoG, SdhA, SdhB, CyoB and CyoC were made as described (Babu et al., 2009). In addition, ΔviaA::cat equivalents were also constructed for the strains expressing NuoA-SPA, NuoCD-SPA and NuoF-SPA via P1 phage transduction (Sternberg & Maurer, 1991).

All plasmids used are also listed in Table 2.1. The vector p11 was obtained from the Toronto Structural Genomics Consortium (SGC). The plasmids p11-ravAp-ravA (pR) and p11- ravAp-ravAviaA (pRV) were constructed by cloning the ravA or the ravAviaA open reading

frame (ORF) along with the native ravA promoter (ravAp; 206 bp immediately upstream of the ravA ORF) into the p11 plasmid. The PCR primers RAVA2_forward (5'- gtggatccgaaatgtgtgcttagtcccttg-3') and RAVA_reverse (5'-tacgtaggatccttagcattgttgtgcctggcg-3') were used to amplify the required DNA fragment for the pR plasmid, and the primers RAVA2- forward and VIAA_reverse (5'-ctatggatccttatcgccgccagcgtctgagc-3') for the pRV plasmid. All fragments were cloned into p11 using the BglII and BamHI restriction sites, which removed the endogenous T7 promoter sequence in the process. To generate the Walker A mutant of RavA, the point mutation K52Q was introduced to the Walker A motif of RavA (GPPGIAKS; mutated residue is underlined) in both pR and pRV, using the QuikChange Site-Directed Mutagenesis kit (Stratagene) and the primers RavA_K52Q_F (5'-cgccaggtattgcccaaagtttgatcgcc-3') and RavA_K52Q_R (5'-ggcgatcaaactttgggcaatacctggcg-3'), which yielded the plasmids p11-ravAp- ravAK52Q (pRK52Q) and p11-ravAp-ravAK52QviaA (pRK52QV), respectively. All plasmids were verified by DNA sequencing.

2.3.2 Quantification of RavA and ViaA levels in cells

WT E. coli MG1655 cells were grown in Luria-Bertani (LB) media (10 g/L bacto-tryptone, 5 g/L yeast extract, and 10 g/L NaCl) at 37 °C aerobically in 2-L culture flasks with vigorous shaking

for 24 hours. Cell growth was tracked by monitoring the changes in OD600 at specific time points. Cells were harvested every two hours by centrifugation and flash-frozen in liquid nitrogen until use. To determine the levels of RavA and ViaA, cell pellets were thawed on ice and then resuspended in a 0.1 M potassium phosphate buffer (pH 7.5) supplemented with 0.1 M NaCl. The volume of each sample was adjusted to achieve a final cell count of approximately 3.8 9 × 10 cells/mL as determined by OD600. Cells were lysed by sonication followed by mixing with 4 × SDS-PAGE sample buffer (200 mM TrisHCl, pH 6.8, 8% SDS, 0.4% bromophenol blue, 40% glycerol, and 400 mM β-mercaptoethanol), and the proteins were separated on 10% or 12% polyacrylamide gels. The amounts of RavA and ViaA were determined by quantitative Western blotting. The numbers of RavA and ViaA molecules expressed per cell were then calculated based on the molecular weights of the two proteins. For comparison, the level of the ClpP protease was also analyzed, while the inner membrane-bound signal peptidase LepB was used as a loading control. The α-RavA, α-Via and α-ClpP rabbit polyclonal antibodies were generated at the Division of Comparative Medicine, University of Toronto. The α-LepB rabbit polyclonal antibody was a generous gift from Dr. Jan Willem de Gier (Stockholm University, Sweden).

Purified RavA, ViaA, and ClpP proteins were used as quantification standards. To estimate the 8 number of proteins per cell, we used the standard conversion assuming 1 OD600 = 5 x 10 cells/mL for E. coli cells.

2.3.3 Subcellular localization of RavA and ViaA

WT E. coli MG1655 cells were grown in LB at 37 °C for 16-18 hours to stationary phase. Subcellular fractionation of the cells was performed as described in (Lee & Ahn, 2000) and (Lemire & Weiner, 1986), with the following modification. After the extraction of periplasmic proteins by osmotic shock, cells were spun down by centrifugation at 4 °C for 30 minutes. Cells were re-suspended in 20 mM TrisHCl (pH 8.0) supplemented with 2 mM EDTA (pH 8.0), and were lysed by French Press. The cytosolic fraction was then cleared of membrane vesicles by ultracentrifugation at ~190000 × g at 4 °C for 1 hour in a Beckman-Coulter Optima TLX bench- top ultracentrifuge. Subcellular localization of RavA and ViaA was then determined by Western blotting. The ClpP protease and the inner membrane-bound LepB signal peptidase were chosen as the localization standards for the cytoplasmic and membrane proteins, respectively. Protein levels were estimated by densitometry using Quantity One v. 4.6.5 (Bio-Rad).

2.3.4 Microarray experiments and data analysis

MG1655 WT, ΔravA::cat, WT + p11 and WT + pRV were grown in LB at 37 °C with a starting

OD600 of ~0.025. Stationary phase cells were harvested when OD600 reached ~3 and total RNA was isolated from 500 μL aliquots of each strain using the Qiagen RNeasy Mini Kit with RNAprotect Bacteria Reagent following the manufacturer’s instructions. Samples were stored at -80°C until use. Total RNA quality was assessed using the Agilent 2100 Bioanalyzer (Agilent Technologies).

All the microarray experiments were carried out at the Centre for Applied Genomics Microarray Facility, Hospital for Sick Children (Toronto). Sample preparation and array processing were performed following standard protocols. cDNA synthesis was performed with Invitrogen Superscript II Reverse Transcriptase enzyme using random primers and 10 μg total RNA template. RNA template was subsequently degraded using NaOH, which was followed by cDNA cleanup using Qiagen MinElute PCR Purification Columns. The purified cDNA was fragmented with DNase I (GE Healthcare) and labeled with biotin at the 3′-end using GeneChip

DNA Labeling Reagent (Affymetrix) and Terminal Deoxynucleotidyl Transferase (Promega). 2 to 5 μg of biotin-labeled cDNA were used in the subsequent hybridization to the E. coli Genome 2.0 Arrays. Hybridization, washing, and staining were performed in the Affymetrix GeneChip Hybridization Oven 640 and Fluidics Station 450. Arrays were scanned using the Affymetrix GeneChip Scanner 3000. Three replicates were prepared for each of the five strains used.

Single array data analysis was performed using the GeneChip Operating Software (GCOS). Array signal intensities were globally scaled using an All Probe Sets Scaling strategy, with a target signal of 150. The presence or absence of signals was determined using default parameters for the GeneChip E. coli Genome 2.0 Array. A signal intensity of zero was automatically assigned to any gene considered as ‘absent’. All details pertaining to the statistical analysis of the raw data can be found in the Affymetrix GeneChip Analysis Manual (Data analysis fundamentals; available on the Affymetrix company website). Both raw and per-assay- normalized data were deposited in the ArrayExpress database of the European Bioinformatics Institute (EMBL-EBI) (Accession number: E-MTAB-2001).

Comparison analysis of the resulting data was performed for ΔravA::cat vs. WT and WT + pRV vs. WT + p11, using a bootstrapping approach for unpaired data. All analyses, based on t- statistics, were performed using in-house software. Changes in gene expression levels having p-

values less than 0.05 were considered significant and the signal log2 ratio of these changes were

calculated. Only significant changes with absolute signal log2 ratios of 0.6 (~1.5 fold absolute change in transcript level) or greater were selected for further analysis. A manual review of the change in gene levels was then performed. All remaining genes were examined using the data currently available in the databases EcoCyc (Keseler et al., 2005) and UniProt (Magrane & Consortium, 2011), and were grouped together into operons whenever possible. Fold-changes in gene expression are represented as heatmaps that are generated with the online software Matrix2png (Pavlidis & Noble, 2003).

2.3.5 E. coli Synthetic Genetic Array (eSGA) analysis

Genes deemed functionally linked to RavA-ViaA by the microarray experiments were validated further by customized E. coli Synthetic Genetic Arrays (Butland et al., 2008; Babu et al., 2011). The double deletion mutants for ΔravA::cat, ΔviaA::cat and ΔravAviaA::cat were constructed via conjugation between the respective Hfr C donor strains carrying the KO cassettes for ravA and/or

viaA and the selected BW25113 recipient strains from the Keio collection of E. coli single-gene deletion mutants (Baba et al., 2006; Yamamoto et al., 2009), following the same protocols as described previously in (Butland et al., 2008) and (Babu et al., 2011). The closest flanking genes upstream and downstream of the genes/operons of interest that show no genetic interaction with either ravA or viaA were used as controls.

2.3.6 Growth of E. coli MG1655 in cultures containing sublethal dosages of different antibiotics

E. coli MG1655 WT, ΔravA, ΔviaA and ΔravAviaA were grown on LB-agar plates overnight at 37 °C to obtain single colonies. Pre-cultures were prepared for each strain by inoculating a single colony into 3 mL of fresh LB and grown with rigorous shaking at 37 °C overnight. Next day, the pre-cultures were used to inoculate fresh LB supplemented with 4 μg/mL kanamycin, 6 μg/mL

streptomycin, 0.5 μg/mL tetracycline, or 1.2 μg/mL chloramphenicol at a starting OD600 of ~0.01. The dosages of antibiotics used were based on similar experiments as reported in (Girgis et al., 2009). Further supplementation to the growth media included the addition of 750 μM reduced L-glutathione (GSH) or 250 μM 2,2'-dipyridyl (DP) where applicable. Growth of cells was monitored via OD600 using a SpectraMax 340PC Plate Reader. Three independent cultures were prepared for each strain and for each growth condition.

Complementation experiments were performed the same way on the following strains:

WT transformed with p11, pR, pRV, pRK52Q or pRK52QV; ΔravA transformed with p11, pR or pRK52Q; and ΔravAviaA transformed with p11, pR, pRV or pRK52QV. 100 μg/mL ampicillin was added to the growth media for plasmid maintenance.

2.3.7 Analysis of intracellular oxidative stress by DHR fluorescence

MG1655 WT + p11, ΔravAviaA + p11, ΔravAviaA + pRV and ΔravAviaA + pRK52QV strains were grown on LB-agar plates supplemented with 100 μg/mL ampicillin overnight at 37 °C to obtain single colonies. Pre-cultures were prepared by inoculating fresh LB + 50 μg/mL ampicillin and grown overnight at 37 °C with rigorous shaking. Next day, the pre-cultures were used to inoculate fresh LB, supplemented with 4 μg/mL kanamycin, 8 mM GSH and/or 250 μM

DP as required, at a starting OD600 of ~0.05. Cells were grown at 37 °C with rigorous shaking to late log phase (4-5 hours). The membrane-permeable reactive oxygen species (ROS) indicator

dihydrorhodamine 123 (DHR) was then added to each culture at 110 μM (8 μg/mL) final concentration from a 5 mg/mL DMSO stock solution, followed by a 30-minute incubation at 37 °C without shaking. Cells that were incubated with DMSO instead of DHR were used as unstained controls. Afterwards, cells were harvested by centrifugation and re-suspended in PBS

(10 mM Na2HPO4, 2 mM KH2PO4, pH 7.4, 137 mM NaCl, and 2.7 mM KCl). 100-μL aliquots of the cell suspensions were then collected in a 96-well plate, and both DHR fluorescence (λex =

500 nm; λem = 530 nm) and OD600 were measured using a Perkin Elmer EnSpire 2300 Multi-

label Reader. The raw DHR fluorescence readings were normalized by their respective OD600 to allow comparison across samples. Autofluorescence was determined from unstained cells and subtracted from the normalized fluorescence readings.

2.3.8 Suppression mutation analysis to identify direct functional targets of RavA and ViaA

E. coli BW25113 single-gene KO’s were selected from the Keio collection (Baba et al., 2006; Yamamoto et al., 2009). Clean KO’s were then generated using the pCP20 plasmid as described above. After confirming the removal of the kanamycin resistance KO cassette and the curing of pCP20, each clean KO was transformed with p11, pRV or pRK52QV. The aerobic growth of the

transformed clean KOs in LB or LB + 4 μg/ml kanamycin was monitored by OD600 over 10 hours. Three independent cultures were prepared for each strain tested. To construct the growth profiles for each strain, the data collected for growth in LB + kanamycin were normalized with the corresponding data collected for growth in LB. This is necessary to exclude any inherent differences in growth due to the KO’s genetic background that are independent of the effects of RavA-ViaA.

2.3.9 Identifying the physical interactors of RavA and ViaA by immunoprecipitation

To confirm the interaction between RavA-ViaA and its downstream targets identified by suppression mutation analysis, DY330 strains expressing endogenous NuoA, NuoB, NuoCD, NuoE, NuoF, NuoG, SdhA, SdhB, CyoB and CyoC that carry C-terminal SPA tags (Babu et al., 2009) were grown aerobically or anaerobically in LB at 30 °C overnight. Cells were harvested by centrifugation, and then re-suspended in immunoprecipitation (IP) buffer (25 mM TrisHCl, pH

7.5, 100 mM KCl, 10 mM MgCl2, 1 mM CaCl2, 0.2 mM EDTA, 1% Triton X-100, 10%

glycerol, and 0.5 mM DTT) supplemented with 1 mg/mL lysozyme and 0.1 U/mL DNase I. Cells were lysed by sonication, and the cell lysate cleared of insoluble debris by centrifugation at 4 °C. The protein complexes carrying the SPA-tagged targets were purified by incubating the cell lysate with α-FLAG M2 affinity gel (Sigma-Aldrich) at 4 °C for 1 hour, followed by three 5- minute washes with IP buffer. The bound complexes were eluted using 3xFLAG peptide re- suspended in IP buffer at 1 mg/mL. The complexes were analyzed by SDS-PAGE and Western blotting for the presence of RavA and ViaA.

To assess the role of ViaA in mediating the interaction between RavA and the SPA- tagged targets, viaA was deleted (ΔviaA::cat) in DY330 strains expressing NuoA-SPA, NuoCD- SPA and NuoF-SPA and the immunoprecipitation experiment were repeated.

2.4 Results

2.4.1 Expression and localization of RavA and ViaA

In an effort to assess the function of RavA and ViaA in E. coli, we first investigated the expression and localization profiles of the two proteins (Fig. 2.1A). For aerobically growing culture in LB media at 37 °C, the optimal expression of both proteins occurred when cells entered stationary phase (6 hours post inoculation; Fig. 2.1A) consistent with our previous observations that the ravAviaA operon is induced by σS (Snider et al., 2006). We estimated that approximately 350 molecules of RavA and 90 molecules of ViaA are present per cell at optimum (Fig. 2.1C). These numbers are considerably lower in comparison to housekeeping proteins such as the molecular chaperone DnaK (11000-12000 molecules per cell (Tomoyasu et al., 1998)), the ClpP subunit of the ClpXP protease complex (approximately 15000 molecules per cell; Fig. 2.1D) or the ribosome-associated trigger factor (approximately 31000 molecules per cell (Ishihama et al., 2008)).

At stationary phase, RavA is mainly localized to the cytoplasm, while ViaA is found in both the cytoplasm and unexpectedly, the inner membrane fraction (Fig. 2.1B). Bioinformatic analysis of ViaA’s primary sequence does not reveal any signal peptides or membrane- associating sequence motifs (data not shown). Thus, the apparent localization of ViaA to the cell membrane is likely an indication of its physical association with a membrane-bound target.

Figure 2.1. Expression and localization of RavA and ViaA in E. coli MG1655

(A) Expression of RavA and ViaA in WT MG1655 grown aerobically in LB at 37 °C profiled over 24 hours by quantitative Western blotting. Both ClpP and LepB were used as loading controls. Different amounts of purified RavA, ViaA, and ClpP were used as indicated to provide the necessary quantification standards.

(B) Total cell lysate and subcellular fractions of WT MG1655 cells grown aerobically to stationary phase in LB at 37 °C were Western-blotted for the presence of RavA and ViaA. ClpP and LepB provide localization standards for cytoplasmic and membrane proteins, respectively. The amount of proteins loaded per lane for each blot is as indicated.

(C) Quantification of RavA and ViaA expressed in WT MG1655 by densitometry, using the Western blots shown in (A). Both OD600 of the culture and the amount of RavA and ViaA expressed per cell at each time point are shown graphically. Dotted lines trace the expression levels of RavA and ViaA.

(D) Quantification of ClpP expressed in WT MG1655 by densitometry, using the Western blots shown in (A). Both OD600 of the culture and the amount of ClpP expressed per cell at each time point are shown graphically. Expression levels of RavA and ViaA are also provided for visual comparison. The dotted line trace the expression level of ClpP.

______

2.4.2 The function of RavA and ViaA is linked to Fe-S cluster assembly and specific respiratory pathways

In order to identify the biological pathways that are functionally linked to RavA and ViaA, we analyzed the gene expression profile of early stationary phase cells having different RavA/ViaA levels using whole-genome microarrays, namely: ΔravA::cat vs. WT (set 1) and WT + pRV vs. WT + p11 (set 2). The pathways associated with RavA and ViaA were determined using the genes and operons with statistically significant changes in expression upon manipulation of RavA and ViaA levels. It should be noted that the ΔravA::cat strain is a ravA KO as well as a ViaA overexpressor (Fig. 2.2), presumably due to a polar effect of the marker on viaA transcription. Such a polar effect is not observed if the marker is removed (Fig. 2.2).

Among a total of 300 different genes showing significant changes in expression in sets 1 or 2 (see Appendix A), 7 respond to both the loss and increase in RavA-ViaA levels, i.e. their mRNA levels change in both sets 1 and 2, namely: asnA; cysC and cysD; feoA, feoB and feoC; and metK. For the genes whose mRNA levels change only in set 1 or set 2, many of them are encoded on the same operons, while others share common biochemical pathways (Appendix A). Some of these genes have potentially greater functional relevance and thus were examined further (see below), and their organization into operons and/or regulons is illustrated in Fig. 2.3.

There are 25 genes in both sets 1 and 2 that are associated with the assembly of Fe-S clusters (Fig. 2.4). These include genes involved in iron uptake and cysteine biosynthesis. In addition, iscR, iscS, hscA and hscB (see ‘Fe-S Clusters Assembly / Repair Genes’ in Fig. 2.4) encode key proteins of the Isc Fe-S clusters assembly pathway (Ayala-Castro et al., 2008), while ytfE gene encodes a di-iron protein important for the repair of oxidative stress-damaged Fe-S cluster proteins (Todorovic et al., 2008).

Several genes related to oxidative stress response were also identified (see ‘Oxidative Stress-induced Genes’ in Fig. 2.4). These include sodA that encodes one of the three superoxide dismutases (Fridovich, 1995) and oxyS that encodes a regulatory small RNA for oxidative stress response (Altuvia et al., 1997). Furthermore, yajL (also known as thiJ) encodes a chaperone that is involved in oxidative stress response (Gautier et al., 2012), and ydeI is important for hydrogen peroxide tolerance (Lee et al., 2010).

Figure 2.2. Expression levels of RavA and ViaA in various strain backgrounds used in this study

The various strains of E. coli MG1655 as shown were grown aerobically to early stationary phase in LB at 37 °C, and the total cell lysate prepared from them were Western-blotted for the presence of RavA and ViaA. The membrane-bound LepB was used as loading control. For WT and the KO mutant strains of ravA and/or viaA, lysate from ~7.4 × 107 cells was loaded per sample, whereas for WT cells transformed with plasmids, lysate from ~ 1.3 × 107 cells was loaded per sample. The tables provide an estimate of the number of RavA and ViaA molecules expressed per cell obtained by densitometry for each strain used. The estimation of RavA and ViaA levels for WT and WT + p11 (indicated by *) was derived from the RavA and ViaA quantification data shown in Fig. 2.1A.

______

Figure 2.3. Genomic organization of genes relevant to Fe-S clusters assembly or bacterial respiration showing statistically significant changes in the microarray experiments

Operons of the same regulon involved in the same biochemical pathways are grouped together. The length of the arrow for each gene corresponds to the size of the gene’s open reading frame. Transcripts detected in the microarray experiments are highlighted in red, and those that were not detected are in gray. All known transcriptional regulators for each operon are boxed. Activators are indicated with a ‘+’ sign and highlighted in green. Repressors are indicated with a ‘−’ sign and highlighted in red. Dual regulators are indicated with ‘+/−’ and highlighted in orange.

______

Figure 2.4. Schematic representation of genes showing significant changes in transcript levels as a result of the deletion or overexpression of RavA/ViaA

Only genes that are functionally relevant to Fe-S clusters assembly and bacterial respiration are shown. Genes that belong to the same functional category are clustered together. In addition, genes that share a common operon are listed, from top to bottom, in the same order as they would appear in the 5'-to-3' direction within the E. coli genome. Changes in gene transcription are represented as heatmaps generated using Matrix2png (Pavlidis & Noble, 2003) expressed as fold-changes with respect to either WT for ΔravA::cat (Set 1), or WT + p11 for WT + pRV (Set 2).

______

Other genes identified are associated with different respiratory processes. fumA and fumC (see ‘Fumarate Metabolism Genes’ in Fig. 2.4) encode two of the three fumarase isozymes found in E. coli, which share the same function in converting (S)-malate to fumarate in the TCA cycle (Tseng et al., 2001). FumA is an Fe-S cluster protein and is expressed during aerobiosis, whereas FumC is iron-independent and is induced primarily under oxidative stress conditions (Park & Gunsalus, 1995). hyaA, hyaB and hyaC (see ‘Hydrogenase 1 Genes’ in Fig. 2.4) encode the small, large and cytochrome b subunits, respectively, of hydrogenase 1, which drives the respiratory hydrogen uptake in the presence of oxygen (Volbeda et al., 2012). The maturation process of hydrogenase 1 requires the accessory proteins encoded by hyaD and hyaF (see ‘Hydrogenase 1 Genes’ in Fig. 2.4) (Menon et al., 1991; Fritsche et al., 1999). The genes napH, napB and napC (see ‘Periplasmic Nitrate Reductase & Cytochrome c Biogenesis Genes’ in Fig. 2.4) encode three of the five subunits of the periplasmic nitrate reductase (Nap) complex (Stewart et al., 2002; Brondijk et al., 2004). In this case, NapH is the Fe-S cluster subunit of the Nap complex (Brondijk et al., 2004). Finally, the ccm genes (see ‘Periplasmic Nitrate Reductase & Cytochrome c Biogenesis Genes’ in Fig. 2.4) share the same operon as the nap genes, and encode proteins that are involved in the biogenesis of c-type cytochromes (Stevens et al., 2011). Although they do not directly participate in bacterial respiration, the Ccm proteins are required for the Nap complex and others that require periplasmic c-type cytochromes for their function (Tanapongpipat et al., 1998; Stevens et al., 2011).

To further confirm the microarray study results, we carried out genetic lethal interaction analysis that was recently developed for E. coli (eSGA) (Butland et al., 2008; Babu et al., 2011). To construct the customized eSGA arrays, specific single-gene KO mutants from the Keio collection (Baba et al., 2006; Yamamoto et al., 2009) were selected based on the genes shown in Fig. 2.4. Genes from the adjacent regions upstream and downstream of the genes being investigated were used as controls. As shown in Fig. 2.5, the isc-hsc-fdx, cys and nap-ccm operons all exhibited synthetic lethal interactions with ravA/viaA. Additional experiments were

69 also performed to validate these genetic interactions as well as others, and the results had been published elsewhere. A modified version of these results is provided in Appendix C.

Figure 2.5. Genetic interactions between ravA/viaA and genes functionally relevant to Fe-S clusters assembly and bacterial respiration

Shown are plates demonstrating that the deletion of ravA, viaA, or ravAviaA results in synthetic lethality when genes belonging to the Isc Fe-S assembly, cysteine biosynthesis, or nap-ccm operons are also deleted. Genes sharing the same operon are grouped together in the same row whenever possible. A total of 2 replicates for each of 2 independent colonies were prepared for each donor-recipient pair, and are arranged into a 2 x 2 configuration as shown. The donors are identified on the left for each row, and the recipients on top of each column. Arrows represent the direction of the genes in each operon (coloured in dark gray) relative to the flanking control genes (coloured in light gray).

______

Taken together, both the microarray and eSGA indicated close functional links between RavA-ViaA and the homeostasis of Fe-S cluster proteins as well as bacterial respiration: from the acquisition of required substrates and the assembly of Fe-S clusters to the expression of specific respiratory enzyme complexes that depend on Fe-S cluster proteins for function. Next, we aimed to identify the potential target(s) of RavA-ViaA activity.

2.4.3 RavA and ViaA sensitize E. coli to aminoglycosides

In a recent whole-genome study, both ravA and viaA were implicated in sensitizing E. coli cells to the presence of sublethal concentrations of aminoglycosides (Girgis et al., 2009). Notably, a large majority of genes that also confer aminoglycoside sensitivity are involved in Fe-S clusters biogenesis and aerobic respiration (Girgis et al., 2009). This closely resembles the results of our high-throughput studies discussed above. To validate the deleterious effects of RavA and ViaA on cell growth in the presence of aminoglycosides, we monitored the aerobic growth of WT, ΔravA, ΔviaA and ΔravAviaA (KOs with marker removed) in LB at 37 °C. The levels of RavA or ViaA is unchanged if viaA or ravA is deleted, respectively (Fig. 2.2). The strains exhibit similar growth behavior in the absence of antibiotics (Fig. 2.6A). In the presence of the aminoglycosides kanamycin or streptomycin (Fig. 2.6B and C), the log-phase growth rate is the same for all the strains, but WT cells reach a lower density of cells in stationary phase compared to the KO cells. This coincides with the fact that RavA-ViaA levels are maximal in early stationary phase. The phenotype is unique to the use of aminoglycosides since WT and the KO strains show similar growth curves when other translation-inhibiting antibiotics, such as tetracycline (Fig. 2.6D) and chloramphenicol (Fig. 2.6E), are used.

To further confirm the role of RavA-ViaA in this phenotype, ΔravA and ΔravAviaA were complemented with plasmids carrying the respective genes under the control of the native promoter for the ravAviaA operon. For ΔravA, complementation with the pR plasmid restores the cell’s sensitivity to kanamycin (Fig. 2.6F). Importantly, no effect is observed when the plasmid carrying the ATPase inactive RavA mutant having the K52Q mutation in the Walker A motif,

pRK52Q, is used instead of pR (Fig. 2.6F). The K52Q mutation replaces the highly conserved lysine residue in the Walker A motif that is crucial for the binding and subsequent hydrolysis of ATP (Walker et al., 1982; Saraste et al., 1990). The same type of mutation was used to abolish the ATPase activity in ClpX, the AAA+ component of the ClpXP protease complex (Wojtyra et al., 2003). This highlights the importance of RavA’s ATPase activity in sensitizing the cells to aminoglycosides.

Figure 2.6. Growth profiles of cells in the presence of sublethal concentrations of aminoglycosides

Growth profiles for MG1655 WT and the KO mutants ΔravA, ΔviaA and ΔravAviaA grown aerobically in LB at 37 °C over 24 hours. Growth of cells was monitored using OD600 readings at the designated time points. The cultures were supplemented as follows: (A) no antibiotics; (B) 4 μg/mL kanamycin; (C) 6 μg/mL streptomycin; (D) 0.5 μg/mL tetracycline; and (E) 1.2 μg/mL chloramphenicol. To confirm the phenotypes observed, ΔravA (F), ΔravAviaA (G) and WT cells (H) were complemented with the plasmids p11 (empty vector control), pR, pRV, pRK52Q, or pRK52QV. All cultures in the complementation experiments were supplemented with 4 μg/mL kanamycin for stress induction, and 100 μg/mL ampicillin for plasmid

maintenance. Error bars were derived from three independent cultures for each strain and for each condition. Details on the E. coli strains and plasmids used are given in Table 2.1.

______

For ΔravAviaA, complementation with pR does not re-sensitize the cells to kanamycin (Fig. 2.6G). Complementation with pRV re-sensitizes the cells (Fig. 2.6G), while

complementation with the pRK52QV plasmid does not (Fig. 2.6G). Thus, both RavA and ViaA are needed for the phenotype, and RavA’s ATPase activity is also required. Interestingly, the pRV plasmid produces a much stronger sensitization effect on ΔravAviaA than the pR plasmid on ΔravA (Fig. 2.6F and G). Given that ViaA expression is unchanged between ΔravA and WT (Fig. 2.2) and that complementation of ΔravAviaA with pRV results in a higher ViaA level than its endogenous expression in WT (Fig. 2.2), we conclude that the manifestation of this phenotype requires RavA’s ATPase activity, with ViaA as a potential regulator of RavA’s function.

To investigate this issue further, WT cells were transformed with plasmids used in the complementation experiments. WT + pR was found to have the same sensitivity towards kanamycin as WT + p11 (empty vector control) (Fig. 2.6H), unlike what is observed for ΔravA + pR versus ΔravA + p11 (Fig. 2.6F); in contrast, WT + pRV is more sensitive to kanamycin (Fig. 2.6H). Importantly, the endogenous expression of ViaA is the same in both WT and ΔravA, and is unaffected by the presence of p11 or pR (Fig. 2.2). Hence, the overexpression of RavA alone is indeed insufficient to increase the cell’s sensitivity towards aminoglycosides, if ViaA levels remain unchanged. Finally, we found that WT + pRK52Q has the same growth profile as the KO

mutants of ravA and/or viaA (Fig. 2.6F, G and H). However, WT + pRK52QV is sensitive to kanamycin (Fig. 2.6H). Evidently, the desensitization effect by RavAK52Q expression is probably caused by the Walker A mutant out-competing its WT counterpart for interaction with ViaA, which manifests into a dominant negative phenotype. This again highlights the critical role of ViaA and of RavA’s ATPase activity in this phenotype.

2.4.4 The RavA-ViaA phenotype is abolished by reduced glutathione and 2,2'-dipyridyl

The exact mechanism behind the bactericidal effects of aminoglycosides remains in dispute. Nevertheless, published works by several different groups on this subject all share the following observations in common: (I) the presence of thiourea (a reducing agent) and/or iron chelators in

the growth media increases the cell’s tolerance to aminoglycosides; (II) the presence of aminoglycosides induces the in vivo oxidation of a fluorescent dye such as hydroxyphenyl fluorescein (HPF) or dihydrorhodamine 123 (DHR) (Goswami et al., 2007; Kohanski et al., 2007; Kohanski et al., 2008; Baharoglu et al., 2013; Keren et al., 2013; Liu & Imlay, 2013).

To determine if the RavA-ViaA phenotype (Fig. 2.6) relies on the same or a similar mechanism, WT and KO mutants of ravA and/or viaA were grown in the presence of kanamycin or streptomycin supplemented with reduced glutathione (GSH) or 2,2'-dipyridyl (DP). GSH is a natural antioxidant utilized by E. coli (Carmel-Harel & Storz, 2000), while DP is a membrane- permeable chelator that sequesters free intracellular Fe2+ ions (Imlay et al., 1988). The presence of GSH (Fig. 2.7A-C) or DP (Fig. 2.7D-F) in the media can effectively rescue the growth reduction of WT cells when exposed to kanamycin (Fig. 2.7B and E) or streptomycin (Fig. 2.7C and F), although their effects on the KO mutants of ravA and/or viaA are minimal by comparison.

In a second experiment, WT + p11, ΔravAviaA + p11, ΔravAviaA + pRV and ΔravAviaA

+ pRK52QV grown in the presence of kanamycin were all treated with DHR (Fig. 2.7G). DHR is a membrane-permeable compound that becomes fluorescent and loses membrane permeability when oxidized. It is commonly used as a probe for intracellular ROS (Gomes et al., 2005), although its specificity for ROS detection has recently been questioned in some studies (Henderson & Chappell, 1993; Crow, 1997; Keren et al., 2013). Nevertheless, as shown in Fig. 2.7G, without kanamycin, only background levels of DHR fluorescence are detectable among the four strains of cells, showing that the activity of RavA and ViaA do not contribute to DHR oxidation. However, with kanamycin, WT + p11, ΔravAviaA + p11, and ΔravAviaA + pRV show 2.9-, 2.2- and 5.6-fold increase in DHR fluorescence, respectively. Furthermore, to highlight the

importance of RavA’s ATPase activity, ΔravAviaA + pRK52QV results in only a 2.7-fold increase in DHR fluorescence, resembling WT + p11. As before, the inclusion of either GSH or DP in the media reduces DHR fluorescence in all four strains to background levels. It should also be noted that, in the presence of kanamycin, ΔravAviaA + p11 exhibits lower DHR fluorescence than WT + p11 (p < 0.05).

Taken together, our results recapture the repeatedly observed effects of antioxidants and iron chelators in conferring greater resistance to the cell against aminoglycosides. This supports

74 the proposition that RavA-ViaA are involved in sensitizing E. coli cells to the presence of sublethal concentrations of aminoglycosides.

Figure 2.7. Effects of glutathione and 2,2'-dipyridyl on the growth profiles of cells in the presence of sublethal concentrations of kanamycin

Growth profiles of MG1655 WT and the KO mutants ΔravA, ΔviaA and ΔravAviaA grown aerobically in LB at 37 °C without (A, and D) or with kanamycin (B, and E) or streptomycin (C and F). Selected cultures were further supplemented with 750 μM of GSH (A-C) or 250 μM DP (D-F). Kanamycin was added at 4 μg/mL final concentration. Error bars were derived from three independent cultures for each strain and for each condition. In some instances, the error bars are smaller than the symbols used and cannot be seen.

(G) DHR fluorescence measurements normalized by OD600 for MG1655 WT + p11, ΔravAviaA + p11, ΔravAviaA + pRV and ΔravAviaA + pRK52QV grown aerobically to late log phase in LB at 37 °C supplemented with 4 μg/mL kanamycin in the presence or absence of 8 mM GSH or 250 μM DP. Error bars were derived from three independent cultures for each strain and for each condition. To highlight the statistical significance, the p-values for ΔravAviaA + p11 vs. WT + p11 (indicated with *) and ΔravAviaA + pRV vs. WT + p11 (indicated with **), in the presence of kanamycin, are given in the upper-right corner of the panel.

______

2.4.5 RavA-ViaA targets specific Nuo subunits and other respiratory proteins in sensitizing E. coli to kanamycin

In order to reveal the direct functional targets of RavA and ViaA, we performed suppression mutation analysis to identify genes that are necessary for the RavA-ViaA phenotype. More specifically, we identified genes whose KO completely abolished the growth suppression induced by RavA-ViaA overexpression in the presence of sublethal concentrations of kanamycin. Candidate genes were chosen based on the results of our high-throughput studies (Fig. 2.4 and 2.5) and the work of Girgis et al. (Girgis et al., 2009). A complete list of genes that were tested is given in Table 2.2. Six genes were identified: nuoB, nuoCD, nuoF, and nuoM that encode 4 of the 13 subunits of NADH:ubiquinone oxidoreductase I (Nuo complex); sdhB that encodes the Fe-S cluster subunit of succinate dehydrogenase (Sdh complex); and cyoB that

encodes subunit I of the cytochrome b0 terminal oxidase (Cyo complex). Examples of the growth profiles of these KO strains are shown for ΔnuoCD, ΔnuoF and ΔcyoB (Fig. 2.8A, B and E, respectively). In all cases, the overexpression of RavA-ViaA fails to sensitize the cell to kanamycin, unlike what was previously observed in ΔravAviaA or WT (Fig. 2.6G and H).

Interestingly, among the 24 genes that did not suppress the RavA-ViaA overexpression phenotype, many encode the other subunits of the Nuo, Sdh and Cyo complexes: ΔnuoI, ΔsdhA, ΔsdhC, ΔsdhD, ΔcyoA, ΔcyoC or ΔcyoD (Table 2.2 and Fig. 2.8). For example, the growth profiles of ΔnuoI (Fig. 2.8C) clearly show that nuoI is not needed in facilitating the RavA-ViaA phenotype, yet all Nuo subunits have been shown to be equally important for maintaining full functionality of the Nuo complex (Erhardt et al., 2012). Thus, the functional role of RavA and

ViaA appears to extend only to specific subunits or subcomplexes, but not the Nuo complex as a whole. This is further supported by the observation that Δndh also fails to suppress the RavA- ViaA phenotype (Fig. 2.8D), despite the fact that ndh encodes an enzyme functionally equivalent to the Nuo complex.

Figure 2.8. Growth profiles of selected single-gene knockouts

Growth profiles of ΔnuoCD (A), ΔnuoF (B), ΔnuoI (C), Δndh (D), ΔcyoB (E) and ΔcyoD (F) transformed with p11, pRV or pRK52QV plasmids as indicated. To account for the inherent differences in growth rates among the knockouts that are independent of the effects of RavA and ViaA, all growth data collected in the presence of kanamycin were normalized by the corresponding data collected in the absence of the antibiotic.

______

TABLE 2.2. Suppression mutation analysis for the RavA-ViaA overexpression-induced sensitization to kanamycin in E. coli MG1655

KO mutations that suppress the RavA-ViaA-overexpression phenotype

Strain Gene Product Description / Function* Associated Cofactors*

ΔnuoB NADH:ubiquinone oxidoreductase I; cytoplasmic subunit B 4Fe-4S ΔnuoCD NADH:ubiquinone oxidoreductase I; fused cytoplasmic subunit CD ΔnuoF NADH:ubiquinone oxidoreductase I; cytoplasmic subunit F FMN, 4Fe-4S ΔnuoM NADH:ubiquinone oxidoreductase I; membrane subunit M ΔsdhB Succinate dehydrogenase; Fe-S cluster subunit 2Fe-2S; 4Fe-4S; 3Fe-4S 2+ ΔcyoB Cytochrome b0 terminal oxidase; subunit I Cytochromes b562, b555; Cu

KO mutations with no effect on the RavA-ViaA-overexpression phenotype

Strain Gene Product Description / Function Associated Cofactors

ΔnuoA NADH:ubiquinone oxidoreductase I; membrane subunit A ΔnuoE NADH:ubiquinone oxidoreductase I; cytoplasmic subunit E 2Fe-2S ΔnuoG NADH:ubiquinone oxidoreductase I; cytoplasmic subunit G 2Fe-2S; 3× 4Fe-4S ΔnuoH NADH:ubiquinone oxidoreductase I; membrane subunit H ΔnuoI NADH:ubiquinone oxidoreductase I; cytoplasmic subunit I 2× 4Fe-4S ΔnuoJ NADH:ubiquinone oxidoreductase I; membrane subunit J ΔnuoK NADH:ubiquinone oxidoreductase I; membrane subunit K ΔnuoL NADH:ubiquinone oxidoreductase I; membrane subunit L ΔnuoN NADH:ubiquinone oxidoreductase I; membrane subunit N Δndh Alternative NADH:ubiquinone oxidoreductase II FAD; Cu+; Mg2+ ΔsdhA Succinate dehydrogenase; flavoprotein FAD

ΔsdhC Succinate dehydrogenase; membrane subunit C Cytochrome b556

ΔsdhD Succinate dehydrogenase; membrane subunit D Cytochrome b556

ΔcyoA Cytochrome b0 terminal oxidase; subunit II ΔcyoC Cytochrome b0 terminal oxidase; subunit III ΔcyoD Cytochrome b0 terminal oxidase; subunit IV ΔiscR DNA-binding transcription regulator for Fe-S cluster assembly, biofilm 2Fe-2S formation & anaerobic respiration ΔiscS Cysteine desulfurase; Isc Fe-S assembly pathway PLP ΔcadA Inducible lysine decarboxylase LdcI PLP ΔfeoB Fe2+ ion uptake transporter ΔcysB DNA-binding transcription dual regulator for cysteine biogenesis & novobicin resistance ΔcysI Sulfite reductase; hemoprotein subunit Siroheme; 4Fe-4S ΔnadA Quinolinate synthase; NAD de novo biogenesis 4Fe-4S ΔnadB L-Aspartate oxidase; NAD de novo biogenesis FAD *Gene annotations were obtained from the online databases EcoCyc (Keseler et al., 2005) and UniProt (Magrane & Consortium, 2011).

Taken together, these results indicate that RavA and ViaA target only specific subunits of the Nuo, Sdh and Cyo complexes when cells are exposed to sublethal concentrations of aminoglycosides during aerobic growth.

2.4.6 RavA and ViaA interact with specific Nuo subunits

To obtain conclusive evidence that RavA and ViaA are physically interacting with specific subunits of the Nuo, Sdh and Cyo respiratory proteins in E. coli, the genes corresponding to these subunits were endogenously tagged at the 3' end with a SPA-tag (Babu et al., 2009), and the tag was used for pull down assays. The SPA-tag consists of three modified FLAG sequences and a calmodulin binding peptide, spaced by a cleavage site for tobacco etch virus protease. The subunits that were successfully tagged are: NuoA, NuoB, NuoCD, NuoE, NuoF, NuoG, SdhA, SdhB, CyoB and CyoC. However, SPA-tagged SdhB and CyoC were not stably expressed and could not be used.

Neither SdhA nor CyoB showed any evidence of physical interaction with RavA and ViaA (data not shown). All of the Nuo subunits tested interacted with RavA and/or ViaA to various degrees (Fig. 2.9A and B). Since the Nuo complex functions aerobically and anaerobically (Tran et al., 1996; Schneider et al., 2008), the pulldowns were carried out under both conditions. In aerobically grown cells, NuoA and NuoF interacted with both RavA and ViaA. NuoE showed weak interaction with only RavA, while NuoB, NuoCD and NuoG all interacted with only ViaA, with NuoB showing weak interaction and NuoG showing moderate interaction (Fig. 2.9A). NuoCD was not pulled down as efficiently as the other subunits (Fig. 2.9A). However, in anaerobically grown cells, NuoCD interacted strongly with both RavA and ViaA (Fig. 2.9B), while the other Nuo subunits exhibited no or weak interactions (Fig. 2.9B). For all the pulldown assays, control experiments are shown for Nuo tagged strains carrying the ΔravAviaA null mutation (Fig. 2.9A and B). In addition, untagged WT DY330 was also used to rule out unspecific binding (Fig. 2.10).

Figure 2.9. Physical interactions between RavA and ViaA with specific subunits of the Nuo complex under different growth conditions

(A-C) WT, ΔravAviaA::cat, ΔviaA::cat DY330 strains having endogenously C-terminally SPA-tagged Nuo subunits were grown to stationary phase in LB under aerobic (A, C) or anaerobic (B, C) conditions. RavA and ViaA that co-purify with the SPA-tagged Nuo subunits were detected by Western blotting using α- RavA and α-ViaA polyclonal antibodies, respectively, whereas the SPA-tagged Nuo subunits were detected using α-FLAG monoclonal antibodies. The ΔravAviaA::cat strain having the SPA-tagged Nuo subunits is shown (A, B) as a control to confirm the identity of the RavA and ViaA bands detected. While the ΔviaA::cat strain having the SPA-tagged Nuo subunits is shown (C) to assess the role of ViaA in facilitating the binding of RavA to SPA-tagged NuoA and NuoF under aerobic conditions, and to SPA- tagged NuoC under anaerobic conditions.

(D) X-ray structure of the NADH:ubiquinone oxidoreductase I from Thermus thermophilus (PDB ID: 3M9S), solved at 3.3 Å (Baradaran et al., 2013). The subunits are identified here using the nomenclature for the E. coli NADH:ubiquinone oxidoreductase I (i.e. the Nuo complex) (Efremov & Sazanov, 2011a; Baradaran et al., 2013). The subunits Nqo15 and Nqo16 unique to T. thermophilus are omitted from the structure for clarity. Physical interactions of specific subunits with RavA and ViaA are indicated by the capital letters R and V, respectively. Red letters denote interactions that were identified in aerobically grown E. coli, and blue letters for the interactions in anaerobically grown cells. Asterisks (*) denote the subunits that are necessary for sensitizing the cell to sub-lethal concentrations of kanamycin upon overexpression of RavA and ViaA.

______

Figure 2.10. Immunoprecipitation experiments on WT DY330 and strains expressing SPA-tagged NuoF or NuoCD

Shown are Western blots for endogenous ViaA and the SPA-tagged NuoF and NuoCD in total soluble proteins (Input) and after immunoprecipitation of the SPA-tagged proteins. DY330 expressing SPA- tagged NuoF under aerobic condition and SPA-tagged NuoCD under anaerobic condition were used. Untagged WT DY330 strain is shown as control.

______

Given the primary function of VWA proteins in mediating protein-protein interactions (Whittaker & Hynes, 2002), it is possible that the interactions of RavA with the Nuo subunits are mediated by ViaA, which is also found to partly reside on the inner membrane (Fig. 2.1B). Hence, experiments were repeated using DY330 ΔviaA::cat strains that express endogenous, C- terminally SPA-tagged NuoA, NuoCD, or NuoF. As shown in Fig. 2.9C, the absence of ViaA results in significantly decreased binding of all three Nuo subunits to RavA. Neither RavA nor the three SPA-tagged Nuo subunits show any noticeable difference in expression between WT and ΔviaA::cat strains.

Taken together, these results strongly indicate that the Nuo complex is a functional target of RavA and ViaA, with NuoA and NuoF being the main subunits targeted under aerobic conditions, and NuoCD under anaerobic conditions. Importantly, ViaA is required for mediating the interaction between RavA and the Nuo subunits, which is reflected in both RavA and ViaA being equally important in their sensitization of the cell to aminoglycosides (Fig. 2.6).

2.5 Discussion

Using a multi-disciplinary approach, we were able to identify novel interactions between RavA- ViaA and specific subunits of the Nuo respiratory complex. A summary of these interactions is illustrated in Fig. 2.9D. Out of the six Nuo subunits tested for physical interactions, NuoF (aerobically) and the fused NuoCD (anaerobically) showed strong interactions with both RavA and ViaA (Fig. 2.9A and B) with both Nuo subunits being necessary for RavA and ViaA to sensitize the cells towards kanamycin (Fig. 2.8A and B). In this regard, it is interesting to note that a recent study on the Nuo proteins in E. coli revealed that the inducible lysine decarboxylase LdcI, which we showed to form a large cage-like structure with RavA (Snider et al., 2006; El Bakkouri et al., 2010), binds specifically to a variant form of the Nuo complex that lacks NuoL (Erhardt et al., 2012). However, in our pulldown assays (Fig. 2.9), we did not observe any significant interaction of LdcI with the SPA-tagged Nuo subunits (data not shown). This seems to suggest that an LdcI-RavA-ViaA complex might interact with a specific Nuo subcomplex, when NuoL is deleted. This subcomplex might contain NuoF and NuoCD.

Our phenotypic data suggest that the interaction of RavA-ViaA with NuoF and NuoCD is likely an important part underlying the sensitization of E. coli towards aminoglycosides by RavA

and ViaA. The exact mechanism behind the bactericidal effects of aminoglycosides is still under debate. Aside from their traditional role in binding ribosomes that causes protein mistranslation (Davis, 1987), one recent model proposes that the bactericidal effects of aminoglycosides may arise from the generation of intracellular reactive oxygen species (ROS) via the Fe2+-mediated Fenton reaction (Kohanski et al., 2007; Kohanski et al., 2008). The source of the free Fe2+ has

been attributed to damaged Fe-S clusters, resulting from increased H2O2 production caused by the upregulated respiratory activities (Kohanski et al., 2007; Kohanski et al., 2008). However,

several groups have recently shown that aminoglycosides can neither increase the level of H2O2 in the cell nor upregulate bacterial respiration (Keren et al., 2013; Liu & Imlay, 2013), nor is ROS necessary for the bactericidal actions of the antibiotics (Goltermann et al., 2012; Keren et al., 2013). Nevertheless, a recent study on the toxicity of protein aggregates generated via aminoglycoside-induced mistranslation has shown that overexpressing AhpF, one of two subunits of the H2O2 scavenger alkyl hydroperoxide reductase, can effectively increase the cell’s tolerance to aminoglycosides by reducing the oxidation and aggregation of mistranslated proteins (Ling et al., 2012). This supports the notion that oxidative damage may still play an important role in the cellular toxicity of protein mistranslation. In contrast, the work by Ezraty et al. (Ezraty et al., 2013) suggests that the bactericidal effect of aminoglycosides is dependent on Fe- S clusters biosynthesis that is independent of ROS. Specifically, the major Isc Fe-S clusters assembly pathway is required for the full maturation and function of the Nuo and Sdh respiratory complexes, which in turn generate proton motive force (PMF) that promotes the uptake of aminoglycosides leading to cell death (Ezraty et al., 2013). The effect of RavA-ViaA might be manifested through such a latter model. Furthermore, the genetic linkage of RavA-ViaA with Fe- S cluster biogenesis genes (Fig. 2.4) may reflect a chaperone-like role of RavA and ViaA for NuoF, for example, and possibly other Fe-S-carrying targets. The physiological implication of the interaction of RavA-ViaA and possibly LdcI with the Nuo complex is the subject of ongoing studies.

Chapter 3 The MoxR AAA+ ATPase RavA and its VWA Cofactor ViaA Interacts and Modulates the Activity of the Fumarate Reductase Complex during Anaerobiosis in E. coli

Data attribution: I performed the majority of the experiments presented in this chapter, with the exception of the co-expression profiling for ravA and viaA, which was contributed by Prof. Sarath C. Janga (Laboratory of Genomics and Systems Biology, Indianapolis, Indiana, US).

3 The MoxR AAA+ ATPase RavA and its VWA Cofactor ViaA Interacts and Modulates the Activity of the Fumarate Reductase Complex during Anaerobiosis in Escherichia coli 3.1 Chapter Summary

In Escherichia coli, the MoxR AAA+ ATPase, RavA, and its VWA cofactor, ViaA, were found to be associated with anaerobic respiration. First, bioinformatic analysis of ravA and viaA showed that they are coexpressed with genes that are regulated by the anaerobic transcriptional regulator Fnr. The expression of ravA and viaA was also revealed to be dependent on Fnr in anaerobically grown cells, and at least two sites for Fnr binding were identified in the ravAviaA promoter region by electromobility shift assay. Next, ViaA was found to physically interact with FrdA, the flavin-containing subunit of the anaerobic fumarate reductase (Frd) complex. Both RavA and the Fe-S-containing subunit of the Frd complex, FrdB, appeared to modulate this interaction. Importantly, both RavA and ViaA were found to be necessary for optimal activity of the Frd complex. This mirrors similar findings for other MoxR proteins and their respective targets. These results strongly support a potential regulatory function of RavA-ViaA on the activity of the Frd complex and the anaerobic respiratory process with fumarate as the terminal electron acceptor.

3.2 Introduction

Among the AAA+ ATPases, the MoxR family is diverse and widespread among bacteria and archaea (Snider & Houry, 2006). The experimental evidence gathered on various MoxR proteins suggests that they have chaperone-like roles in the maturation of specific protein complexes that participate in a variety of biological processes, from metabolism, cell morphology and development, tolerance against various types of stress, to pathogenesis (Snider & Houry, 2006; Wong & Houry, 2012). A characteristic of the MoxR AAA+ ATPases is the co-occurrence of the AAA+ ATPase with one or more cofactors that carry the von Willebrand factor A (VWA) domain (Snider & Houry, 2006). An important feature of the VWA domain is the highly conserved MIDAS (Metal Ion-dependent Adhesion Site) motif, which binds a single divalent cation, usually Mg2+, and is important for mediating protein-protein interactions (Whittaker &

Hynes, 2002). Genes encoding the AAA+ ATPase and the VWA protein are usually positioned in close proximity with each other (Snider & Houry, 2006).

The primary focus of our research is the functional characterization of the MoxR AAA+ ATPase, RavA (Regulatory ATPase variant A), and its corresponding VWA protein, ViaA (VWA interacting with AAA+ ATPase), in Escherichia coli. RavA belongs to the MoxR subfamily RavA (Snider & Houry, 2006). As such, the ravA and viaA genes are organized in a pattern that is typical of the RavA subfamily, with ravA positioned immediately upstream of viaA, forming a single operon (Snider & Houry, 2006). Under aerobic condition, the ravAviaA operon is induced during stationary phase, with the alternative sigma factor, σS (RpoS) serving as the primary transcriptional regulator (Snider et al., 2006).

RavA has been characterized extensively from both a biochemical and biophysical perspective. In vitro, the ATPase activity of RavA is optimal at neutral pH at 37 °C, which is enhanced further by the presence of the VWA protein ViaA (Snider et al., 2006). RavA can also hydrolyze GTP, albeit at a slower rate (Snider et al., 2006). Structurally, RavA has three well- defined domains – the N-terminal AAA+ domain, the LARA (LdcI associating domain of RavA) domain, and a triple-helical bundle that connects the AAA+ and LARA domains (El Bakkouri et al., 2010). In the presence of nucleotides, RavA forms a hexamer (Snider et al., 2006; El Bakkouri et al., 2010), which is the typical oligomeric state for AAA+ ATPases (Hanson & Whiteheart, 2005). Aside from ViaA, RavA also interacts and modulates the activity of the inducible lysine decarboxylase LdcI (Snider et al., 2006; El Bakkouri et al., 2010; Kanjee et al., 2011), which is a major acid stress response protein in E. coli (Park et al., 1996; Kanjee et al., 2011). This interaction, which is mediated by the LARA domain, leads to the formation of a large cage-like complex consisting of five RavA hexamers and two LdcI decamers (Snider et al., 2006; El Bakkouri et al., 2010).

RavA and ViaA are also functionally associated with bacterial respiratory processes. A recent high throughput study showed that both ΔravA and ΔviaA null mutations can protect E. coli from the bactericidal effects of aminoglycosides (e.g. kanamycin, streptomycin, etc.) under aerobic conditions (Girgis et al., 2009). Follow-up experiments performed by our group confirmed that both RavA and ViaA are required to sensitize the cell to aminoglycosides. Furthermore, this phenotype is dependent on both the ATPase activity of RavA and the correct

stoichiometry of RavA and ViaA (see Section 2.4.3). Importantly, the identification of null mutations that suppressed the RavA-ViaA-induced phenotype and subsequent immunoprecipitation experiments revealed that RavA-ViaA interacts with specific subunits of the NADH:ubiquinone oxidoreductase I (Nuo), commonly known as Complex I. The Nuo complex is a major player in the aerobic respiration of E. coli (Unden & Bongaerts, 1997; Price & Driessen, 2010). It is also important in the anaerobic respiration of fumarate and dimethylsulfoxide, using a non-fermentable carbon source such as glycerol (Tran et al., 1996). Under aerobic condition, RavA-ViaA was found to interact primarily with the FMN (flavin mononucleotide)-containing subunit NuoF. Under anaerobic condition, the fused NuoCD subunit became the primary target for interaction instead (see Section 2.4.6). In addition, our high- throughput studies also revealed functional links between RavA-ViaA and a number of pathways that are directly or indirectly related to bacterial respiration. These include iron-sulfur (Fe-S) clusters biosynthesis, iron transport, and anaerobic respiration (see Section 2.4.2).

In this chapter, additional evidence will be presented that supports a regulatory role of RavA-ViaA over the activity of the fumarate reductase (Frd) complex, which works in conjunction with the Nuo complex in the electron transfer from NADH to fumarate during the anaerobic respiration of fumarate, with glycerol as a non-fermentable carbon source (Tran et al., 1996; Cecchini et al., 2002). Together with our previous results, we propose that RavA-ViaA may regulate or facilitate the electron transfer from NADH to fumarate through its interaction with both the Nuo and Frd complexes.

3.3 Materials and Methods

3.3.1 Bacterial strains and plasmids used

All bacterial strains and plasmids used are listed in Table 3.1. All knockout (KO) mutants of the ravA/viaA open reading frames were constructed as previously described (Snider et al., 2006) by employing lambda red recombination (Datsenko & Wanner, 2000) and P1 phage transduction (Miller, 1992). MG1655 ΔfrdA::kanR and ΔrpoS::kanR were also constructed via P1 phage transduction; the required frdA and rpoS KO cassettes that carry the kanR gene were obtained from BW25113 ΔfrdA::kanR and ΔrpoS::kanR, respectively, both of which came from the KEIO collection (Baba et al., 2006; Yamamoto et al., 2009). All DY330 strains expressing C-terminal SPA-tagged proteins were a generous gift from Dr. Andrew Emili (University of Toronto) and

87 were constructed using the protocols described in Zeghouf et al. (Zeghouf et al., 2004). DY330 FrdA-SPA ΔviaA::cat was constructed by P1 phage transduction; the required viaA KO cassette was obtained from MG1655 ΔviaA::cat that was used in a previous study (Snider et al., 2006).

TABLE 3.1. List of bacterial strains and plasmids used

Bacterial Strains Genotype Reference

MG1655 (WT) F-, rph-1, λ- (Guyer et al., 1981) MG1655 ΔravAviaA MG1655, ΔravAviaA This study MG1655 Δfnr::kanR MG1655, Δfnr::kanR This study MG1655 ΔrpoS::kanR MG1655, ΔrpoS::kanR This study PK22 BL21(DE3), Δcrp-bs990, rpsL, Δfnr, zcj-3061::Tn10 (Lazazzera et al., 1993) DY330 (WT) W3110, ΔlacU169, gal490, pglΔ8λ, [ρ]cI857 (cro-bioA) (Yu et al., 2000) DY330 RavA-SPA DY330, ravA-SPA::kanR This study DY330 ViaA-SPA DY330, viaA-SPA::kanR This study DY330 LdcI-SPA DY330, ldcI-SPA::kanR This study DY330 FrdA-SPA DY330, frdA-SPA::kanR This study DY330 FrdA-SPA DY330, frdA-SPA::kanR, ΔviaA::cat This study ΔviaA::cat DY330 HemC-SPA DY330, hemC-SPA::kanR This study DY330 HemX-SPA DY330, hemX-SPA::kanR This study DY330 CysA-SPA DY330, cysA-SPA::kanR This study DY330 CysB-SPA DY330, cysB-SPA::kanR This study DY330 CysI-SPA DY330, cysI-SPA::kanR This study DY330 CysJ-SPA DY330, cysJ-SPA::kanR This study DY330 CysM-SPA DY330, cysM-SPA::kanR This study DY330 CysN-SPA DY330, cysN-SPA::kanR This study DY330 CysP-SPA DY330, cysP-SPA::kanR This study DY330 NapA-SPA DY330, napA-SPA::kanR This study DY330 NapD-SPA DY330, napD-SPA::kanR This study DY330 NapH-SPA DY330, napH-SPA::kanR This study DY330 HypA-SPA DY330, hypA-SPA::kanR This study DY330 HypB-SPA DY330, hypB-SPA::kanR This study DY330 HypC-SPA DY330, hypC-SPA::kanR This study DY330 HypD-SPA DY330, hypD-SPA::kanR This study DY330 HycE-SPA DY330, hycE-SPA::kanR This study DY330 HycG-SPA DY330, hycG-SPA::kanR This study

Plasmids Description Reference p11 Cloning vector derived from pET15b(+) (Zhang et al., 2001) pR p11-ravAp-ravA, for overexpression of RavA regulated by the (Snider et al., 2006) native ravA promoter pRK52Q p11-ravAp-ravA(K52Q), for overexpression of RavA Walker This study

A mutant regulated by the native ravA promoter pRV p11-ravAp-ravAviaA, for RavA and ViaA overexpression (Snider et al., 2006) regulated by the native ravA promoter pRK52QV p11-ravAp-ravA(K52Q)viaA, for the overexpression of RavA This study Walker A mutant and wild-type ViaA regulated by the native ravA promoter pPK824 pET11a-fnrD154A, for IPTG-induced expression of the (Lazazzera et al., 1993) mutant FnrD154A p11-frdp p11-frdp; frd promoter control for similar plasmids carrying This study genes encoding the subunits of fumarate reductase pfrdA p11-frdp-frdA, for overexpressing FrdA regulated by the This study native frd promoter pfrdAB p11-frdp-frdAB, for overexpressing FrdA and FrdB regulated This study by the native frd promoter pfrdABCD p11-frdp-frdABCD, for overexpressing all fumarate reductase This study subunits regulated by the native frd promoter pfrdB p11-frdp-frdB, for overexpressing FrdB regulated by the This study native frd promoter pfrdBCD p11-frdp-frdBCD, for overexpressing FrdB, FrdC and FrdD This study regulated by the native frd promoter cat = chloramphenicol acetyltransferase gene; confers resistance to chloramphenicol. kanR = kanamycin resistance gene

The plasmids pR (p11-ravAp-ravA) and pRV (p11-ravAp-ravAviaA) were constructed as

described in our previous work (Snider et al., 2006). The plasmids pRK52Q and pRK52QV were generated by QuikChange site-directed mutagenesis (Stratagene), using the primers RavA K52Q F (5'-CGCCAGGTATTGCCCAAAGTTTGATCGCC-3') and RavA K52Q R (5'- GGCGATCAAACTTTGGGCAATACCTGGCG-3'). For the plasmids p11-frdp (control vector), pfrdA (p11-frdp-frdA), pfrdAB (p11-frdp-frdAB), and pfrdABCD (p11-frdp-frdABCD), all inserts were PCR-amplified using the common forward primer FrdABCD BamHI F (5'- GATTATTATTGGATCCGGCTGCCAGGATGC-3'). The reverse primer frdp NheI R (5'- CATTATTATTGCTAGCCCTCCAGATTGTTTTTATCCCAC-3'), FrdA NheI R (5'- CATTATTATTGCTAGCTCAGCCATTCGCCTTCTCCTTC-3'), FrdB NheI R (5'- CATTATTATTGCTAGCTTAGCGTGGTTTCAGGGTCG-3') and FrdD NheI R (5'- CATTATTATTGCTAGCTTAGATTGTAACGACACCAATCAGCGTG-3') were used for p11- frdp, pfrdA, pfrdAB and pfrdABCD, respectively. These inserts were cloned into the p11 vector using the restriction enzymes BamHI and NheI (New England Biolabs). Similarly, for the plasmids pfrdB (p11-frdp-frdB) and pfrdBCD (p11-frdp-frdBCD), all inserts were PCR- amplified using the common forward primer FrdB NheI F (5'-

GATTATTATTGCTAGCATGGCTGAGATGAAAAACCTGAAAATTG-3'). The reverse primers FrdB XbaI R (5'-CATTATTATTTCTAGATTAGCGTGGTTTCAGGGTCG-3') and FrdD XbaI R (5'-CATTATTATTTCTAGATTAGATTGTAACGACACCAATCAGCGTG-3') were used for pfrdB and pfrdBCD, respectively. The inserts were then cloned into p11-frdp using the restriction enzymes NheI and XbaI (New England Biolabs). All constructs were verified by DNA sequencing.

3.3.2 Co-expression profiling of ravA and viaA in E. coli

The expression levels of ravA and viaA were compared across different experimental conditions to identify genes with similar expression profile. To this end, a large compendium composed of 445 microarray datasets was obtained from the M3D public database (Build 4 of E. coli expression data) (Faith et al., 2008). These data were available in the form of Robust Multi Array (RMA) normalized profiles (Irizarry et al., 2003), which enables the direct comparison of the expression profiles of different protein-encoding genes across multiple experimental conditions. The Pearson correlations, used for comparing the similarity of expression profiles, were computed for all 4,125 genes present on the Affymetrix chip against both ravA and viaA. This allowed the identification of genes that exhibit the most similar expression profiles to the seed set of genes. Due to the large number of conditions in the compendium, a conservative cut- off of 0.5 was adopted as the correlation threshold to identify the functional links to the seed genes. All functional annotations were obtained from publicly available online databases, such as EcoCyc (Keseler et al., 2005), UniProtKB {Magrane, 2011 #245}, and RegulonDB (Gama- Castro et al., 2011).

3.3.3 Expression analysis of RavA and ViaA in E. coli under aerobic and anaerobic conditions

E. coli MG1655 WT, Δfnr::kanR, ΔrpoS::kanR, and ΔravAviaA were grown in Luria-Bertani (LB) media (10 g/L bacto-tryptone, 5 g/L yeast extract, and 10 g/L sodium chloride) at 37 °C either aerobically in 200-mL culture flasks with vigorous shaking, or anaerobically in 60-mL disposable syringes (sealed with sterile end caps) with gentle agitation. All cultures were inoculated with single colonies grown overnight on LB-agar plates. Growth of cells was tracked

by monitoring the changes in OD600 at specific time points. For each time point, an aliquot of the cells was harvested by centrifugation and flash-frozen in liquid nitrogen until use. To determine

the levels of RavA and ViaA, the cell pellets collected were thawed on ice and then resuspended in 0.1 M potassium phosphate (pH 7.5) supplemented with 0.1 M sodium chloride. The volume of each sample was adjusted to give a final cell count of approximately 3.8 × 109 cells/mL as

determined by OD600. The cells were lysed by sonication, followed by treatment with 4 × SDS- PAGE sample buffer (200 mM TrisHCl, pH 6.8, 8% SDS, 0.4% bromophenol blue, 40% glycerol, and 400 mM β-mercaptoethanol) and separated on 10% or 12% polyacrylamide gels. The levels of RavA and ViaA were analyzed by Western blotting. A 70-kDa cross-reacting band in the α-ViaA blot was used as the loading control.

3.3.4 Electromobility shift assay

The E. coli Fnr mutant, FnrD154A, was expressed from the plasmid pPK824 (pET11a-fnrD154A) in strain PK22 lacking fnr (Table 3.1) and purified as described in (Lazazzera et al., 1993), except that SP sepharose (GE Health Sciences) was used in place of BioRex-70 during the first round of

purification. FnrD154A was used in this assay because it retains the same specificity and affinity as WT Fnr for binding to the Fnr consensus DNA sequence even under aerobic conditions (Lazazzera et al., 1993). All DNA substrates required were PCR-amplified using the appropriate primers listed in Table 3.2.

TABLE 3.2. List of primers used for generating the DNA substrates used in EMSA

Primer Name Sequence (5' → 3') Description

ravAp NcoI F ATTCCATGGCACGGCATCGCGTTCAAC Forward primer for R-1 ravAp BamHI R ATTGGATCCGTGGCGTCCTTTCGTCAAAAG Reverse primer for R-1 ravAp(fnr1-2+) NcoI F ATTCCATGGTGCTCATAGACTAGTCTTTCGTTGAA Forward primer for R-2 ATATGAAATG ravAp(fnr1-2+) BamHI R ATTGGATCCAGGAGGAACACACTTTCACCACTTA Reverse primer for R-2 ATG ravAp(fnr1-2-) NcoI F ATTCCATGGAGAAAAATACCCCCCCTTTGAGAC Forward primer for R-3 ravAp(fnr1-2-) BamHI R ATTGGATCCAATAGAAAGGGGACCAAAAACTTCT Reverse primer for R-3 TCCG fepDp NcoI F ATTTATTCCATGGCATCATCTGGATCTGCACCG Forward primer for F-1 fepDp BamHI R ATTTATTGGATCCGGCCTCCAGCACTACGGAAGC Reverse primer for F-1 GG hypBp NcoI F ATTCCATGGCGACGTGTCATTTCGACATCATCGAC Forward primer for H-1

91 hypBp BamHI R ATTGGATCCGACACTGTGGACAGCGGC Reverse primer for H-1 hypBp(fnr1-2+) NcoI F ATTCCATGGGGCCGCAAAACACGGCGCAAAAC Forward primer for H-2 hypBp(fnr1-2+) BamHI R ATTGGATCCAAACGCGGAATGAGGGTTATGTTCA Reverse primer for H-2 TCACC hypBp(fnr1-2-) NcoI F ATTCCATGGCGCGGCAGCGTGGCGGAAG Forward primer for H-3 hypBp(fnr1-2-) BamHI R ATTGGATCCAATGCAGGTCGCCTTCTTCAGTCTGG Reverse primer for H-3

To detect the binding of FnrD154A to the DNA substrates, 3 nM of DNA substrate was incubated with 60 nM FnrD154A in 20 mM Tris-acetate (pH 7.5) supplemented with 40 mM KCl,

1 mM MgCl2 and 5% (v:v) glycerol for 30 minutes at 37 °C. All samples were electrophoresed at 4 °C in a 4% polyacrylamide native gel supplemented with 10% polyethylene glycol, with 20 mM Tris-acetate (pH 8.0) as the running buffer. The gel was then incubated in 20 mM Tris- acetate (pH 8.0) supplemented with the RedSafe DNA stain (Chembio) with gentle agitation at room temperature to stain the DNA bands contained within. Visualization of the DNA bands was done using the GelDoc 2000 (BioRad).

3.3.5 Immunoprecipitation by SPA-tagged bait proteins

Endogenous Sequential Peptide Affinity (SPA)-tagging of proteins was carried out in E. coli DY330 using the protocols described in Zeghouf et al. (Zeghouf et al., 2004). Cells with confirmed incorporation of the SPA-tags at the C-terminus of targeted proteins were grown in LB media at 30°C in sealed sterile 50-mL centrifuge tubes for over 24 hours. Cells were then harvested by centrifugation at 4 °C and resuspended in lysis buffer (25 mM TrisHCl, pH 7.5, 100 mM KCl, 10 mM MgCl2 1 mM CaCl2, 0.2 mM EDTA, 1% Triton X-100, 10% glycerol, and 0.5 mM DTT) supplemented with 1 mg/mL lysozyme (BioShop) and 0.1 U/mL DNaseI (Fermentas). After incubation on ice for 15 minutes, the cells were lysed by sonication. Total soluble proteins were isolated from the crude cell lysate by centrifugation. The SPA-tagged proteins and the stably associated proteins were purified using ANTI-FLAG M2 Affinity Beads (Sigma-Aldrich) following the manufacturer’s protocols, which were then analyzed by SDS-PAGE and Western blotting.

3.3.6 Western Blotting

Samples to be analyzed were first separated using a 10% or 12% SDS-PAGE gels. The protein bands were then transferred onto an Amersham Hybond-P PVDF membrane (GE Healthcare)

using the TE77X Semi-dry Transfer Unit (Hoefer Inc.) following the manufacturer’s instructions. The membrane was then blocked, washed and incubated with the appropriate antibodies as required, using standard protocols. The polyclonal rabbit antibodies against RavA and ViaA were generated at the Division of Comparative Medicine, University of Toronto. The polyclonal rabbit antibodies against FrdA and FrdB in E. coli were generously provided by Professor Joel Weiner (University of Alberta, Edmonton). The polyclonal rabbit antibody against the inner membrane protein LepB in E. coli was a gift from Professor Jan Willem de Gier (Stockholm University, Sweden). The monoclonal mouse antibody against the FLAG tag was purchased from Sigma-Aldrich.

3.3.7 Fumarate reductase activity assay

E. coli MG1655 WT, ΔravAviaA, ΔravAviaA + p11, ΔravAviaA + pR, ΔravAviaA + pRV, R ΔravAviaA + pRK52QV and BW25113 ΔfrdA::kan were grown anaerobically inside sealed, sterilized containers in LB supplemented with 1% (v:v) glycerol and 50 mM sodium fumarate at 37 °C over 16 hours. Cells were then harvested by centrifugation, re-suspended in 0.1 M sodium phosphate buffer (pH 7), and lysed by two passages through a French Press (Thermo Spectronic) at 18000 lb/in2. Cell lysis by French Press generates inside-out membrane vesicles. Cell debris was removed by centrifugation. To isolate the membrane vesicles, the cleared cell lysate was subjected to ultracentrifugation at 150000 × g at 4 °C for 1.5 hours. The pelleted membrane vesicles were resuspended in 0.1 M sodium phosphate buffer (pH 7), flash-frozen with liquid nitrogen, and stored at -80 °C until use.

To measure the activity of endogenously expressed fumarate reductase in the isolated inside-out membrane vesicles, a modified version of the benzyl viologen (BV) colorimetric assay described in (Bilous & Weiner, 1985) was used. Briefly, 0.125 mM (final concentration) BV was

first reduced with 3 mM (final concentration) Na2S2O4 in 0.1 M sodium phosphate (pH 7), followed by the addition of 0.18 mg/mL of membrane vesicles. The mixture was incubated at room temperature for 1 minute. To initiate the reaction, 20 mM (final concentration) sodium fumarate was added and the mixture was homogenized by gentle pipetting to minimize oxidation of BV by air. Fumarate reductase activity was tracked by monitoring the loss of the purple colour as BV was oxidized in the presence of fumarate. This was done by measuring absorbance at 500 nm of the reaction mixture in a standard 1-cm cuvette using the CARY300 UV-Vis

Spectrophotometer (Agilent Technologies). Measurements were taken every second for 3 minutes at room temperature. Fumarate reductase activity was calculated using the equation 1 U = 1 μmol BV oxidized/min, with the extinction coefficient of BV = 7.8 × 103 M-1⋅cm-1 (Bilous & Weiner, 1985). The results were then normalized to the amount of membrane vesicles used to allow comparison between samples.

3.4 Results

3.4.1 ravA and viaA display similar co-expression profiles as those of the Fnr-inducible genes

Co-expression profiling was performed to identify genes that co-express with both ravA and viaA. This approach is based on the principle that genes are organized in a network of distinct, functional modules or hubs with highly coordinated expression patterns that correspond to specific biological processes (Carlson et al., 2006; Luo et al., 2007; Costanzo et al., 2010; Hume et al., 2010). Thus, genes that are functionally associated have a higher likelihood of sharing common transcriptional regulatory elements and of displaying similar expression profiles in response to the same physiological signals or external environmental stimuli.

The co-expression profiles for ravA and viaA genes were constructed by data-mining a public collection of 445 E. coli microarray datasets collected across multiple experimental conditions and by identifying genes that display highly similar co-expression patterns with ravA and viaA (Fig. 3.1). Our analysis yielded a total of 62 genes that co-express with ravA, and 56 that co-express with viaA. Of these, 32 genes co-express with both ravA and viaA. Given that ravA and viaA are in the same operon (Snider et al., 2006; Snider & Houry, 2006), genes that are co-expressed with both ravA and viaA are considered as the most reliable candidates for functional association and were studied further. A complete list of these genes and their correlation scores is given in Appendix B.

One important trend uncovered in our analysis is that many of the genes that co-express with both ravA and viaA are involved in anaerobic respiration (Fig. 3.1). These include frdA, frdB, and frdC, which encode three of the four subunits of the fumarate reductase complex (FrdABCD); nirB and nirD, which encode the large and small subunits, respectively, of the

nitrite reductase complex (NirDB2); hybO, which encodes the small subunit of hydrogenase 2

(HybABOC); and nrfA, which is the structural gene for cytochrome c552 and a component of the formate-dependent nitrite reductase complex (NrfDCBA).

Figure 3.1. Co-expression profiles of ravA and viaA

Shown are genes that have similar co-expression profiles as ravA and viaA. Single or multiple genes enclosed in rectangular boxes denote the constituents of monocistronic (only one gene included in its own box) or polycistronic operons (multiple genes in one box), respectively. Genes from the same polycistronic operon that are classified to a different co-expression category are linked with broken lines. All genes that are under the control of the transcriptional regulator Fnr are denoted with an asterisk (*).

______

A second group of genes falls under protein maturation and modification, all of which – hypA, hypB, hypC and hypD – are involved in the insertion of Ni2+ ion for the maturation of the membrane-bound hydrogenase 3 (HycDCFGBE), and, thus, are also associated with anaerobiosis. Hydrogenase 3 works in conjunction with formate dehydrogenase H (FdhF) in both mixed acid fermentation and anaerobic respiration. Other genes that co-express with both ravA and viaA participate in various metabolic pathways (gpmM, mtlD, pfkA, ansB, aspA, selA, pepE, pldB, and udp), biosynthesis of cofactors and prosthetic groups (hemC, hemX, and menD), and transport of metabolites across the cell membrane (dcuA, dcuB, and nikA).

Importantly, 14 out of the 32 genes that co-express with ravA and viaA are inducible by the transcriptional regulator Fnr (marked with * in Fig. 3.1). In E. coli, Fnr regulates the expression of a large number of genes during the transition from aerobiosis to anaerobiosis (Salmon et al., 2003; Kang et al., 2005). This suggests that Fnr is also likely to regulate the expression of both ravA and viaA.

3.4.2 Fnr enhances the expression of RavA and ViaA during anaerobiosis

To determine whether the expression of RavA and ViaA is indeed regulated by Fnr, WT and the null mutants Δfnr::kanR and ΔrpoS::kanR were grown in liquid media to various phases under

aerobic or anaerobic conditions. Cell growth was monitored via measuring OD600 at specific time points. When cells were grown aerobically, all three strains shared almost identical growth profiles (Fig. 3.2A). On the other hand, during anaerobic growth, Δfnr::kanR exhibited a growth lag from early log to late log phase, and both Δfnr::kanR and ΔrpoS::kanR had a lower cell count per unit volume compared to WT upon reaching stationary phase (Fig. 3.2A).

The expression levels of RavA and ViaA in each strain were then analyzed by Western blotting. Under aerobic condition, WT cells displayed the expected RavA expression profile as reported previously (Snider et al., 2006), with minimal expression during log phase that increases to optimum in stationary phase (Fig. 3.2B). Interestingly, under anaerobic condition, RavA expression was significantly enhanced in WT at all growth phases, which indicates that oxygen starvation is likely an important stimulus for protein expression. Importantly, unlike WT, the expression of RavA in Δfnr::kanR did not show any increase by growing the cells anaerobically (Fig. 3.2B). Thus, Fnr is necessary for the enhanced expression of RavA during anaerobic growth. Furthermore, in aerobically grown ΔrpoS::kanR cells, the expression of RavA was severely compromised due to the loss of σS (Fig. 3.2B), as previously reported (Snider et al., 2006). In contrast, under anaerobic condition, the expression of RavA in ΔrpoS::kanR closely resembles the WT, despite that σS was deleted (Fig. 3.2B). This strongly suggests that the expression of RavA during anaerobic growth is largely dependent on Fnr, not σS.

Figure 3.2. Expression analysis of RavA and ViaA in E. coli MG1655 WT, Δfnr::kanR and ΔrpoS::kanR by Western blotting

(A) Growth profiles for MG1655 wild-type (WT), Δfnr::kanR and ΔrpoS::kanR that were grown aerobically (black data points) or anaerobically (white data points) in LB over 24 hours. Growth curves for WT (diamonds) Δfnr::kanR (triangles) and ΔrpoS::kanR (squares) are as indicated in the figure legend.

(B) Western blots for RavA and ViaA for aerobically grown cells (shown on the left), and anaerobically grown cells (shown on the right). All strains were grown in LB media. The time points at which cells were harvested are as indicated at the top. ΔravAviaA cells harvested after 24 hours of growth were used to provide a reference for the RavA and ViaA expressed. A cross-reacting band (70-kDa) in the α-ViaA blots that remains consistent at all time points was used as the loading control.

______

Like RavA, the expression of ViaA shows a similar dependence on Fnr in anaerobically grown cells. In WT, ViaA expression was significantly enhanced during anaerobic growth and the deletion of σS did not affect this enhancement; however, in the absence of Fnr, no such enhancement was observed (Fig. 3.2B). It is interesting to note that ViaA expression under anaerobic conditions was higher in log phase compared to stationary phase – the reverse of RavA. This was most apparent in WT and Δfnr::kanR cells grown under anaerobic conditions. Since both ravA and viaA are on the same operon (Snider et al., 2006), the difference in the expression levels of RavA and ViaA likely reflects the presence of additional regulatory elements affecting mRNA and/or protein levels.

Overall, our results clearly illustrate that, under anaerobic condition, Fnr functionally replaces σS and becomes the primary regulator for the expression of both RavA and ViaA, which makes both the ravA and viaA genes novel constituents of the Fnr regulon.

3.4.3 Identification of potential Fnr-binding sites in the native promoter region of ravAviaA

The Fnr-induced expression of RavA and ViaA in anaerobically grown cells indicates the existence of regulatory elements in the ravAviaA promoter region for Fnr binding. Using knowledge-based sequence motifs recognition software such as Virtual Footprint (Munch et al., 2005), SCOPE (Chakravarty et al., 2007), and PromoScan (Studholme & Dixon, 2003), our initial analysis of the genomic sequence of this region to identify potential binding sites of various transcriptional regulators revealed two potential Fnr binding sites – one centred at -72.5 (TTAACCTGGCTCAA; bolded bases represent perfect matches to the Fnr consensus sequence) and another one located further upstream at -188.5 (TTGCTTATTATCAG) (Fig. 3.3). Both sites are similar to the Fnr consensus sequence TTGATnnnnATCAA (n can be any base) (Green

et al., 2001), but they lack the characteristic palindromic sequences that flank the two ends.

Figure 3.3. Genomic DNA sequence in E. coli K-12 MG1655 corresponding to the ravAviaA promoter region

The ravA and kup (encodes the K+ transporter; shown here in reverse complement) open reading frames are coloured in purple and green, respectively. The σS consensus sequence (in red) (Typas et al., 2007), transcription start site (in blue) (Mendoza-Vargas et al., 2009) and Shine-Dalgarno sequence (in black box) are indicated as shown. The potential Fnr binding sites are highlighted with red boxes. The absolute genomic coordinates are given to the left of each line.

______

To further examine these two potential binding sites for Fnr, three linear DNA substrates of the same length encompassing different parts of the ravAviaA promoter region (R-1, R-2 and

R-3; Fig. 3.4A) were synthesized by PCR for use in EMSA with FnrD154A. The FnrD154A mutant, which has the same affinity and specificity for the Fnr consensus sequence as its WT counterpart (Lazazzera et al., 1993), was used so that the experiment could be performed under aerobic

condition. As shown in Fig. 3.4B, the inclusion of FnrD154A in the sample induced significant band shifts for both DNA substrates R-1, which has both the -72.5 and -188.5 Fnr binding sites,

99 and R-2, which retains only the -72.5 site (Fig. 3.4A). In contrast, band shift was reduced for substrate R-3 (Fig. 3.4B), which lacks both Fnr binding sites (Fig. 3.4A). Furthermore, as both

R-1 and R-2 displayed very similar band shifts in the presence of FnrD154A (Fig. 3.4B), the – 188.5 site did not appear to be necessary for binding. In other words, the interaction between

FnrD154A and the ravAviaA promoter appears to be largely mediated through the -72.5 site.

Figure 3.4. Electromobility shift assays (EMSA) on the ravAviaA promoter region and its variants using the mutant transcriptional regulator FnrD154A

(A, C, D) Schematic representations of the DNA substrates R-1, R-2 and R-3 for the ravAviaA promoter region (A), F-1 for the fepD promoter region (C), and H-1, H-2 and H-3 for the hypBCDE promoter region (D). Both putative and confirmed binding sites for Fnr are indicated with black boxes as illustrated, and their DNA sequences are shown below. Bases that are underlined represent the half-sites (both putative and confirmed) that are crucial for Fnr binding (Green et al., 2001; Messenger & Green, 2003). The genomic region covered by each DNA substrate is indicated at both ends with the corresponding E. coli K-12 genome coordinates. Bent arrows represent the transcriptional start sites (+1).

100

(B, D) EMSA results using substrates R-1, R-2 and R-3 (B), and F-1, H-1, H-2 and H-3 (E). The absence (-) and presence (+) of FnrD154A in the reaction mixture are as indicated at the top. The origins of all DNA substrates used are as shown. The molecular weights of the DNA ruler used and their in-gel positions are shown on the left.

______

To provide additional proof that the interaction between FnrD154A and the DNA substrates R-1 and R-2 was specific, EMSA was repeated using the substrates F-1 as negative control, and H-1, H-2 and H-3 as positive controls. Substrate F-1 encompasses the entire fepDGC promoter region, which carries four binding sites for the Fe2+-sensing Fur (Chenault & Earhart, 1991; Lavrrar et al., 2002; Chen et al., 2007) but none for Fnr (Fig. 3.4C). In contrast, substrate H-1 encompasses the hypBCDE promoter region that is internal to the hypA gene. It contains two Fnr binding sites – one centred at -43.5 (TTGATCTGGTTTGC; bolded bases represent perfect matches to the Fnr consensus sequence) and the other one further upstream at -149.5 (TTGATCGAACAGCA) (Messenger & Green, 2003) (Fig. 3.4D). Following the same scheme in removing the Fnr binding sites from the ravAviaA promoter (Fig. 3.4A), substrates H-2 (only the -43.5 site is retained) and H-3 (all Fnr binding sites removed) were also synthesized (Fig.

3.4D). As shown in Fig. 3.4E, FnrD154A did not interact with either F-1 or H-3, as neither substrate carries Fnr binding sites. On the other hand, H-1 contains both Fnr binding sites and it

interacted strongly with FnrD154A, resulting in a significant band shift; while, for substrate H-2,

removal of the upstream -149.5 Fnr binding site severely diminished its interaction with FnrD154A

(Fig. 3.4E). Thus, the binding of FnrD154A to substrates R-1, R-2 and H-1 were indeed specific.

3.4.4 ViaA physically interacts with FrdA in anaerobically growing cells

Next, we aimed to identify physical interactors of RavA and ViaA under anaerobic conditions, using strains that express endogenous C-terminally SPA-tagged proteins (Babu et al., 2009), followed by the detection of RavA and/or ViaA bound to the purified protein complexes by Western-blotting. The choices of proteins to be tagged were based on the results of our co- expression profiling for ravA and viaA, as well as, on the several high-throughput studies that we previously carried out (see Section 2.4.2).

Among the 22 proteins that were successfully SPA-tagged, ViaA (untagged; MW = 56 kDa) was observed to interact strongly with SPA-tagged FrdA (Fig. 3.5A). FrdA is the FAD- binding component of the fumarate reductase complex (FrdABCD) that is involved in anaerobic

101

respiration (Cecchini et al., 2002). No ViaA was observed when FrdA-SPA pull-down was repeated on ΔviaA cells (Fig. 3.5B), which confirms the validity of the observed ViaA-FrdA interaction. In contrast, none of the SPA-tagged proteins interacted with RavA, with the only exception being LdcI-SPA (data not shown), which agrees with our previous report (Snider et al., 2006). In addition, we had previously shown that RavA only interacts weakly with ViaA (Snider et al., 2006), and accordingly, no ViaA was brought down with RavA-SPA (Fig. 3.5A), and vice versa (data not shown).

Figure 3.5. Physical interactors of RavA and ViaA by SPA-tag immunoprecipitation

(A) Initial screen to identify physical interactors of ViaA in DY330 cells grown anaerobically in LB. SPA tags were added to the endogenous proteins. The identities of bait proteins are as given at the top. “∗” denotes ViaA-SPA; “†” denotes cross-reacting bands in the α-ViaA blot. The ViaA band is indicated by arrow.

102

(B) Confirmation of the interaction between ViaA and FrdA-SPA. Note that FrdB expression was abolished due to the introduction of the SPA tag. All cells were grown anaerobically in LB. Identities of the strains are as given at the top. Total = soluble proteins from total cell lysate; IP = immunoprecipitation.

______

3.4.5 RavA-ViaA bind free FrdA

In the FrdA-SPA strain, the expression of the Fe-S clusters-containing subunit of fumarate reductase, FrdB, was compromised due to the introduction of the SPA tag (Fig. 3.5B, 3.6A and B). Similarly, given that both frdC and frdD are located downstream of frdB, expression of FrdC and FrdD were presumed to be compromised as well (however antibodies against the two proteins is not currently available for us). Hence, the interaction between RavA/ViaA and FrdA- SPA does not require endogenous FrdBCD and RavA/ViaA bind free FrdA.

To further establish this observation, the FrdA-SPA strain was transformed with plasmids that overexpress the subunits of the Frd complex in various combinations, and the immunoprecipitation experiments were repeated. Interestingly, the expression of FrdB, FrdBCD, FrdAB, FrdABCD under the frd promoter (frdp) did not affect the total levels of RavA and ViaA, but resulted in less ViaA that is solubilized with 1% Triton X-100. (Fig. 3.6A and B input). In contrast, the levels of detergent-soluble RavA were not affected (Fig. 3.6A and B input). The inner membrane protein LepB is shown as control for proper cell lysis and loading (Fig. 3.6A). Interestingly, the levels of detergent-soluble FrdB were also compromised during the expression of FrdB or FrdBCD under the control of frdp, although significant improvements were observed with the expression of FrdAB or FrdABCD (Fig. 3.6A). This suggests a potential stabilization effect of FrdA on FrdB. Importantly, no interaction was observed between FrdA and ViaA in the presence of FrdB or FrdBCD (Fig. 3.6B IP). We interpret these results to mean that ViaA binds free FrdA and not FrdA in an FrdAB or FrdABCD complex.

3.4.6 RavA modulates the binding of ViaA to FrdA in an ATP-dependent manner

In light of the interaction of FrdA-SPA with ViaA, the potential role of RavA in modulating this interaction was also investigated. The immunoprecipitation experiments were repeated for FrdA-

SPA strain transformed with the plasmids pRV or pRK52QV to better detect the interactions.

pRK52QV expresses WT ViaA but a mutant RavA in which the Walker A Lys is mutated to Gln

103

rendering the protein ATPase inactive. As shown in Figure 3.6C, protein levels were the same

(input), however, more ViaA and also RavA bound FrdA-SPA in the presence of pRK52QV compared to pRV. Hence, we speculate that RavA ATPase activity might be required to reduce the ViaA-FrdA interaction.

Figure 3.6. RavA and FrdB modulate the interaction between ViaA and FrdA-SPA

104

(A) Western blots for RavA, ViaA, FrdA, and FrdB in anaerobically grown DY330 FrdA-SPA Confirmation of the loss of soluble ViaA from overexpression of FrdB. Cells express different constructs as shown. The membrane protein marker LepB is used to monitor solubilization of membrane proteins. Details on the strains and plasmids used are given in Table 3.1.

(B) Same as (A) but also include the immunoprecipitation of FrdA. Sol. Ptn. = soluble proteins; IP = Immunoprecipitation.

(C) Western blots for RavA, ViaA and FrdA in anaerobically grown DY330 FrdA-SPA transformed with pRV or pRK52QV. Details on the strains and plasmids used are given in Table 3.1.

______

3.4.7 RavA-ViaA stimulate the activity of the Frd complex in E. coli

Both the physical interaction between ViaA and FrdA-SPA and the antagonistic effect of RavA on the FrdA-ViaA interaction suggest that RavA-ViaA might functionally modulate the activity of the Frd complex. To test this possibility, membrane vesicles isolated from anaerobically grown MG1655 cells expressing different levels of RavA and/or ViaA were examined for differences in fumarate reductase activity in vitro by measuring the oxidation of benzyl viologen (BV) in the presence of fumarate.

As shown in Fig. 3.7A, vesicles isolated from ΔravAviaA + p11 and ΔravAviaA + pR both showed a decrease in Frd activity compared to WT vesicles. In contrast, ΔravAviaA + pRV vesicles showed a ~46% increase in activity relative to WT. Evidently, both RavA and ViaA are required for the anaerobically grown cell to achieve optimal Frd activity, although neither protein is essential. The drastic difference observed between ΔravAviaA + pR and ΔravAviaA + pRV vesicles provides additional supporting evidence for our proposed role of ViaA in mediating the interaction between RavA and the Frd complex. As control, vesicles from ΔfrdA::kanR did not show any fumarate reductase activity, which highlights the specificity of this assay in capturing only the activity of the Frd complex. Importantly, neither RavA nor ViaA had any observable effects on the expression levels of FrdA or FrdB (and presumably, FrdC and FrdD) (Fig. 3.7B). Thus, the increase in Frd activity observed thus far could only be attributed to RavA-ViaA stimulating the activity of the Frd complexes that were already expressed. To further determine if the ATPase activity of RavA is required in stimulating Frd activity, the assay was also performed on vesicles isolated from anaerobically grown ΔravAviaA + pRK52QV. As shown in Fig. 3.7A,

ΔravAviaA + pRK52QV vesicles exhibited less Frd activity compared to ΔravAviaA + pRV vesicles (p < 0.02), although the difference was not as significant as with ΔravAviaA + pR

105 vesicles. Evidently, the overexpression of ViaA alone is sufficient to stimulate the Frd complex, although RavA is required for optimal stimulation.

Figure 3.7. RavA-ViaA enhance fumarate reductase activity in anaerobically grown E. coli MG1655

(A) Benzyl viologen (BV) assay on membrane vesicles isolated from cells expressing different levels of RavA and ViaA. The strains used are identified below the graph. Three independent experiments were

106

conducted for each strain. 1 U = 1 μmol BV oxidized per minute. P-values comparing the ΔravAviaA transformants with WT (indicated with *) and between ΔravAviaA + pRV and ΔravAviaA + pRK52QV are as shown. Details on the strains and plasmids used are available in Table 3.1.

(B) Western blotting confirms that protein levels of FrdA and FrdB remain unchanged upon changes in RavA and ViaA levels. Incidentally, loss of FrdB expression was also detected in ΔfrdA::kanR.

______

3.5 Discussion

The results we have presented here reveal a novel interaction between the RavA-ViaA system and the Frd complex in E. coli, in which RavA and ViaA are required for optimal activity of the complex during anaerobic growth. Accordingly, the expression of RavA-ViaA was also revealed to be inducible by the anaerobic transcriptional regulator Fnr, which also regulates the expression of the Frd complex as well as other proteins involved in anaerobiosis (Kang et al., 2005; Grainger et al., 2007).

In a previous report, we have shown that RavA and ViaA interact both physically and functionally with specific subunits of the NADH:ubiquinone oxidoreductase I (Nuo complex) (see Section 2.4.6). Specifically, RavA-ViaA interacts primarily with the FMN (flavin mononucleotide)-binding NuoF subunit under aerobic conditions, and with the fused NuoCD subunit under anaerobic conditions (see Section 2.4.6). Importantly, the Nuo complex is known to be involved in both aerobic and anaerobic respiration of E. coli (Unden & Bongaerts, 1997; Price & Driessen, 2010). It has also been shown that the coupling between the Nuo complex and the Frd complex is important for the electron transfer from NADH to fumarate during the anaerobic respiration of E. coli (Tran et al., 1996). Taken together, our results support a potential regulatory role of RavA-ViaA over the anaerobic utilization of fumarate via its interaction with both the Nuo and Frd complexes. Considering that MoxR proteins generally act as chaperones for specific targets (Snider & Houry, 2006; Wong & Houry, 2012), it is conceivable that RavA- ViaA can fulfill its regulatory function by regulating or facilitating the maturation process of specific Nuo and Frd subunits and/or the assembly of these subunits into functional respiratory units. Further investigation is in progress to decipher the molecular mechanism underlying this proposed regulatory role of RavA-ViaA in the fumarate-dependent anaerobic respiratory process in greater detail.

107

Chapter 4 General Conclusion & Future Directions

108

4 General Conclusion & Future Directions 4.1 General Conclusion of Thesis

My research on the function of the AAA+ ATPase RavA and the associated VWA protein ViaA in Escherichia coli has revealed their physical and functional interactions with specific subunits of the NADH:ubiquinone oxidoreductase I (NuoA-N) and fumarate reductase (FrdA-D), both of which being the major protein complexes in the anaerobic respiration of fumarate in the absence of nitrate. This is also the first instance of a MoxR protein in E. coli or a similar bacterial species showing a functional linkage to the anaerobic respiratory process (Snider & Houry, 2006; Wong & Houry, 2012). Importantly, after careful examination of the experimental evidence presented, two key premises were conceived regarding the potential biological function of RavA-ViaA. These are as follows.

4.1.1 Interaction of RavA-ViaA with the target subunits occurs prior to the assembly of the full complex

The first premise is that the interaction of RavA-ViaA with specific subunits of the Nuo and Frd complexes likely occurs prior to their full assembly. This is supported by the following observations:

Firstly, with respect to the Nuo complex, only the null mutations ΔnuoB, ΔnuoCD and ΔnuoF can fully suppress the sensitization of the cell to kanamycin induced by RavA-ViaA overexpression (Fig. 2.8 and Table 2.2; see Section 2.4.5), despite that null mutations of any single nuo genes will abolish the activity of the full complex (Erhardt et al., 2012). This strongly indicates that RavA-ViaA target only specific Nuo subunits rather than the fully assembled Nuo complex. Furthermore, the physical interaction profile of RavA-ViaA with the targeted Nuo subunits that are SPA-tagged differs significantly between aerobically and anaerobically grown cells (Fig. 2.9; see Section 2.4.6). Coincidentally, the expression of RavA-ViaA is significantly increased under anaerobic condition (Fig. 3.2; see Section 3.4.2), which is possibly related to the change in RavA-ViaA’s interactions with the Nuo subunits. This further substantiates RavA- ViaA’s preference for specific Nuo subunits over the full complex.

Secondly, with respect to the Frd complex, ViaA physically interacts with the free SPA- tagged FrdA subunit in the absence of endogenous FrdB, and presumably FrdC and FrdD (Fig.

109

3.5; see Section 3.4.5). Furthermore, the re-introduction of FrdB alone or together with FrdC and FrdD via plasmids antagonizes the ViaA-FrdA interaction by inducing the aggregation of the VWA protein (Fig. 3.6A and B; see Section 3.4.6). These two observations provide direct evidence that the stable interaction between ViaA and FrdA can only occur in the absence of the other Frd subunits. Importantly, the gene arrangement of the frdABCD operon dictates that FrdA must first be expressed prior to FrdB, followed by FrdC and FrdD (Cecchini et al., 2002). Also, the assembly of the full complex is proposed to initiate with the incorporation of FrdC and FrdD into the inner membrane, followed by the anchoring of the hydrophilic FrdA and FrdB (Cecchini et al., 2002). Thus, considering all of the factors above, it is conceivable that the interaction must occur prior to the maturation of the Frd complex. Hypothetically, the same assumption can be made for the interaction between RavA-ViaA and the targeted Nuo subunits.

4.1.2 RavA-ViaA facilitates the assembly of the full complex

The second premise that was conceived from the data presented is that RavA-ViaA may facilitate the assembly and maturation of the full Nuo and Frd complex via the prior interaction with their respective subunits.

Although the initial attempts to assess the effects of RavA-ViaA on the activity of the Nuo complex were unsuccessful (data not shown), the effects on the Frd complex were significant. Namely, the null mutant lacking both RavA and ViaA (i.e. ΔravAviaA) showed a decrease in endogenous Frd activity, while the overexpression of RavA-ViaA induces a large increase (Fig. 3.7A; see Section 3.4.7). The modulation of Frd activity also appears to be dependent on RavA’s ATPase activity (compare “ΔravAviaA + pRV” and “ΔravAviaA + pRK52QV” in Fig. 3.7A). Furthermore, examination of the expression of endogenous FrdA and FrdB in wild type, the null mutant and the RavA-ViaA overexpressor confirms that the observed changes in Frd activity was not the result of changes in the complex’s expression. This reflects the hypothetical role of RavA-ViaA in facilitating the maturation of the Frd complex, in which the ViaA-FrdA interaction is a significant step. Notably, this is the classical biological function that has been reported repeatedly for the many MoxR proteins characterized to date (Snider & Houry, 2006; Wong & Houry, 2012). These include MoxR (MRP subfamily) from Paracoccus denitrificans and Methylobacterium extorquens (van Spanning et al., 1991; Toyama et al., 1998), NirQ/NorQ (CGN subfamily) from Pseudomonas aeruginosa, Pseudomonas stutzeri,

110

Paracoccus denitrificans, and Rhodobacter sphaeroides 2.4.3 (Jungst & Zumft, 1992; de Boer et al., 1996; Bartnikas et al., 1997; Arai et al., 1999), and CoxD (APE2220) from Oligotropha carboxidovorans OM5 (Pelzmann et al., 2009), among others.

Based on the two premises discussed above, RavA-ViaA can hypothetically serve a chaperone-like function for specific subunits of the Nuo and Frd complexes, which in turn facilitates their assembly into mature functional enzymes. Using the assembly of the Frd complex as the basis, a model for the molecular actions of RavA-ViaA is proposed, as illustrated in Fig. 4.1. Briefly, ViaA physically interacts and stabilizes a newly synthesized FrdA, while the translation, folding and post-translational processing of FrdB takes place (Fig. 4.1A). Once FrdB has reached its native state, ViaA is displaced from FrdA (Fig. 4.1B) and the free ViaA forms insoluble aggregates (Fig. 4.1C). The release of ViaA is likely to be facilitated by RavA in an ATP-dependent manner (Fig. 4.1B). The now stabilized FrdAB sub-complex then binds the membrane subunits FrdC and FrdD upon their proper insertion into the inner membrane (Fig. 4.1D), leading to the fully assembled Frd complex (Fig. 4.1E). A similar scenario may also be envisioned for the Nuo complex.

Figure 4.1. Hypothetical Role of RavA-ViaA as Molecular Chaperone for FrdA in the Assembly and Maturation of the Fumarate Reductase Complex

111

(A) ViaA physically interacts and stabilizes the newly synthesized FrdA.

(B) A newly synthesized FrdB displaces ViaA from FrdA to form the stable FrdAB dimer. The release of ViaA is likely facilitated by the RavA hexamer in an ATP-dependent manner. The free ViaA may participate in another round of FrdA-binding, although no experimental evidence is available at this point (indicated with dotted arrow).

(D) The FrdAB dimer docks with the newly synthesized FrdCD membrane anchor, upon the complete folding and insertion of the two membrane subunits.

(E) The docking of FrdAB with FrdCD completes the assembly process of the mature fumarate reductase complex.

______

4.2 Future Directions

The identification of specific Nuo and Frd subunits as the physical and functional interactors of RavA-ViaA, as well as the initial evidence of RavA-ViaA in modulating the activity of the Frd complex via a potential chaperone-like function, both of which have provided invaluable clues towards understanding the molecular and biological function of RavA-ViaA in Escherichia coli. Importantly, these results bring focus onto specific functional aspects of RavA-ViaA that can now be systematically tested and validated through the strategic design of future experiments, as follows.

4.2.1 In-depth characterization of the interactions between RavA-ViaA and the subunits of the Nuo and Frd respiratory complexes

This section outlines experiments that aim at deciphering the molecular details underlying the interactions between RavA-ViaA and specific subunits of the Nuo and Frd respiratory complexes, as well as the dynamics of these interactions.

4.2.1.1 Analyzing the composition of Nuo sub-complexes that co-purify with RavA and/or ViaA

Although specific Nuo subunits that are SPA-tagged have been identified as specific interactors of RavA and/or ViaA (Fig. 2.9A and B), it remains unclear if the interactions occur directly or indirectly via additional proteins that are also co-purified. Furthermore, the introduction of endogenous SPA-tagged to one Nuo subunit may disrupt the expression of the other subunits, as observed in the loss of endogenous expression of FrdB, C and D due to FrdA being SPA-tagged

112

(Fig. 3.5B). The preferred binding of RavA and/or ViaA with incomplete Nuo complexes lacking specific subunits (i.e. Nuo sub-complexes) may have important functional relevance. For example, Erhardt and co-workers have reported the formation of Nuo sub-complexes in the E. coli null mutants, ΔnuoK and ΔnuoL (Erhardt et al., 2012). Importantly, they found two forms of Nuo sub-complexes in ΔnuoL – one being enzymatically active and the other is not – such that the acid-stress response protein LdcI (inducible lysine decarboxylase) binds specifically to the inactive form of the ΔnuoL sub-complex (Erhardt et al., 2012). LdcI is a known interactor of RavA (Snider et al., 2006; El Bakkouri et al., 2010; Kanjee et al., 2011).

To analyze the composition of the Nuo sub-complexes that are co-purified with RavA and/or ViaA, the immunoprecipitation experiments on SPA-tagged Nuo subunits (see Section 2.3.9) will be repeated, and the co-purified protein complexes will be analyzed using conventional methods to identify their composition. These include SDS-PAGE followed by in- gel trypsin digest and protein identification by mass spectrometry, as well as detection of specific Nuo subunits by Western blotting. The co-purified protein complexes will also be subjected to in vitro experimentation, such as various enzymatic assays (e.g. NADH dehydrogenase assay (Erhardt et al., 2012)) for assessing specific respiratory processes, and the application of EPR (Electron Paramagnetic Resonance) to analyze the bound redox cofactors (e.g. FMN, Fe-S clusters, etc.) (Euro et al., 2008; Verkhovskaya et al., 2008; Erhardt et al., 2012) for any changes induced by RavA and/or ViaA.

4.2.1.2 Characterizing the ViaA-FrdA interaction and the antagonistic role of FrdB at the molecular level

Compared to the Nuo subunits, more details about the interaction of ViaA and FrdA were obtained from the experiments presented in this thesis (see Sections 3.4.4 to 3.4.6). Thus, aside from refining the in vitro assays for assessing the role of RavA-ViaA in the activity of the Frd complex (see Section 3.4.7), the experimental focus will also be placed on analyzing the biophysical aspects of the ViaA-FrdA interaction. In addition, the displacement of ViaA from FrdA and the induction of ViaA aggregation by FrdB will be re-examined in vitro. The potential role of RavA in releasing ViaA from FrdA will also be analyzed in detail.

In order to assess the ViaA-FrdA interaction at the molecular level, the contribution of individual ViaA domains will first be examined. Based on primary sequence analysis, ViaA can

113

be divided into two domains – an N-terminal domain of unknown function consisted mostly of α-helices, and a C-terminal VWA domain (Dr. Jamie Snider; personal communications). Thus, the two ViaA domains will be cloned into protein overexpression vectors and purified using Ni- NTA in conjunction with conventional columns used in FPLC (Fast Performance Liquid Chromatography). Also, FrdA will be overexpressed from the plasmid pFAB-HT (a generous gift from Prof. Gary Cecchini, Veterans Affairs Medical Centre, San Francesco, US) and purified as described in reference (Leger et al., 2001). The purified ViaA domains and FrdA will then be used in established in vitro assays designed to analyze biomolecular interactions quantitatively, such as ELISA (Enzyme-linked Immunosorbent Assay) (Engvall & Perlman, 1971; van Weemen & Schuurs, 1971), SPR (Surface Plasmon Resonance)(Jason-Moller et al., 2006) or ITC (Isothermal Titration Calorimetry) (Ghai et al., 2012). The same techniques can also be applied for the full-length ViaA. In addition, for the C-terminal VWA domain of ViaA, point mutations can also be introduced systematically to the conserved and functionally important MIDAS motif (Whittaker & Hynes, 2002; Springer, 2006) via QuikChange site-directed mutagenesis (Papworth et al., 1996). This will allow the critical assessment of the VWA domain’s role in mediating the ViaA-FrdA interaction.

The displacement of ViaA from FrdA by FrdB (Fig. 3.6A and B; see Section 3.4.5) is intriguing as it reflects the potential role of ViaA in stabilizing free FrdA prior to the formation of the FrdAB dimer (Fig. 4.1). To assess the antagonistic role of FrdB on the ViaA-FrdA interaction, the Fe-S clusters-carrying subunit will be cloned and purified using conventional protein purification techniques as discussed previously. The purified FrdB will then be used in modified versions of the in vitro biomolecular assays listed above to monitor the displacement of ViaA from FrdA induced by FrdB. Similarly, RavA will be overexpressed and purified as described in reference (El Bakkouri et al., 2010). The purified RavA will then be used in similar in vitro assays to validate its role in facilitating the release of ViaA from FrdA in the presence of ATP (Fig. 3.6C and 4.1).

In addition to the aforementioned experiments, the interaction interface between ViaA and FrdA will also be characterized from a structural perspective. The co-crystallization of ViaA and FrdA and X-ray crystallography will be attempted to obtain the best possible data on the interaction interface. Alternatively, X-ray structures of full-length ViaA and its two domains will be solved. At the same time, electron microscopy (EM) techniques on biomolecules (e.g. cryo-

114

EM (Milne et al., 2013)) will be performed on the purified oligomers of ViaA (or its domains) and FrdA to obtain the necessary structural data. The EM structures can then be used for docking the X-ray structures of FrdA (Iverson et al., 2002), ViaA and its two domains to construct structural models of the ViaA-FrdA interaction interface (see (Allen & Stokes, 2013). These experiments will provide invaluable information on key surfaces and amino acid residues that mediate the ViaA-FrdA interaction.

4.2.2 Re-examination of the RavA-ViaA-induced sensitization to aminoglycosides independent of ROS

Aside from providing the necessary clues for the identification of specific Nuo and Frd subunits as the physical and functional interactors, the sensitization of the E. coli cell towards aminoglycosides (e.g. kanamycin, streptomycin, etc.) induced by the activity of RavA and ViaA (Fig. 2.6 and 2.7; see Sections 2.4.3 and 2.4.4) highlights their involvement in the mechanisms behind the bactericidal effects of these antibiotics.

Recent studies have shown that the cell’s sensitivity towards aminoglycosides can be attributed to specific metabolic genes, many of which encode proteins that participate respiratory electron transport (e.g. Nuo, Sdh, Cyo, etc.) or in the biosynthesis of the required cofactors (e.g. Ubi, Isc, etc.) (Kohanski et al., 2008; Girgis et al., 2009; Ezraty et al., 2013; Mahoney & Silhavy, 2013). Although the details behind bactericidal effects of aminoglycosides remain under debate (see Section 2.5), one of the latest models suggests that the influx of aminoglycosides into the cell is dependent on the proton translocation activities of the Nuo complex and the succinate dehydrogenase (SdhA-D), and the disruption to their maturation process by impairing the Isc Fe- S clusters assembly pathway results in greater aminoglycoside resistance of the cell (Ezraty et al., 2013). Importantly, unlike several other models that postulates the formation of ROS (reactive oxygen species) as the primary means of bacterial killing (Kohanski et al., 2007; Kohanski et al., 2008; Ling et al., 2012; Mahoney & Silhavy, 2013), the ROS-less model is applicable to both aerobically and anaerobically grown E. coli, given the experimental evidence showing that ROS is not generated and therefore is not necessary for cell death by aminoglycosides under anaerobic condition (Keren et al., 2013; Liu & Imlay, 2013). Furthermore, the ROS-less model has greater pathological relevance to E. coli when discussing the effects of aminoglycosides during the course of an infection, which occurs in the anaerobic environment of the gut (Kaper et al., 2004; Croxen & Finlay, 2010).

115

Thus, using the ROS-less model as the basis, the RavA-ViaA-induced sensitization to aminoglycosides will be re-examined. The focus will now be placed on studying the effects of RavA-ViaA on the influx of aminoglycosides. Specifically, the influx of aminoglycosides into the cell will be monitored via radiolabeled versions of these antibiotics (e.g. 3H-kanamycin) for wild type E. coli and the null mutants of ravA and/or viaA, using protocols similar to the one described in (Ezraty et al., 2013). Similar experiments will then be performed on strains complemented with plasmids expressing RavA and/or ViaA from the native ravA promoter for validation purposes. Furthermore, these experiments can be adopted for specific E. coli mutants that rely on a specific respiratory chain (e.g. Nuo-Frd) for growth, in order to determine the mechanical details of the proposed link between RavA-ViaA and the influx of aminoglycosides.

4.2.3 Investigation of a Potential Role of RavA-ViaA in Stress-induced Mutagenesis

Stress-induced mutagenesis (SIM) refers to the mutagenic events that occur as part of the bacterial stress response process, which in turn may facilitate evolution (Foster, 2007). In E. coli, the repair of double-strand breaks (DSBs) in DNA under stressful conditions results in SIM due to the use of error-prone DNA polymerases II, IV and V in a σS-dependent manner (Rosenberg et al., 2012). Recently, Al Mamun and co-workers identified a network of 93 genes that facilitates SIM, in which the largest class of genes encodes proteins associated with the electron transport chain (ETC) (Al Mamun et al., 2012). These include the subunits of several major aerobic respiratory complexes (e.g. Nuo, Sdh and Cyo), the ubiquinone biosynthetic pathway (Ubi), and proteins involved in the biosynthesis of various cofactors (e.g. HemL and HscB) (Al Mamun et al., 2012). Importantly, these proteins promote SIM by signaling the activation of σS-dependent starvation and general stress response via the signaling proteins ArcB, ArcA and the σS regulator RssB (Al Mamun et al., 2012).

The interactions between RavA-ViaA and subunits of the Nuo complex (see Chapter 2) raise the possibility that RavA and ViaA are functionally linked to SIM as well. This is partially supported by results from the microarray experiments, such that the expression of umuC and umuD, which encode the two subunits of the error-prone DNA polymerase V, both increased in ΔravA::cat (see Appendix A, “Set 1 (ΔravA::cat vs. WT) increases”). Thus, the loss of RavA in

116

the cell is likely to increase SIM, possibly via the interactions with Nuo subunits or some other means that are yet to be determined.

To confirm the link between RavA-ViaA and SIM, DSB-dependent mutagenesis will be examined in MG1655 WT, the KO mutants of ravA and/or viaA, as well as strains carrying plasmids that overexpress RavA and/or ViaA under the control of the native ravAviaA promoter (see Tables 2.1 and 3.1). Changes in mutagenesis rate will first be identified using the qualitative colony-colour papillation assay, which can then be validated and quantified using the DSB- dependent chromosomal stress-induced mutation assays, all of which are described in (Al Mamun et al., 2012). Furthermore, the interactions between RavA-ViaA and subunits of the Nuo complex will be re-examined in the context of SIM. For example, null mutations of ravA, viaA and specific nuo genes can be introduced in various combinations. The mutagenic rates of the resultant digenic or multigenic mutants under stress will then be measured and compared to WT and the monogenic KO of ravA, viaA or specific nuo genes.

4.3 Closing Remarks

Suffice to say, the experimental data presented has provided but a first glimpse of the biological properties of RavA-ViaA that are yet to be explored. Our understanding of the molecular basis behind the interaction between RavA-ViaA and its targets is still at its infancy. Similarly, the sensitization of the cell towards aminoglycosides induced by the activity of RavA-ViaA has yet to be investigated completely, albeit the obvious clinical implications.

Thus, the road remains long ahead, and we must continue our efforts in understanding the nature of RavA-ViaA and the MoxR family in general.

117

References

Al Mamun, A. A. M., Lombardo, M., Shee, C., Lisewki, A. M., Gonzalez, C., Lin, D., Nehring, R. B., Saint-Ruf, C., Gibson, J. L., Frisch, R. L., Lichtarge, O., Hastings, P. J. and Rosenberg, S. M. (2012). Identity and Function of a Large Gene Network Underlying Mutagenic Repair of DNA Breaks. Science. 338(6112): 1344-8. Alexopoulos, E., Kanjee, U., Snider, J., Houry, W. A. and Pai, E. F. (2008). Crystallization and preliminary X-ray analysis of the inducible lysine decarboxylase from Escherichia coli. Acta Crystallogr Sect F Struct Biol Cryst Commun. 64(Pt 8): 700-6. Allen, G. S. and Stokes, D. L. (2013). Modeling, docking, and fitting of atomic structures to 3D maps from cryo-electron microscopy. Methods Mol Biol. 955: 229-41. Altuvia, S., Weinstein-Fischer, D., Zhang, A., Postow, L. and Storz, G. (1997). A small, stable RNA induced by oxidative stress: role as a pleiotropic regulator and antimutator. Cell. 90(1): 43-53. Alvarez, A. F., Rodriguez, C. and Georgellis, D. (2013). Ubiquinone and menaquinone electron carriers represent the yin and yang in the redox regulation of the ArcB sensor kinase. J Bacteriol. 195(13): 3054-61. Ammelburg, M., Frickey, T. and Lupas, A. N. (2006). Classification of AAA+ proteins. J Struct Biol. 156(1): 2-11. Ansaldi, M., Theraulaz, L., Baraquet, C., Panis, G. and Mejean, V. (2007). Aerobic TMAO respiration in Escherichia coli. Mol Microbiol. 66(2): 484-94. Aono, R. and Sanada, T. (1994). Hyper-autolysis of the faculatative alkaliphile Bacillus sp. C- 125 cells grown at neutral pH: culture-pH dependent cross-linking of the peptide moieties of the peptidoglycan. Biosci Biotechnol Biochem. 58: 2015-9. Arai, H., Kodama, T. and Igarashi, Y. (1999). Effect of nitrogen oxides on expression of the nir and nor genes for denitrification in Pseudomonas aeruginosa. FEMS Microbiol Lett. 170(1): 19-24. Ayala-Castro, C., Saini, A. and Outten, F. W. (2008). Fe-Cluster Assembly Pathways in Bacteria. Microbiol Mol Biol Rev. 72(1): 110-25. Baba, T., Ara, T., Hasegawa, M., Takai, Y., Okumura, Y., Baba, M., Datsenko, K. A., Tomita, M., Wanner, B. L. and Mori, H. (2006). Construction of Escherichia coli K-12 in-frame, single-gene knockout mutants: the Keio collection. Mol Systems Biol. Babu, M., Butland, G., Pogoutse, O., Li, J., Greenblatt, J. F. and Emili, A. (2009). Sequential peptide affinity purification system for the systematic isolation and identification of protein complexes from Escherichia coli. Methods Mol Biol. 564(373-400). Babu, M., Diaz-Mejia, J. J., Vlasblom, J., Gagarinova, A. G., Phanse, S., Graham, C., Yousif, F., Ding, H., Xiong, X., Nazarians-Armavil, A., Alamgir, M., Ali, M., Pogoutse, O., Pe'er, A., Arnold, R., Michaut, M., Parkinson, J., Golshani, A., Whitfield, C., Wodak, S. J., Moreno-Hagelsieb, G., Greenblatt, J. and Emili, A. (2011). Genetic Interaction Maps in Escherichia coli Reveal Functional Crosstalk among Cell Envelope Biogenesis Pathways. PLoS Genet. 7(11): e1002377.

118

Bachmann, B. J. (1972). Pedigrees of some mutant strains of Escherichia coli K-12. Bacteriol Rev. 36(4): 525-57. Baharoglu, Z., Krin, E. and Mazel, D. (2013). RpoS Plays a Central Role in the SOS Induction by Sub-Lethal Aminoglycoside Concentrations in Vibrio cholerae. PLoS Genet. 9(4): e1003421. Baichoo, N. and Helmann, J. D. (2002). Recognition of DNA by Fur: a reinterpretation of the Fur box consensus sequence. J Bact. 184(21): 5826-32. Baradaran, R., Berrisford, J. M., Minhas, G. S. and Sazanov, L. A. (2013). Crystal structure of the entire respiratory complex I. Nature. 494(7483): 443-8. Bartnikas, T. B., Tosques, I. E., Laratta, W. P., Shi, J. and Shapleigh, J. P. (1997). Characterization of the nitric oxide reductase-encoding region in Rhodobacter sphaeroides 2.4.3. J Bacteriol. 179(11): 3534-40. Bekker, M., Alexeeva, S., Laan, W., Sawers, G., Teixeira de Mattos, J. and Hellingwerf, K. (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(3): 746-54. Bilous, P. T. and Weiner, J. H. (1985). Dimethyl sulfoxide reductase activity by anaerobically grown Escherichia coli HB101. J Bacteriol. 162(3): 1151-5. Bochtler, M., Hartmann, C., Song, H. K., Bourenkov, G. P., Bartunik, H. D. and Huber, R. (2000). The structures of HsIU and the ATP-dependent protease HsIU-HsIV. Nature. 403(6771): 800-5. Bradley, M. D., Beach, M. B., de Koning, A. P., Pratt, T. S. and Osuna, R. (2007). Effects of Fis on Escherichia coli gene expression during different growth stages. Microbiology. 153(Pt. 9): 2922-40. Braun, M., Bungert, S. and Friedrich, T. (1998). Characterization of the overproduced NADH fragment of the NADH:ubiquinone oxidoreductase (complex I) from Escherichia coli. Biochemistry. 37(7): 1861-7. Brewster, A. S., Wang, G., Yu, X., Greenleaf, W. B., Carazo, J. M., Tjajadi, M., Klein, M. G. and Chen, X. S. (2008). Crystal structure of a near-full-length archaeal MCM: functional insights for an AAA+ hexameric helicase. PNAS. 105(51): 20191-6. Brondijk, T. H., Nilavongse, A., Filenko, N., Richardson, D. J. and Cole, J. A. (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(Pt. 1): 47-55. Brzoska, P., Rimmele, M., Brzostek, K. and Boos, W. (1994). The pho regulon-dependent Ugp uptake system for glycerol-3-phosphate in Escherichia coli is trans inhibited by Pi. J Bacteriol. 176(1): 15-20. Bueno, E., Mesa, S., Bedmar, E. J., Richardson, D. J. and Delgado, M. J. (2012). Bacterial adaptation of respiration from oxic to microoxic and anoxic conditions: redox control. Antioxid Redox Signal. 16(8): 819-52. Butland, G., Babu, M., Diaz-Mejia, J. J., Bohdana, F., Phanse, S., Gold, B., Yang, W., Li, J., Gagarinova, A. G., Pogoutse, O., Mori, H., Wanner, B. L., Lo, H., Wasniewski, J.,

119

Christopoulos, C., Ali, M., Venn, P., Safavi-Naini, A., Sourour, N., Caron, S., Choi, J., Laigle, L., Nazarians-Armavil, A., Deshpande, A., Joe, S., Datsenko, K. A., Yamamoto, N., Andrews, B. J., Boone, C., Ding, H., Sheikh, B., Moreno-Hagelsieb, G., Greenblatt, J. and Emili, A. (2008). eSGA: E. coli synthetic genetic array analysis. Nat Methods. 5(9): 789-795. Butland, G., Peregrin-Alvarez, J. M., Li, J., Yang, W., Yang, X., Canadien, V., Starostine, A., Richards, D., Beattie, B., Krogan, N., Davey, M., Parkinson, J., Greenblatt, J. and Emili, A. (2005). Interaction network containing conserved and essential protein complexes in Escherichia coli. Nature. 433: 531-537. Calhoun, M. W. and Gennis, R. B. (1993). Demonstration of separate genetic loci encoding distinct membrane-bound respiratory NADH dehydrogenases in Escherichia coli. J Bacteriol. 175(10): 3013-9. Carlson, M. R., Zhang, B., Fang, Z., Mischel, P. S., Horvath, S. and Nelson, S. F. (2006). Gene connectivity, function and sequence conservation: predictions from modular yeast co- expression networks. BMC Genomics. 7: 40. Carmel-Harel, O. and Storz, G. (2000). Roles of the glutathione- and thioredoxin-dependent reduction systems in the Escherichia coli and Saccharomyces cerevisiae responses to oxidative stress. Annu Rev Microbiol. 54: 439-61. Cashel, M., Gentry, D. R., Hernandez, V. J. and Vinella, D. (1996). The Stringent Response. Escherichia coli and Salmonella : cellular and molecular biology. Curtiss, R. and Neidhardt, F. C. Washington, D.C., ASM Press: pp. 1458-1496. Cecchini, G., Schroder, I., Gunsalus, R. P. and Maklashina, E. (2002). Succinate dehydrogenase and fumarate reductase from Escherichia coli. Biochim Biophys Acta. 1553(1-2): 140-57. Chakravarty, A., Carlson, J. M., Khetani, R. S. and Gross, R. H. (2007). A novel ensemble learning method for de novo computational identification of DNA binding sites. BMC Bioinformatics. 8: 249. Chen, Z., Lewis, K. A., Shultzaberger, R. K., Lyakhov, I. G., Zheng, M., Doan, B., Storz, G. and Schneider, T. D. (2007). Discovery of Fur binding site clusters in Escherichia coli by information theory models. Nucleic Acids Res. 35(20): 6762-77. Chenault, S. S. and Earhart, C. F. (1991). Organization of genes encoding membrane proteins of the Escherichia coli ferrienterobactin permease. Mol Microbiol. 5(6): 1405-13. Chib, S. and Mahadevan, S. (2012). Involvement of the global regulator H-NS in the survival of Escherichia coli in stationary phase. J Bacteriol. 194(19): 5285-93. Cole, S. T., Eiglmeier, K., Ahmed, S., Honore, N., Elmes, L., Anderson, W. F. and Weiner, J. H. (1988). Nucleotide sequence and gene-polypeptide relationships of the glpABC operon encoding the anaerobic sn-glycerol-3-phosphate dehydrogenase of Escherichia coli K-12. J Bacteriol. 170(6): 2448-56. Costanzo, M., Baryshnikova, A., Bellay, J., Kim, Y., Spear, E. D., Sevier, C. S., Ding, H., Koh, J. L., Toufighi, K., Mostafavi, S., Prinz, J., St. Onge, R. P., van der Sluis, B., Makhnevych, T., Vizeacoumar, F. J., Alizadeh, S., Bahr, S., Brost, R. L., Chen, Y., Cokol, M., Deshpande, R., Li, Z., Lin, Z. Y., Liang, W., Marback, M., Paw, J., San Luis, B. J., Shuterigi, E., Tong, A. H., van Dyk, N., Wallace, I. M., Whitney, J. A., Weirauch,

120

M. T., Zhong, G., Zhu, H., Houry, W. A., Brudno, M., Ragibizadeh, S., Papp, B., Pal, C., Roth, F. P., Giaever, G., Nislow, C., Troyanskaya, O. G., Bussey, H., Bader, G. D., Gingras, A. C., Morris, Q. D., Kim, P. M., Kaiser, C. A., Myers, C. L., Andrews, B. J. and Boone, C. (2010). The genetic landscape of a cell. Science. 327(5964): 425-31. Crow, J. P. (1997). Dichlorodihydrofluorescein and dihydrorhodamine 123 are sensitive indicators of peroxynitrite in vitro: implications for intracellular measurement of reactive nitrogen and oxygen species. Nitric Oxide. 1(2): 145-57. Croxen, M. A. and Finlay, B. B. (2010). Molecular mechanisms of Escherichia coli pathogenicity. Nat Rev Microbiol. 8: 26-38. Datsenko, K. A. and Wanner, B. L. (2000). One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. PNAS. 97(12): 6640-6645. Davey, M. J., Fang, L., McInerney, P., Georgescu, R. E. and O'Donnell, M. (2003). Reconstitution of the Mcm2-7p heterohexamer, subunit arrangement, and ATP site architecture. J Biol Chem. 278(7): 4491-9. Davidson, A. L. and Chen, J. (2004). ATP-binding cassette transporters in bacteria. Annu Rev Biochem. 73(241-68). Davis, B. D. (1987). Mechanism of bactericidal action of aminoglycosides. Microbiol Rev. 51: 341-50. de Boer, A. P., van der Oost, J., Reijnders, W. N., Westerhoff, H. V., Stouthamer, A. H. and van Spanning, R. J. (1996). Mutational analysis of the nor gene cluster which encodes nitric- oxide reductase from Paracoccus denitrificans. Eur J Biochem. 242(3): 592-600. Dieppedale, J., Gesbert, G., Ramond, E., Chhuon, C., Dubail, I., Dupuis, M., Guerrera, I. C. and Charbit, A. (2013). Possible links between stress defense and the tricarboxylic acid cycle in Francisella pathogenesis. Mol Cell Proteomics. [Epub ahead of print]. Dieppedale, J., Sobral, D., Dupuis, M., Dubail, I., Klimentova, J., Stulik, J., Postic, G., Frapy, E., Meibom, K. L., Barel, M. and Charbit, A. (2011). Identification of a putative chaperone involved in stress resistance and virulence in Francisella tularensis. Infect Immun. 79(4): 1428-39. Dobbek, H., Gremer, L., Meyer, O. and huber, R. (1999). Crystal structure and mechanism of CO dehydrogenase, a molybdo iron-sulphur flavoprotein containing S-selanylcysteine. PNAS. 96(16): 8884-9. Dominguez, D. C. (2004). Calcium signalling in bacteria. Mol Microbiol. 54(2): 291-7. Edgar, R. C. (2004). MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res. 32(5): 1792-7. Efremov, R. G., Baradaran, R. and Sazanov, L. A. (2010). The architecture of respiratory complex I. Nature. 465(7297): 441-5. Efremov, R. G. and Sazanov, L. A. (2011a). Respiratory complex I: 'steam engine' of the cell? Curr Opin Struct Biol. 21(4): 532-40. Efremov, R. G. and Sazanov, L. A. (2011b). Structure of the membrane domain of respiratory complex I. Nature. 476(7361): 414-20.

121

Efremov, R. G. and Sazanov, L. A. (2012). The coupling mechanism of respiratory complex I - A structural and evolutionary perspective. Biochim Biophys Acta. 1817(10): 1785-95. El Bakkouri, M., Gutsche, I., Kanjee, U., Zhao, B., Yu, M., Goret, G., Schoehn, G., Burmeister, W. P. and Houry, W. A. (2010). Structure of RavA MoxR AAA+ protein reveals the design principles of a molecular cage modulating the inducible lysine decarboxylase activity. PNAS. 107: 22499-504. Engel, P., Kramer, R. and Unden, G. (1994). Transport of C4-dicarboxylates by anaerobically grown Escherichia coli. Energetics and mechanism of exchange, uptake and efflux. Eur J Biochem. 222(2): 605-14. Engvall, E. and Perlman, P. (1971). Enzyme-linked immunosorbent assay (ELISA). Quantitative assay of immunoglobulin G. Immunochemistry. 8(9): 871-4. Erhardt, H., Steimle, S., Muders, V., Pohl, T., Walter, J. and Friedrich, T. (2012). Disruption of individual nuo-genes leads to the formation of partially assembled NADH:ubiquinone oxidoreductase (complex I) in Escherichia coli. Biochim Biophys Acta. 2012(1817): 6. Euro, L., Bloch, D. A., Wikstrom, M., Verkhovsky, M. I. and Verkhovskaya, M. (2008). Electrostatic interactions between FeS clusters in NADH:ubiquinone oxidoreductase (Complex I) from Escherichia coli. Biochemistry. 47(10): 3185-93. Ezraty, B., Vergnes, A., Banzhaf, M., Duverger, Y., Huguenot, A., Brochado, A. R., Su, S., Espinosa, L., Loiseau, L., Py, B., Typas, A. and Barras, F. (2013). Fe-S Cluster Biosynthesis Controls Uptake of Aminoglycosides in a ROS-Less Death Pathway. Science. 340(6140): 1583-7. Faith, J. J., Driscoll, M. E., Fusaro, V. A., Cosgrove, E. J., Hayete, B., Juhn, F. S., Schneider, S. J. and Gardner, T. S. (2008). Many Microbe Microbe Microarrays Database: uniformly normalized Affymetrix compendia with structured experimental metadata. Nucleic Acids Res. 36: D866-870. Fodje, M. N., Hansson, A., Hansson, M., Olsen, J. G., Gough, S., Willows, R. D. and Al- Karadaghi, S. (2001). Interplay between an AAA module and an integrin I domain may regulate function of magnesium chelatase. J Mol Biol. 311(1): 111-22. Foster, P. L. (2007). Stress-induced mutagenesis in bacteria. Crit Rev Biochem Mol Biol. 42(5): 373-97. Fridovich, I. (1995). Superoxide radical and superoxide dismutases. Annu Rev Biochem. 64: 97- 112. Fritsche, E., Paschos, A., Beisel, H. G., Bock, A. and Huber, R. (1999). Crystal structure of the hydrogenase maturating endopeptidase HYBD from Escherichia coli. J Mol Biol. 288(5): 989-98. Fuhrmann, S., Ferner, M., Jeffke, T., Henne, A., Gottschalk, G. and Meyer, O. (2003). Complete nucleotide sequence of the circular megaplasmid pHCG3 of Oligotropha carboxidovorans: function in the chemolithoautotrophic utilization of CO, H2 and CO2. Gene. 322: 67-75. Gama-Castro, S., Salgado, H., Peralta-Gil, M., Santos-Zavaleta, A., Muniz-Rascado, L., Solano- Lira, H., Jimenez-Jacinto, V., Weiss, V., Garcia-Sotelo, J. S., Lopez-Fuentes, A., Porron- Sotelo, L., Alquicira-Hernandez, S., Medina-Rivera, A., Martinez-Flores, I., Alquicira-

122

Hernandez, K., Martinez-Adame, R., Bonavides-Martinez, C., Miranda-Rios, J., Huerta, A. M., Mendoza-Vargas, A., Collado-Torres, L., Taboada, B., Vega-Alvarado, L., Olvera, M., Olvera, L., Grande, R., Morett, E. and Collado-Vides, J. (2011). RegulonDB version 7.0: transcriptional regulation of Escherichia coli K-12 integrated within genetic sensory response units (Gensor Units). Nucleic Acids Res. 39(Database issue): D98- D105. Gautier, V., Le, H. T., Malki, A., Messaoudi, N., Caldas, T., Kthiri, F., Landoulsi, A. and Richarme, G. (2012). YajL, the prokaryotic homolog of the Parkinsonism-associated protein DJ-1, protects cells against protein sulfenylation. J Mol Biol. 421(4-5): 662-70. Gennis, R. B. and Stewart, V. (1996). Respiration. Escherichia coli and Salmonella typhimurium: Cellular and Molecular Biology. Neidhardt, F. C. a. o. Washington, DC, American Society for Microbiology. Ghai, R., Falconer, R. J. and Collins, B. M. (2012). Applications of isothermal titration calorimetry in pure and applied research - survey of the literature from 2010. J Mol Recognit. 25(1): 32-52. Girgis, H. S., Hottes, A. K. and Tavazoie, S. (2009). Genetic Architecture of Intrinsic Antibiotics Susceptibility. PLoS ONE. 4(5): e5629. Goltermann, L., Good, L. and Bentin, T. (2012). Chaperonins Fight Aminoglycoside-induced Protein Misfolding and Promote Short-term Tolerance in Escherichia coli. J Biol Chem. 288(15): 10483-9. Gomes, A., Fernandes, E. and Lima, J. (2005). Fluorescence probes used for detection of reactive oxygen species. J Biochem Biophys Methods. 65(2-3): 45-80. Goswami, M., Mangoli, S. H. and Jawali, N. (2007). Effects of Glutatione and Ascorbic Acid on Streptomycin Sensitivity of Escherichia coli. Antimicrob Agents Chenomther. 51(3): 1119-22. Grainger, D. C., Aiba, H., Hurd, D., Browning, D. F. and Busby, S. J. W. (2007). Transcription factor distribution in Escherichia coli: studies with FNR protein. Nucleic Acids Res. 35(1): 269-78. Green, J., Scott, C. and Guest, J. R. (2001). Functional versatility in the CRP-FNR superfamily of transcription factors: FNR and FLP. Adv Microb Physiol. 44: 1-34. Guenther, B., Onrust, R., Sali, A., O'Donnell, M. and Kuriyan, J. (1997). Crystal structure of the δ' subunit of the clamp-loader complex of E. coli DNA polymerase III. Cell. 91(3): 335- 45. Gutowski, S. J. and Rosenberg, H. (1975). Succinate uptake and related proton movements in Escherichia coli K12. Biochem J. 152(3): 647-54. Guyer, M. S., Reed, R. E., Steitz, T. and Low, K. B. (1981). Identification of a sex-factor-affinity site in E. coli as gamma delta. Cold Spr Harb Symp Quant Biol. 45: 135-40. Hammes, W., Schleifer, K. H. and Kandler, O. (1973). Mode of Action of Glycine on the Biosynthesis of Peptidoglycan. J Bacteriol. 116(2): 1029-53. Hanson, P. I. and Whiteheart, S. W. (2005). AAA+ Proteins: have engine, will work. Nat Rev. 6: 519-529.

123

Hayashi, M., Miyoshi, T., Takashina, S. and Unemoto, T. (1989). Purification of NADH- ferricyanide dehydrogenase and NADH-quinone reductases from Escherichia coli membranes and their roles in the respiratory chain. Biochim Biophys Acta. 977(1): 62-9. Henderson, L. M. and Chappell, J. B. (1993). Dihydrorhodamine 123: a fluorescent probe for superoxide generation? Eur J Biochem. 217(3): 973-80. Hume, D. A., Summers, K. M., Raza, S., Baillie, J. K. and Freeman, T. C. (2010). Functional clustering and lineage markers: insights into cellular differentiation and gene function from large-scale microarray studies of purified primary cell populations. Genomics. 95(6): 328-38. Imlay, J. A., Chin, S. M. and Linn, S. (1988). Toxic DNA damage by hydrogen peroxide through the Fenton reaction in vivo and in vitro. Science. 240: 640-2. Ippen-Ihler, K. A. and Minkley, E. G. J. (1986). The conjugation system of F, the fertility factor of Escherichia coli. Annu Rev Genet. 20: 593-624. Irizarry, R. A., Hobbs, B., Collin, F., Beazer-Barclay, Y. D., Antonellis, K. J., Scherf, U. and Speed, T. P. (2003). Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics. 4: 249-264. Ishihama, Y., Schmidt, T., Rappsilber, J., Mann, M., Hartl, F. U., Kerner, M. J. and Frishman, D. (2008). Protein abundance profiling of the Escherichia coli cytosol. BMC Genomics. 9: 102-18. Iverson, T. M., Luna-Chavez, C., Croal, L. R., Cecchini, G. and Rees, D. C. (2002). Crystallographic studies of the Escherichia coli quinol-fumarate reductase with inhibitors bound to the quinol-binding site. J Biol Chem. 277(18): 16124-30. Iyer, L. M., Leipe, D. D., Koonin, E. V. and Aravind, L. (2004). Evolutionary history and higher order classification of AAA+ ATPases. J Struct Biol. 146(1-2): 11-31. Jain, V., Kumar, M. and Chatterji, D. (2006). ppGpp: Stringent Response and Survival. J Microbiol. 44(1): 1-10. Jason-Moller, L., Murphy, M. and Bruno, J. (2006). Overview of Biacore systems and their applications. Curr Protoc Protein Sci. 19(19). Johnson, A. and O'Donnell, M. (2003). Ordered ATP hydrolysis in the gamma complex clamp loader AAA+ machine. J Biol Chem. 278(16): 14406-13. Jungst, A. and Zumft, W. G. (1992). Interdependence of respiratory NO reduction and nitrite reduction revealed by mutagenesis of nirQ, a novel gene in the denitrification gene cluster of Pseudomonas stutzeri. FEBS Lett. 314(3): 308-14. Kachlany, S. C., Planet, P. J., Bhattacharjee, M. K., Kollia, E., DeSalle, R., Fine, D. H. and Figurski, D. H. (2000). Nonspecific adherence by Actinobacillus actinomycetemcomitans requires genes widespread in bacteria and archaea. J Bacteriol. 182(21): 6169-76. Kang, Y., Weber, K. D., Qiu, Y., Kiley, P. J. and Blattner, F. R. (2005). Genome-wide expression analysis indicates that FNR of Escherichia coli K-12 regulates a large number of genes of unknown function. J Bacteriol. 187(3): 1135-60. Kanjee, U., Gutsche, I., Alexopoulos, E., Zhao, B., El Bakkouri, M., Thibault, G., Liu, K., Ramachandran, S., Snider, J. D., Pai, E. F. and Houry, W. A. (2011). Linkage between

124

the bacterial acid stress and stringent responses: the structure of the inducible lysine decarboxylase. EMBO J. 30(5): 931-44. Kaper, J. B., Nataro, J. P. and Mobley, H. L. (2004). Pathogenic Escherichia coli. Nat Rev Microbiol. 2(2): 123-40. Karata, K., Inagawa, T., Wilkinson, A. J., Tatsuta, T. and Ogura, T. (1999). Dissecting the role of a conserved motif (the second region of homology) in the AAA family of ATPases: site-directed mutagenesis of the ATP-dependent protease FtsH. J Biol Chem. 274(37): 26225-32. Katerov, V., Lindgren, P. E., Totolian, A. A. and Schalen, C. (2000). Streptococcal opacity factor: a family of bifunctional proteins with lipoproteins and fibronectin-binding activities. Curr Microbiol. 40(3): 149-56. Keren, I., Wu, Y., Inocencio, J., Mulcahy, L. R. and Lewis, K. (2013). Killing by bactericidal antibiotics does not depend on reactive oxygen species. Science. 339(6124): 1213-6. Keseler, I. M., Collado-Vides, J., Gama-Castro, S., Ingraham, J., Paley, S., Paulsen, I. T., Peralta-Gil, M. and Karp, P. D. (2005). EcoCyc: a comprehensive database resource for Escherichia coli. Nucleic Acids Res. 33(Database issue): D334-7. King, P. W. and Przybyla, A. E. (1999). Response of hya expression to external pH in Escherichia coli. J Bacteriol. 181(17): 5250-6. Kohanski, M. A., Dwyer, D. J., Hayete, B., A., L. C. and Collins, J. J. (2007). A Common Mechanism of Cellular Death Induced by Bactericidal Antibiotics. Cell. 130: 797-810. Kohanski, M. A., Dwyer, D. J., Wierzbowski, J., Cottarel, G. and Collins, J. J. (2008). Mistranslation of Membrane Proteins and Two-Component System Activation Trigger Antibiotic-Mediated Cell Death. Cell. 135: 679-90. Konto-Ghiorghi, Y., Mairey, E., Mallet, A., Dumenil, G., Caliot, E., Trieu-Cuot, P. and Dramsi, S. (2009). Dual role for pilus in adherence to epithelial cells and biofilm formation in Streptococcus agalactiae. PLoS Pathog. 5(5): e1000422. Kredich, N. M. (1992). The molecular basis for positive regulation of cys promoters in Salmonella typhimurium and Escherichia coli. Mol Microbiol. 6(19): 2747-53. Lavrrar, J. L., Christoffersen, C. A. and McIntosh, M. A. (2002). Fur-DNA interactions at the bidirectional fepDGC-entS promoter region in Escherichia coli. J Mol Biol. 322(5): 983- 95. Lazazzera, B. A., Bates, D. M. and Kiley, P. J. (1993). The activity of the Escherichia coli transciption factor FNR is regulated by a change in oligomeric state. Genes Dev. 7: 1993- 2005. Lee, J., Hiibel, S. R., Reardon, K. F. and Wood, T. K. (2010). Identification of stress-related proteins in Escherichia coli using the pullutant cis-dichloroethylene. J Appl Microbiol. 108(6): 2088-102. Lee, J. E. and Ahn, T. I. (2000). Periplasmic localization of a GroES homologue in Escherichia coli transformed with groESx cloned from Legionella-like endosymbionts in Amoeba proteus. Res Microbiol. 151: 605-18.

125

Leger, C., Heffron, K., Pershad, H. R., Maklashina, E., Luna-Chavez, C., Cecchini, G., Ackrell, B. A. C. and Armstrong, F. A. (2001). Enzyme electrokinetics: energetics of succinate oxidation by fumarate reductase and succinate dehydrogenase. Biochemistry. 40(37): 11234-45. Leif, H., Sled, V. D., Ohnishi, T., Weiss, H. and Friedrich, T. (1995). Isolation and characterization of the proton-translocating NADH:ubiquinone oxidoreductase from Escherichia coli. Eur J Biochem. 230(2): 538-48. Lemire, B. D. and Weiner, J. H. (1986). Fumarate reductase of Escherichia coli. Methods Enzymol. 126: 377-86. Li, Z. F., Li, B., Liu, Z. G., Wang, M., Gu, Z. B., Du, G. C., Wu, J. and Chen, J. (2009). Calcium leads to further increase in glycine-enhanced extracellular secretion of recombinant α- cyclodextrin glycosyltransferase in Escherichia coli. J Agric Food Chem. 57(14): 6231-7. Ling, J., Cho, C., Guo, L., Aemi, H. R., Rinehart, J. and Soll, D. (2012). Protein Aggregation Caused by Aminoglycoside Action is Prevented by a Hydrogen Peroxide Scavenger. Mol Cell. 48: 713-22. Liu, Y. and Imlay, J. A. (2013). Cell death from antibiotics without the involvement of reactive oxygen species. Science. 339(6124): 1210-3. Lo, T. C. (1977). The molecular mechanism of dicarboxylic acid transport in Escherichia coli K 12. J Supramol Struct. 7(3-4): 463-80. Lodish, H., Berk, A., Zipursky, L., Matsudaira, P., Baltimore, D. and Darnell, J. (2000). Molecular Cell Biology. New York, W. H. Freeman. Lundqvist, J., Elmlund, H., Wulff, R. P., Berglund, L., Elmlund, D., Emanuelsson, C., Hebert, H., Willows, R. D., Hansson, M., Lindahl, M. and Al-Karadaghi, S. (2010). ATP-induced conformational dynamics in the AAA+ motor unit of magnesium chelatase. Structure. 18(3): 354-65. Luo, F., Yang, Y., Zhong, J., Gao, H., L., K., Thompson, D. K. and Zhou, J. (2007). Constructing gene co-expression networks and predicting functions of unknown genes by random matrix theory. BMC Bioinformatics. 8: 299. Magrane, M. and Consortium, U. (2011). UniProt Knowledgebase: a hub of integrated protein data. Database. 2011(bar009). Mahoney, T. F. and Silhavy, T. J. (2013). The Cpx Stress Response Confers Resistance to Some, but Not All, Bactericidal Antibotics. J Bacteriol. 195(9): 1869-74. Malpica, R., Sandoval, G. R., Rodriguez, C., Franco, B. and Georgellis, D. (2006). Signaling by the arc two-component system provides a link between the redox state of the quinone pool and gene expression. Antioxid Redox Signal. 8(5-6): 781-95. Matsushita, K., Ohnishi, T. and Kaback, H. R. (1987). NADH-ubiquinone oxidoreductases of the Escherichia coli aerobic respiratory chain. Biochemistry. 26(24): 7732-7. McBroom, A. J. and Kuehn, M. J. (2007). Release of outer membrane vesicles by Gram-negative bacteria is a novel envelope stress response. Mol Microbiol. 63(2): 545-558. Mendoza-Vargas, A., Olvera, L., Olvera, M., Grande, R., Vega-Alvarado, L., Taboada, B., Jimenez-Jacinto, V., Salgado, H., Juarez, K., Contreras-Moreira, B., Huerta, A. M.,

126

Collado-Vides, J. and Morett, E. (2009). Genome-wide identification of transcription start sites, promoters and transcription factor binding sites in E. coli. PLoS One. 4(10): e7526. Meng, S. Y. and Bennett, G. N. (1992). Nucleotide sequence of the Escherichia coli cad operon: a system for neutralization of low extracellular pH. Journal of Bacteriology. 174(8): 2659-69. Menon, N. K., Robbins, J., Wendt, J. C., Shanmugam, K. T. and Przybyla, A. E. (1991). Mutational analysis and characterization of the Escherichia coli hya operon, which encodes [NiFe] hydrogenase 1. J Bacteriol. 173(15): 4851-61. Messenger, S. L. and Green, J. (2003). FNR-mediated regulation of hyp expression in Escherichia coli. FEMS Microbiol Lett. 228(1): 81-6. Meyer, O., Gremer, L., Ferner, R., Ferner, M., Dobbek, H., Gnida, M., Meyer-Klaucke, W. and Huber, R. (2000). The role of Se, Mo and Fe in the structure and function of carbon monoxide dehydrogenase. Biol Chem. 381((9-10)): 865-76. Miller, J. H. (1972). Experiments in Molecular Genetics, Cold Sping Harbor Laboratories. Miller, J. H. (1992). A short course in bacterial genetics: a laboratory manual and handbook for Escherichia coli and related bacteria. Plainview, NY, Cold Spring Harbor Laboratory Press. Milne, J. L., Borgnia, M. J., Bartesaqhi, A., Tran, E. E., Earl, L. A., Schauder, D. M., Lengyel, J., Pierson, J., A., P. and Subramaniam, S. (2013). Cryo-electron microscopy - a primer for the non-microscopist. FEBS J. 280(1): 28-45. Munch, R., Hiller, K., Grote, A., Scheer, M., Kelin, J., Schobert, M. and Jahn, D. (2005). Virtual Footprint and PRODORIC: an integrative framework for regulon prediction in prokaryotes. Bioinformatics. 21: 4187-9. Neuwald, A. F., Aravind, L., L., S. J. and Koonin, E. V. (1999). AAA+: A Class of Chaperone- Like ATPases Associated with the Assembly, Operation, and Disassembly of Protein Complexes. Genome Res. 9(1): 27-43. Newton, W. E. (2000). Nitrogen fixation in perspective. Nitrogen Fixation: From Molecules to Crop Productivity. Pedrosa, F. O., Hungria, M., Yates, G. and Newton, W. E., Springer: 3-8. Nystrom, T. (2004). Stationary-phase physiology. Annual Review of Microbiology. 58: 161-81. Ogura, T. and Wilkinson, A. J. (2001). AAA+ superfamily ATPases: common structure - diverse function. Genes Cells. 6(7): 575-97. Oldroyd, G. E. D. and Downie, J. A. (2008). Coordinating nodule morphogenesis with rhizobial infection in legumes. Ann Rev Plant Biol. 59: 519-46. Papworth, C., Bauer, J. C., Braman, J. and Wright, D. A. (1996). Site-directed mutagenesis in one day with >80% efficiency. Strategies. 9(3): 3-4. Park, S. J. and Gunsalus, R. P. (1995). Oxygen, iron, carbon and superoxide control of the fumarase fumA and fumC genes of Escherichia coli: role of the arcA, fnr, and soxR gene products. J Bacteriol. 177(21): 6255-62.

127

Park, Y. K., Bearson, B., Bang, S. H., Bang, I. S. and Foster, J. W. (1996). Internal pH crisis, lysine decarboxylase and the acid tolerance response of Salmonella typhimurium. Mol Microbiol. 20(3): 605-11. Pavlidis, P. and Noble, W. S. (2003). Matrix2png: A Utility for Visualizing Matrix Data. Bioinformatics. 19: 295-6. Pelzmann, A., Ferner, M., Gnida, M., Meyer-Klaucke, W., Maisel, T. and Meyer, O. (2009). The CoxD Protein of Oligotropha carboxidovorans Is a Predicted AAA+ ATPase Chaperone Involved in the Biogenesis of the CO Dehydrogenase [CuSMoO2] Cluster. J Biol Chem. 284: 9578-9586. Prangishvili, D., Vestergaard, G., Haring, M., Aramayo, R., Basta, T., Rachel, R. and Garrett, R. A. (2006). Structural and Genomic Properties of the Hyperthermophilic Archaeal Virus ATV with an Extracellular Stage of the Reproductive Cycle. J Mol Biol. 359(5): 1203-16. Price, C. E. and Driessen, A. J. M. (2010). Biogenesis of membrane bound respiratory complexes in Escherichia coli. Biochim Biophys Acta. 1803(6): 748-66. Puustinen, A., Finel, M., Haltia, T., Gennis, R. B. and Wikstrom, M. (1991). Properties of the two terminal oxidases of Escherichia coli. Biochemistry. 30(16): 3936-42. Rabin, R. S. and Stewart, V. (1993). Dual response regulators (NarL and NarP) interact with dual sensors (NarX and NarQ) to control nitrate- and nitrite-regulated gene expression in Escherichia coli K-12. J Bacteriol. 175(11): 3259-68. Radzi Noor, M. and Soulimane, T. (2012). Bioenergetics at extreme temperature: Thermus thermophilus ba(3)- and caa(3)-type cytochrome c oxidases. Biochim Biophys Acta. 1817(4): 638-49. Rappas, M., Schumacher, J., Niwa, H., Buck, M. and Zhang, X. (2006). Structural basis of the nucleotide driven confomrational changes in the AAA domain of transcription activator PspF. J Mol Biol. 357(2): 481-92. Richardson, D. J. (2000). Bacterial respiration: a flexible process for a changing environment. Microbiology. 146(Pt. 3): 551-71. Richardson, D. J., Berks, B. C., Russell, D. A., Spiro, S. and Taylor, C. J. (2001). Functional, biochemical and genetic diversity of prokaryotic nitrate reductases. Cell Mol Life Sci. 58(2): 165-78. Roger, M., Castelle, C., Guiral, M., Infossi, P., Lojou, E., Giudici-Orticoni, M. T. and Ilbert, M. (2012). Mineral respiration under extreme acidic conditions: from a supramolecular organization to a molecular adaptation in Acidithiobacillus ferrooxidans. Biochem Soc Trans. 40(6): 1324-9. Rombel, I., Peters-Wendisch, P., Mesecar, A., Thorgeirsson, T., Shin, Y. K. and Kustu, S. (1999). MgATP binding and hydrolysis determinants of NtrC, a bacterial enhancer- binding protein. J Bacteriol. 181(15): 4628-38. Rosenberg, S. M., Shee, C., Frisch, R. L. and Hastings, P. J. (2012). Stress-induced mutation via DNA breaks in Escherichia coli: A molecular mechanism with implications for evolution and medicine. Bioessays. 34(10): 885-92.

128

Rowbury, R. J. (1997). Regulatory components, including integration host factor, CysB and H- NS, that influence pH responses in Escherichia coli. Lett Appl Microbiol. 24(5): 319-28. Sabo, D. L., Boeker, E. A., Byers, B., Waron, H. and Fischer, E. H. (1974). Purification and physical properties of inducible Escherichia coli lysine decarboxylase. Biochemistry. 13(4): 662-70. Salmon, K., Hung, S. P., Mekjian, K., Baldi, P., Hatfield, G. W. and Gunsalus, R. P. (2003). Global gene expression profiling in Escherichia coli K12. The effects of oxygen availability and FNR. J Biol Chem. 278(32): 29837-55. Santic, M., Al-khodor, S. and Kwaik, Y. A. (2010). Cell biology and molecular ecology of Francisella tularensis. Cell Microbiol. 12(2): 129-39. Saraste, M., Sibbald, P. R. and Wittinghofer, A. (1990). The P-loop - a common motif in ATP- and GTP-binding proteins. Trends Biochem Sci. 15(11): 430-4. Sargent, F., Ballantine, S. P., Rugman, P. A., Palmer, T. and Boxer, D. H. (1998). Reassignment of the gene encoding the Escherichia coli hydrogenase 2 small subunit - identification of a soluble precursor of the small subunit in a hypB mutant. Eur J Biochem. 255(3): 746- 54. Scheele, U., Erdmann, S., Ungewickell, E. J., Felisberto-Rodrigues, C., Ortiz-Lombardia, M. and Garrett, R. A. (2011). Chaperone Role for Proteins p618 and p892 in the Extracellular Tail Development of Acidianus Two-tailed virus ATV. J Virol. Schneider, D., Pohl, T., Walter, J., Dorner, K., Kohlstadt, M., Berger, A., Spehr, V. and Friedrich, T. (2008). Assembly of the Escherichia coli NADH:ubiquinone oxidoreductase (complex I). Biochimica et Biophysica Acta. 1777(7-8): 735-9. Schryvers, A. and Weiner, J. H. (1981). The anaerobic sn-glycerol-3-phosphate dehydrogenase of Escherichia coli. Purification and characterization. J Biol Chem. 256(19): 9959-65. Schweizer, H. and Larson, T. J. (1987). Cloning and characterization of the aerobic sn-glycerol- 3-phosphate dehydrogenase structural gene glpD of Escherichia coli K-12. J Bacteriol. 169(2): 507-13. Senior, A. E., Nadanaciva, S. and Weber, J. (2002). The molecular mechanism of ATP synthesis by F1F0 synthase. Biochim Biophys Acta. 1553(3): 188-211. Shestopalov, A. I., Bogachev, A. V., Murtazina, R. A., Viryasov, M. B. and Skulachev, V. P. (1997). Aeration-dependent changes in composition of the quinone pool in Escherichia coli. Evidence of post-transcriptional regulation of the quinone biosynthesis. FEBS Lett. 404(2-3): 272-4. Snider, J. D., Gutsche, I., Lin, M., Baby, S., Cox, B., Butland, G., Greenblatt, J., Emili, A. and Houry, W. A. (2006). Formation of a Distinctive Complex between the Inducible Bacterial Lysine Decarboxylase and a Novel AAA+ ATPase. J Biol Chem. 281(3): 1532- 1546. Snider, J. D. and Houry, W. A. (2006). MoxR AAA+ ATPases: A novel family of molecular chaperones? J Struct Biol. 156(1): 200-209. Snider, J. D. and Houry, W. A. (2008). AAA+ proteins: diversity in function, similarity in structure. Biochem Soc Trans. 36(Pt 1): 72-7.

129

Snider, J. D., Thibault, G. and Houry, W. A. (2008). The AAA+ superfamily of functionally diverse proteins. Genome Biol. 9(4): 216. Soballe, B. and Poole, R. K. (1999). Microbial ubiquinones: multiple roles in respiration, gene regulation and oxidative stress management. Microbiology. 145(Pt. 8): 1817-30. Song, Y. and Buettner, G. R. (2010). Thermodynamic and kinetic considerations for the reaction of semiquinone radicals to form superoxide and hydrogen peroxide. Free Radic Biol Med. 49(6): 919-62. Soupene, E., He, L., Yan, D. and Kustu, S. (1998). Ammonia acquisition in enteric bacteria: physiological role of the ammonium/methylammonium transport B (AmtB) protein. PNAS. 95(12): 7030-4. Springer, T. A. (2006). Complement and the multifaceted functions of VWA and integrin I domains. Structure. 14(11): 11-6. Sternberg, N. L. and Maurer, R. (1991). Bacteriophage-mediated generalized transduction in Escherichia coli and Salmonella typhimurium. Methods Enzymol. 204: 18-43. Stevens, J. M., Mavridou, D. A., Hamer, R., Kritsiligkou, P., Goddard, A. D. and Ferguson, S. J. (2011). Cytochrome c biogenesis System I. FEBS J. 278(22): 4170-8. Stewart, V. (1993). Nitrate regulation of anaerobic respiratory gene expression in Escherichia coli. Mol Microbiol. 9(3): 425-34. Stewart, V. (1994). Dual interacting two-component regulatory systems mediate nitrate- and nitrite-regulated gene expression in Escherichia coli. Res Microbiol. 145(5-6): 450-4. Stewart, V., Lu, Y. and Darwin, A. J. (2002). Periplasmic nitrate reductase (NapABC enzyme) supports anaerobic respiration by Escherichia coli K-12. J Bacteriol. 184(5): 1314-23. Studholme, D. J. and Dixon, R. (2003). Domain architectures of sigma54-dependent transcriptional activators. J Bacteriol. 185: 1757-67. Takahashi, Y. and Nakamura, M. (1999). Functional Assignment of the ORF2-iscS-iscU-iscA- hscB-hscA-fdx-ORF3 Gene Cluster Involved in the Assembly of the Fe-S Clusters in Escherichia coli. J Biochem. 126: 917-26. Tanapongpipat, S., Reid, E., Cole, J. A. and Crooke, H. (1998). Transcriptional control and essential roles of the Escherichia coli ccm gene products in formate-dependent nitrite reduction and cytochrome c synthesis. Biochem J. 334(Pt. 2): 355-65. Tang, Y. P., Dallas, M. M. and Malamy, M. H. (1999). Characterization of the BatI (Bacteroides aerotolerance) operon in Bacteroides fragilis: isolation of a B. fragilis mutant with reduced aerotolerance and impaired growth in in vivo model systems. Mol Microbiol. 32(1): 139-49. Todorovic, S., Justino, M. C., Wellenreuther, G., Hildebrandt, P., Murgida, D. H., Meyer- Klaucke, W. and Saraiva, L. M. (2008). Iron-sulfur repair YtfE protein from Escherichia coli: structural characterization of the di-iron center. J Biol Inorg Chem. 13(5): 765-70. Tomoyasu, T., Ogura, T., Tatsuta, T. and Bukau, B. (1998). Levels of DnaK and DnaJ provide tight control of heat shock gene expression and protein repair of Escherichia coli. Mol Microbiol. 30(3): 567-81.

130

Toyama, H., Anthony, C. and Lidstrom, M. E. (1998). Construction of insertion and deletion mxa mutants of Methylobacterium extorquens AM1 by electroporation. FEMS Microbiol Lett. 166(1): 1-7. Tran, Q. H., Bongaerts, J., Vlad, D. and Unden, G. (1996). Requirement for the proton-pumping NADH dehydrogenase I of Escherichia coli in respiration of NADH to fumarate and its bioenergetic implications. Eur J Biochem. 244: 155-60. Tseng, C. P., Yu, C. C., Lin, H. H., Chang, C. Y. and Kuo, J. T. (2001). Oxygen- and growth rate-dependent regulation of Escherichia coli fumarase (FumA, FumB, and FumC) activity. J Bacteriol. 183(2): 461-7. Typas, A., Becker, G. and Hengge, R. (2007). The molecular basis of selective promoter activation by the sigmaS subunit of RNA polymerase. Mol Microbiol. 63(5): 1296-1306. Unden, G., Achebach, S., Holighaus, G., Tran, H. G., Wackwitz, B. and Zeuner, Y. (2002). Control of FNR function of Escherichia coli by O2 and reducing conditions. J Mol Microbiol Biotechnool. 4(3): 263-8. Unden, G. and Bongaerts, J. (1997). Alternative respiratory pathways of Escherichia coli: energetics and transcriptional regulation in response to electron acceptors. Biochim Biophys Acta. 1320(3): 217-34. Vaara, M. (1992). Agents that increase the permeability of the outer membrane. Microbiol Rev. 56(3): 395-411. van der Ploeg, J. R. and Eichhorn, E. (2001). Sulfonate-sulfur metabolism and its regulation in Escherichia coli. Arch Microbiol. 176: 1-8. van Spanning, R. J., Wansell, C. W., de Boer, T., Hazelaar, M. J., Anazawa, H., Harms, N., Otlmann, L. F. and Stouthamer, A. H. (1991). Isolation and characterization of the moxJ, moxG, moxI, and moxR genes of Paracoccus denitrificans: inactivation of moxJ, moxG, and moxR and the resultant effect on methylotrophic growth. J Bacteriol. 173(21): 6948- 61. van Weemen, B. K. and Schuurs, A. H. (1971). Immunoassay using antigen-enzyme conjugates. FEBS Letters. 15(3): 232-6. Vanderlinde, E. M., Magnus, S. A., Tambalo, D. D., Koval, S. F. and Yost, C. K. (2011). Mutation of a broadly conserved operon from Rhizobium leguminosarum biovar viciae (RL349-RL3502) causes defects to cell morphology and envelope intergrity. J Bacteriol. [Epub ahead of print]. Verkhovskaya, M. L., Belevich, N., Euro, L., Wikstrom, M. and Verkhovskaya, M. I. (2008). Real-time electron transfer in respiratory complex I. PNAS. 105(10): 3763-7. Vignais, P. M. and Colbeau, A. (2004). Molecular biology of microbial hydrogenases. Curr Issues Mol Biol. 6(2): 159-88. Volbeda, A., Amara, P., Darnault, C., Mouesca, J. M., Parkin, A., Roessler, M. M., Armstrong, F. A. and Fontecilla-Camps, J. C. (2012). X-ray cystallographic and computational studies of the O2-toleranct [NiFe]-hydrogenase 1 from Escherichia coli. PNAS. 109(14): 5305-10.

131

Walker, J. E., Saraste, M., Runswick, M. J. and Gay, N. J. (1982). Distantly related sequences in the α- and β-subunits of ATP synthase, myosin, kinases and other ATP-requiring enzymes and a common nucleotide binding fold. EMBO J. 1(8): 945-51. Wang, J., Song, J. J., Franklin, M. C., Kamtekar, S., Im, Y. J., Rho, S. H., Seong, I. S., Lee, C. S., Chung, C. H. and Eom, S. H. (2001). Crystal structures of the HslVU peptidase- ATPase complex reveal an ATP-dependent proteolysis mechanism. Structure. 9(2): 177- 84. Whittaker, C. A. and Hynes, R. O. (2002). Distribution and evolution of von Willebrand / integrin A domains: widely dispersed domains with roles in cell adhesion and elsewhere. Mol Biol Cell. 13(10): 3369-87. Willows, R. D. (2003). Biosynthesis of chlorophylls from protoporphyrin IX. Nat Prod Rep. 20(3): 327-41. Wissenbach, U., Kroger, A. and Unden, G. (1990). The specific functions of menaquinone and demethylmenaquinone in anaerobic respiration with fumarate, dimethylsulfoxide, trimethylamine N-oxide and nitrate by Escherichia coli. Arch Microbiol. 154(1): 60-6. Wissenbach, U., Ternes, D. and Unden, G. (1992). An Escherichia coli mutant containing only demethylmenaquinone, but no menaquinone: effects on fumarate, dimethylsulfoxide, trimethylamine N-oxide and nitrate respiration. Arch Microbiol. 158(1): 68-73. Wojtyra, U. A., Thibault, G., Tuite, A. and Houry, W. A. (2003). The N-terminal Zinc Binding Domain of ClpX is a Dimerization Domain that Modulates the Chaperone Function. J Biol Chem. 278(49): 48981-90. Wolf, T., Janzen, W., Blum, C. and Schnetz, K. (2006). Differential dependence of StpA on H- NS in autoregulation of stpA and in regulation of bgl. J Bacteriol. 188(19): 6728-38. Wong, K. and Houry, W. A. (2012). Novel structural and functional insights into the MoxR family of AAA+ ATPases. J Struct Biol. 179(2): 211-21. Xi, H., Schneider, B. L. and Reitzer, L. (2000). Purine catabolism in Escherichia coli and function of xanthine dehydrogenase in purine salvage. J Bacteriol. 182(19): 5332-41. Yamamoto, N., Nakahigashi, K., Nakamichi, T., Yoshino, M., Takai, Y., Touda, Y., Furubayashi, A., Kinjyo, S., Dose, H., Datsenko, K. A., Nakayashiki, T., Tomita, M., Wanner, B. L. and Mori, H. (2009). Update on the Keio collection of Escherichia coli single-gene deletion mutants. Mol Sys Biol. 5: 335. Yankovskaya, V., Horsefield, R., Tornroth, S., Luna-Chavez, C., Miyoshi, H., Leger, C., Byrne, B., Cecchini, G. and Iwata, S. (2003). Architecture of succinate dehydrogenase and reactive oxygen species generation. Science. 299(5607): 700-4. Yu, D., Ellis, H. M., Lee, E., Jenkins, N. A., Copeland, N. G. and Court, D. L. (2000). An efficient recombination system for chromosome engineering in Escherichia coli. PNAS. 97(11): 5978-83. Zeghouf, M., Li, J., Butland, G., Borkowska, A., Canadien, V., Richards, D., Beattie, B., Emili, A. and Greenblatt, J. (2004). Sequential Peptide Affinity (SPA) System for the Identification of Mammalian and Bacterial Protein Complexes. J Proteome Res. 3(3): 463-8.

132

Zhang, B., Crack, J. C., Subramanian, S., Green, J., Thomson, A. J., Le Brun, N. E. and Johnson, M. K. (2012). Reversible cycling between cysteine persulfide-ligated [2Fe-2S] and cysteine-ligated [4Fe-4S] clusters in the FNR regulatory protein. PNAS. 109(39): 15734- 9. Zhang, R. G., Skarina, T., Beasley, S., Khachatryan, A., Vyas, S., Clarke, S., Edwards, A., Joachimiak, A. and Savchenko, A. (2001). Structure of Thermotoga maritima stationary phase survival protein SurE: a novel acid phosphatase. Structure. 9(11): 1095-1106. Zheng, H., Wisedchaisri, G. and Gonen, T. (2013). Crystal structure of a nitrate/nitrite exchanger. Nature. 497(7451): 647-51.

133

Appendix A

Genes showing statistically significant changes in transcript levels caused by the deletion or overexpression of RavA/ViaA

This is a complete list of genes showing significant changes in transcript levels that were detected by the microarray analyses. All changes in expression are shown as fold-changes with respect to WT for ΔravA::cat, set 1, or WT + p11 for WT + pRV, set 2. Genes showing an increase in expression are listed separately from those showing a decrease. “++” represents a fold-increase that cannot be calculated, and “--” for a fold-decrease that cannot be calculated, due to the corresponding transcript being undetectable in WT or WT + p11. Genes are sorted by their b-numbers.

Set 1 (ΔravA::cat vs. WT) decreases

b Number Gene Name Description Fold Change b0014 dnaK Chaperone Hsp70; DNA biosynthesis; autoregulated heat shock -1.6 proteins b0032 carA Carbamoyl phosphate synthetase -1.7 b0033 carB Carbamoyl phosphate synthetase -1.5 b0162 sdaR SdaR transcriptional regulator -1.6 b0175 cdsA CDP-diglyceride synthetase -1.5 b0221 fadE Acyl-CoA dehydrogenase -1.5 b0473 htpG HtpG monomer -1.8 b0484 copA YbaR -2.2 b0548 ninE DLP12 prophage, conserved protein similar to phage 82 and -6.1 lambda proteins b0621 dcuC DcuC dicarboxylate transporter -1.5 b0674 asnB Asparagine synthetase B -3.2 b0898 ycaD YcaD MFS transporter -1.5 b0934 ssuC YcbE/YcbM ABC transporter -1.5 b1014 putA PutA bifunctional enzyme and transcriptional regulator -1.5 b1223 narK NarK nitrite MFS transporter -4.3 b1415 aldA Aldehyde dehydrogenase A -1.6 b1507 hipA HipA transcriptional activator -1.5 b1508 hipB HipB transcriptional activator -1.7 b1525 sad NAD-dependent succinate semialdehyde dehydrogenase -1.7 b1611 fumC Fumarase C monomer -1.7 b1612 fumA Fumarase A monomer -1.5 b1616 uidB UidB glucuronides GPH transporter -2.4 b1617 uidA β-glucuronidase -6.5 b2014 plaP Putrescine:H+ symporter -1.6 b2046 wzxC Uncharacterizaed polysaccharide transporter -1.6 b2049 cpsB Mannose-1-phosphate guanylyltransferase-(GDP) -3.6 b2143 cdd Cytidine deaminase -1.7 b2220 atoC AtoC-Phosphorylated transcriptional activator -1.6 b2387 fryB Predicted enzyme IIB component of PTS -1.7 b2579 yfiD Stress-induced alternate pyruvate formate-lyase subunit -1.8 b2592 clpB ClpB chaperone -1.9

134 b2750 cysC Adenylylsulfate kinase -1.8 b2752 cysD Sulfate adenylyltransferase -2.5 b2799 fucO Propanediol oxidoreductase monomer -1.9 b2800 fucA L-fuculose-phosphate aldolase -1.8 b2871 ygeX 2,3-diaminopropionate ammonia-lyase monomer -1.6 b2873 hyuA Phenylhydantoinase monomer -1.6 b2888 ygfU YgfU NCS2 transporter -1.7 b3005 exbD ExbD uptake of enterochelin; tonB-dependent uptake of B -1.5 colicins b3006 exbB ExbB protein; uptake of enterochelin; tonB-dependent uptake of -1.6 B colicins b3060 ttdR Dan-L-tartrate DNA-binding activator -1.9 b3092 uxaC D-glucuronate isomerase / D-galacturonate isomerase -1.8 b3112 tdcG L-serine deaminase 3 -1.6 b3113 tdcF Predicted L-PSP (mRNA) endoribonuclease -1.6 b3114 tdcE 2-ketobutyrate formate-lyase / pyruvate formate-lyase 4 -2.3 b3115 tdcD Propionate kinase / acetate kinase C -2.5 b3116 tdcC TdcC threonine STP transporter -2.2 b3117 tdcB Threonine dehydratase (catabolic) -1.8 b3118 tdcA TdcA transcriptional activator -2.1 b3212 gltB Glutamate synthase (NADPH) large chain precursor -1.7 b3213 gltD Glutamate synthase (NADPH) small chain -1.6 b3408 feoA Ferrous iron transport protein A -2.1 b3409 feoB FeoB ferrous iron transporter -3.2 b3410 feoC Putative transcriptional regulator -2.8 b3469 zntA Zinc, cobalt and lead efflux system -1.7 b3616 tdh Threonine dehydrogenase -1.6 b3617 kbl 2-amino-3-ketobutyrate CoA ligase -1.6 b3656 yicI α-xylosidase -1.5 b3691 dgoT YidT galactonate MFS transporter -4.6 b3707 tnaC tna operon leader peptide -1.9 b3709 tnaB TnaB tryptophan ArAAP transporter -1.7 b3744 asnA Aspartate-ammonia ligase -11.2 b3746 ravA MoxR AAA+ ATPase interacting with LdcI -- b3763 pssR HdfR transcriptional regulator -1.6 b3846 fadB Multifunctional enzyme involved in fatty acid oxidation -1.7 b3863 polA DNA polymerase I -1.5 b3870 glnA Adenylyl-glutamine synthetase -1.7 b3908 sodA Superoxide dismutase (Mn) -2.3 b3927 glpF GlpF - glycerol MIP channel -1.9 b3947 ptsA PEP-protein phosphotransferase system enzyme I -2.3 b4002 zraP Zinc homeostasis protein -3.1 b4021 pepE Peptidase E, a dipeptidase where amino-terminal residue is -1.6 aspartate b4032 malG Maltose/Maltodextrin Transport System -2.1 b4033 malF Maltose/Maltodextrin Transport System -2.1 b4034 malE Maltose/Maltodextrin Transport System -2.2 b4035 malK Maltose/Maltodextrin Transport System -2.2 b4077 gltP GltP glutamate/aspartate DAACS transporter -1.6 b4119 melA α-galactosidase monomer -1.6 b4209 ytfE Protein for repairing stress-damaged Fe-S clusters -1.5 b4213 cpdB 2',3'-cyclic nucleotide 2'-phosphodiesterase / 3'-nucleotidase -1.8 b4252 tabA Toxin-antitoxin biofilm protein -1.6 b4304 sgcC Putative PTS permease component -1.9 b4321 gntP GntP Gluconate Gnt transporter -1.7 b4323 uxuB Mannonate oxidoreductase -2.1 b4431 rprA Small RNA regulator of RpoS -3.4

135

b4458 oxyS OxyS RNA; oxidative stress regulator -2.1

b0013 yaaI Predicted protein -4.2 b0364 yaiS Conserved protein -1.8 b0629 ybeF Predicted DNA-binding transcriptional regulator, LYSR-type -1.5 b0801 ybiC Predicted dehydrogenase -2.0 b1321 ycjX Conserved protein -4.3 b1322 ycjF Putative membrane protein -1.6 b1571 ydfA Qin prophage; predicted protein -1.8 b1772 ydjH Predicted kinase -1.6 b1862 yebB Conserved hypothetical protein -4.7 b2172 yeiQ Putative oxidoreductase -1.6 b2681 ygaY Predicted transporter -1.6 b2848 yqeJ Predicted protein -1.6 b2870 ygeW Predicted carbamoyltransferase -1.7 b2875 yqeB Conserved protein with NAD(P)-binding Rossman fold -1.6 b2887 ygfT Fused predicted oxidoreductase, Fe-S subunit and nucleotide- -1.6 binding subunit b3074 ygjH Putative tRNA synthetase -1.6 b3602 yibL Conserved protein -1.6 b3881 yihT Putative aldolase -1.6 b4048 yjbM Conserved hypothetical protein -2.3 b4185 yjfM Conserved hypothetical protein -4.1 b4357 yjjM Predicted DNA-binding regulator -1.6 b4358 yjjN Predicted L-galactonate oxidoreductase -- b4379 yjjW Putative pyruvate formate lyase activating enzyme -1.8 b4504 ykfH Predicted protein -- b4541 YehK Predicted protein -1.9

Set 1 (ΔravA::cat vs. WT) increases

b Number Gene Name Description Fold Change

b0012 htgA Predicted DNA-binding transcriptional regulator ++ b0019 nhaA Sodium/proton NhaA transporter 1.7 b0060 polB DNA polymerase II 1.7 b0285 paoB Aldehyde dehydrogenase, FAD-binding domain 1.8 b0353 mhpT MhpT MFS transporter 1.5 b0399 phoB PhoB-Phosphorylated transcriptional dual regulator 1.6 b0416 nusB Transcription antitermination protein NusB 1.6 b0424 yajL Chaperone in response to oxidative stress 1.6 b0443 fadM Thioesterase III 2.1 b0581 ybdK γ-glutamyl:cysteine ligase YbdK 1.8 b0600 ybdL Methionine aminotransferase, PLP-dependent 1.9 b0692 potE Putrescine/proton symporter: putrescine/ornithine antiporter 3.3 b0708 phrB Deoxyribodipyrimidine photolyase (photoreactivation) 1.6 b0752 zitB Zn2+/Cd2+/Ni2+/Cu2+ efflux transporter 1.7 b0908 aroA 3-phosphoshikimate-1-carboxyvinyltransferase 1.6 b0966 yccV Heat shock protein, hemimethylated DNA-binding protein 1.7 b0972 hyaA Hydrogenase 1, small subunit 1.7 b0973 hyaB Hydrogenase 1, large subunit 1.8 b0974 hyaC Hydrogenase 1, b-type cytochrome subunit 2.7 b0975 hyaD Protein involved in processing of HyaA and HyaB proteins 1.9 b0977 hyaF Protein involved in nickel incorporation into hydrogenase 1 2.0 proteins b0980 appA 6-phytase / pH 2.5 acid phosphatase 1.7

136 b1089 rpmF 50S ribosomal subunit protein L32 1.9 b1102 fhuE Outer membrane porin for ferric-coprogen uptake 1.6 b1183 umuD SOS mutagenesis and repair 1.6 b1184 umuC SOS mutagenesis and repair 2.5 b1264 trpE Anthranilate synthase component I 3.9 b1265 trpL Trp operon leader peptide 7.7 b1285 yciR RNase II modulator 1.5 b1353 sieB Rac prophage; phage superinfection exclusion protein 2.0 b1416 gapC Interrupted glyceraldehyde-3-dehydrogenase 1.6 b1465 narV Nitrate reductase Z, g-subunit 1.8 b1536 ydeI Oxidative stress tolerance protein 1.5 b1562 hokD Polypeptide destructive to membrane potential 2.5 b1609 rstB Sensor kinase-phosphotransferase 2.5 b1660 ydhC YdhC drug MFS transporter 1.6 b1721 arpB Hypothetical protein 2.4 b1806 yeaY YeaY-outer membrane lipoprotein 1.8 b1916 sdiA SdiA transcriptional activator 1.5 b1973 yodA Cadmium-induced metal binding protein 1.8 b2086 yegS Lipid kinase 1.6 b2194 ccmH Cytochrome c biogenesis protein CcmH 1.8 b2195 ccmG Thioredoxin-like protein, cytochrome c biogenesis 1.9 b2196 ccmF Cytochrome c-type biogenesis protein 2.1 b2197 ccmE CcmABCDEFGH cytochrome c biogenesis system 1.7 b2198 ccmD CcmABCDEFGH cytochrome c biogenesis system 1.5 b2199 ccmC CcmABCDEFGH cytochrome c biogenesis system 2.1 b2200 ccmB CcmABCDEFGH cytochrome c biogenesis system 1.8 b2201 ccmA CcmABCDEFGH cytochrome c biogenesis system 1.8 b2202 napC Cytochrome c protein 1.6 b2203 napB Small subunit of periplasmic nitrate reductase, cytochrome c550 1.5 protein b2204 napH Ferredoxin-type protein 1.8 b2253 arnB UDP-L-Ara4O C-4" transaminase 1.8 b2254 arnC Undecaprenyl phosphate-L-Ara4FN transferase 1.6 b2255 arnA UDP-L-Ara4N formyltransferase / UDP-GlcA C-4"- 1.7 decarboxylase b2353 tfaS CPS-53 (KpLE1) prophage; tail fiber assembly protein fragment 1.8 b2365 dsdX D-serine transporter 1.5 b2370 evgS Sensor kinase phosphotransferase 1.6 b2393 nupC NupC nucleoside NUP transporter 2.4 b2507 guaA GMP synthase / GMP synthase (ammonia dependent) 1.7 b2508 guaB IMP dehydrogenase 1.7 b2526 hscA Chaperone, member of Hsp70 protein family 1.9 b2527 hscB Hsc20 co-chaperone that acts with Hsc66 in IscU iron-sulfur 1.7 cluster assembly b2530 iscS Cysteine desulfurase monomer 2.0 b2531 iscR IscR transcriptional regulator 1.9 b2838 lysA Diaminopimelate decarboxylase ++ b2881 xdhD Putative oxidoreductase; possible selenate reductase / role in 1.7 purine salvage b2937 speB Agmatinase 1.7 b2942 metK MetK S-adenosylmethionine synthetase monomer 2.2 b3093 exuT ExuT hexuronate MFS transporter 1.6 b3096 mzrA Regulator of EnvZ/OmpR osmoregulatory signalling 1.5 b3161 mtr Mtr tryptophan ArAAP transporter 4.6 b3324 gspC Putative protein secretion protein for export 1.8 b3393 hofO Protein involved in utilization of DNA as a carbon source 3.4 b3475 acpT Holo-[acyl carrier protein] synthase 2 1.5

137 b3513 mdtE MdtEF multidrug transporter 2.7 b3514 mdtF MdtEF multidrug transporter 2.1 b3517 gadA Glutamate decarboxylase A subunit 2.2 b3528 dctA C4-dicarboxylate / orotate:H+ symporter 1.6 b3714 purP Adenine:H+ symporter ++ b3717 cbrC Colicin E2 tolerance protein 3.5 b3729 glmS L-glutamine:D-fructose-6-phosphate aminotransferase 1.5 b3745 viaA VWA-containing protein associated with RavA 3.7 b3747 kup TrkD potassium KUP transporter 16.8 b3826 yigL Sugar phosphatase 1.7 b3869 ntrB NtrB sensory histidine kinase 1.7 b3981 secE Sec Protein Secretion Complex 1.5 b4309 nanS Probable 9-O-acetyl-N-acetylneuraminate esterase 3.2 b4410 ecnA Entericidin A, antidote to lipoprotein entericidin B 1.7 b4412 hokC HokC, Gef toxin; interferes with membrane function when in 1.5 excess b4416 rybA Small RNA 1.7 b4417 rybB Small RNA that interacts with Hfq 1.7 b4420 rdlA Antisense RNA, trans-acting regulator of ldrA 1.6 b4422 rdlB Antisense RNA, trans-acting regulator of ldrB 1.8 b4424 rdlC Antisense RNA, trans-acting regulator of ldrC 2.1 b4526 ydaE Rac prophage; zinc-binding protein 4.8 b0135 yadC Putative fimbrial-like protein 4.5 b0137 yadL Putative adhesin-like protein 1.5 b0289 yagV Conserved protein 1.8 b0453 ybaY Predicted outer membrane lipoprotein 1.5 b0546 ybcM DLP12 prophage; predicted DNA-binding transcriptional 2.0 regulator b0603 ybdO Predicted DNA-binding transcriptional regulator LYSR-type 2.8 b0637 ybeB Predicted protein 1.5 b0648 ybeU Predicted tRNA ligase 4.1 b0689 ybfP Predicted protein ++ b0773 ybhB Predicted kinase inhibitor 1.7 b0943 ycbV Predicted fimbrial-like adhesin protein ++ b1168 ycgG Conserved protein 1.6 b1242 ychE Predicted inner membrane protein 2.1 b1358 ydaT Rac prophage; hypothetical protein 1.6 b1359 ydaU Rac prophage; hypothetical protein 1.6 b1510 ydeK Predicted lipoprotein 1.6 b1541 ydfZ Conserved protein 2.4 b1685 ydiH Hypothetical protein 1.6 b1688 ydiK Hypothetical protein; transcription may be purine regulated 1.6 b1707 ydiV Conserved protein 1.6 b1856 yebA Predicted peptidase 1.6 b1932 yedL Predicted acyltransferase 1.7 b1956 yedQ Predicted diguanylate cyclase 1.6 b1964 yedS Putative outer membrane protein 1.7 b1965 yedS 2 Putative outer membrane protein 1.8 b2332 yfcO Conserved hypothetical protein 1.9 b2345 yfdF Hypothetical protein ++ b2602 yfiL Predicted lipoprotein 1.6 b2629 yfjM CP4-57 prophage; predicted protein 2.2 b3036 ygiA Predicted protein 1.7 b3220 yhcG Conserved hypothetical protein 1.8 b3508 yhiD Predicted Mg2+ transport ATPase 1.9 b3524 yhjG Predicted outer membrane biogenesis protein 1.9

138

b3527 yhjJ Predicted Zn-dependent peptidase 4.1 b4269 yjgB Predicted alcohol dehydrogenase; Zn- and NAD(P)-dependent 1.7 b4274 yjgW KpLE2 phage-like element 3.7 b4275 yjgX KpLE2 phage-like element; putative transmembrane protein 1.7 b0457 ylaB Conserved inner membrane protein 1.7 b1144 ymfJ e14 prophage; predicted protein 1.5 b1330 ynaI Conserved inner membrane protein 1.7 b1451 yncD Probable TonB-dependent receptor 1.6 b1588 ynfF Oxidoreductase subunit paralog of DmsA 1.5 b1589 ynfG Oxidoreductase, Fe-S subunit paralog of DmsB 1.5 b1808 yoaA Conserved protein 1.6 b1815 yoaD Predicted phosphodiesterase 1.8 b1843 yobB Conserved protein 1.7 b2211 yojI Fused predicted multidrug ABC transporter; ATP-binding 1.6 component b2689 yqaA Conserved inner membrane protein 1.5 b2941 yqgD Hypothetical protein 1.6 b3099 yqjE Conserved inner membrane protein 1.9

Set 2 (WT + pRV vs. WT + p11) decreases

b Number Gene Name Description Fold Change b0342 lacA Galactoside O-acetyltransferase monomer -5.5 b0344 lacZ β-galactosidase monomer -- b1490 dosC Diguanylate cyclase -2.1 b2414 cysK Cysteine synthase -2.3 b2421 cysM Cysteine synthase B -2.1 b2422 cysA Sulfate ABC transporter -4.3 b2423 cysW Sulfate ABC transporter -2.8 b2424 cysU Sulfate ABC transporter -5.1 b2425 cysP Thiosulfate ABC transporter -4.8 b2750 cysC Adenylylsulfate kinase -3.2 b2751 cysN Sulfate adenylyltransferase -3.7 b2752 cysD Sulfate adenylyltransferase -5.9 b2762 cysH 3'-phospho-adenylylsulfate reductase -4.4 b2763 cysI Sulfite reductase hemoprotein subunit -3.7 b2764 cysJ Sulfite reductase flavoprotein subunit -5.0 b2942 metK MetK S-adenosylmethionine synthetase monomer -1.6 b3408 feoA Ferrous iron transport protein A -2.1 b3409 feoB FeoB ferrous iron transporter -1.8 b3410 feoC Putative Fe-S-dependent transcriptional regulator -1.6

b1257 yciE Conserved protein associated with stress response -1.7 b1287 yciW Predicted oxidoreductase -3.7 b1729 ydjN Predicted transporter -2.5 b2012 yeeD Conserved hypothetical protein -4.0 b2013 yeeE Putative transport system permease protein -4.6

Set 2 (WT + pRV vs. WT + p11) increases

b Number Gene Name Description Fold Change b0314 betT BetT choline BCCT transporter 1.9

139 b1851 edd Phosphogluconate dehydratase 1.7 b2105 rcnR RcnR DNA-binding transcriptional regulator ++ b2155 cirA Outer membrane ferric siderophore, colicin receptor 4.9 b2557 purL Phosphoribosylformylglycinamide synthase 3.8 b2980 glcC GlcC transcriptional dual regulator ++ b3238 yhcN Hydrogen peroxide stress-induced protein 2.1 b3372 frlC Fructoselysine 3-epimerase 1.6 b3395 hofM Protein involved in utilization of DNA as a carbon source 4.3 b3743 asnC AsnC transcriptional repressor 1.9 b3744 asnA Aspartate-ammonia ligase 2.0 b3745 viaA VWA-containing protein associated with RavA 12.2 b3746 ravA MoxR AAA+ ATPase interacting with LdcI 6.4 b3823 rhtC RhtC threonine Rht efflux transporter 1.7 b0427 yajR YajR putative MFS transporter 2.6 b0960 yccS Hypothetical protein 1.5 b1315 ycjS NADH-dependent dehydrogenase 1.7 b1457 ydcD Hypothetical protein 1.6 b1671 ydhX Putative oxidoreductase, Fe-S subunit 2.1 b1774 ydjJ Predicted oxidoreductase, Zn-dependent and NAD(P)-binding 1.7 b1959 yedA Putative transmembrane subunit 1.5 b2973 yghJ Predicted inner membrane lipoprotein 1.6 b3108 yhaM Conserved protein of unknown function 2.6 b3110 yhaO YhaO putative STP transporter 3.4

140

Appendix B

Pearson correlation scores for ravA- and viaA-co-expressing genes and their functional annotations

The following list contains genes that meet the stringent cut-off (correlation score ≥ 0.5) and are considered to co-express with ravA and/or viaA. Their Pearson correlation scores corresponding to each of these genes are provided on the right. The b numbers and functional descriptions (if available) are also provided as shown. The genes are divided into three categories, depending on whether the co-express with both ravA and viaA (Category I), with ravA only (Category II), or with viaA only (Category III). A Venn diagram of the data is given in Fig. 3.1.

Category I: Genes co-expressing with both ravA and viaA Correlation with Correlation with Gene Name b Number Description ravA viaA ansB b2957 Asn metabolism 0.65051 0.60753 aspA b4139 Asp metabolism 0.54194 0.57059 dcuA b4138 C4-dicarboxylate transport 0.66535 0.54888 dcuB b4123 C4-dicarboxylate transport 0.56817 0.51744 frdA b4154 Anaerobic respiration; Fermentation 0.69038 0.63405 frdB b4153 Anaerobic respiration; Fermentation 0.64379 0.60204 frdC b4152 Anaerobic respiration; Fermentation 0.61917 0.5964 gpmM b3612 Glycolysis 0.59776 0.52958 hemC b3805 Hem biosynthesis 0.51487 0.50138 hemX b3803 Porphyrin biosynthesis 0.5281 0.54115 hybO b2997 Anaerobic respiration 0.62005 0.58672 hypA b2726 Protein modification; Anaerobic respiration 0.64622 0.5954 hypB b2727 Protein maturation 0.69607 0.62498 hypC b2728 Protein maturation; Anaerobic respiration 0.67458 0.63329 hypD b2729 Protein modification; Anaerobic respiration 0.65885 0.59676 menD b2264 Menaquinone biosynthesis 0.54297 0.50761 mtlD b3600 Carbohydrate catabolism 0.50802 0.56832 nikA b3476 Ni2+ transport 0.54464 0.55937 nirB b3365 Anaerobic respiration; nitrate assimilation 0.57724 0.51632 nirD b3366 Anaerobic respiration; nitrate assimilation 0.53498 0.51578 nrfA b4070 Anaerobic respiration 0.53363 0.51896 pepE b4021 Glycopeptide catabolism 0.65684 0.56658 pfkA b3916 Glycolysis 0.56294 0.51026 pldB b3825 Lipid biosynthesis 0.52289 0.59484 selA b3591 Selenocysteine incorporation 0.69996 0.62336 udp b3831 Nucleoside metabolism 0.50767 0.53799 yhhN b3468 0.54876 0.5435 yieE b3712 0.5822 0.53172 yieF b3713 Xenobiotic metabolism 0.55289 0.55248 yieP b3755 Transcription regulation 0.59788 0.57585 yjjI b4380 0.70975 0.69532 ysaA b3573 Electron transport chain 0.61717 0.56855

Category II: Genes co-expressing with ravA only

141

Gene Name b Number Description Correlation with ravA cpxR b3912 DNA-binding response regulator in two-component 0.56566 regulatory system with CpxA cydA b0733 cytochrome d terminal oxidase, subunit I 0.51738 dcuC b0621 anaerobic C4-dicarboxylate transport 0.52104 dmsA b0894 dimethyl sulfoxide reductase, anaerobic, subunit A 0.5548 elbB b3209 isoprenoid biosynthesis protein with amidotransferase-like 0.50812 domain epd b2927 D-erythrose 4-phosphate dehydrogenase 0.50287 focA b0904 formate transporter 0.52822 frdD b4151 fumarate reductase (anaerobic), membrane anchor subunit 0.50535 galT b0758 galactose-1-phosphate uridylyltransferase 0.50055 glgX b3431 glycogen debranching enzyme 0.52472 lysU b4129 lysine tRNA synthetase, inducible 0.53954 menF b2265 isochorismate synthase 2 0.52301 nadK b2615 NAD kinase 0.54078 nagA b0677 N-acetylglucosamine-6-phosphate deacetylase 0.55783 nrfB b4071 nitrite reductase, formate-dependent, penta-heme 0.51511 cytochrome c pck b3403 phosphoenolpyruvate carboxykinase 0.50262 pepP b2908 proline aminopeptidase P II 0.5116 pflB b0903 pyruvate formate lyase I (inactive) 0.52252 pgi b4025 glucosephosphate isomerase 0.53519 pgk b2926 phosphoglycerate kinase 0.53232 srlR b2707 GutR glucitol repressor 0.58033 tpiA b3919 triosephosphate isomerase 0.57139 viaA b3745 VWA-containing protein associated with RavA 0.84586 yfbB b2263 (1R,6R)-2-succinyl-6-hydroxy-2,4-cyclohexadiene-1- 0.50662 carboxylate synthase ygdH b2795 hypothetical protein 0.5543 yhbT b3157 predicted lipid carrier protein 0.52801 yhdH b3253 predicted oxidoreductase, Zn-dependent and NAD(P)- 0.54479 binding yihY b3886 predicted inner membrane protein 0.5313 yjdK b4128 hypothetical protein 0.56851 yqhD b3011 alcohol dehydrogenase, NAD(P)-dependent 0.55211

Category III: Genes co-expressing with viaA only

Gene Name b Number Description Correlation with viaA amiB b4169 N-acetylmuramoyl-l-alanine amidase II 0.52138 cpdA b3032 cyclic 3',5'-adenosine monophosphate phosphodiesterase 0.53408 envZ b3404 sensory histidine kinase in two-component regulatory system 0.53976 with OmpR glmU b3730 fused N-acetyl glucosamine-1-phosphate uridyltransferase 0.52324 and glucosamine-1-phosphate acetyl transferase gntR b3438 DNA-binding transcriptional repressor 0.5719 gntX b3413 gluconate periplasmic binding protein with 0.50584 phosphoribosyltransferase domain, GNT I system hemD b3804 uroporphyrinogen III synthase 0.50487 hypE b2730 carbamoyl phosphate phosphatase, hydrogenase 3 0.51627 maturation protein kdgK b3526 2-keto-3-deoxygluconokinase 0.54256 murI b3967 glutamate racemase 0.51627 nikC b3478 nickel transporter subunit 0.52011

142 nikD b3479 nickel transporter subunit 0.50367 pcm b2743 L-isoaspartate protein carboxylmethyltransferase type II 0.53129 ravA b3746 MoxR AAA+ ATPase interacting with LdcI 0.84586 tdcB b3117 catabolic threonine dehydratase, PLP-dependent 0.51152 ubiH b2907 2-octaprenyl-6-methoxyphenol hydroxylase, FAD/NAD(P)- 0.51134 binding yhbU b3158 predicted peptidase, collagenase-like 0.54514 yhiR b3499 Protein utilizing DNA as a carbon source 0.53345 yiaF b3554 hypothetical protein 0.52386 yicN b3663 hypothetical protein 0.62638 yidF b3674 predicted DNA-binding transcriptional regulator 0.5513 yiiM b3910 protein involved in base analog detoxification 0.53678 yjjW b4379 predicted pyruvate formate lyase activating enzyme 0.5497 zraS b4003 sensory histidine kinase in two-component regulatory system 0.54079 with ZraR

143

Appendix C

Experimental Contribution to Collaborative Work

This section contains a re-write of excerpts concerning RavA-ViaA from the following article, which includes myself as a co-first author. The article is published in PLoS Genetics:

Babu, M.*, Arnold, R.*, Bundalovic-Torma, C.*, Gagarinova, A.*, Wong, K. S.* et al. (2013) Quantitative Genome-Wide Genetic Interaction Screens Reveal the Global Epistatic Relationship of Soluble Protein Complexes in Escherichia coli. PLoS Genetics. 10(2): e1004120. [PMID: 24586182]

*These are co-first authors who contributed equally to this work.

Data Attribution: I performed the majority of the experiments shown in Fig. C.1, except those shown in panels B and F. For the eSGA data shown in panel A, I constructed the necessary Hfr C ΔravA, ΔviaA and ΔravAviaA donor strains for conjugation with the single-gene knockout recipient strains.

Background:

The E. coli Synthetic Genetic Arrays (eSGA) is a high-throughput genetic screening technology that is designed to investigate genetic interactions (GI’s) at the genomic level (Butland et al., 2008; Babu et al., 2011), and is analogous to the technology that is widely used in studying the GI’s in Saccharomyces cerevisiae (Butland et al., 2008). The eSGA involves the systematic construction of digenic mutants (i.e. double KO’s) by mating a donor strain that carries the null mutation with the chloramphenicol resistance marker (cat) for the query gene, against the KEIO collection (Baba et al., 2006; Yamamoto et al., 2009) of BW25113 recipient strains that carry defined null mutations (for non-essential genes; 3968 in total) or hypomorphic mutations (insertion of KO cassette in 3’UTR to alter transcript abundance or stability; 149 in total) with the kanamycin resistance marker (kanR) for the other genes in the E. coli genome (Butland et al., 2008; Babu et al., 2011). This is achieved via bacterial conjugation, as the donor strain is constructed in the Hfr Cavalli (Hfr C; Hfr stands for “High frequency of recombination”)

144

background, which has a genomically integrated F factor (fertility factor) that enables the cell to transfer its genomic DNA, starting with regions adjacent to the oriT origin of transfer (Ippen- Ihler & Minkley, 1986). The genomic DNA transferred from donor is incorporated into the recipient’s genome via DNA recombination, and the amount of DNA transferred varies depending on the duration of conjugation before being disrupted (Ippen-Ihler & Minkley, 1986). To facilitate the construction of digenic mutants, the conjugation of Hfr C donor and the KEIO collection of recipients is carried out by the automated pinning of an array of Hfr C donor colonies grown of LB-agar (96-well format) directly onto another array of recipient colonies (96- well format) in quadruple (384-well format) for the entire KEIO collection, followed by overnight incubation of the plates (Butland et al., 2008; Babu et al., 2011). Selection of digenic mutants is then carried out by automated pinning of the 384 colonies onto LB-agar supplemented with chloramphenicol and kanamycin in a second quadruple (1536-well format) (Butland et al., 2008; Babu et al., 2011). GI’s were determined by imaging the plates, quantifying the colony size of the digenic mutants, normalizing the data to account for experimental artifacts (e.g. plate- edge effects), and assignment of S scores as quantitative measure of the GI’s. A negative S score represents a putative aggravating GI (i.e. synthetic sick or lethal) that suggests essential functional compensation between the two genes; a positive S score represents a putative alleviating GI suggests possible genetic suppression by the two genes on other pathways (Butland et al., 2008; Babu et al., 2011).

Experimental Rationale:

To identify genetic interactors of ravA and/or viaA in E. coli, Hfr C donor strains harboring the null mutations ΔravA::cat, ΔviaA::cat and ΔravAviaA::cat were constructed (see Section , and eSGA was performed following the protocols as described in references (Butland et al., 2008) and (Babu et al., 2011). A second eSGA screen using customized mini arrays (Fig. 2.5) was performed subsequently to validate the GI’s detected in the first round. The confirmed GI’s were validated further by additional experiments that target specific biological processes.

Results:

The experimental results are summarized in Fig. C.1. First, both ravA and viaA were found to be synthetic lethal with genes from the primary Isc Fe-S clusters biosynthetic pathway (iscA, iscS,

145

iscU, hscA, hscB, and fdx) and the alternative Suf pathway (sufA, sufB, sufC, sufD and sufE) (Fig. C.1A; see also Fig. 2.5). A stringent cut-off of ⏐S score⏐≥ 3 (Butland et al., 2008; Babu et al., 2011) was applied to ensure confidence in the data. The synthetically lethal growth observed on

LB-agar was also reproducible in liquid LB culture, where growth (measured by OD600 at stationary phase) of the triple mutant ΔravAviaA ΔhscA showed a statistically significant growth

suppression compared to WT, ΔravAviaA, ΔhscA, and the statistically derived OD600 if the observed GI was absent (i.e. expected triple) (Fig. C.1B). Furthermore, tracking the growth of WT and ΔravAviaA transformed with the plasmid pRKISC that overexpresses proteins in the Isc pathway (Takahashi & Nakamura, 1999) or the control plasmid pRKNMC, reveals that an Isc overexpression suppresses the growth of WT E. coli, but not of ΔravAviaA, in the presence of kanamycin at a sub-lethal dosage of 4 μg/ml (Fig. C.1C, panels II and III). Details on the effects of RavA-ViaA on cell growth in the presence of sub-lethal concentrations of kanamycin are discussed in Sections 2.4.3 and 2.4.4. Importantly, in the absence of kanamycin, both WT and ΔravAviaA grow equally well despite the overexpression of Isc proteins (Fig. C.1C, panel I). As further proof of the GI’s between ravA, viaA and genes of the Isc pathway, immunoprecipitation experiments using DY330 strains expressing Isc proteins carrying C-terminal SPA-tags (Babu et al., 2009) showed that both RavA and ViaA physically interact with the majority of the Isc proteins tested (Fig. C.1F). This strongly suggests that RavA-ViaA may have an important role in the Fe-S clusters biosynthesis of E. coli.

Secondly, both ravA and viaA were also found to be synthetic lethal with genes involved in cysteine biosynthesis (cysB, cysE, cysI, cysJ, cysK and cysM) (Fig. C.1A; see also Fig. 2.5). Importantly, despite that cysteine is available in rich media (and thus the cysteine biosynthetic pathway presumably becomes non-essential), the growth of the null mutant ΔcysB, which has impaired expression of the cys regulon (Kredich, 1992), remains significantly slower than WT (Fig. C.1D). Evidently, the presence of cysteine in the media was insufficient to fully compensate for the loss of cysteine biosynthesis, which in turn justifies the aggravating GI’s observed. To validate the GI’s between ravA, viaA and the cys genes, WT transformed with the RavA-ViaA-overexpressing plasmid pRV (see Table 2.1 and Section 2.3.1) or the control plasmid p11 were grown in W-salts minimal media (100 mM potassium phosphate, pH 7.5, 2.1 mM MgSO4, 0.66 μM thiamine, 22 μM glucose, 19 μM NH4Cl) supplemented with different sources of inorganic or organic sulfur sources (Fig. C.1E). Interestingly, the overexpression of

146

RavA suppresses the growth of WT past mid-log phase only if inorganic sulfur sources (i.e. 2- 2- SO4 and S2O3 ), which depends on the (CysP/Sbp)CysUWA2 ABC transporter for their import, were used (Fig. C.1E, panels I and II). No growth defect was observed when organic sulfur sources such as taurine, 2-(4-pyridyl)-ethanesulfonate (PESF) or cystine (Cys-S-S-Cys), which utilize alternative ABC transporters for import, were used (Fig. C.1E, panels III, IV and V). Thus, RavA overexpression, which presumably disrupts the normal RavA-to-ViaA stoichiometry and the activity of RavA-ViaA in vivo, appears to affect only the cys genes, which are required for the assimilation of inorganic sulfur to generate cysteine (van der Ploeg & Eichhorn, 2001). This provides supporting evidence that RavA-ViaA is functionally linked to cysteine biosynthesis and the closely associated Fe-S clusters biosynthesis in E. coli.

147

Figure C.1. RavA and ViaA are linked to Fe-S assembly

(A) Sub-network of GIs of two unannotated genes with Fe-S cluster assembly and cysteine biosynthesis components.

148

(B) Differential growth of select single, double and triple mutants in rich medium (LB) at 32 °C over 24 h; expected fitness derived using multiplicative model, p-value calculated using Student’s t-test.

(C) Impact of ectopic over-expression of Isc Fe-S cluster assembly proteins (pRKISC expression plasmid vs. pRKNMC control vector) on growth of ravA-viaA double mutants vs. wild-type (WT) E. coli before (I) and after (II) oxidative stress (sub-lethal concentrations of kanamycin, Kan); OD600 readings at 11-hr time point (III) highlight differential responses. Tetracycline (Tet) included in media for plasmid maintenance. Asterisks represent significant (p ≤ 0.01; Student’s t-test) difference between WT+ pRKISC vs. WT+ pRKNMC.

(D) Slow growth of cysB deletion mutants on liquid LB medium at 32 ºC. Each data point shows the mean ± SD (error bars) of three independent biological measurements.

(E) Growth inhibition profiles of ectopic over-expression of RavA vs. WT on W-salt medium supplemented with sub-lethal concentration of inorganic (I and II) and organic (III-V) sources of sulfur.

(F) Co-immunoprecipitation analysis of endogenous ViaA (I) and RavA (II). Immunoblots show chromosomally tagged Isc assembly proteins, expressed at native levels, in input whole cell extract (WCE) and anti-FLAG immunoprecipitates (IP) as indicated. Untagged parental strain and an irrelevant bait protein (ATP-dependent iron hydroxamate transporter, FhuB), served as negative controls. Molecular masses (kDa) of marker proteins by SDS-PAGE are indicated.

149

Appendix D

Miscellaneous Experiments

This section describes additional experiments performed for the characterization of RavA and ViaA that are not included in the main body of this thesis. The results of each experiment are provided in summary.

D.1. Phenotypic Studies on Potential Effects of RavA on asparagine biosynthesis

Objective: Determine if RavA functions in parallel with AsnB in asparagine biosynthesis, as the expression of asnB decreased in ΔravA::cat, while the expression of both asnA and asnC increased in WT + pRV.

Method Description: MG1655 WT, ΔravA, ΔasnA and the digenic mutant ΔravA ΔasnA were grown in MOPS minimal media (pH 7.5) + 0.1% glucose / glycerol + 5 mM L-Asn / L-Gln. Growth was monitored using the BioScreen-C (Oy Growth Curves Ab Ltd.).

Results Summary: No difference observed in growth between the single and digenic mutants.

D.2. Phenotypic Studies on Potential Effects of RavA on xanthine dehydrogenase in purine salvage pathway

Objective: Determine if RavA is functionally linked to xanthine dehydrogenase and thus the purine salvage pathway, as xdhD expression increased in ΔravA::cat. ΔxdhD is sensitive to adenine toxicity due to possible defects in purine salvage (Xi et al., 2000).

Method Description: MG1655 WT and ΔravA were grown in W-salts minimal media (pH 7.5)

+/- 0.1% NH4Cl + 0.03%/0.06%/0.1%/0.15% adenine. Growth was monitored using the BioScreen-C (Oy Growth Curves Ab Ltd.).

150

Result Summary: No significant difference in growth observed between WT and ΔravA in the presence of adenine.

D.3. Phenotypic Studies on Potential Effects of RavA on H2O2 Tolerance

Objective: Determine if RavA is required for tolerance against oxidative stress by

hydrogen peroxide (H2O2).

Method Description: MG1655 WT and ΔravA were grown aerobically in W-salts minimal

media (pH 5 or 7.5) + 0.1% NH4Cl + 0.04% glucose to stationary phase,

and then exposed to 0/2.5/7.5/20 mM H2O2 for 30 minutes. W-salts media + 0.1% gelatin was added to neutralize the oxidative stress. Cells were diluted in serial dilution series and pinned onto W-salts-agar or LB- agar plates for recovery. Phenotypic trends were then determined via visual examination of the colony growth.

Result Summary: ΔravA shows marginally higher H2O2 tolerance than WT in several experiments, although the results were inconsistent in others.

D.4. Phenotypic Studies on Potential Effects of RavA on HOCl Tolerance

Objective: Determine if RavA is required for tolerance against oxidative stress by hypochlorous acid (HOCl).

Method Description: MG1655 WT and ΔravA were grown aerobically in LB or W-salts

minimal media (pH 7.5) + 0.1% NH4Cl + 0.04% glucose to stationary

phase. The cultures were diluted to OD600 ~2.0, and 0/50/100/200 μM NaOCl was added. Cell survival/growth was monitored using the BioScreen-C (Oy Growth Curves Ab Ltd.).

Result Summary: No difference in survival/growth observed between WT and ΔravA.

151

D.5. Phenotypic Studies on Effects of RavA-ViaA on E. coli Outer Membrane Vesiculation

Description: DH5α ΔviaA::Tn5 was reported to show an increase in outer membrane vesicle (OMV) formation compared to WT (McBroom & Kuehn, 2007). These results were validated by repeating the experiment on MG1655 WT, ΔravA, ΔviaA, as well as WT + p11/pR/pRV. DH5α ΔravA::cat and ΔviaA::cat were also constructed in attempt to reproduce the reported phenotype. The original DH5α ΔviaA::Tn5 was obtained from the authors of the paper and was tested as well.

Method Description: Cells were grown aerobically in LB to stationary phase. The cells were removed, and the media containing the secreted OMV’s was filtered with a 0.45-μm sterilized filter to remove residual cells. The OMV’s were then isolated by ultra-centrifugation and re-suspended in PBS (pH 7.5) + 0.2 M NaCl, followed by SDS-PAGE analysis of vesicular proteins. Vesiculation was quantified via densitometry on major OMV proteins, such as OmpA or OmpF (McBroom & Kuehn, 2007). All data was

quantified against OD600 of the respective cultures.

Result Summary: DH5α ΔviaA::Tn5 produced ~5x more OMV’s than WT DH5α. However, no difference was observed between WT and the mutants in MG1655 or in the DH5α strains prepared in-house. In contrast, WT + pR and WT + pRV both produced more OMV’s than WT + p11.

D.6. Construction of ΔravA::kanR and ΔravAviaA::kanR Mutants in Salmonella typhimurium CS401 and SL1344

Description: Construction of null mutants ΔravA::kanR and ΔravAviaA::kanR in Salmonella typhimurium CS401 and SL1344 using the λRed recombination system (Datsenko & Wanner, 2000).

152

Method Description: WT CS401 and SL1344 were transformed with pKD46 (λRed plasmid). After inducing λRed expression, KO cassettes generated by PCR using pKD4 plasmid as the template were introduced to pKD46 transformants by electroporation. Successful KO’s were selected on LB-agar + kanamycin. The pKD46 plasmid was later removed by heat shocking the KO mutants. All KO mutants were verified via colony PCR.

Result Summary: Both ΔravA::kanR and ΔravAviaA::kanR in CS401 and SL1344 were created successfully.

D.7. Phenotypic Studies on Effects of RavA-ViaA on Oxidative Stress by H2O2 in Salmonella typhimurium CS401

Description: The same experiments performed on E. coli MG1655, as described in Experiment D.3, were repeated on S. typhimurium CS401 WT, ΔravA::kanR and ΔravAviaA::kanR.

Method Description: See Experiment D.3.

Result Summary: No difference in H2O2 tolerance was observed between WT and the KO mutants.

D.8. Outer Membrane Vesiculation Studies on Salmonella typhimurium CS401

Description: The same experiments performed on E. coli MG1655 and DH5α, as described in Experiment D.5, were repeated on S. typhimurium CS401 WT, ΔravA::kanR and ΔravAviaA::kanR.

Method Description: See Experiment D.5.

Result Summary: No difference in OMV production was observed between WT and the KO mutants.

153

D.9. Construction of lacZ Reporter Plasmids for Monitoring the Effects of RavA-ViaA on Expression of the Fur Regulon

Objective: Construct lacZ reporter plasmids by cloning a Fur-regulated promoter region directly upstream of the lacZ gene in pETM60 (-T7p -ABD)-lacZ. Fur is the iron-sensitive transcriptional regulator that modulates various iron acquisition pathways in E. coli (Baichoo & Helmann, 2002; Chen et al., 2007). A lacZ reporter for the ravAviaA native promoter region was also created to provide reference for its activity.

Method Description: The native promoter regions upstream of selected genes of the Fur regulon were PCR-amplified and cloned into pETM60 (-T7p -ABD)-lacZ using NcoI and BamHI. The genes selected are as follows: fes, fepD, entC, sufA, ftnA, feoA, fhuF, narG, sodB and fieF. Standard cloning and sequence validation procedures were employed.

Result Summary: All lacZ reporter plasmids were constructed as planned, except for narG promoter, which has an internal NcoI cut site in the middle.

D.10. Effects of RavA-ViaA on the Activity of Fur-inducible Promoters

Objective: Determine if RavA-ViaA modulates the activity of Fur-inducible promoters via interaction with the regulator Fur.

Method Description: EDCM367 (MG1655 ΔlacZY) was first transformed with the Fur- inducible lacZ reporter plasmids discussed in Experiment D.9. Successful transformants were then transformed a second time with p11, pR, pRV or

pRK52QV. Cells carrying both plasmids were selected and grown aerobically in LB + kanamycin + ampicillin. They were harvested at different growth phases, and the expression of lacZ in these cells was determined by the Miller assay that measures β-galactosidase activity

154

(Miller, 1972; Miller, 1992). The promoter regions of the following genes were tested: fieF, fhuF, fepD, ftnA, entC and feoA.

Result Summary: In log phase cells, no effect was observed in the promoters’ activity from RavA-ViaA. In stationary phase cells, no effect was observed in the activity of promoters for fieF and feoA. For the promoters of fhuF and fepD, overexpression of RavA-ViaA from pRV induces a significant increase in activity. Importantly, this increase is dependent on RavA’s ATPase activity. For the entC promoter, a large increase was observed in

cells transformed with pRK52QV, while only marginal increases were seen for pR and pRV transformants.

D.11. Effects of Iron Starvation and Excess Iron on the Expression of RavA-ViaA

Objective: Determine if the expression of RavA-ViaA is influenced by the cell facing iron starvation or exposed to excess solubilized iron in the media.

Method Description: EDCM367 transformed with pETM60 (-T7p -ABD)-ravAp-lacZ reporter plasmid was grown in LB + 0/0.5/2/5 mM 2,2’-dipyridyl (DP) for iron

starvation, or in LB + 0/0.25/1/2.5 mM FeSO4 for excess iron exposure. Cells were then grown to early stationary phase and harvested for use in the Miller assay to determine activity of the ravAviaA promoter.

Result Summary: No change in RavA-ViaA expression was observed in either iron starvation or exposure to excess solubilized iron.

D.12. Effects of RavA and ViaA on the DNA-binding Activity of Fur

Objective: Determine potential effects of RavA-ViaA on the DNA-binding activity of the transcriptional regulator Fur by Electromobility Shift Assay (EMSA).

155

Method Description: RavA was purified as described in reference (Snider et al., 2006). To obtain purified ViaA, the fusion protein NusA-HV*-HV-ViaA was overexpressed from pETM60m-viaA and purified using Ni-NTA following standard procedures. The NusA tag was then cleaved by TEV protease. ViaA was then isolated using a HiTrap Heparin HP column (GE Health Sciences) with a linear 10 mM – 1 M KCl gradient on the AKTA FPLC. Fur was cloned, overexpressed and purified using the Profinity eXact Purification System (Bio-Rad). An inactive form of Fur was also prepared by treatment with 20 mM EDTA to remove the bound Zn2+, followed by dialysis to remove any excess EDTA. The promoter region for fepD was used as the DNA substrate, and was prepared by PCR. Details of the EMSA protocol were provided in Section 3.3.4 of this thesis. 1 mM ATP was used in the reaction mixture as required.

Result Summary: Neither RavA nor ViaA showed any effect on the binding of Fur-Zn2+ (active form) to the fepD promoter substrate. In addition, neither RavA nor ViaA showed any DNA-binding activity.