The Regulation of Elongation Factor P Post-Translational Modification in Maintenance of Gene Expression in Bacillus Subtilis

The Regulation of Elongation Factor P Post-Translational Modification in Maintenance of Gene Expression in Bacillus subtilis

Dissertation

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University

Anne Marie Witzky, BS

Graduate Program in Molecular Genetics

The Ohio State University

2019

Dissertation Committee:

Dr. Michael Ibba, Adviser

Dr. Anita Hopper

Dr. Guramrit Singh

Dr. Kurt Fredrick

Anne Marie Witzky

2019

Abstract

Elongation Factor P (EF-P) is a ubiquitous translation factor that facilitates translation of polyproline motifs. In order to perform this function, EF-P generally requires post-translational modification (PTM) on a conserved residue. Although the position of the modification is highly conserved, the structure can vary widely between organisms, ranging from R-β-lysine in Escherichia coli to cyclic L-rhamnose in Pseudomonas areuginosa. Loss of EF-P or its modification in either of these organisms results in a wide range in pleiotropic phenotypes. Recently, characterization of EF-P has been expanded to

Gram-positive organisms. In Bacillus subtilis, EF-P is modified at Lys32 with a 5- aminopentanol moiety, and the only major phenotype associated with loss of 5- aminopentanolylated EF-P is aberrant swarming motility. Due to this stark phenotypic contrast between B. subtilis and gamma-proteobacteria efp mutants, characterizing the functional and physiological role of EF-P in B. subtilis is of great importance for understanding the significance of EF-P in translation.

Although it has been determined that EF-P in B. subtilis is modified with 5- aminopentanol, the genes required for synthesis and ligation of 5-aminopentanol are unknown. Here, we began by screening B. subtilis mutants that display aberrant swarming motility for altered EF-P modification state. We determined that YmfI catalyzes the

i reduction of 5-aminopentanone to 5-aminopentanol. In the absence of YmfI, accumulation of 5-aminopentanonylated EF-P is inhibitory to swarming motility. Suppressor mutations that enhanced swarming in the absence of YmfI were found at two positions on EF-P, including one that changed the conserved modification site (Lys 32) and abolished PTM.

Thus, while modification of EF-P is thought to be essential for EF-P activity, here we show that in some cases it can be dispensable.

In light of these results, we hypothesized that deletion of genes upstream of YmfI in the modification pathway would similarly suppress the ΔymfI swarming phenotype. To identify additional genes required for PTM of EF-P, we screened for ΔymfI swarming suppressors. Tandem mass spectrometry analysis of the PTM mutant strains indicated that ynbB, gsaB, and ymfI are required for modification and that yaaO, yfkA, and ywlG influence the level of modification. Structural analyses also showed that EF-P can retain unique intermediate modifications, suggesting that 5-aminopentanol is likely directly assembled on EF-P through a novel modification pathway. Phenotypic characterization of these PTM mutants showed that each mutant does not strictly phenocopy the efp mutant, as has previously been observed in other organisms. Rather, each mutant displays phenotypes consistent with either the efp mutant or wild-type B. subtilis depending on the growth condition. In vivo polyproline reporter data indicates that the observed phenotypic differences result from variation in both the severity of polyproline translation defects as well as altered EF-P context dependence in each mutant.

Although loss of efp is known to induce ribosomal queuing on transcripts that code for polyproline motifs, it is unclear how this queuing impacts transcript stability. We

ii investigated the relationship between polyproline induced ribosomal queuing and transcript stability in B. subtilis. In the absence of efp, target transcripts displayed increased stability that is dependent on the polyproline motif, indicating that ribosomal queuing directly inhibits mRNA decay. In addition, 5’-3’ exonuclease RNaseJ1 was downregulated in the absence of EF-P, which further exacerbated this effect. As transcripts impacted by this mechanism are elevated under glucose exhaustion, we investigated the regulation of EF-P under this condition. When glucose is limited, cells negatively regulate

EF-P activity through modulation of 5-aminopentanol. This negative regulation correlates with enhanced stability of target transcripts under nutrient limitation. Taken together, these findings establish a novel PTM pathway for EF-P and demonstrate how regulation of this

PTM can be used to modulate gene expression.

iii

Dedication

This document is dedicated to my family, especially my husband Tyler for his constant

love, support, and sacrifice through my education.

Acknowledgements

Since beginning graduate school, I have been fortunate enough to be guided by a committee of talented scientists. Dr. Anita Hopper, Dr. Guramrit Singh, and Dr. Kurt

Fredrick have all had significant input on this work. I would have undoubtedly spent years floundering without their criticisms and suggestions that made this document what it is today.

I would also like to thank Dr. Michael Ibba. Mike has always given me the freedom to take this project in any direction, even if that direction led to a very expensive dead end.

The idea behind chapter 4 was initially conceived on the back of a receipt in a bar, and none of that work would have come to fruition if Mike had not always given me the freedom to explore. He has always been incredibly supportive of my science, career aspirations, and personal growth. I could not imagine completing a PhD with any other adviser.

In addition to these mentors, this work has also been greatly influenced by the work of all of our collaborators, who I commonly refer to as the “EF-People.” Dr. Andrei

Rajkovic laid the groundwork for this project, and was a wonderful mentor to me in my early years. Rodney Tollerson II always provided helpful discussions and was there to let me know that whenever I thought I had a good idea, I in fact did not. Dr. Daniel B. Kearns and Katherine R. Hummels at Indiana University taught me everything I know about

Bacillus subtilis. It is difficult to believe that five years ago I did not even know how to pronounce “Bacillus subtilis,” and today I have written a 130+ page document on it. None of this would have been possible without their guidance and expertise.

Last but not least, I would like to acknowledge my family and friends. Rebecca

Steiner, Paul Kelly, and Rodney Tollerson II have always been there to support me on a personal and professional level. My family has always been there to remind me that the world would in fact not end if I had to repeat another mass spectrometry experiment. I would finally like to acknowledge my husband, Tyler, who has spent countless Sunday evenings waiting in the car while I run in to start cultures or brought me dinner on late nights. His continued support and sacrifice over the past five years has been instrumental in my completion if this work.

Vita

2010…………………..Granville High School

2013…………………..B.S. Biology, Furman University

2014 to present……….Graduate Teaching and Research Associate, Department of

Molecular Genetics, The Ohio State University

Publications

Tollerson II R, Witzky A, Ibba M. (2018) Elongation factor P is required to maintain proteome homeostasis at high growth rate. PNAS. 115(43): 11072-11077.

Witzky A, Hummels KR, Tollerson II R, Rajkovic A, Jones LA, Kearns DB, Ibba M. (2018) EF-P post-translational modification has variable impact on polyproline translation in Bacillus subtilis. mBio. 9(2): e00306-18.

Hummels KR, Witzky A, Rajkovic A, Tollerson II R, Jones LA, Ibba M, Kearns DB. (2017) Carbonyl reduction by YmfI completes the modification of EF-P in B. subtilis to prevent accumulation of an inhibitory modification state. Mol Micro. 106(2): 236-251.

Tollerson II R, Witzky A, Ibba M. (2017) Elongation factor P interactions with the ribosome are independent of pausing. mBio. 8(4): e01056-17.

Rajkovic A, Hummels KR, Witzky A, Erickson S, Gafken PR, et al. (2016) Translation control of swarming proficiency in Bacillus subtilis by 5-amino-pentanolylated elongation factor P. J Biol Chem. 291(21): 10976-85.

Rajkovic A, Witzky A, Navarre W, Darwin AJ, Ibba M. (2015) Elongation factor-P at the crossroads of the host- endosymbiont interface. Microb Cell. 2(10): 360-2.

Rajkovic A, Erickson S, Witzky A, Branson OE, Seo J, Gafken PR, Frietas MA, Whitelegge JP, Faull KF, Navarre W, Darwin AJ, Ibba M. (2015) Cyclic Rhamnosylated

vii

Elongation Factor P Establishes Antibiotic Resistance in Pseudomonas aeruginosa. mBio. 6(3): e00823.

Fields of Study Major Field: Molecular Genetics

viii

Table of Contents

Abstract ...... i

Dedication ...... iv

Acknowledgments...... v

Vita ...... vii

Table of contents ...... ix

List of tables ...... xiv

List of figures ...... xvi

List of symbols and abbreviations ...... xix

Chapter 1 ...... 1

1. The functional and physiological significance of elongation factor P ...... 1

1.1 The functional role of EF-P ...... 1

1.1.1 Introduction ...... 1

1.1.2 The context dependence of polyproline pausing ...... 4

1.1.3 Uncoupling of transcription and translation ...... 5

1.1.4 eIF5A in translation elongation and termination ...... 6

1.1.5 Ribonucleolytic activity ...... 6

1.1.6 Growth rate dependence of EF-P activity ...... 7

1.2 Post-translational modification of EF-P...... 8

1.2.1 Hypusine and deoxyhypusine ...... 10

1.2.2 R-β-Lysine ...... 11

1.2.3 L-Rhamnose ...... 12

1.2.4 5-Aminopentanol ...... 13

1.2.5 The role of EF-P PTM ...... 14

Chapter 2 ...... 16

2. Carbonyl reduction by YmfI in Bacillus subtilis prevents accumulation of an inhibitory

EF-P modification state...... 16

2.1 Introduction ...... 16

2.2 Results ...... 18

2.2.1 The absence of YmfI impairs swarming motility ...... 18

2.2.2 YmfI is required for 5-aminopentanolylation of EF-P ...... 28

2.2.3 Modification independent alleles of EF-P bypass the need for

YmfI ...... 35

2.3 Discussion ...... 40

2.4 Materials and Methods ...... 45

2.4.1 Strains and growth conditions ...... 45

2.4.2 The absence of YmfI impairs swarming motility ...... 45

2.4.3 Strain construction ...... 45

2.4.4 Mass spectrometry ...... 56

2.4.5 Isotope distribution analysis ...... 57

2.4.6 Phylogenetic analyses ...... 57

2.4.7 In vitro reactions ...... 58

Chapter 3 ...... 60

3. EF-P post-translational modification has variable impact on polyproline translation in

Bacillus subtilis ...... 60

3.1 Introduction ...... 60

3.2 Results ...... 62

3.2.1 A forward genetic screen to identify genes required for modification of

EF-P ...... 62

3.2.2 Mutants that suppress the absence of YmfI display aberrant

5-aminopentanolylation ...... 70

3.2.3 Phenotypic characterization of modification mutants...... 87

3.2.4 Altered EF-P modification state has variable impact on

polyproline translation ...... 92

3.3 Discussion ...... 95

3.3.1 EF-P is modified through multistep assembly that is reminiscent of

FAB ...... 95

3.3.2 5-aminopentanolylation as an EF-P modification strategy in other

organisms ...... 96

3.3.3 EF-P dependent pausing is variably impacted by altered modification

state ...... 98

3.4 Materials and Methods ...... 99

3.4.1 Growth Conditions and strain construction ...... 99

3.4.2 YmfI suppressor screen...... 108

3.4.3 Swarming motility assay ...... 109

3.4.4 Isoelectric focusing ...... 110

3.4.5 Semi-native gel electrophoresis ...... 110

3.4.6 Mass spectrometry ...... 111

3.4.7 GFP reporter assay ...... 113

3.4.8 Phenotypic microarray ...... 113

3.4.9Antibiotic sensitivity assay ...... 114

3.4.10 Phylogenetic analysis ...... 114

Chapter 4 ...... 115

4. Regulation of EF-P post-translational modification increases transcript stability in

Bacillus subtilis ...... 115

4.1 Introduction ...... 115

4.2 Results ...... 117

4.2.1 RNaseJ1 degraded transcripts display increased stability in the absence

of EF-P ...... 117

4.2.2 Increased stability is the result of ribosomal queuing and downregulation

of RNaseJ1 ...... 123

4.2.3 Regulation of EF-P PTM correlates with increased transcript

stability ...... 125

4.3 Discussion ...... 132

4.3.1 EF-P relieved ribosomal queuing can increase stability of RNAseJ1

xii

targets ...... 132

4.3.2 Increase in transcript stability correlates with regulation of EF-P . 133

4.3 Materials and Methods ...... 137

4.4.1 Growth Conditions and strain construction ...... 137

4.4.2 qRT-PCR analysis ...... 139

4.4.3 Semi-native and SDS polyacrylamide gel electrophoresis ...... 139

4.4.4 Mass spectrometry ...... 140

4.4.5 Flow cytometry ...... 140

Chapter 5 ...... 144

5. Conclusions and outlook ...... 144

5.1 The diversity of EF-P PTMs ...... 144

5.2 Regulation of EF-P PTM ...... 145

5.3 Conclusions ...... 144

References ...... 149

xiii

List of Tables

Table 1. Summary of known EF-P PTMs ...... 9

Table 2. Mutations in efp that improve swarming in the absence of YmfI ...... 37

Table 3. Strains used in chapter 2 ...... 48

Table 4. Plasmids used in chapter 2 ...... 49

Table 5. Primers used in chapter 2 ...... 50

Table 6. Transposon insertion suppressors of ymfI ...... 66

Table 7. Summary of modifications identified in mutant strains...... 74

Table 8. Phenotype microarray conditions where WT displayed enhanced respiration ...... 89

Table 9. Phenotype microarray conditions where Δefp displayed enhanced respiration ...... 90

Table 10. Strains used in chapter 3 ...... 100

Table 11. Plasmids used in chapter 3 ...... 101

Table 12. Primers used in chapter 3 ...... 102

Table 13. Transcripts elevated in the absence of RNaseJ1 that encode a polyproline motif in the first 150 amino acids ...... 119

Table 14. Transcript half-lives in WT and Δefp ...... 122

xiv

Table 15. Transcript half-lives in WT and Δefp after mutation of the polyproline motif ...... 124

Table 16. Plasmids used in chapter 4 ...... 141

Table 17. Strains used in chapter 4 ...... 142

Table 18. Primers used in chapter 4 ...... 143

List of Figures

Figure 1. Elongation factor P stimulates translation of polyproline motifs ...... 3

Figure 2. EF-P resolves as two species on a semi-native gel in a YmfI-depedent manner...... 20

Figure 3. Cells mutated for ymfI are defective in swarming motility and swarming can be restored by mutations in efp ...... 21

Figure 4. YmfI is a paralog of FabG ...... 24

Figure 5. YmfI is not required for growth ...... 25

Figure 6. Genetic architecture and phylogenetic distribution of ymfI locus ...... 26

Figure 7. Genes adjacent to ymfI and efp are not required for swarming motility ...... 27

Figure 8. Deletion of ymfI results in 5-aminopentatonation of EF-P and can be suppressed by abolishing EF-P post-translational modification ...... 29

Figure 9. EF-P is 5-aminopentatonated in the absence of YmfI ...... 30

Figure 10. Isotope distribution indicates that EF-P is not 5-aminopentatonated in the absence of YmfI ...... 32

Figure 11. Overexpression of EF-P results in hyper-accumulation of the unmodified form and partially bypasses the need for YmfI ...... 34

Figure 12. Model of EF-P activation by YmfI ...... 43

xvi

Figure 13. Distribution of EF-P modification enzymes and conservation of residues 29 and 32 ...... 44

Figure 14. Quantitative swarm expansion assay ...... 65

Figure 15. Deletion of gsaB, ynbB, yaaO, yfkA, or ywlG suppresses the swarming defect of a ymfI mutant ...... 67

Figure 16. Isoelectric focusing gel of mutants ...... 68

Figure 17. Semi-native gel of mutants ...... 69

Figure 18. MS/MS spectra of EF-P peptide containing Lys32 ...... 75

Figure 19. Extracted ion chromatograms of unmodified and modified peptides containing

Lys32...... 84

Figure 20. Proposed modification pathway based on tandem mass spectrometry analysis ...... 85

Figure 21. Overexpression or knockdown of fatty acid biosynthesis factors does not influence EF-P modification state ...... 86

Figure 22. Antibiotic sensitivity in efp and PTM mutants ...... 91

Figure 23. PPX-GFP reporter in PTM mutants ...... 94

Figure 24. Phylogenetic tree predicting other bacteria that will employ 5- aminopentanolylation as a modification strategy ...... 97

Figure 25. Transcript stability in the absence of efp ...... 121

Figure 26. EF-P expression and PTM in M9 minimal media ...... 127

Figure 27. MS/MS spectra of EFP peptide containing Lys32 in exponential and stationary phase in M9 minimal media ...... 128

xvii

Figure 28. Flow cytometry analysis in exponential or stationary phase ...... 131

Figure 29. A model for EF-P dependent regulation of transcript stability ...... 136

xviii

List of symbols and abbreviations

AA amino acid

ACP acyl carrier protein

ATP adenosine triphsophate aIF5A archaeal initiation factor 5A

β beta

CaCl2 calcium chloride

CCA cytosine cytosine adenine

CID collision induced dissociation

Da daltons

ºC degrees celsius

Δ delta

DHS deoxyhypusine synthase

DNA deoxyribonucleic acid

DOHH deoxyhypusine hydroxylase

EarP elongation factor P maturation arginine rhamonsyltransferase

EDTA ethylenediaminetetraacetic acid

EF-P elongation factor P eIF5A eukaryotic initiation factor 5A

xix

ESI electrospray ionization

ETD electron transfer dissociation

EpmA elongation factor P modification protein A

EpmB elongation factor P modification protein B

EpmC elongation factor P modification protein C

FAB fatty acid biosynthesis

FT fourier-transform g gram

GFP green fluorescent protein

IPTG isopropyl β-D-1-thiogalactopyranoside

KCl potassium chloride

KH2PO4 monopotassium phosphate

L liter

LB luria broth

LC liquid chromatography

Lys lysine

M molar

MgSO4 magnesium sulfate

MLS lincomycin and erythromycin mRNA messenger ribonucleic acid

µg microgram

µl microliter

µM micromolar ml mililiter min minute

MS mass spectrometry mM milimolar

MS/MS tandem mass spectrometry

NADPH nicotinamide adenine dinucleotide phosphate

NaCl sodium chloride

Na2HPO4 disodium phosphate

NH4Cl ammonium chloride

NMR nuclear magnetic resonance

OD600 optical density at 600 nm

PBS phosphate buffered saline

PCR polymerase chain reaction

PTC peptidyl transfer center

PPE proline proline glutamate

PPP proline proline proline

PPW proline proline tryptophan

PPX polyproline motif with the Z position unspecified

Pro proline

PTM post-translational modification

RmlC dTDP-4-dehydrorhamnhose 3,5-epimerase

xxi rRNA ribosomal ribonucleic acid

RUT Rho utilization

SDS sodium dodecyl sulfate

SDS-PAGE sodium dodecyl sulfate polyacrylamide gel electrophoresis

SILAC stable isotope labeling by amino acids in cell culture

SUMO small ubiquitin-like modifier tRNA transfer ribonucleic acid

UTR untranslated region

V Volts

WT wild-type

ZPPX polyproline motif

3x-Pro triproline motif

xxii

Chapter 1

The functional and physiological significance of elongation factor P

1.1 The functional role of EF-P

1.1.1 Introduction

During translation, the ribosome employs aminoacyl-tRNAs as well as a number of translation factors in order to decode an mRNA and synthesize a polypeptide. The rate of translation can depend on a number of factors such as codon usage, mRNA structure, and amino acid (AA) structure (1, 2). For example, proline has a unique pyrrolidine ring structure that creates significant steric constraints, making proline both a poor peptidyl acceptor and donor (3). Due to this limitation, polyproline translation can be substantially slower than other AA motifs and often results in translational pausing (4, 5). In order to alleviate polyproline-induced translational pausing, a universally conserved translation factor, elongation factor P (EF-P, eIF5A in eukaryotes and aIf5A in archaea) binds the ribosome between the P and E sites and entropically stimulates peptide bond formation by stabilizing the P-site tRNA in a favorable conformation (Figure 1) (6-11). EF-P continuously probes translating ribosomes independent of pausing and then stimulates peptide bond formation upon binding a paused ribosome (12, 13). In bacteria, productive

1 binding is dependent on the identity of the P-site tRNA, with the D-arm of tRNAPro as the major determinant (14). In contrast, activity of the eukaryotic homolog, eIF5A, is independent of the P-site tRNA identity (15).

Figure 1. Elongation factor P stimulates translation of polyproline motifs. The ribosome (grey) pauses on a polyproline motif (red). Elongation factor P (blue) binds the ribosome and stimulates peptide bond formation. Figure adapted from (16).

1.1.2 The context dependence of polyproline pausing

The consequences of polyproline induced ribosomal pausing in the absence of EF-

P are highly context dependent (4, 5, 17-19). The strength of the pause is first dependent on the encoded polyproline motif itself (4, 5, 19). Within the context of EF-P regulation, a polyproline motif is defined as ZPPX, where Z and X can be any AA and the ribosome is paused with the second Proline in the P-site. The AAs in both the Z and X position will alter the EF-P dependence of the encoded motif, although in general, the AA in the X position has more of an impact on polyproline induced ribosomal pausing than the AA in the Z position (4, 5). For example, ribosomal profiling studies in wild-type and Δefp

Escherichia coli have shown that in general, a PPW motif results in the strongest pause in the absence of efp while a PPF motif often does not require EF-P activity for efficient translation (4, 5).

When a ribosome pauses on any ZPPX motif in the absence of EF-P, trailing ribosomes can build up behind the lead ribosome through ribosomal queuing (4, 5, 18).

Although queuing does not directly impact the strength of the pause, it does have a significant impact on net protein production and is highly context dependent. Polyproline induced ribosomal pausing will only have a significant impact on net protein production if it becomes the rate limiting step of translation. Transcripts encoding a strong Shine-

Dalgarno sequence in the 5’-UTR and AUG start codon have a high rate of translation initiation that can result in heavy ribosomal traffic and exacerbate ribosomal queuing. In this instance, the polyproline induced pause becomes the rate limiting step of translation, and net protein production is reduced in the absence of EF-P (5, 18). If a polyproline motif

4 is encoded near the 5’ end of a transcript, queued ribosomes can also inhibit translation initiation by blocking the start codon (5). This compounds the translational defects already observed from inefficient polyproline elongation in the absence of efp. Polyproline induced ribosomal pausing is heavily dependent on the nature of the encoded ZPPX motif, location of the motif, and translation initiation rate. These three factors explain why strictly bioinformatic approaches only have moderate success in identification of proteins that require EF-P for efficient synthesis (4, 5).

1.1.3 Uncoupling of transcription and translation

During bacterial gene expression, it is generally accepted that transcription and translation are coupled (20). This coupling is necessary to prevent the premature transcription termination induced by formation of intrinsic hairpin terminators or transcription termination factor Rho (20). In the absence of EF-P, polyproline induced ribosomal pausing can uncouple transcription and translation, allowing for premature transcription termination to occur (21). The role of EF-P in maintenance of coupling was first established in E. coli with fluorescence based reporter assays, where a polyproline motif and Rho utilization (RUT) site or intrinsic hairpin terminator were artificially integrated into the construct (21). Results with natural gene expression also confirmed a role for EF-P in coupling but were more modest (21). Bioinformatic analyses indicated that this model likely only holds true for a small portion of transcripts. However, given that

Rho utilization sites (RUT) are difficult to definitively identify, it is unclear if the role of

EF-P in maintenance of coupling is truly limited to a select set of operons or if it could have broader implications.

1.1.4 eIF5A in translation elongation and termination

Although the role for bacterial EF-P is restricted to facilitating translation of polyproline motifs, the eukaryotic homolog, eIF5A, has a much broader significance for eukaryotic translation. Ribosomal profiling in a Saccharomyces cerevisiae eIF5A conditional depletion strain revealed that while reduction of eIF5A does induce ribosomal pausing at polyproline motifs, it also induces widespread pausing at other motifs that lack proline (22, 23). This result corroborates another report that indicated that unlike EF-P, eIF5A does not specifically require tRNAPro for activity (15). In addition to this broad role in translation elongation, ribosomal profiling also revealed that knockdown of eIF5A results in an increase in ribosomal occupancy at stop codons, suggesting that eIF5A also has a role in efficient translation termination (22, 23). In vitro analyses confirmed that this increased occupancy is in fact the result of inefficient translation termination in the eIF5A knockdown (22). These studies indicate that eIF5A has a much broader role in translation elongation and termination than EF-P, which is restricted to facilitating polyproline translation.

1.1.5 Ribonucleolytic activity

In addition to these direct roles in translation, eIF5A/aIF5A has also been implicated in mRNA decay. The most direct evidence for eIF5A/aIF5A ribonucleolytic activity was found in Halobacterium sp. NRC-1 and Sulfolobus solfataricus, where purified aIF5A displayed ribonucleolytic activity in vitro and a small number of transcripts co-purified with the protein in vivo (24, 25). In eukaryotic organisms, the evidence for a role for eIF5A in mRNA decay is less well defined and perhaps suggests an indirect effect.

6 eIF5A has been shown to directly bind RNA in vitro, and depletion of eIF5A results in an increase in abundance of select nonsense mediated decay targets in vivo (26-29). However, it is possible that this in vivo result is the indirect result of inefficient translation termination in the absence of eIF5A. In any case, given that eIF5A/aIF5A have only been shown to interact with a small number of transcripts, it is likely that any ribonucleolytic activity maintained by either enzyme plays a relatively minor role in its function.

1.1.6 Growth rate dependence of EF-P activity

In initial investigations of the role of EF-P in vivo, analyses have generally been performed in optimal culture conditions. Under such conditions, EF-P activity is necessary to meet the high translational demands required to sustain rapid growth. As these conditions pose the greatest requirement for EF-P, this allowed for initial studies to fully reveal the functional and physiological role of EF-P in vivo. However, given that organisms in the environment are often growing more slowly than in artificial laboratory conditions, it is important to also consider the role of EF-P in alternative growth environments. A recent study revealed that when E. coli is grown under conditions that induce slow growth, such as low temperature or nutrient limitation, EF-P becomes dispensable for translation

(30). As growth is slowed, the rate of translation is also reduced in parallel. In this circumstance, EF-P relieved ribosomal pausing is no longer the rate limiting step in translation, and net protein production is not impacted by the presence or absence of EF-P

(30). Although EF-P still maintains the aforementioned functional roles under conditions of rapid growth, these results highlights the variability in the importance of EF-P in different settings.

1.2 Post-translational modification of EF-P

In order to stimulate translation of polyproline motifs, EF-P requires post-translational modification (PTM) at a highly conserved residue (31, 32). The structure of the modification can vary substantially between organisms (Table 1). Although a fully modified EF-P is universally required to maintain proper proteome homeostasis, there are many organism-specific features of EF-P and its diverse modifications that would potentially allow for more specialized translational control and regulation.

Modification Organisms Associated Phenotypes

Hypusine Eukaryotes/Archaea Lethal

Deoxyhypusine Archaea Lethal

R-β-Lysine Escherchia coli Loss of motility Salmonella enterica Hypersensitivity to antibiotics Shigella flexneri Hypersensitivity to detergents Agrobacterium tumefaciens Moderate growth defects Erwinia amylovora Osmoloarity Defects Loss of virulence

L-Rhamnose Shewanella oneidensis Loss of motility Pseudonmonas aeruginosa Hypersensitivity to antibiotics Neisseria meningitidis Severe growth defects Often lethal Decreased pathogenicity

5-Aminopentanol Bacillus subtilis Loss of swarming motility

Table 1. Summary of known EF-P PTMs, organisms in which they have been identified, and phenotypes associated with loss of modification.

1.2.1 Hypusine and Deoxyhypusine

As in the case of bacteria, eukaryotes and archaea also require an EF-P homolog

(eIF5A and aIF5A respectively) for efficient polyproline translation (10). Structurally, both of these proteins resemble the C-terminal portion of EF-P, and the position of the PTM on

EF-P, eIF5A and aIF5A is conserved (6, 33). In eukaryotes, deoxyhypusine synthase

(DHS) uses spermidine as a substrate for initial ligation of deoxyhypusine to eIF5A (34,

35). Deoxyhypusine hydroxylase (DOHH) then hydroxylates C2 on deoxyhypusine to form the final hypusine modification (36). Many archaea use a similar modification strategy, but some omit the final hydroxylation step (37, 38). Similar to bacteria that use

R-β-lysylation of EF-P, the hydroxylation step can be dispensable for some Archaea (37).

However, in eukaryotes, addition of the hydroxyl group is required to maintain modification stability by preventing the reversal of DHS activity (39). Cryo-EM analysis of eIF5A bound to the ribosome indicates that as in the case of R-β-lysine, the terminal amine on hypusine interacts with the CCA end of the P-site tRNA, positioning it in a conformation favorable for peptide bond formation (33). Although there is little structural data available for the functional role of deoxyhypusine, it likely performs a similar role to hypusine and R-β-lysine given their structural similarities.

In the majority of eukaryotes and Archaea, eIF5A (or aIF5A) and the required modification machinery are essential (32). This is likely due to the high polyproline content maintained in these organisms as well as the additional role that eIF5A has in translation termination. One notable exception to this essentiality is Saccharomyces cerevisiae, which does not require DOHH for survival (40). This is likely due to the fact that the

10 hydroxylation of deoxyhypusine aids in stability rather than function of the modification

(39). As the hydroxylation is dispensable in many other organisms, it is not surprising that some eukaryotic organisms can survive without this step. In higher eukaryotes, misregulation of eIF5A or DHS has been linked to diabetes and cancer (41, 42). Inhibitors of DHS activity have been identified as potential therapeutics for such ailments (43).

However, given the essentiality of hypusinated eIF5A, it is unclear how effective these will be in a clinical setting.

1.2.2 R-β-Lysine

In many gammaproteobacteria, EF-P is post-translationally modified at a conserved lysine residue with an R-β-lysine moiety (44-48). This modification strategy requires a three step process mediated by EpmA, EpmB, and EpmC (44, 46-50). First, EpmB, a 2,3- lysine aminomutase, uses alpha-lysine for the synthesis of R-β-lysine (44, 46-48). EpmA then directly ligates R-β-lysine onto a conserved lysine reside on EF-P (44, 46-48).

Following this ligation, EpmC hydroxylates either C4 or C5 of the now modified lysine residue (49, 50). R-β-lysine enhances EF-P activity through increasing the affinity of EF-

P for the ribosome and also stabilizing the CCA end of the P-site tRNA into a conformation that is favorable for peptide bond formation (8, 11). Although the terminal amine in R-β- lysine is critical for this stabilization, both in vitro and in vivo analyses have indicated that the additional hydroxyl group added by EpmC is dispensable for EF-P activity (51). It has been speculated the hydroxylation step instead aids in the overall stability of the modification. However, as a ΔepmC Escherichia coli mutant does not phenotypically resemble a ΔepmA mutant, it is unlikely that loss of the hydroxylation results in an overall

11 decrease of R-β-lysine modification on EF-P (51). The true role of EpmC mediated hydroxylation of EF-P remains elusive.

The requirement for R-β-lysylated EF-P has been well documented in a range of human and plant pathogens including Escherichia coli, Salmonella enterica, and Shigella

Flexneri, Agrobacterium tumefaciens, and Amylovora erwinia (31, 32, 52, 53). Within these organisms, EF-P and EpmA (and in some instances EpmB) have been shown to be required for a wide variety of pleiotropic phenotypes, many of which are highly relevant in a clinical setting. For example, Δefp and ΔepmA mutants display hypersensitivity to antibiotics and detergents, decreased motility, moderate growth defects and loss of virulence (45, 54-57). Although EF-P does require EpmA mediated modification to maintain full activity, ΔepmA mutants do not strictly phenocopy Δefp mutants, indicating that EF-P can retain low levels of activity in the absence of modification (46, 54).

1.2.3 L-Rhamnose

After the discovery of the R-β-lysine modification pathway, it was apparent that the vast majority of bacteria do not maintain EpmA, EpmB, or EpmC and that the conserved modification residue is often an arginine instead of a lysine. This suggested that alternative modification strategies could exist. Both Shewanella oneidensis and Pseudomonas aeruginosa were originally identified as organisms that lack the R-β-lysine modification machinery and maintain an arginine at the conserved modification residue (58, 59). Within both of these organisms, mass spectrometry analyses indicated that EF-P was modified with a cyclic rhamnose moiety. The rhamnose substrate is synthesized by RmlC and then ligated onto the conserved EF-P arginine residue by EarP, a glycosyltransferase that is

12 frequently found in the same genomic neighborhood as efp. Although the rhamnose modification is highly divergent from R-β-lysine, in vivo analyses indicate that it is equally important for EF-P activity (58, 59). Given that the terminal amine in R-β-lysine is critical for enhancing EF-P activity, it is unclear how a cyclic compound such as rhamnose could perform the same function. It is likely that rhamnose enhances EF-P activity through an alternative mechanism.

As in the case of organisms that employ R-β-lysine as a modification strategy, rhamnosylated EF-P is required for a wide variety of cellular processes. In S. oneidensis and P. aeruginosa, loss of EF-P or EarP is associated with aberrant motility, severe growth defects, hypersensitivity to antibiotics and decreased pathogenicity (58, 59). In Neisseria meningitidis, both EF-P and EarP are essential (60). Unlike bacteria that use R-β-lysylation of EF-P as a modification strategy, those that use rhamnosylation tend to have more severe phenotypic defects associated with the loss of a fully active EF-P. Although in general this does correlate with a higher polyproline content in these organisms, we cannot rule out the possibility that functional differences between R-β-lysine and rhamnose account for this discrepancy.

1.2.4 5-Aminopentanol

Even with the discovery of modification pathways for R-β-Lysine and rhamnose,

EF-P PTM state can still only be predicted in approximately 30% of all bacteria (31). This suggests that either the majority of prokaryotic organisms do not require EF-P PTM, or that we have merely seen the tip of the iceberg in the actual diversity of EF-P PTMs. We have recently expanded the search for EF-P PTMs to Gram-positive bacteria. In Bacillus

13 subtilis, EF-P is modified with a 5-aminopentanol moiety at Lys32 (61). Initial characterization of 5-aminopentanol indicates that it is required for efficient translation of polyproline motifs, though little is known of the precise functional role of 5-aminopentanol

(61). Given the structural similarities between hypusine, R-β-lysine and 5-aminopentanol, it is likely that 5-aminopentanol also stabilizes the P-site tRNA and increases the affinity of EF-P for the ribosome. The genes required for the synthesis and ligation of 5- aminopentanol are unknown. In the case of R-β-lysine and rhamnose, discovery of the genes required for modification was relatively straightforward, as the modification genes are found within the same genomic neighborhood as efp (57-59). This is not the case in B. subtilis, where efp is surrounded by genes known to be required for sporulation rather than modification.

Despite the structural similarities between 5-aminopentanol and other EF-P PTMs, the physiological importance of a fully modified EF-P is distinct from other characterized organisms. In B. subtilis, loss of EF-P or its modification does not result in the severe viability and growth defects observed in other bacteria. Instead, the major requirement for

EF-P in B. subtilis is for swarming motility (61, 62). This is in direct contrast to the pleiotropic phenotypes associated with loss of EF-P in other organisms and suggests that while EF-P generally has a housekeeping function, it plays a much more specialized role in B. subtilis.

1.2.5 The role of EF-P PTM

Although the functional role of EF-P PTM is well established, it is unclear why evolution has selected for divergent EF-P PTM pathways rather than an EF-P that functions

14 in the absence of modification. This question becomes even more perplexing considering that EF-P appears to be constitutively modified in all examples characterized to date. For many other proteins, organisms will use PTMs as a means of functional regulation. It is possible that EF-P also maintains its conserved PTMs as means to regulate activity. Here, we uncover the EF-P PTM pathway in B. subtilis and show how regulation of EF-P PTM can be used to fine tune translation and mRNA decay.

Chapter 2

Carbonyl reduction by YmfI in Bacillus subtilis prevents accumulation of an

inhibitory EF-P modification statea

2.1 Introduction

The protein Elongation Factor P (EF-P) is a translation elongation factor highly conserved in all living things (eIF5A in Eurkayotes and aIF5A in Archaea). Recent work has shown that EF-P enhances translation of particular codon combinations encoding prolines in both Bacteria and Eukaryotes (7, 8, 10). The mechanism by which EF-P potentiates translation is poorly understood but EF-P structurally resembles a tRNA molecule, has been shown to bind between the E and P sites of the ribosome, and may function catalytically by increasing entropy (6, 9, 33, 64). In nearly all systems studied to date, depletion of EF-P is either lethal or results in severe growth limitation, presumably due to the impaired ability to translate target sequences in essential genes (60, 65-67). The

a The work in this chapter was done in collaboration with Katherine R. Hummels, Rodney Tollerson II, Andrei Rajkovic, Lisa A. Jones, Daniel B. Kearns, and Michael Ibba. It was originally published in Molecular Microbiology [63, Hummels KR, Witzky A, Rajkovic A, Tollerson R, Jones LA, Ibba M, Kearns DB. 2017. Carbonyl reduction by YmfI in Bacillus subtilis prevents accumulation of an inhibitory EF-P modification state. Mol Microbiol.] Molecular Microbiology explicitly grants authors the rights to reuse full articles in a dissertation or thesis. Experiments performed by co-authors are credited in the corresponding figure legend of this chapter. All other experiments were performed by Anne M. Witzky. 16 one exception thus far is the Gram positive firmicute bacterium Bacillus subtilis where EF-

P is not required for growth and is instead specifically required for swarming motility, a flagellar-mediated multicellular behavior in which cells rapidly move across a semi-solid surface (61, 62)

In all organisms tested to date, EF-P is post-translationally modified on a conserved lysine/arginine residue at a position analogous to the amino acid acceptor site of tRNAs

(68). The chemical nature of the EF-P post-translational modification varies between species. Within the proteobacteria, Escherichia coli and Salmonella enterica modify EF-

P lysine 34 with a β-lysine group (57), whereas Neisseria meningitidis, Pseudomonas aeruginosa and Shewanella oneidensis modify EF-P arginine 32 with a rhamnose moiety

(58-60). The EF-P ortholog in Eukaryotes, eIF5A, is modified by the addition of hypusine

(69). In each case mentioned thus far the function of the modification is thought to be essential for EF-P activity because mutations in the modification system result in growth defects similar to efp deletion mutants and reduce the translation rate of consecutive prolines (8, 48, 60, 67, 70). The mechanism by which the various modification groups enhance EF-P activity is unknown.

B. subtilis modifies EF-P lysine 32 with yet another chemical variant, 5- aminopentanol. Substitution of the conserved modified lysine to alanine (EF-PK32A) was shown to impair surface motility but the enzymes required for 5-aminopentanolylation were unknown (61). Here, we revisit the swarming mutants from the same screen that first discovered EF-P as a swarming motility activator and we identify YmfI as a protein required for EF-P 5-aminopentanolylation. YmfI is paralogous to the fatty acid

17 biosynthesis gene FabG and catalyzes the reduction of 5-aminopentanone to 5- aminopentanol as, in the absence of YmfI, EF-P was modified with a 5-aminopentanone group and cells exhibit attenuated swarming motility. Finally, in a forward genetic screen, we identify mutations in efp that bypass the requirement for YmfI in swarming motility and in one of the mutants, EF-PK32R, post-translational modification was abolished. Thus,

YmfI is the first example of an enzyme involved in EF-P modification in B. subtilis, and the B. subtilis EF-P is the first example that retains activity in the absence of modification

2.2 Results

2.2.1 The absence of YmfI impairs swarming motility.

EF-P proteins are thought to be activated by post-translational modification and B. subtilis EF-P is post-translationally modified by a 5-aminopentanol group attached to lysine 32 (61). B. subtilis EF-P resolved as two bands in semi-native gel electrophoresis and Western blot analysis, and we hypothesized that the two bands could represent different states of EF-P modification and/or activity (Fig 2A left, lane T0). To determine whether the two species of EF-P were differentiated by a post-translational mechanism, translation was inhibited by the addition of spectinomycin and EF-P electrophoretic mobility in semi-native gel electrophoresis was monitored over time. Within 15 minutes of inhibiting translation, the lower band species of EF-P diminished in abundance (Fig 2A left, lanes T0-T240). We hypothesize that the two bands of EF-P represent different post- translational modification states.

To find proteins involved in EF-P post-translational modification, whole cell lysates of transposon mutants taken from the same screen that identified the swarming

18 motility defect of an efp mutant were resolved by semi-native gel electrophoresis and

Western blot analysis (Kearns and Losick, 2004). Mutation of one gene, ymfI, resulted in a single species of EF-P (Fig 2B). When resolved side-by-side, EF-P of the ymfI mutant exhibited an electrophoretic mobility similar to the lower band found in the wild type (Fig

2C, compare lanes 1 and 2). Moreover, the band intensity of ymfI EF-P did not change over time after spectinomycin treatment (Fig 2A middle). We conclude that the protein encoded by the ymfI gene, YmfI, alters EF-P mobility in semi-native gels and we hypothesize that YmfI might control EF-P post-translational modification.

Whereas mutation of efp results in a severe defect in swarming motility, mutation of ymfI resulted in periodic cessation and reinitiation of swarming motility to create a terraced colony appearance on swarm plates (Fig 3A,B) (62). The terracing pattern of swarming was reminiscent of an EF-P mutant defective in post-translation modification in which the modification site lysine 32 was mutated to an alanine (Fig 3A,C) (61). In both qualitative and quantitative analyses, the swarming defect of the efpK32A mutant was more severe than that of a ymfI mutant (Fig 3A,C). Furthermore, the swarming motility of a ymfI efpK32A double mutant resembled that of the efpK32A mutant alone, suggesting that the efpK32A mutation was epistatic (Fig 3A,C). As further evidence of epistasis, Western blots of EF-P resolved by semi-native gel electrophoresis indicated an altered mobility of the mutant EF-PK32A protein that was not further altered by the absence of YmfI (Fig 2C, compare lanes 3 and 4). We conclude that YmfI requires EF-P lysine residue 32 to promote swarming motility and alter EF-P electrophoretic mobility.

Figure 2. EF-P resolves as two species on a semi-native gel in a YmfI-dependent manner. (A) Translation was inhibited in mid-log phase cultures by the addition of spectinomycin, lysates were subsequently harvested at the indicated time points, resolved by semi-native (top panel, "SN") or denaturing (middle and bottom panels) polyacrylamide gel electrophoresis, electroblotted, and probed with anti-EF-P or anti-SigA polyclonal antisera (used as a loading control) as indicated. The following strains were used to generate the samples: wild type (DK1042), ymfI (DK3621), and efpK29N (DK4282). For denaturing gels, EF-P resolved at approximately 22 kDa and SigA resolved at approximately 46 kDa. The black arrow indicates the upper band and the white arrow indicates the lower band. (B) Lysates of mid-log phase cultures were resolved by semi- native (top panel, “SN”) or denaturing (middle and bottom panels) polyacrylamide gel electrophoresis, electroblotted, and probed with anti-EF-P or anti-SigA polyclonal antisera as indicated. The following strains were used to generate the samples: wild type (3610), cheC (DS1045), cheD (DS1064), comP (DS1028), efp (DS1124), flhG (DS1164), rrnB-16S (DS1146), srfAA (DS1102), srfAB (DS1044), srfAC (DS1122), swrA (DS1026), swrB (DS1107), swrC (DS1113), yabR (DS1078), and ymfI (DS1029) (C) Lysates of mid-log phase cultures were resolved by semi-native (top panel, “SN”) or denaturing (middle and bottom panels) polyacrylamide gel electrophoresis, electroblotted, and probed with anti-EF-P or anti-SigA polyclonal antisera as indicated. Lanes are numbered at the bottom of the panels for clarity in text. The presence and absence of a wild type copy of the ymfI gene is indicated by (+) and (-) respectively at the top of the panels. The following strains were used to generate samples: wild type (DK1042), ymfI (DK3621), efpK32A (DK3235), efpK32A ymfI (DK3712), efpK29N (DK4282), efpK29N ymfI (DK4396), efpK32R (DK4359), efpK32R ymfI (DK4397), efpK29N,K32R (DK4420), and efpK32A,K32R ymfI (DK4436). (D) EF-P-FLAG purified from a ymfI mutant (EF-P- FLAGymfI) was incubated alone, with YmfI, or with YmfI and 150 mM NADPH for 30 min at 37C. Reactions were subsequently resolved by semi-native gel electrophoresis, electroblotted, and probed with anti-EF-P antisera. (E) Lysates of mid-log phase cultures were resolved by semi-native (top panel, “SN”) or denaturing (middle and bottom panels) polyacrylamide gel electrophoresis, electroblotted, and probed with anti-EF-P or anti-SigA polyclonal antisera as indicated. The following strains were used to generate the samples: wild type (DK1042), efp (efp) (DK3780), ymfI (DK3621), ymfI efp (efp) (DK3789), efpK32A (DK3235), efp (efpK32A) (DK2248), efpK29N (DK4282), efp (efpK29N) (DK4043), efpK32R (DK4359), efp (efpK32R) (DK4072). Experiments in panels A-C and E were performed by Katherine R. Hummels. Experiment in panel D was performed by Anne M. Witzky. 20

Figure 3. Cells mutated for ymfI are defective in swarming motility and swarming can be restored by mutations in efp. (A) Top views of centrally-inoculated swarm plates incubated overnight at 37°C were imaged against a black background. Zones of colonization appear light grey. The plate inoculated with the ymfI mutant has internal rings that mark the locations at which the population stopped moving and restarted at a later time point. Note: a comparable cessation can be seen at hour 5 in the quantitative swarm expansion assay in panel 2B. The following strains were used to generate the panels: wild type (DK1042), efp (DK2050), ymfI (DK3621), efpK32A (DK3235), and ymfI efpK32A (DK3712).(B-K) Quantitative swarm expansion assays in which mid-log phase cultures were concentrated and used to inoculate swarm plates. Swarm expansion was monitored along the same axis every 30 min for 5-6 hours. Each data point represents the average of three replicates and error bars represent the standard deviation. The following strains were used as the inoculum (B) wild type (DK1042), efp (DK2050), and ymfI (DK3621). (C) efpK32A (DK3235) and efpK32A ymfI (DK3712). (D) ymfI (ymfIY150A) (DK4233), ymfI (ymfI) (DK3969), and ymfI (DK3621). (E) efp (DK2050) and efp (efp) (DK3780). (F) efp ymfI (DK2886), and efp ymfI (efp) (DK3789). (G) efp ymfI (efp) DK3789, efp ymfI (efpK29N) (DK4043), and , efp ymfI (efpK32R) (DK4072). (H) efp ymfI (efp) (DK3789) and efp ymfI (efpK32A) (DK2889). (I) efpK29N (DK4282), efpK29N ymfI (DK4396), and ymfI (DK3621). (J) efpK32R (DK4359), efpK32R ymfI (DK4397), and ymfI (DK3621). (K) efpK29N,K32R (DK4420), efpK29N,K32R ymfI (DK4436), and ymfI (DK3621). Experiments were performed by Katherine R. Hummels.

YmfI is homologous to FabG, a 3-keto-acyl ACP reductase essential for fatty acid biosynthesis in E. coli, and contains conserved active site residues (Fig 4A) (71). Whereas

FabG is essential for growth, cells mutated for ymfI exhibited only a 5% decrease in growth rate resembling the slight growth rate reduction found in cells mutated for efp (Fig 5). To determine whether the YmfI active site residues are required for promoting swarming motility and altering EF-P semi-native gel mobility, we first complemented the ymfI mutation in trans. The ymfI gene appeared to be the third gene in a putative four gene operon (ymfFHIJ) (Fig 6A), and swarming motility was restored to the ymfI mutant when ymfI was cloned downstream of the PymfF promoter and inserted at an ectopic site in the chromosome (amyE::PymfF-ymfI) (Fig 3D). Thus, the ymfI mutant defect was due to the absence of ymfI, not due to a polar effect on other genes in the operon (Fig 7A-E). The ectopic ymfI complementation construct did not restore swarming motility to the ymfI mutant, however, when the ymfI putative active site residue tyrosine 150 was mutated to

Y150A an alanine (amyE::PymfF-ymfI ) (Fig 3D, Fig 4A). We conclude that the active site residue of YmfI is required for either YmfI activity or stability and we infer that YmfI may activate EF-P enzymatically.

To determine if YmfI directly participates in the alteration of EF-P electrophoretic mobility, recombinant YmfI was purified from E. coli and EF-P-FLAG was purified from the B. subtilis ymfI mutant (EF-P-FLAGymfI). EF-P-FLAGymfI resolved as a single band on semi-native gel electrophoresis in the absence of YmfI or in the presence of YmfI but the absence of NADPH (Fig 2D, left and center lanes). Upon incubation of EF-P-FLAGymfI in the presence of purified YmfI and NADPH, however, a faint upper band resolved on semi-

22 native gel electrophoresis (Fig 2D, right lane). We conclude that YmfI is an enzyme that requires the NADPH cofactor and is sufficient in vitro to alter EF-P electrophoretic mobility on a semi-native gel.

Figure 4. YmfI is a paralog of FabG. (A) A multiple sequence alignment of YmfI from B. subtilis (YmfI), FabG from B. subtilis (FabG-Bsu) and FabG from E. coli (FabG-Eco). The location of B. subtilis YmfI conserved active site residue Tyr150 is indicated by a caret. (B) Cartoon of the chemical reactions catalyzed by YmfI and FabG. The cartoon indicates that the hydroxyl/carbonyl group is at the C3 position but it is possible that the hydroxyl group is at the C4 position instead. The precise position of the carbonyl can only be unequivocally determined with the use NMR or other high resolution structural methods. Figure was generated by Katherine R. Hummels.

Figure 5. YmfI is not required for growth. (A) Growth curve of cells grown in LB at 37°C. OD600 was monitored every 30 min for 7 hours. Each data point represents the average of three biological replicates and error bars represent standard deviation. (B) Exponential growth rate of cells grown at 37°C. The average of 3 biological replicates are reported. Error bars represent standard deviation. The following strains were used to inoculate the cultures: wild type (DK1042), efp (DK2050), and ymfI (DK3621). Experiment was performed by Katherine R. Hummels.

Figure 6. Genetic architecture and phylogenetic distribution of the ymfI locus. (A) Predicted genetic architecture of the ymfI locus and the design of the PymfF-ymfI complementation construct. Both ymfF and ymfH are predicted to encode peptidases and ymfJ encodes a DUF3235-domain containing protein. (B) The distribution of homologs of the EF-P modification enzymes deoxyhypusine synthase, DHS (dark blue), EpmA (light blue), EarP (yellow), and YmfI (red) across the three domains of life. YmfI is a member of the large family of alcohol dehydrogenases and the proximity to neighboring genes as indicated in Fig 6A was used to aid in the identification of YmfI homologs. Where multiple enzymes are encoded in the same genome, the bar is split accordingly. White space indicates the absence of a homolog to any EF-P modification enzyme. Numbers indicate the following clades (1) Flavobacterium-Cytophaga-Bacteroides group, (2) Chlamydiales, (3) Planctomycetes, (4) Spirochaetes, (5) Actinobacteria, (6) Deinococcus-Thermus group, and (7) Cyanobacteria. Bacillus subtilis is highlighted in pink. Figure was created by Katherine R. Hummels.

Figure 7. Genes adjacent to ymfI and efp are not required for swarming motility. (A-D) Quantitative swarming motility assays. Each point is the average of three replicates. The wild type (DK1042), ymfI (DK3621), and efp (DK2050) mutant data are reused for each panel as indicated. Strains used to make the panels are as follows: ymfF (DK3726), ymfH (DK3727), ymfFH (DK3728), ymfJ (DS326), yqhS (DK1962), and papA (DK3159). Note, that a ymfFH double mutant was generated and tested in panel (C) because ymfF and ymfH appeared to encode proteins of similar function and thus the two gene products could potentially have been redundant. (E) Lysates of mid-log phase cultures were resolved by semi-native (top panel, "SN”) or denaturing (middle and bottom panels) polyacrylamide gel electrophoresis electroblotted, and probed with anti-EF-P or anti-SigA polyclonal antisera (used as a loading control) as indicated. Note, EF-P from the papA mutant appears as a faint band in native gel for reasons unrelated to levels as the amount of EF-P from the papA mutant in denaturing gels is comparable to wild type. The following strains were used to generate the samples: wild type (DK1042), ymfI (DK3621), ymfF (DK3726), ymfH (DK3727), ymfFH (DK3728), ymfJ (DS326), yqhS (DK1962), and papA (DK3159). (F) Predicted genetic architecture of the efp locus and the design of the PyqhS-efp complementation construct. yqhS is predicted to encode a dehydroquinate dehydratase and papA has been shown to encode a proline aminopeptidase. Shaded arrows indicate open reading frames and size is proportional to gene length. Bent arrows indicate promoters. G-H) Quantitative swarming motility assays. Each point is the average of three replicates. Strains used to make the panels are as follows: wild type (DK1042), yqhS (DK1962), and papA (DK3159). Experiments were performed by Katherine R. Hummels.

2.2.2 YmfI is required for 5-aminopentanolylation of EF-P

One way in which YmfI might alter semi-native gel electrophoretic mobility in a manner dependent on its enzymatic activity is if YmfI is involved in 5- aminopentanolylation of EF-P. In order to assess the modification status of EF-P, EF-P was affinity tagged and ectopically overexpressed from an IPTG-inducible promoter in efp and efp ymfI mutant backgrounds and subjected to chymotrypsin in-gel digestion. The digested protein was then resolved on an orbitrap elite mass spectrometer, and subjected to electron transfer dissociation fragmentation to generate a complete MS2 spectrum of the peptide containing Lys32 (QHVKPGKGAAF). In the wild type, both unmodified (Fig 8A,

1138.624 Da) and 5-aminopentanolylated (Fig 8B, 1239.709 Da) EF-P peptides were detected. The additional mass of 101.085 Da is consistent with previous reports for 5- aminopentanol C5H12NO (61). In the absence of YmfI, unmodified peptides were detected but instead of a 101.085 Da mass, a mass of 99.068 Da was identified on Lys32 (Fig 9).

From the mass difference of 2 Da we determined that in the absence of YmfI, the post- translational modification on Lys32 is C5H10NO. Although C5H10NO could represent multiple different structures, based on previous MS3 data of 5-aminopentanol and the proposed reductase activity of ymfI, we suggest it likely corresponds to 5-aminopentanone

(Fig 8F). Further, while the precise position of the carbonyl can only be unequivocally determined with the use NMR or other high resolution structural methods, we suggest it likely corresponds to the position of the hydroxyl group in 5-aminopentanol at either the

C3 or C4 position (61).

Figure 8. Deletion of ymfI results in 5-aminopentanonation of EF-P and can be suppressed by abolishing EF-P post-translational modification. Extracted ion chromatograms of chymotrypsin-digested EF-P peptide from wild type (A-C), ymfI mutant (D-F), efpK29N mutant (G-I), and efpK32R mutant (J-L). Three different species were detected for the peptide corresponding to the wild type sequence QHVKPGKGAAF containing including unmodified (A,D,G,J), 5-aminopentanolylated (B,E,H,K), and 5-aminopentanonated (C,F,I,L) lysine residue 32. All chromatograms for ymfI represent the 2+ ion for the indicated peptide. All other chromatograms represent the 3+ ion for the indicated peptide. Point mutations are indicated in bold red lettering. Predicted chemical structures are indicated. The cartoon indicates that the hydroxyl/carbonyl group is at the C3 position but it is possible that the hydroxyl group is at the C4 position instead. The precise position of the carbonyl can only be unequivocally determined with the use NMR or other high resolution structural methods.

Figure 9. EF-P is 5-aminopentanonated in the absence of YmfI. MS2 spectrum generated from ETD fragmentation of the QHVKPGKGAAF peptide. Both (A) Unmodified (precursor m/z of 380.5492, 3+) and (B) 5-aminopentanonated (precursor m/z of 310.4310, 4+) peptide were detected in the absence of ymfI. z ions are indicated in blue, and c ions are indicated in green. Site of 5-aminopentanonation (Lys32) is marked in red. Parent ions were excluded from spectra when exported from Proteome Discoverer.

Very low levels of 5-aminopentanolylated EF-P also appeared to be detected in the absence of ymfI (Fig 8E). As the reduced and oxidized forms differ by two hydrogen atoms, however, it is possible that the parent ion predicted to be the 5-aminopentanolylated peptide is in reality a heavy isotope of the 5-aminopentanonated peptide. In order to address this possibility, a predicted isotope distribution was calculated for QHVKPGKGAAF with either modification, and compared to the measured isotope distribution (Fig 10). Actual measured abundances for each isotope of the 5-aminopentanonated peptide in the ymfI background corresponded with predicted abundances precisely, but not for the 5- aminopentanolylated peptide. Thus, the apparent 5-aminopentanolylated peptide in the ymfI mutant is likely an artifact of heavy isotope incorporation. Although the possibility that extremely low levels of 5-aminopentanol exist in the absence of ymfI cannot be ruled out, the differences between the predicted and actual isotope distribution for 5- aminopentanolylated EF-P makes this possibility remote. We conclude that YmfI is an enzyme that directly catalyzes the reduction of 5-aminopentanone to 5-aminopentanol. We further conclude, based on YmfI’s ability to alter EF-P electrophoretic mobility in vitro, that YmfI alone is capable of catalyzing the reduction reaction when the substrate is ligated to EF-P.

Figure 10. Isotope distribution indicates that EF-P is not 5-aminopentanolylated in the absence of YmfI. Actual (A,C) and predicted (B,D) isotope distributions for 5-aminopentanonated (A,B) and 5- aminopentanolylated (C,D) EF-P in the absence of ymfI. Distributions are shown for the 3+ precursor ions of the QHVKPGKGAAF peptide with a precursor m/z of 413.572 for the 5-aminopentanonated peptide and 414.244 for the 5-aminopentanolylated peptide.

During the mass spectrometry analyses, we noted that there appeared to be a relative excess of unmodified EF-P that did not seem consistent with the relative band intensity in the wild type (Compare Fig 2A left and Fig 8A,B). As the EF-P-FLAG construct was expressed from an IPTG-inducible promoter, we wondered whether 1 mM

IPTG induction was causing over-representation of the unmodified form. Thus, an IPTG- titration was conducted for both the swarming motility phenotype and the semi-native gel

Western blots. The Physpank-efp-FLAG construct restored swarming motility to the efp mutant even in the absence of IPTG, suggesting that leakiness of the promoter was sufficient to rescue the phenotype (Fig 11A, upper) though EF-P protein was not detectable by Western blot analysis until 0.01 mM IPTG was added (Fig 11A, lower). Moreover, the

EF-P upper band saturated in intensity while the lower band continued to accumulate (Fig

11A, lower). We conclude that the affinity-tagged version of EF-P is functional at very low levels in the efp mutant and that IPTG-induction resulted in hyper-accumulation of the unmodified form. In the efp ymfI double mutant, however, 0.03 mM IPTG induction of

EF-P-FLAG was required to restore partial swarming (Compare first four lanes of Fig

11A,B upper). Further, 0.05 mM IPTG and greater allowed the efp ymfI double mutant to swarm at a faster rate than the ymfI mutant alone (Fig 11B, upper). Regardless, in the absence of YmfI, only one EF-P band was detected in Western blot analysis at all IPTG concentrations tested (Fig 11B, lower). We conclude that the efp ymfI double mutant requires more EF-P for swarming motility and that EF-P overexpression could partially suppress the absence of YmfI.

Figure 11. Overexpression of EF-P results in hyper-accumulation of the unmodified form and partially bypasses the need for YmfI. A) An efp mutant and B) a efp ymfI double mutant each with an ectopically integrated Physpank-EF-P-FLAG construct was induced with the indicated concentrations of IPTG and measured for swarming motility radius after 4 hours of incubation (upper panel) and Western blot analysis following semi-native gel (SN) (top row) and denaturing (middle and bottom rows) electrophoresis of cell lystates (lower panel). Note, dashed line on 4 hr swarm expansion graph indicates the extent of expansion of either wild type (A) or a ymfI mutant (B). The following strains were used to generate samples: efp Physpank- efp-flag (DK2448) and efp ymfI Physpank-efp-flag (DK3828). Wild type (DK1042) and ymfI (DK3621) were used to generate the dashed lines. Experiments were performed by Katherine R. Hummels.

2.2.3 Modification independent alleles of EF-P bypass the need for YmfI

To explain the swarming defect of the ymfI mutant, we hypothesized that either the reduced 5-aminopentanol modification was essential for EF-P activity, or the oxidized 5- aminopentanone modification was inhibitory. To distinguish the two possibilities, we randomly mutagenized the efp gene and screened for mutations that bypass the need for

YmfI. If the 5-aminopentanol group was essential for activity, we predicted that there would be no mutation in efp sufficient to bypass the absence of YmfI. To set up the screen, we first complemented the efp mutation in trans. The efp gene appears to be the third gene in a putative three gene operon (yqhS-papA-efp) (Fig 7F), and swarming motility was restored to the efp mutant when efp was cloned downstream of the PyqhS promoter region and inserted at an ectopic site in the chromosome (amyE::PyqhS-efp) (Fig 3E). Thus the efp mutant defect was due to the absence of EF-P, not due to a polar effect on other genes in the region which, when mutated had no effect on swarming motility or on EF-P semi-native gel electrophoretic mobility (Fig 7E,G,H). The PyqhS-efp complementation construct, however, did not restore swarming to the efp ymfI double mutant (Fig 3F). The failure to complement the swarming defect of efp ymfI similar to that of ymfI alone appeared to be due a reduced level of ectopic EF-P expression relative to that of the native locus (Fig 2E), consistent with the observation that ymfI rescue depends on EF-P levels (Fig 11B). We conclude that transcription of efp from PyqhS is sufficient under most conditions but additional promoters may be needed to amplify expression when ymfI is mutated. We note that the upstream PyqhR may contribute to EF-P expression as transcriptomics indicates that there is no transcriptional terminator between the yqhR and yqhS genes (Fig 7F) (72). We

35 conclude that, in the absence of YmfI, a threshold of EF-P protein level and thus activity is required to initiate swarming motility.

Next, the PyqhS-efp complementation construct was randomly mutagenized using an error prone polymerase in seven separate tubes and each PCR product was separately transformed into an efp ymfI double mutant background to generate seven parallel pools of transformants. Each pool of transformants was collected and spotted as a mixed population onto swarming motility agar. Cells initially grew as a tight colony in the center of the plate but, in each pool, motile cells emerged with a prolonged lag period and swarmed outward from the origin. One colony was isolated from each motile swarm and was confirmed to have restored swarming motility. Sequencing of the PyqhS-efp construct indicated that each isolate contained either a mutation of Lys29 to an asparagine (EF-PK29N) or a mutation of

Lys32 to an arginine (EF-PK32R) (Table 2). In three swarming proficient isolates, K29N or

K32R was the only mutation present (Fig 3G). Mutation of K32R was allele-specific for the rescue of swarming motility in an efp ymfI background as ectopic introduction of K32A was not sufficient (Fig 3H). Furthermore, neither ectopic integrant of K29N nor K32R exhibited increased protein levels relative to ectopic integrant of the wild type allele in otherwise wild type cells indicating that suppression was not simply due to increased stability or expression of the mutant proteins (Fig 2E). We conclude that particular substitutions of Lys29 and Lys32 yielded EF-P at least partially YmfI-independent and sufficient to restore partial swarming in the absence of YmfI.

Strain Number Genotype Nucleotide substitution(s) DK4043 efpK29N AAA > AAT (residue 29) DK4978 efpK29N AAA > AAT (residue 29) DK4072 efpK32R AAA > AGG (residue 32) DK4075 efpK32R,E101V AAA > AGA (residue 32) GAA > GTA (residue 101) DK4076 efpK32R,V111M AAA > AGA (residue 32) GTG > ATG (residue 111) DK4073 efpK32R,E124G AAA > AGA (residue 32) GAA > GGA (residue 124)

Table 2. Mutations in efp that improve swarming in the absence of YmfI. Experiment was performed by Katherine R. Hummels.

To further characterize the YmfI-independent alleles of EF-P, each single mutant was introduced by allelic replacement at the native locus. Cells expressing EF-PK29N at the native site swarmed like the wild type and migrated as two bands on a semi-native gel (Fig

3I, Fig 2C lane 5). The majority of the EF-PK29N resolved as the lower species, unlike wild type EF-P which resolved mostly as the upper species (Fig 2C, compare lane 1 to lane 5).

One way in which the ratio of upper/lower bands could be altered in this way is if EF-PK29N affected the rate of post-translational modification. To test this hypothesis, translation was inhibited in an efpK29N mutant and EF-P electrophoretic mobility on a semi-native gel was monitored over time. The dominant EF-PK29N lower band appeared to decrease slowly with a concomitant increase in the intensity of the upper band indicative of a slow conversion of the oxidized to reduced modification (Fig 2A, right). Further, EF-PK29N expressed from the native site was not fully independent of YmfI as mutation of ymfI diminished swarming migration and abolished the faint upper band in semi-native gels (Fig 2C, compare lanes 5 and 6; Fig 3I). We conclude that EF-PK29N has a decreased rate of post-translational modification thereby preventing accumulation of 5-aminopentanonated EF-P.

Cells expressing EF-PK32R at the native site arrested swarming prematurely and resolved as a single band on native gels (Fig 2C lane 7; Fig 3J). Unlike efpK29N, mutation of ymfI did not diminish swarming migration of efpK32R or alter mobility on semi-native gel electrophoresis of the EF-PK32R protein (Fig 2C compare lanes 7 and 8; Fig 3J). Finally, wild type cells expressing the combined double mutant allele EF-PK29N,K32R exhibited a premature swarming arrest and resolved as a single band on native gels regardless of the presence of YmfI, similar to the EF-PK32R single mutant alone (Fig 2C compare lanes 9 and

10; Fig 3K). We conclude that both lysine residues contribute to the same function as the phenotype of the double mutant was not additive, and that mutation of K32R yields EF-P

YmfI-independent and is epistatic to K29N.

To determine the consequence of the suppressor mutations on post-translational

EF-P modification, the EF-P alleles were fused to a FLAG tag, overexpressed, purified from B. subtilis, and analyzed by mass spectrometry. For the wild type and EF-PK29N protein, both unmodified and 5-aminopentanolylated EF-P were detected (Fig 8G-I). By contrast, only the unmodified fragment was detected for the EF-PK32R protein and no additional masses could be detected on the peptide (Fig 8J-L). We conclude that one way to bypass the need for YmfI and restore EF-P activity is to abolish EF-P modification on

Lys32. We also conclude that cells mutated for ymfI exhibit a swarming defect because they accumulate an inhibitory 5-aminopentanone group on EF-P. Swarming inhibition may be relieved either by converting the group to 5-aminopentanol or by the accumulation of unmodified EF-P. Unmodified EF-P can accumulate by mutation of K32R, by decreasing the rate of post-translation modification as seen in K29N mutants, or by simply overexpressing EF-P protein. Finally, we conclude that, to the best of our knowledge, B. subtilis EF-P is the only member of its family that functions in the absence of post- translational modification.

2.3 Discussion

Elongation factor P is a highly conserved protein found in all domains of life that structurally resembles a tRNA, associates with the ribosome, and assists in translating particular primary sequences such as polyprolines (6, 8, 10, 33, 64). Furthermore, the EF-

P position analogous to the site of amino acid attachment on tRNA is post-translationally modified by species-specific, structurally-divergent molecules (58). Here we identify YmfI of Bacillus subtilis as the first protein shown to be involved in the post-translational modification of EF-P by 5-aminopentanol. YmfI is a protein of previously unknown function that is a paralog of the fatty acid biosynthesis protein FabG and requires a conserved active site residue for its function (Fig 4). Furthermore, YmfI catalyzes the

NADPH-dependent alteration of EF-P electrophoretic mobility in vitro and the reduction of EF-P 5-aminopentanone to 5-aminopentanol in vivo (Fig 12A). To date, B. subtilis is the only species known to modify EF-P with 5-aminopentanol but YmfI homologs were found by sequence analysis and synteny with upstream genes in the Firmicutes order of

Bacilliales (Fig 6A,B). Based on our analyses, we predict that the genera of Listeria and

Staphylococcus also modify EF-P with 5-aminopentanol.

In Bacillus subtilis, EF-P is required for a flagellar-mediated form of surface motility called swarming, and mutation of YmfI conferred a swarming defect. The swarming motility defect was not due to the inability to synthesize the mature 5- aminopentanol group and was rather due to the accumulation of the inhibitory oxidized 5- aminopentanone group as indicated by suppressor mutations that altered the EF-P primary sequence. One suppressor mutation changed the modified EF-P residue Lys32 to an

40 arginine, and while the allele impaired swarming in a wild type cell, it bypassed the requirement for YmfI by abolishing modification altogether. Both a swarming impairment and lack of modification were previously reported in a site-directed mutation that changed

Lys32 to Ala (61), but unlike the EF-PK32R allele, the EF-PK32A allele was unable to bypass the swarming requirement for YmfI. We infer that a positive charge at position 32 assists in the function of EF-P and that a lysine residue in particular is ideal in B. subtilis. The partial phenotypes of the EF-P K32R and K32A mutations however, indicate that 5- aminopentanolylation per se is not explicitly required for EF-P activity. To the best of our knowledge, the EF-P protein of B. subtilis is the first of its kind to retain substantial biological activity in the absence of post-translational modification possibly due to the below-expected number of genomic polyprolines encoded in B. sutilis (61).

Another suppressor mutation changed the highly conserved EF-P residue Lys29 to an asparagine. EF-PK29N appears to reduce the rate of 5-aminopentanolylation on Lys32 as indicated by a dramatic reduction in upper EF-P band on semi-native gels. The lower band accumulated in the EF-PK29N mutant but could represent at least two different molecular species. 5-aminopentanonated EF-P resolves as the lower band as it was the only band observed in the absence of YmfI. The lower band also accumulated when EF-P was overexpressed, and unmodified EF-P predominated in mass spectrometry analyses. Thus, the lower band on semi-native gels may represent two molecular species, one in which

Lys32 is unmodified and one in which Lys32 is 5-aminopentanonated. We infer that EF-

PK29N restores swarming in the absence of YmfI by reducing the rate of post-translational modification thereby accumulating EF-P unmodified at Lys32. EF-P Lys29 could play a

41 role in the catalysis of 5-aminopentanoylation as a binding site for modification proteins or as a residue for allosteric regulation. While we cannot distinguish whether Lys29 is an active or allosteric site, we note that Lys29 appears to be as highly conserved as Lys32 in

EF-P proteins (Fig 12B, Fig 13).

Post-translational modification is thought to be essential for EF-P activity for the organisms in which it has been studied and is generally assumed to directly participate in the synthesis of certain primary sequences, but how the modification participates in potentiating translation is unknown. If the modification does directly participate in synthesis, it seems unusual that a conserved EF-P function can tolerate a wide variety of different chemical modifications including but likely not limited to lysine, hypusine, rhamnose, and 5-aminopentanol groups depending on the organism. Furthermore, some modifications such as deoxyhypusine in the fungus Fusarium alter EF-P function and here we show in B. subtilis that 5-aminopentanonation has similar consequences, perhaps for regulatory reasons (73). Consistent with regulation, we show that in B. subtilis mutation of

Lys32 to an Arg (a residue naturally found in other EF-P proteins; Fig 12B, Fig 13) partially preserves EF-P function in the absence of the 5-aminopentanol group. Moreover, overexpression of EF-P and mutation of Lys29 to an asparagine results in the hyper- accumulation of unmodified EF-P and bypasses the swarming defect caused by the absence of YmfI. Thus, we hypothesize an unmodified Lys32 may be fully functional in Bacillus subtilis. We speculate that post-translational modification of EF-P may serve a regulatory function in B. subtilis and perhaps other organisms.

Figure 12. Model of EF-P activation by YmfI. (A) The crystal structure of EF-P isolated from the close relative of B. subtilis, Clostridium thermocellum (PDB ID 1YBY) with the location of Lys 29 (K29) and Lys 32 (K32) indicated. YmfI catalyzes the converstion of 5-aminopentanone (left) to 5-aminopentanol (right) groups on Lys32. In the absence of YmfI, 5-aminopentanone accumulates on EF-P and inhibits swarming motility (T-bar). (B) A multiple sequence alignment of EF-P orthologs from B. subtilis (Bsu), C. thermocellum (Cth), E. coli (Eco), Pseudomonas aeruginosa (Pae), Thermus thermophilus (Tth), the archaeon Haloferax volcanii (Hvo), and the eukaryote Homo sapiens (Hsa). The location of B. subtilis EF- P conserved residues K29 and K32 are indicated by carets. Figure was generated by Katherine R. Hummels.

Figure 13. Distribution of EF-P modification enzymes and conservation of residues 29 and 32. Phylogenetic tree of organisms from all domains of life onto which the presence of a homolog to DHS, EpmA, EarP, or YmfI (inner circle), the identity of EF-P residue 32 (middle circle) and the identity of EF-P residue 29 (outer circle) have been mapped. In the case of multiple homologs encoded within the same genome, the band is split accordingly. Figure was generated by Katherine R. Hummels.

2.4 Materials and Methods

2.4.1 Strains and growth conditions B. subtilis and E. coli strains were grown in lysogeny broth (LB) (10 g tryptone, 5 g yeast extract, 5 g NaCl per L) or on LB plates fortified with 1.5% Bacto agar at 37˚C.

When appropriate, antibiotics were included at the following concentrations: 10 µg/ml tetracycline, 100 µg/ml spectinomycin, 5 µg/ml chloramphenicol, 5 µg/ml kanamycin, and

1 µg/ml erythromycin plus 25 µg/ml lincomycin (mls). Isopropyl -D- thiogalactopyranoside (IPTG, Sigma) was added to the medium at the indicated concentration when appropriate.

For quantitative swarm assays, strains were grown to mid-log phase (OD600 0.3-

1.0) concentrated to an OD600 of 10 in PBS pH 7.4 (0.8% NaCl, 0.02% KCl, 100 mM

Na2HPO4, and 17.5 mM KH2PO4) plus 0.5% India ink. LB plates fortified with 0.7% agar were dried for 10 min open-faced in a laminar flow hood and subsequently inoculated by spotting 10 uL cell resuspensions onto the center of the plate. Plates were dried an additional 10 min open-faced in a laminar flow hood and then incubated at 37°C in a humid chamber. Swarm radius was measured along the same axis every 30 minutes.

Images of swarm plates were obtained by toothpick-inoculating a colony into the center of an LB plate fortified with 0.7% agar. Plates were dried open-faced in a laminar flow hood for 12 min and incubated at 37°C in a humid chamber for 20 hrs. Images were taking using a BioRad Gel Doc.

2.4.2 Western Blotting

Strains were grown to mid-log phase, concentrated to an OD600 of 10 in lysis buffer

(17.2 mM Tris pH 7.0, 8.6 mM EDTA pH 8.0, 1 mg/mL Lysozyme, 0.1 mg/mL RNaseA,

20 µg/mL DNase I, and 50 µg/mL phenylmethane sulfonyl fluoride) and incubated at 37°C for 30 min. SDS sample buffer (500 mM Tris pH 6.8, 22% glycerol, 10% SDS, and 0.12% bromophenol blue) was added, and samples were boiled for 5 min. 12 µl boiled samples were loaded onto 10% polyacrylamide native (with no added SDS) or 15% polyacrylamide denaturing (with 0.1% SDS) gels. Lysates were resolved at 150 V for 1.25 hours, transferred onto nitrocellulose membranes, and subsequently probed with a 1:40,000 dilution of anti-EF-P or a 1:80,000 dilution anti-SigA polyclonal antiserum. Following incubation with the primary antibodies, nitrocellulose membranes were probed with horseradish peroxidase conjugated goat anti-rabbit immunoglobulin G. Blots were developed using Pierce ECL substrate (Thermo Fisher Scientific). To inhibit translation,

200 ug/mL spectinomycin was added to mid-log phase cultures.

2.4.3 Strain construction

All constructs were introduced into DK1042, a competent derivative of ancestral strain

3610 (Konkol et al., 2013) or laboratory strain PY79 and then transferred to the 3610 background using SPP1-mediated generalized phage transduction (Yasbin and Young,

1974). All strains used in this study are listed in Table 3. All plasmids used in this study are listed in Supplemental Table 4. All primers used in this study are listed in Supplemental

Table 5.

Complementation constructs- To create the efp complementation construct, the PyqhS promoter was amplified using primers 477/478 and the efp gene was amplified using primers 483/486 using B. subtilis 3610 genomic DNA as a template. Fragments were

46 digested with EcoRI/XhoI and XhoI/BamHI respectively and subsequently ligated into the

EcoRI and BamHI restriction sites of the plasmid pDG364, containing a polylinker and chloramphenicol resistance cassette between two arms of the amyE gene (Guérout-Fleury et al., 1996), to produce pDP85. pDP85 was transformed into DK2050 to produce

DK3780. Site directed mutation of efp to change lysine 32 to an alanine was performed by amplifying DK3780 genomic DNA with primer pairs 3177/4039 and 3180/4038 (primers

4038 and 4039 encode the efpK32A mutation). The resulting fragments were subjected to

Gibson assembly (Gibson, 2009), and transformed into DK2050 followed by selection for chloramphenicol resistance to produce DK2248.

Strain Genotype Reference 3610 Wild type DK1042 comIQ12L Konkol, 2013 DK1962 comIQ12L ∆yqhS DK2050 comIQ12L ∆efp Rajkovic, 2016 Q12L K32A DK2248 comI ∆efp amyE::PyqhS-efp cat Q12L DK2448 comI ∆efp amyE::Physpank-efp-flag spec Rajkovic, 2016 DK2886 comIQ12L ∆efp ymfI::tet Q12L K32A DK2889 comI ∆efp ymfI::tet amyE::PyqhS-efp cat DK3159 comIQ12L ∆papA DK3235 comIQ12L efpK32A Rajkovic, 2016 DK3621 comIQ12L ymfI::tet DK3712 comIQ12L efpK32A ymfI::tet DK3726 comIQ12L ymfF::tet DK3727 comIQ12L ymfH::tet DK3728 comIQ12L ymfFH::tet Q12L DK3780 comI ∆efp amyE::PyqhS-efp cat Q12L DK3789 comI ∆efp ymfI::tet amyE::PyqhS-efp cat Q12L DK3828 comI ∆efp ymfI::tet amyE::Physpank-efp-flag spec Q12L DK3969 comI ymfI::tet amyE::PymfF-ymfI spec Q12L K29N DK4043 comI ∆efp ymfI::tet amyE::PyqhS-efp cat Q12L K32R DK4072 comI ∆efp ymfI::tet amyE::PyqhS-efp cat Q12 L K32R,E124G DK4073 comI ∆efp ymfI::tet amyE::PyqhS-efp cat Q12 L K32R,E101V DK4075 comI ∆efp ymfI::tet amyE::PyqhS-efp cat Q12L K32R,V111M DK4076 comI ∆efp ymfI::tet amyE::PyqhS-efp cat Q12L Y150A DK4233 comI ymfI::tet amyE::PymfF-ymfI spec Q12L K29N DK4246 comI ∆efp amyE::Physpank-efp -flag spec Q12L K32R DK4247 comI ∆efp amyE::Physpank-efp -flag spec DK4282 comIQ12L efpK29N DK4359 comIQ12L efpK32R DK4396 comIQ12L efpK29N ymfI::tet DK4397 comIQ12L efpK32R ymfI::tet DK4420 comIQ12L efpK29N,K32R DK4436 comIQ12L efpK29N,K32R ymfI::tet DS235 ymfI::tet Kearns, 2004 DS236 ymfJ::tet DS1026 swrA::Tn10 spec Kearns, 2004 DS1028 comP::Tn10spec Kearns, 2004 DS1029 ymfI::Tn10spec Kearns, 2004 DS1044 srfAB::Tn10spec Kearns, 2004 DS1045 cheC::Tn10spec Kearns, 2004 DS1064 cheD::Tn10spec Kearns, 2004 DS1078 yabR::Tn10spec Kearns, 2004 DS1102 srfAA::Tn10spec Kearns, 2004 DS1107 swrB::Tn10spec Kearns, 2004 DS1113 swrC::Tn10spec Kearns, 2004 DS1122 srfAC::Tn10spec Kearns, 2004 DS1124 efp::Tn10spec Kearns, 2004 DS1146 rrnB::Tn10spec Kearns, 2004 DS1164 flhG::Tn10spec Kearns, 2004 PY79 sfp0 swrA

Table 3. Strains used in chapter 2. 48

Plasmid Genotype Reference pDG364 amyE::cat amp (Guérout-Fleury et al., 1996) pDG1515 tet amp (Guérout-Fleury et al., 1995) pAH25 amyE spec amp Gift from Amy Camp, Mt. Holyoke College pMimiMAD oriPt.s mls amp (Patrick and Kearns, 2008) pTB146 pT7-His6-SUMO amp (Bendezu et al, 2009) pDP85 amyE::PyqhS-efp cat amp pKRH1 Ω∆yqhS mls amp pKRH2 Ω∆papA mls amp pKRH50 amyE::PymfF-ymfI spec amp pKRH59 ΩefpK29N mls amp pKRH60 ΩefpK32R mls amp pKRH61 ΩefpK29N,K32R mls amp Y150A pKRH67 amyE::PymfF-ymfI spec amp pKRH76 pT7-His6-SUMO-ymfI amp

Table 4. Plasmids used in chapter 2.

Primer Sequence 105 TCTTGGATATGAGCCTGTTTGATGT 106 GAACAACCTGCACCATTGCAAGAAGTTTTCTAATACTGACATGCTTGTTT 107 TTGATCCTTTTTTTATAACAGGAATTCAAGCTTGTTGAAAATAACAGCACTCA 108 TCCAATGACATTGCTTTTTCCTCTC 477 AGCTGAATTCGGAAAAGCTTGCAC 478 CTCCTCTCGAGTTTCGCGAGGGTTTCTCATGTAAAA 483 CTCCTGGATCCTCTATGCTCTTGAAACGTAAGAACC 486 GAGGACTCGAGAAGAACTTATTATTTTGTGATTGGAATAT 861 AGGAGTGGCGGTGCTTGCCGCCAGCG 862 CTCCTTCTACTAATTCAAGGCGTGTCTCACCA 954 AAGAGAACCGCTTAAGCCCGAGT 3177 CTAATTCAAGGCGTGTCTCAC 3180 GCGGTATTCCGTATGTCAAG 3250 ACGACTCACTATAGGGCGAATTG 3251 CTCACTAAAGGGAACAAAAGCTGG 4023 GTCGACTCTAGAGGATCCCCGAAAAAGATACAGAACAAAATGAA 4024 GCTGGCTTAATAAGTAGCGAACATTCGGTCCGTTCAAAATCAAAAA 4025 TTTTTGATTTTGAACGGACCGAATGTTCGCTACTTATTAAGCCAGC 4026 GTGAATTCGAGCTCGGTACCCGTTTGTCACTGGCCACACC 4027 GTCGACTCTAGAGGATCCCCCCTCATTTTTTGATTTTGAACGG 4028 GGTGATTGTCCGGTTGCCTGTAATCAGCATGCCGTCAAT 4029 ATTGACGGCATGCTGATTACAGGCAACCGGACAATCACC 4030 GTGAATTCGAGCTCGGTACCCGCTGTATCACCTTTAATACCAGGT 4031 GTCGACTCTAGAGGATCCCCGCCTTCGTTCAAGCCTTCC 4034 GTGAATTCGAGCTCGGTACCCGGAGAACCTCAGCGATTTCA 4038 CAGCACGTAAAACCTGGAGCAGGTGCGGCATTTGTCCGC 4039 GCGGACAAATGCCGCACCTGCTCCAGGTTTTACGTGCTGGA 4721 TCCTCAAGCTTGCGGAAAGAAGGAAAGAGGA 4722 AGGAGGCTAGCTCAGCAGTGCCAGCCGCC 4724 AGGAGAAGCTTGTTCATTTCTTAATTAGTATGTATT 4788 CATGTTTCTTTGTACATACTTTC 4789 CAATTCGCCCTATAGTGAGTCGTCGTGCTTTGATTTGAGTTCATT 4790 CCAGCTTTTGTTCCCTTTAGTGAGCTTGCCAATATTGAGCATGTC 4791 CTGCAAGGGTTTCGCTGATG 4792 GCGCTTTCTTGCACAGTTTAC 4793 CAATTCGCCCTATAGTGAGTCGTATCAAGGCCGTTGGCCGTT 4795 CTACTTTAATTGTTTCCATCGA 4832 CCAGCTTTTGTTCCCTTTAGTGAGGATGTCGTAACGGTACTCGA 4884 AGGAGGAATTCCCTGCTTTAGATGCATTGATAA 4891 GGAATATAGGAGGACATTAAAC 4892 GTTTAATGTCCTCCTATATTCC 4970 GAGCATCCTGTGAAGTGTTGGCAAGCATGGCAAAAGGAGCTCA 4971 TGAGCTCCTTTTGCCATGCTTGCCAACACTTCACAGGATGCTC 5132 TCCAGCACGTAAACCCTGGA 5133 TCCAGGGTTTACGTGCTGGA 5142 TAAAACCTGGAAGAGGTGCG 5143 CGCACCTCTTCCAGGTTTTA 5144 TAAACCCTGGAAGAGGTGCG 5145 CGCACCTCTTCCAGGGTTTA 5439 AGGAGGCTCTTCCGGTATGAACAAAACAGCACTAATCA 5440 TCCTCCTCGAGAAGAGAAACTATACACCGAATA

Table 5. Primers used in chapter 2.

To create the ymfI complementation construct, the PymfF promoter was amplified using primers 4724/4884 and the ymfI gene was amplified using primers 4721/4722 using

B. subtilis 3610 chromosomal DNA as a template. Fragments were digested with

EcoRI/HinDIII and HinDIII/NheI respectively were subsequently ligated into the EcoRI and NheI restriction sites of the plasmid pAH25, containing a polylinker and spectinomycin resistance cassette between two arms of the amyE gene, to produce pKRH50. pKRH50 was transformed into DK3621 followed by selection for spectinomycin resistance to produce

DK3969. Site directed mutation of ymfI to change tyrosine 150 to an alanine was performed by amplifying DK3969 genomic DNA with primer pairs 4884/4971 and 4722/4970. The resulting fragments were subjected to Gibson assembly (Gibson, 2009) followed by PCR amplification by primers 4722/4884. The resulting DNA fragment was subsequently ligated into the EcoRI and NheI restriction sites of the plasmid pAH25 to produce pKRH67.

Plasmid pKRH67 was transformed into DK3621 followed by selection on spec to produce

DK4233.

Site directed mutation of the Physpank-efp-flag construct was performed by amplifying DK2448 genomic DNA (Rajkovic, 2015) with primer pairs 861/5133 and

862/5132 for efpK29N and 861/5143 861/5142 for efpK32R . Cognate fragments were subjected to Gibson assembly (Gibson, 2009) and transformed into DK2050 followed by selection for spectinomycin resistance to yield DK4246 and DK4247.

YmfI expression construct - To create the His-Sumo-YmfI expression construct, ymfI was amplified using primers 5439/5440 using B. subtilis 3610 chromosomal DNA as a

51 template. The resulting fragment was digested with XhoI/SapI and subsequently ligated into the XhoI and SapI restriction sites of the plasmid pTB146 to produce pKRH76.

Native site mutants- Site directed mutation of efp at the native site was performed by allelic replacement. 3610 genomic DNA was amplified using primer pairs 4031/5133 4034/5132 for efpK29N, 4031/5143 4034/5042 for efpK32R, and 4031/5145 4034/5144 for efpK29N,K32R.

Cognate fragments were introduced into the SmaI restriction site of pMiniMAD using

Gibson assembly (Patrick and Kearns, 2008; Gibson, 2009) to produce pKRH59, pKRH60, and pKRH61 which were subsequently passaged through the recA+ Escherichia coli strain

TG1 and transformed into DK1042. Plasmid pMiniMAD (and its derivatives pKRH59, pKRH60, and pKRH61) encodes an mls resistance cassette and a temperature sensitive origin that is active at room temperature but not at 37°C. Thus, mls-resistant colonies were isolated at 37°C (ensuring integration of the plasmid into the chromosome) and subsequently grown in overnight at room temperature, thereby activating the plasmid- encoded temperature sensitive origin and inciting plasmid excision. Mls-sensitive colonies were isolated and the efp locus was sequenced to determine the allele present at that site.

Random mutagenesis of efp - Mutagenesis of the efp complementation construct was achieved by amplification of DK3789 genomic DNA with primer pair 4891/3180 using the error prone Taq polymerase (New England BioLabs Inc.). To decrease the frequency of obtaining mutations that increase expression of EF-P through altering its promoter, the portion of the complementation construct upstream of efp (including PyqhS) was amplified

52 with primers 3177/4892 using the high fidelity Phusion polymerase (New England

BioLabs Inc.). The two fragments were ligated together using Gibson assembly (Gibson,

2009) and the products subsequently PCR amplified with primers 954/3180. The resulting

7 pools of mutagenized amyE::PyqhS-efp were transformed into DK2886 and chloramphenicol-resistant colonies were selected.

Approximately 1700 colonies for each of 7 pools were scraped off of selection plates as a mixture, diluted into LB, and grown to mid-log phase at 37°C. Cell mixtures were then subjected to swarm assays as described above. Following a 9-10 hour incubation on swarm plates, swarming-competent cells began to emerge from the site of inoculation and one colony per pool was isolated from the edge of the swarm front. The amyE::PyqhS- efp complementation construct from each of the ymfI suppressors was backcrossed into

DK2886 by SPP1-mediated transduction and the retention of swarming motility was confirmed. For each ymfI suppressor, the efp complementation construct was sequenced to determine the mutations present.

Insertion/deletion mutants - ymfF::tet. Mutation of ymfF was performed by amplifying the regions upstream (with primers 4788/4789) and downstream (with primers 4790/4791) of ymfF using 3610 chromosomal DNA as a template, and amplifying the tetracycline resistance cassette from pDG1515 with primers 3250/3251 (Geurot-Fleury, 1995). The three fragments were ligated by Gibson assembly and transformed into DK1042 to produce

DK3726. The mutation was confirmed by PCR length polymorphism analysis.

ymfH::tet. Mutation of ymfH was performed by amplifying the regions upstream

(with primers 4792/4793) and downstream (with primers 4832/4795) of ymfH using 3610 chromosomal DNA as a template, and amplifying the tetracycline resistance cassette from pDG1515 with primers 3250/3251. The three fragments were ligated using Gibson assembly and transformed into DK1042 to produce DK3727. The mutation was confirmed by PCR length polymorphism analysis.

ymfFH::tet. Simultaneous mutation of ymfF and ymfH was performed by amplifying the regions upstream of ymfF (with primers 4788/4789) and downstream of ymfH (with primers 4832/4795) using 3610 chromosomal DNA as a template, and amplifying the tetracycline resistance cassette from pDG1515 with primers 3250/3251.

The three fragments were ligated using Gibson assembly and transformed into DK1042 to produce DK3727. The mutation was confirmed by PCR length polymorphism analysis.

ymfJ::tet. Mutation of ymfJ was performed by amplifying the regions upstream

(with primers 105/106) and downstream (with primers 107/108) of ymfJ using 3610 chromosomal DNA as a template, the fragments were purified and used to amplify the tetracycline resistance cassette from HaeII-digested pDG1515. The amplification product was used as a template to cement the fragment using primers 105/108. The final long- flanking homology disruption cassette (Wach, 1996) was transformed into laboratory strain

PY79 and transduced to 3610 by SPP1-mediated phage transduction to produce DS236.

The mutation was confirmed by PCR length polymorphism analysis.

∆yqhS. Deletion of yqhS was achieved through allelic replacement. 3610 genomic

DNA was amplified using primer pairs 4023/4024 and 4025/4026 and the resulting

54 fragments were introduced into the SmaI restriction site of pMiniMAD using Gibson assembly (Gibson, 2009) to produce pKRH1. Plasmid pKRH1 was passaged through the recA+ Escherichia coli strain TG1, transformed into DK1042, integrated by selecting mls resistant colonies at 37°C and evicted by passage at room temperature. Mls-sensitive colonies were isolated and confirmed to encode the deletion by PCR length polymorphism analysis.

∆papA. Deletion of papA was achieved through allelic replacement. 3610 genomic

DNA was amplified using primer pairs 4027/4028 and 4029/4030 and the resulting fragments were introduced into the SmaI restriction site of pMiniMAD using Gibson assembly (Gibson, 2009) to produce pKRH2. Plasmid pKRH2 was passaged through the recA+ Escherichia coli strain TG1, transformed into DK1042, integrated by selecting mls resistant colonies at 37°C and evicted by passage at room temperature. Mls-sensitive colonies were isolated and confirmed to encode the deletion by PCR length polymorphism analysis.

SPP1-mediated transduction

SPP1-mediated transductions were performed as described previously (Yasbin,

2004). In short, lysates were created on B. subtilis strains grown in TY (1% Tryptone, 0.5% yeast extract, 0.5% NaCl, 10mM MgSO4, and 1mM MnSO4). Recipient strains were grown in TY to stationary phase, 1 mL diluted into 9 mL TY, and 10 µL (for tetracycline selection) or 25 μL (for spectinomycin and chloramphenicol selection) lysates were added, followed by incubation at room temperature for 30 min and selection for the respective antibiotic

55 resistance at 37°C overnight. For transductions in which spectinomycin or chloramphenicol-resistance was selected for, 10 mM sodium citrate was added to the selection plates.

2.4.4 Mass Spectrometry

FLAG-tagged EF-P was overexpressed and purified from each mutant strain

(DK2448, DK3828, DK4246, and DK4247) as previously described (Rajkovic, 2016).

Following purification, the eluate was concentrated and resolved on a 13% SDS-PAGE gel. Protein was visualized with colloidal Coomassie, excised from the gel, and in-gel digested with chymotrypsin. The generated peptide samples were brought up in 2% acetonitrile in 0.1% formic acid (20 μL) and analyzed (4 μL) by LC/ESI MS/MS with a

Thermo Scientific Easy-nLC II (Thermo Scientific, Waltham, MA) coupled to a hybrid

Orbitrap Elite ETD (Thermo Scientific, Waltham, MA) mass spectrometer using an instrument configuration as described (Yi et al). In-line de-salting was accomplished using a reversed-phase trap column (100 μm × 20 mm) packed with Magic C18AQ (5-μm 200Å resin; Michrom Bioresources, Auburn, CA) followed by peptide separations on a reversed- phase column (75 μm × 250 mm) packed with Magic C18AQ (5-μm 100Å resin; Michrom

Bioresources, Auburn, CA) directly mounted on the electrospray ion source. A 40-minute gradient from 2% to 40% acetonitrile in 0.1% formic acid at a flow rate of 400 nL/minute was used for chromatographic separations. A spray voltage of 2750 V was applied to the electrospray tip and the Orbitrap Elite instrument was operated in the data-dependent mode, switching automatically between MS survey scans in the Orbitrap (AGC target value

1,000,000, resolution 120,000, and injection time 250 milliseconds) with MS/MS spectra

56 detected in the Orbitrap (AGC target value of 50,000, 15,000 resolution and injection time

250 milliseconds). The 3 most intense ions from the Fourier-transform (FT) full scan were selected for fragmentation in the Orbitrap using ETD 100 ms activation time with supplemental CID activation with normalized collision energy of 35%. Selected ions were dynamically excluded for 10 seconds.

2.4.5 Isotope Distribution Analysis

Predicted isotope distributions were calculated using the Scientific Instrument Services,

Inc. Isotope Distribution Calculator and Mass Spec Plotter. Actual abundances of each ion were quantified using Thermo Xcalibur software with a Genesis peak picking algorithm and an m/z range of 0.02. To compare the predicted and actual distributions, the measured abundances were set relative to the most abundant ion for the peptide with the indicated modification.

2.4.6 Phylogenetic Analyses

All phylogenetic trees are presented using the Interactive Tree of Life visualization software (Letunic, 2016). To identify the conservation of the different EF-P post- translational modification systems and B. subtilis EF-P residues 29 and 32 (analogous to residues 31 and 34 respectively in E. coli) across all domains of life, the genomes of 191 organisms identified by Ciccarelli, et al. were annotated with the Pfam 29 library using the software hmmer v 3.1b2 and an E value threshold of 1e-5 (Ciccarelli, 2006; Finn, 2016;

Eddy, 1998). The presence of EarP was established by the annotation of a DUF2331 domain and the presence of DHS was established by the annotation of a DS domain. The presence of EpmA was established by the annotation of a protein that contained a tRNA-

57 synt_2 domain with homology beginning at residue 15-30 of the profile hidden Markov model (HMM) and without a tRNA_anti-codon domain. The presence of YmfI was established by the presence of 3 co-oriented open reading frames separated by no more than 500 bp that encode proteins with domains found in the YmfFHI proteins. Specifically,

YmfF and YmfH were identified by the presence of Peptidase_M16 or Peptidase_M16_C domains, and YmfI homologs were identified by the presences of adh_short or adh_shortC2 domains.

To identify EF-P and eIF5a homologs, proteins that most closely aligned to the eIF-

5a, EFP_N, EFP, or Elongation-fact-P_C were identified and subsequently aligned to the

EFP_N domain. The identity of EF-P residue 32 was determined by the residue aligning to position 30 of the EFP_N profile HMM and the identity of EF-P residue 29 was determined by the residue aligning to position 27 of the EFP_N profile HMM.

To gain further clarity on which bacteria are likely to encode a YmfI homolog and thereby 5-aminopentanolylated EF-P, all species taxIDs with a completed bacterial genome available on the NCBI reference sequence database were analyzed for the presence of a

YmfFHI operon structure as described above.

2.4.7 In vitro Reactions

The His-Sumo-YmfI expression construct, pKRH76, was transformed into E. coli

Rosetta gami cells and grown in Terrific broth (12 g tryptone, 24 g yeast extract, 4 mL glycerol, 2.31 g monobasic potassium phosphate and 12.54 g dibasic potassium phoasphate) to mid-log phase. 1 mM IPTG was then added and the culture was grown overnight at 16°C. Cells were pelleted, resuspended in lysis buffer (50 mM Na2HPO4, 300

58 mM NaCl, and 10 mM imidazole) and lysed by sonication. Cell debris was pelleted by centrifugation at 31,000 x g for 30 min and Ni-nitrolotriacetic acid resin (Novagen) was added to the clarified supernatant. The bead-lysate mixture was incubated at 4°C overnight.

Beads were sedimented, the supernatant was removed, and the beads were washed 3 times with wash buffer (50 mM Na2HPO4, 300 mM NaCl, and 30 mM imidazole). Beads were resuspended in wash buffer, applied to a 1-cm separation column (Bio-Rad), and His-

SUMO-YmfI was eluted with wash buffer containing 250 mM imidazole. To cleave the

His-SUMO tag from the purified YmfI, ubiquitin ligase/protease was added and the reaction was incubated at 4°C overnight. To remove the free His-SUMO and any remaining

His-SUMO-YmfI from the cleavage reaction, Ni-nitrolotriacetic acid resin (Novagen) was added and incubated at 4°C for 1 hr. Beads were pelleted by centrifugation and the supernatant, containing untagged YmfI, was dialyzed into PBS pH 7.4 plus 50% glycerol and stored at 4°C. EF-PymfI-FLAG was purified from DK3828 as described previously

(Rajkovic, 2016).

10 µL reactions containing 5 µg EF-PymfI-FLAG was incubated in 100 mM sodium phosphate buffer at 37°C for 30 min. 4 μg YmfI with or without 150 μM NADPH was added to the reaction, where appropriate. Reactions were stopped by the addition of 50% glycerol, resolved by native gel electrophoresis, electroblotted, and probed with anti-EF-P polyclonal antisera.

Chapter 3

EF-P post-translational modification has variable impact on polyproline translation

in Bacillus subtilisb

3.1 Introduction

In order to stimulate translation of polyproline motifs, EF-P requires post- translational modification (PTM) at a highly conserved residue (31, 32). The structure of the modification can vary substantially between organisms. In eukaryotes, deoxyhypusine synthase (DHS) and deoxyhypusine hydroxylase (DOHH) coordinately modify the EF-P

3homolog eIF5A with a hypusine moiety, and hypusinated eIF5A is essential for viability

(69). In Pseudomonas aeruginosa and Neisseria meningitidis, EF-P is modified by EarP with a rhamnose moiety. Loss of EF-P or EarP can cause severe growth defects or, in some cases, lethality (58-60). In Escherichia coli and Salmonella enterica, EpmA, EpmB, and

EpmC modify EF-P with R-β-lysine (44, 49, 57). Although R-β-lysylated EF-P is not required for viability, mutants display a wide range of pleiotropic phenotypes including

b The work in this chapter was done in collaboration with Katherine R. Hummels, Rodney Tollerson II, Andrei Rajkovic, Lisa A. Jones, Daniel B. Kearns, and Michael Ibba. It was originally published in mBio [74, Witzky A, Hummels KR, Tollerson II R, Rajkovic A, Jones LA, Kearns DB, Michael I. 2018. EF-P post-translational modification has variable impact on polyproline translation in Bacillus subtilis. mBio 9:e00306-18.] mBio explicitly grants authors the rights to reuse full articles in a dissertation. Experiments performed by co-authors are credited in the corresponding figure legend of this chapter. All other experiments were performed by Anne M. Witzky. 60 growth defects and loss of virulence (54, 57). Recently, the search for EF-P PTMs has been expanded to Gram-positive bacteria. In Bacillus subtilis, EF-P is modified with a 5- aminopentanol moiety at Lys32 (61). However, in B. subtilis loss of EF-P does not result in the severe viability and growth defects observed in other bacteria. Instead, the major requirement for EF-P in B. subtilis is for swarming motility (61, 62).

We have recently shown that YmfI reduces 5-aminopentanone to 5-aminopentanol in the final step of EF-P modification (63). Further, we showed that 5-aminopentanone but not unmodified EF-P is inhibitory to swarming motility, as abolishing post-translational modification of EF-P altogether by mutation of Lys32 to an arginine suppressed the ymfI swarming defect (63). Here, we take advantage of EF-P activity in the absence of modification to identify other enzymes that act upstream of YmfI in the EF-P modification pathway (63). Mass spectrometry analyses of EF-P purified from wild-type (WT) B. subtilis and each of the modification mutants revealed that EF-P can retain incomplete modifications. In order to investigate the physiological consequences of producing EF-P with an intermediate modification, we phenotypically characterized these mutants for swarming proficiency and antibiotic resistance. Through this screen, we observed that PTM mutants display similar phenotypes as an efp mutant under some growth conditions but not others. In vivo polyproline reporter data revealed that this phenotypic discrepancy is due to variability in polyproline translation defect severity and context dependence in each mutant. Together, these results establish the 5-aminopentanol modification pathway and reveal a relationship between EF-P modification and motif specific EF-P dependence.

3.2 Results

3.2.1 A forward genetic screen to identify genes required for modification of EF-P

We have previously shown that, in the absence of YmfI, EF-P accumulates an intermediate modification on Lys32 that is inhibitory for swarming motility (63). One way to bypass this inhibition was to prevent PTM of EF-P through mutation of the modification site,

Lys32, to an arginine (63). We hypothesized that another way to bypass inhibition was to abolish PTM of EF-P through deletion of enzymes required for modification. A non- swarming ymfI efp (efp+) sensitized mutant background was transposon mutagenized in 26 parallel replicates, approximately 10,000 colonies were separately combined to form each pool, and the 26 separate pools were used to inoculate swarm agar. After a prolonged lag period, a subpopulation of cells were able to move out from the site of inoculation of each pool. One swarming proficient clone from each pool was isolated and backcrossed to the ymfI efp (efp+) parent by SPP1-mediated phage transduction. Each transposon insertion improved swarming motility, suggesting that the transposon was linked to the phenotype and suppressed the swarming defect of the ymfI mutation alone (Fig 14). The locations of the transposons were determined by inverse PCR.

Of the 26 transposon insertion suppressors of ymfI, 9 had a mutation in gsaB, 8 had a mutation in the ynbAB operon, 7 had a mutation in yaaO, 1 had a mutation in yfkA, and

1 had a mutation in ywlG (Table 6). To determine whether each gene disrupted by the transposon insertions was directly responsible for the phenotype, gsaB, ynbA, ynbB, yaaO, yfkA, and ywlG, were disrupted by in-frame markerless deletion. Deletion of gsaB, ynbB, yaaO, yfkA, and ylwG increased swarming of the ymfI mutant but deletion of ynbA did not

(Fig 15a-g). The gsaB, ynbB, yfkA, and ywlG deletion mutants could each be complemented by integrating a construct containing the corresponding gene cloned downstream of the putative native promoter at the ectopic amyE site in the chromosome (Fig 15a-c, e, f). While the yaaO deletion mutant could not be complemented by cloning yaaO expressed from its putative native promoter, it was complemented when yaaO was expressed under the control of an IPTG-inducible promoter (Fig 14b and Fig 15d). We infer that suppression of the ymfI swarming defect was caused by deletion of gsaB, ynbB, yaaO, yfkA, and ywlG rather than polar effects caused by the transposon insertion. By contrast, transposon insertions in ynbA were likely polar on ynbB as swarming motility inhibition was restored to the transposon mutants when the ynbB complementation construct was ectopically integrated

(Fig 15h). We conclude that deletion of gsaB, yaaO, ynbB, yfkA or ywlG but not ynbA suppresses the ymfI mutant swarming defect.

To determine if the identified genes impact post-translational modification of EF-

P, each single deletion strain was screened for EF-P electrophoretic mobility by isoelectric focusing and semi-native gel electrophoresis. In isoelectric focusing, EF-P in wild-type B. subtilis cells resolved as one band. EF-P in the yfkA, ynbB, gsaB, and yaaO mutants migrated as one band with an isoelectric focusing point lower than wild-type EF-P suggesting an alteration in modification state (Fig 16). Unlike the other mutants, EF-P from the ywlG and ymfI mutants migrated as two bands (Fig 16). In semi-native gel electrophoresis, EF-P resolved as two bands in the wild-type background (Fig 17). In each of the mutant strains, except ywlG, EF-P resolved as one band with electrophoretic mobility similar to the lower band in the wild-type strain (Fig 17). We infer that mutations that

63 restore swarming in the absence of YmfI alter EF-P electrophoretic mobility and may do so by different mechanisms.

Figure 14. Quantitative swarm expansion assay. Mid-log phase cultures were used to inoculate swarm plates. Swarm radius was monitored along the same axis every 30 minutes for 6 hours. Data points represent the average of 3 technical replicates. For panel B, swarm plates were supplemented with 1 mM IPTG. The following strains were used as the inoculum (A) ymfI efp (efp) (DK3789), ymfI efp (efp) gsaB::Tn (DK4304), ymfI efp (efp) yaaO::Tn (DK300), ymfI efp (efp) ynbB::Tn (DK4301), ymfI efp (efp) yfkA::Tn (DK4303), and ymfI efp (efp) ywlG::Tn (DK4368). (B) yaaO ymfI (DK4077) and yaaO ymfI PxpcA-yaaO (DK5305). (C) yaaO (DK3894), and yaaO (yaaO) (DK5327). (D) yfkA (DK4564) and yfkA (yfkA) (DK5165). (E) ywlG (DK4612) and ywlG (ywlG) (DK5166). Experiments were performed by Katherine R. Hummels.

Table 6. Transposon insertion suppressors of ymfI. The insertion tag indicates the 9 basepairs immediately adjacent to the transposon insertion element upstrtea of the kanamycin resistance cassette. Experiments were performed by Katherine R. Hummels.

Figure 15. Deletion of gsaB, ynbB, yaaO, yfkA, or ywlG suppresses the swarming defect of a ymfI mutant. Quantitative swarm expansion assay in which mid-log phase cultures were used to inoculate swarm plates. Swarm radius was monitored along the same axis every 30 minutes for 6 hours. Data points represent the average of 3 technical replicates. For panel D, swarm plates were supplemented with 1 mM IPTG. The following strains were used as the inoculum (A) WT (DK1042), efp (DK2050), and ymfI (DK3621). (B) gsaB ymfI (DK5171) and gsaB ymfI (gsaB) (DK5320). (C) ynbB ymfI (DK5172) and ynbB ymfI (ynbB) (DK5321). (D) yaaO ymfI (DK4077) and yaaO ymfI (yaaO) (DK5328). (E) yfkA ymfI (DK5170) and yfkA ymfI (yfkA) (DK5304). (F) ywlG ymfI (DK5174) and ywlG ymfI (ywlG) (DK5326). (G) ynbA ymfI (DK5173) and ynbA ymfI (ynbA) (DK5322). (H) ynbA::Tn ymfI (DK5339) and ynbA::Tn ymfI (ynbB) (DK5340). (I) gsaB (DK4601), gsaB (gsaB) (DK5308), ynbB (DK4604), and ynbB (ynbB) (DK5309). Experiments were performed by Katherine R. Hummels.

Figure 16. Isoelectric focusing gel of mutants. Lysates from WT (DK1042), ymfI (DK3621), yaaO (DK3894), gsaB (DK4601), ynbB (DK4604), yfkA (DK4564), ywlG (DK4612) were resolved via isoelectric focusing. Purified recombinant EF-P was run as an unmodified control. Blots were probed with anti-EFP polyclonal antisera.

Figure 17. Semi-native gel of mutants. Lysates from WT (DK1042), ymfI (DK3621), gsaB (DK4554), yaaO (DK4325), ynbB (DK4389), yfkA (DK4555), or ywlG (DK4556) were resolved by semi-native (top) or denaturing (middle and bottom) polyacrylamide gel electrophoresis and probed with anti-EFP (top and middle) or anti-SigA (bottom) polyclonal antisera. Experiment was performed by Katherine R.Hummels.

3.2.2 Mutants that suppress the absence of YmfI display aberrant 5-

aminopentanolylation.

In order to determine the modification status of EF-P in the absence of GsaB, YaaO,

YnbB, YfkA or YwlG, EF-P-FLAG was overexpressed and affinity-tag purified from each mutant background, in-gel digested with trypsin, and then resolved in an Oribtrap Elite or

Orbitrap Fusion mass spectrometer. In tandem mass spectrometry (MS/MS) analyses, the most abundant ions were selected for electron transfer dissociation (ETD). In the wild-type control, ions corresponding to unmodified EF-P (m/z=314.181, z=4+), as well as 5- aminopentanolylated EF-P (m/z=489.785, z=4+), with an additional mass of 101.084 Da on Lys32 were identified (Table 7, Fig 18a-b). Furthermore, additional masses on Lys32 of

82.042 (C5H7O) and 100.052 (C5H9O2) were also identified in the wild-type strain (Table

7, Fig 18c-d). Given the structure of 5-aminopentanol and its precursor, 5-aminopentanone, it is likely that the additional mass of 82.042 corresponds to pentenone and 100.052 corresponds to hydroxypentanone.

In an effort to assign each enzyme to a step in 5-aminopentanolylation, we searched for all identified intermediates in each mutant dataset as well as the ymfI dataset (63). In the ynbB and gsaB mutants, 5-aminopentanol was not detected on Lys32 (Fig 18e, 18g).

However, in the absence of ynbB and gsaB, Lys32 was acetylated (Table 7, Fig 18f, 18h).

Reanalysis of the MS/MS data for the ymfI mutant (63) identified pentenone, but not hydroxypentanone or the acetylation (Table 7, Fig 18i). We propose that deletion of ymfI restricts the completion of modification leading to the accumulation of intermediates immediately upstream of 5-aminopentanone. This is consistent with pentenone synthesis

70 immediately preceding 5-aminopentanone reduction in the EF-P modification pathway. In the absence of yaaO, yfkA, and ywlG, Lys32 retained low levels 5-aminopentanol, indicating that these genes are not strictly required for 5-aminopentanolylation (Table 7,

Fig 18j-k, 18m-n, 18p-q). An acetylation could also be identified on Lys32 in each of these mutants (Table 7, Fig 18L, 18O, 18R).

The presence of 5-aminopentanol on Lys32 in the absence of yaaO, yfkA, and ywlG, indicates that these genes are not essential for modification. Although the experimental design here did not allow for precise quantitative analysis, it is evident in the extracted ion chromatograms for the unmodified and modified peptides that the level of modification is substantially lower in the absence of each of these genes (Fig 19). This together with isoelectric focusing data of the native protein (Fig 16) suggests that while these genes are not essential for modification, they do influence it, potentially through synthesis of the initial precursor. It has previously been observed that low levels of EF-P modification can be achieved through alternate metabolic routes or environmental acquisition in the absence of canonical modification synthesis genes (59). The results here are consistent with yaaO, ywlG, and yfkA playing a role in synthesis of the substrate rather than direct modification on EF-P itself.

In order to identify the initial substrate, we investigated the proposed activities of

PTM genes that have clear homologs with known activities. GsaB is a paralog of HemL, a glutamate-1-semialdehyde aminotransferase that synthesizes 5-aminolevulinic acid, and

YaaO is predicted to synthesize cadaverine through lysine decarboxylation. Both 5- aminolevulinic acid and cadaverine bare striking resemblance to 5-aminopentanol.

However, supplying either of these substrates into the growth media of the yaaO and gsaB mutants failed to restore modification of EF-P as assessed by isoelectric focusing (data not shown), indicating that these molecules are not the initial precursor in 5-aminopentanol formation.

5-aminopentanol was not detected in the absence of ynbB, gsaB, and ymfI, indicating that these genes are strictly required for modification. YmfI has previously been shown to catalyze reduction of 5-aminopentanone to 5-aminopentanol in the final step of modification (63). As the acetylation was the only modification found on Lys32 in the absence of both ynbB and gsaB, it is difficult to definitively assign each protein to a position in the EF-P PTM pathway. Based on sequence similarity to genes of known function, GsaB is predicted to have aminotransferase activity, and YnbB is predicted to have carbon-sulfur lyase activity. Based on these proposed activities, it is likely that GsaB facilitates addition of the final amine group onto pentenone, and that YnbB removes a FAB substrate from an acyl carrier protein and forms hydroxypentanone on EF-P, possibly forming the first intermediate in 5-aminopentanol formation. Although an acetyl group was identified in the

PTM mutants, the absence of this modification in the wild-type sample suggests that it is not a true intermediate in 5-aminopentanol formation, but rather a spurious side reaction that occurs in the presence of high levels of unmodified EF-P. Taken together, these data suggest a pathway for the 5-aminopentanolylation of EF-P in B. subtilis (Fig 20).

Based on the proposed pathway, 5-aminopentanol is assembled directly on EF-P in a series of dehydration/reduction reactions, a type of synthesis that resembles fatty acid biosynthesis (FAB). Given these similarities, we hypothesized that 5-aminopentanol is

72 derived from FAB and that assembly could be impacted by changes in this process. To investigate this possibility, EF-P modification state was assessed via isoelectric focusing after knockdown or overexpression of three essential FAB factors (fabF, fabG, accB, Fig

21, (75)). Genetic manipulations of these factors did not influence the isoelectric focusing point of EF-P, likely indicating that the FAB and PTM pathway function independently in

B. subtilis.

Strain Mass Change Name Proposed Structure

101.084 5-aminopentanol

WT 100.052 Hydroxypentanone

82.042 Pentenone

99.068 5-aminopentanone

ymfI

82.042 Pentenone

ynbB 42.011 Acetylation

gsaB 42.011 Acetylation

101.084 5-aminopentanol

yaaO

42.011 Acetylation

101.084 5-aminopentanol

yfkA

42.011 Acetylation

101.084 5-aminopentanol

ywlG

42.011 Acetylation

Table 7. Summary of modifications identified in mutant strains.

Figure 18. MS/MS spectra of EF-P peptide containing Lys32. EF-P was purified from DK2448 (A-D), DK4310 (E-F), DK4313 (G-H), DK3908 (J-L), DK4572 (M-O), and DK4573 (P-R). Following purification, protein was in-gel digested with trypsin, and then resolved with tandem mass spectrometry on Orbitrap Elite mass spectrometer (A-H, M-R) or Orbitrap Fusion mass spectrometer (J-L). Modification at Lys32 inhibits trypsin cleavage, resulting in differences in digested peptide products depending on the modification state of Lys32. Parent ions are z=4+ (A-N, P-Q) or z=3+ (O, R). MS/MS for unmodified peptide (A, E, G, J, M, P, m/z = 314.1809), or peptide modified with 5-aminopentanol (B, K, N, Q m/z=489.7853), pentenone (C, m/z=485.0248), hydroxypentanone (D, m/z=489.5274), or an acetylation (F, H, L, O, R, m/z=475.0169, z=4+ or m/z=633.0200, z=3+) are shown. Modifications on Lys32 are marked as 5-aminopentanol (AP), pentenone (P), hydroxypentanone (HP), or acetylation (Ac). Data in (I) was obtained from Hummels et al. 2017. EF-P was purified from DK3828 (ymfI), in-gel digested with chymotrypsin, and then resolved with tandem mass spectrometry. MS/MS spectra of chymotrypsin digested EF-P with pentenone (P) on Lys32 (m/z=407.8966, z=4+). Continued

Figure 18. Continued

Figure 19. Extracted ion chromatograms of unmodified (A, C, E, G) and modified (B, D, F, H) EF-P peptides containing Lys32. EF-P-Flag was purified from wild-type (DK2448), ∆yfkA (DK 4572), ∆ywlG (DK4573), ∆yaaO (DK3908) in gel digested with trypsin, and resolved on an Orbitrap Elite mass spectrometer (A-F) or Orbitrap fusion mass spectrometer (G-H). Trypsin digestion is inhibited by modification at Lys32, resulting in the peptide VVDFQHVKPGK (m/z = 314.1809) when Lys32 is unmodified and VVDFQHVKPGKGAAFVR (m/z = 489.7852) when modified. Intensity values are relative to the intensity of the unmodified peptide for the given sample. All chromatograms represent the 4+ ion.

Figure 20. Proposed modification pathway outline based on tandem mass spectrometry analysis. Hummels et al. 2017 established that YmfI reduces 5-aminopentanone to 5-aminopentanol in the final step of modification. Additional intermediates identified here represent most likely structure based on the structure of 5-aminopentanol.

Figure 21. Overexpression (A) or knockdown (B) of fatty acid biosynthesis factors does not influence EF-P modification state. WT, fabF OE (AW146), fabG OE (AW147), accB OE (AW148), fabF KD (BEC11340), fabG KD (BEC15910), and accB KD (BEC24350) were grown to mid-log in LB or LB supplemented with 1 mM IPTG (A) or 0.1% xylose (B). Lysates were resolved via isoelectric focusing. Purified recombinant EF-P was run as an unmodified control. Blots were probed with polyclonal EF-P antisera.

3.2.3 Phenotypic characterization of modification mutants

Given that stable PTM intermediates were identified in each mutant strain, it is possible that EF-P associated phenotypes are also altered in each of these strains. Consistent with this possibility, swarming motility arrested prematurely in the gsaB and ynbB mutants (but not yaaO, yfkA, or ywlG) and both mutants could be complemented by wild-type alleles

(Fig 15i, Fig 14c-e). To investigate the impact of the PTM gene deletions in alternative EF-

P associated phenotypes, growth of WT and efp were compared in a Biolog phenotype microarray that tests for 1,920 different metabolic and chemical sensitivities. For the majority of metabolic stressors, WT and efp exhibited no significant growth differences.

WT displayed relatively enhanced respiration in 2,3-butanone (carbon source stress) and efp displayed enhanced respiration in D-glucose-1-phosphate (organic phosphate source stress) (Tables 8 and 9). Conversely, efp displayed reduced respiration under 60 different chemical stressors with diverse mechanisms of action, indicating that EF-P activity is required under these chemical stressors (Table 8). Many chemicals reduced the growth of the efp mutant relative to wild-type, but efp grew better than wild-type in the presence of folate biosynthesis inhibitors like sulfanilamide (Table 9). To determine if sulfonamide sensitivity is impacted by EF-P modification state, all modification mutants were grown in the presence of 1 mg/ml to 5 mg/ml sulfanilamide, and growth was measured after 18 hours. The majority of modification mutants (yaaO, gsaB, ynbB, yfkA, ywlG) did not display the same sulfanilamide resistance seen in the efp mutant (Fig 22A). The exception was the ymfI mutant, which was somewhat resistant to sulfanilamide, though not to the full extent of the efp mutant (Fig 22A).

In addition to sulfonamide resistance, efp also displayed enhanced respiration in the presence of puromycin, a translation inhibitor that results in premature termination of polypeptide synthesis (Table 9). To determine if puromycin sensitivity is impacted by the

EF-P modification state, all modification mutants were grown in the presence of 1 μg/ml to 20 μg/ml puromycin, and growth was measured after 8 hours. In contrast to the sulfanilamide phenotype, the majority of modification mutants (yaaO, gsaB, ynbB, ymfI) phenocopied the efp mutant, while both the ywlG and yfkA mutants were sensitive to the puromycin treatment (Fig 22B).

Growth Condition PM Mode (WT Enhanced Resipration) 2,3-Butanone 2A C-Source, Alcohol Sodium Benzoate, pH 5.0 9 Toxicity, Benzoate Sodium Sulfate 9 Osmotic Sensitivity Sodium Formate 9 Osmotic Sensitivity Urea 9 Osmotic Sensitivity Sodium Nitrate 9 Toxicity Sodium Nitrite 9 Toxicity pH 4.5 10 pH pH 4.5 + AA 10 pH, decarboxylase pH 4.5 + Ornithine 10 pH, decarboxylase pH 4.5 + Cysteic acid 10 pH, decarboxylase pH 4.5 + Diamino-Pimelic Acid 10 pH, decarboxylase pH 4.5 + Urea 10 pH, decarboxylase Tobramycine 12B Protein Synthesis, 30S Nickel Chloride 13B Toxic Cation 2, 2' Dipyridyl 13B Chelator, Lipophilic Potassium Chomate 13B Toxic Cation Oxolinic Acid 13B DNA Topoisomerase Doxycycline 13B Protein Synthesis, 30S Cesium Chloride 13B Toxic Cation Thallium (I) Acetate 13B Toxic Cation Cobalt Chloride 13B Toxic Cation Manganese (II) Chloride 13B Toxic Cation Cupric Chloride 13B Toxic Cation Acriflavine 14 DNA intercalator Furaltadone 14 Nitro Compound 1-Hydroxypyridine-2-thione 14 Biofilm Inhibitor Sodium Cyanate 14 Toxic Anion Iodoacetate 14 Oxidizes Sulfhydryls Sodium Dichromate 14 Toxic Anion Cefoxitin 14 Wall Sodium Metaborate 14 Toxic Anion Carbenicillin 14 Wall Nitrite 14 Toxic Anion EDTA 15 Chelator, Hydrophilic 5,7-Dichloro-8-hydroxy-quinalone 15 Chelator, Lipophilic 5,7-Dichloro08-hydroxyquinoline 15 Chelator, Lipophilic Nitrofurazone 15 Nitro Compound Menadione 15 Respiration 2-Nitroimidazole 15 Nitro Compound Norfloxacin 16 DNA topoisomerase Protamine Sulfate 16 Membrane Cinoxacin 16 DNA Topoisomerase Streptomycin 16 Protein Synthesis, 30S Sorbic Acid 16 Respiration D-Serine 17 Inhibits 3PGA denydrogenase Sodium salicylate 17 Biofilm Inhibitor Aminotriazole 17 Inhibits Catalse, Histidine Compound 48/80 17 cyclic AMP phosphodiesterase inhibitor D,L-Methionine hydroxamate 17 tRNA synthetase Phenylarsine Oxide 17 Tyrosine Phosphatase Inhibitor Ketoprofen 18C Biofilm Inhibitor Pyrophosphate 18C Chelator, Hydrophilic Thiamphenicol 18C Protein Synthesis, 30S Sodium m-arsenite 18C Toxic Anion Sodium Bromate 18C Toxic Anion Lidocaine 18C Ion Channel Inhibitor Sodium Periodate 18C Toxic Anion Hexamminecobalt (III) Chloride 19 DNA Syntehsis Ornidazole 20 Nitro Compound Patulin 20 microtubulin polymerization inhibitor

Table 8. Phenotype microarray conditions where WT displayed enhanced respiration.

Growth Condition PM Mode (∆efp Enhanced Resipration)

D-Glucose-1-Phosphate 3 P-Source, Organic Sulfadiazine 12B Folate Antagonist, PABA analog Sulfamethazine 12B Folate Antagonist, PABA analog Sulfathiazole 12B Folate Antagonist, PABA analog Sulfamethoxazole 12B Folate Antagonist, PABA analog Cefuroxime 13B Wall, cephalosporin Puromycin 15 Protein Synthesis, 30S Sulfanilamide 16 Folate Antagonist, PABA analog Trimethoprim 16 Folate Antagonist, PABA analog 1-Chloro-2,4-dinitrobenzene 16 Oxidizes sulfhydryls, depletes gluathione Sulfachloropyridazine 17 Folate Antagonist, PABA analog Sulfamonomethoxine 17 Folate Antagonist, PABA analog Sulfisoxazole 17 Folate Antagonist, PABA analog Plumbagin 18 Oxidizing Agent

Table 9. Phenotype microarray conditions where ∆efp displayed enhanced respiration.

Figure 22. Antibiotic senstivity in efp and PTM mutants. WT (DK1042), efp (DK2050), yfkA (DK4564), ywlG (DK4612), yaaO (DK3894), gsaB (DK4601), ynbB (DK4604), and ymfI (DK3621) were grown in the presence of the indicated amount of 2 mg/ml sulfanilamide (A) or 15 μg/ml puromycin (B), and the final OD600 was measured after 18 hours (A) or 8 hours (B) of growth.

3.2.4 Altered EF-P modification state has variable impact on polyproline translation.

In previously characterized instances of EF-P PTM, EF-P has been shown to require modification for activity, and PTM mutants display similar phenotypes as efp mutants (54,

57, 59). Here, phenotypic characterization of the PTM mutants has revealed an apparent discrepancy between the requirement for EF-P itself and requirement for EF-P modification. One way to explain the discrepancy is if each of the PTM mutants have variation in polyproline translation defects. In order to address this possibility, all mutants were transformed with a polyproline-GFP reporter construct containing one of three polyproline motifs (PPP, PPW, and PPE) (61). In the absence of ywlG, there was no significant defect in polyproline translation for any motif (Fig 23B-D). In contrast, mutants that lacked gsaB, ynbB, or yaaO displayed significant polyproline translation defects for all three polyproline motifs, although this defect was not as significant as seen in an efp mutant (Fig 23B-D). Loss of ymfI or yfkA has variable impact on polyproline translation.

In the ymfI mutant, there is no significant decrease in PPW-GFP levels, whereas there is a significant decrease in PPP-GFP and PPE-GFP levels (Fig 23B-D). In the yfkA mutant, there is a small defect in PPW-GFP levels and more significant decrease in PPP-GFP and

PPE-GFP levels. It is also noteworthy that while there is a decrease in polyproline translation efficiency for the ymfI and yfkA mutants, it is less severe than for gsaB, ynbB, and yaaO mutants.

To characterize the relative severity of the polyproline translational defect for each motif, the average PPX-GFP fluorescence in the wild-type strain was compared to each mutant (Fig 23E). Consistent with previous studies, PPW-GFP displayed the highest EF-P

92 dependence in the efp mutant (5, 61). In contrast, in each of the PTM mutants, PPP-GFP displayed the highest EF-P dependence. Each PTM mutant also displayed variability in the relative EF-P dependence for each motif. Many mutants exhibited a more significant defect in PPW translation than in PPE, but in the case of ymfI and ywlG, the translational defect with both of these motifs was essentially indistinguishable. This indicates that modulation of EF-P modification state not only impacts the strength of EF-P dependent pausing, but it also alters the context dependence of EF-P dependent pausing.

Figure 23. PPX-GFP reporter in PTM mutants. A GFP (A), PPW-GFP (B), PPP-GFP (C), or PPE-GFP (D) reporter construct was chromosomally inserted into amyE in each mutant (Rajkovic et al. 2015). After a 1 hour induction with 1 mM IPTG, GFP fluorescence was measured. Fluorescence levels were normalized to OD600. Error bars represent mean +/- SD from three biological replicates. Statistical significance was determined with an ANOVA and tukey post-hoc test (*P<0.05, **p<0.01). (E) Ratio of WT/mutant for each PPX-GFP reporter in each mutant background.

3.3 Discussion

3.3.1 EF-P is modified through multistep assembly that is reminiscent of FAB

In order to efficiently stimulate translation of polyproline motifs, EF-P requires

PTM of a conserved residue. In Gram-negative bacteria and eukaryotes, several diverse modifications have been identified, and the enzymes that facilitate modification are known.

Recently, a novel EF-P PTM has been identified in the Gram-positive bacterium, B. subtilis, but many of the genes required for modification have remained unknown (61).

Mass spectrometry analyses revealed that intermediate modifications (5-aminopentanone, hydroxypentanone, pentenone, and acetylation) are detectable on Lys32 in WT and modification deficient samples. Based on the intermediates identified, we propose an outline for EF-P PTM in B. subtilis (Fig 20). It should be noted that intermediates would not have be detected in this study if they are unstable or rapidly turned over. Therefore, additional intermediates in the pathway cannot be excluded. Nevertheless, our results indicate that B. subtilis employs a novel multistep method of modification that produces intermediates with distinct similarities to those found in FAB (Fig 20). Although 5- aminopentanolylation does appear to be derived from FAB, genetic manipulation of FAB did not impact the isoelectric focusing point of EF-P (Fig 21). This could indicate that the two processes are not biologically linked. It is also likely that the metabolic burden that

EF-P modification imposes on the cell is minor in comparison to that of FAB. Therefore, altering the levels of several FAB factors would not be sufficient to impact the overall EF-

P modification substrate pool to the point that changes in EF-P modification status could be readily observed.

Several of the intermediates identified here were not detected in previous studies, possibly because in this study EF-P-FLAG was purified using milder purification conditions to help with the detection of less stable intermediates. Furthermore, in our mass spectrometry procedure, the top ten most abundant ions from each cycle were selected for

MS/MS. In previous studies, only the top four were selected. Together, these differences in methodology allowed for the detection of ions and thus corresponding intermediates that would be less stable or present at much lower abundance.

3.3.2 5-aminopentanolylation as an EF-P modification strategy in other organisms

Here, we employed a forward genetic screen to identify genes required for 5- aminopentanolylation of EF-P. In order to predict which other bacteria modify EF-P with

5-aminopentanol, we searched for organisms that maintain all ymfI, yaaO, gsaB, ynbB, yfkA, and ywlG. Although many bacteria contained at least one of the modification genes, organisms that maintain the entire core set of modification genes could only be found within firmicutes (Fig 24). A number of proteobacteria also appear to maintain the core set of 5-aminopentanolylation genes. As these species also contain the machinery for R-β- lysylation, it is unlikely that 5-aminopentanolylation is the EF-P modification strategy for these organisms. This result instead likely stems from the fact that the genes required for

5-aminopentanol assembly are homologous to other broadly conserved genes.

Figure 24. Phylogenetic tree predicting other bacteria that will employ 5-aminopentanolylation as a modification strategy. Organisms that maintain ymfI, ynbB, gsaB, yfkA, ywlG and yaaO are marked in blue. Phylogenetic classifications are marked by corresponding colors in key. Taxonomic tree was generated using iTol. Figure was generated by Andrei Rajkovic.

3.3.3 EF-P dependent pausing is variably impacted by altered modification state

In all known instances of EF-P and eIF5A PTM, EF-P requires modification for function, and PTM mutants essentially phenocopy efp mutants (57-60). Conversely in B. subtilis, there is substantial phenotypic variability between the efp mutant and each PTM mutant. The complete 5-aminopentanol PTM is required for puromycin sensitivity, but it is dispensable for swarming motility and sulfanilamide sensitivity in the majority of the

PTM mutants. In vivo polyproline reporter data indicates that each mutant displays a context dependent defect in polyproline translation, with each motif being impacted unequally by the altered modification state of EF-P. This result likely accounts for the phenotypic variability observed between the efp mutant and the PTM mutants. It is of note that the majority of the modification mutants displayed a polyproline translation defect equal to or more severe than the ymfI mutant, yet they restored swarming to the ymfI mutant in the initial screen. This suggest result that PPP, PPE, and PPW are not relevant motifs for swarming proficiency. Of the proteins that are known to be required for motility, 23 contain polyproline motifs, with 13 different PPX motifs distributed among them (61). It is likely that in the absence of ymfI, there is a significant polyproline translation defect in one of these other EF-P dependent proteins, resulting in the swarming defect that can be alleviated by blocking 5-aminopentanone formation.

It has recently been established that in the case of R-β-lysylated EF-P, the modification interacts with the CCA end of the p-site tRNA, stabilizing it in a conformation that is favorable for peptide bond formation (11). Here, we have shown that alteration of

EF-P modification not only decreases polyproline translation efficiency, but it also

98 modulates the relative impact that EF-P has on each PPX motif. This result suggests that the relative EF-P dependence of each PPX motif is directly impacted by the nature of the modification itself. It would therefore be expected that organisms with a high content in a specific PPX motif would require a modification suited for a stabilizing conformation for that motif, perhaps justifying the structural diversity found in all known EF-P PTMs.

Ribosomal profiling in organisms that employ alternative modification strategies will be required to directly address this possibility.

While EF-P modification is critical for efficient polyproline translation, both the structure and the source of the modification are variable between organisms. This speaks to the nature of EF-P post-translational modification, in that although modifications are structurally diverse, they may all be simply siphoned out of natural metabolic processes.

Through this work, we have expanded the possible metabolic sources for EF-P PTM, and have shown how alteration of this modification modulates EF-P activity.

3.4 Materials and Methods

3.4.1 Growth conditions and strain construction.

Strains were grown in luria broth (5g NaCl, 5g yeast extract,10g tryptone per 1L with 100

μg/ml spectinomycin, 12.5 μg/ml tetracyclin, 0.5 μg/ml erythromycin, 5 μg/ml chloramphenicol, 5 μg/ml kanamycin, 1 µg/ml erythromycin plus 25 µg/ml lincomycin

(mls) and 100 μg/ml ampicillin when appropriate. Primers, strains, and plasmids used in this study are listed in tables 10-12.

Table 10. Strains used in chapter 3.

100

Table 11. Plasmids used in chapter 3.

101

Table 12. Primers used in chapter 3.

102

∆yaaO – 3610 genomic DNA was PCR amplified using primers 4933/4934 and

4935/4936. The resulting fragments were inserted into the SmaI restriction site of pMiniMAD2 using Gibson assembly to create pKRH58. pKRH58 was passaged through the recA+ Escherichia coli strain TG1 and subsequently transformed into DK1042 or

DK4601. pMiniMAD2 encodes mls resistance and a temperature sensitive origin of replication that is active at room temperature, but not 37°C. Thus, colonies that integrated the plasmid into their genome were selected for by growth in the presence of mls at 37C.

Plasmid eviction was promoted by inducing the temperature sensitive origin through overnight growth at room temperature. The resulting mls-sensitive colonies were isolated and confirmed to encode the deletion by PCR-length polymorphism analysis to create

DK3894 (∆yaaO) and DK4750 (∆gsaB ∆yaaO). The transposon from DK4300 was transduced into DK1042 to produce DK4325.

∆yfkA – 3610 genomic DNA was PCR amplified using primers 5279/5281 and

5280/5282. The resulting fragments were inserted into the SmaI restriction site of pMiniMAD2 using Gibson assembly to create pKRH69. pKRH69 was passaged through the recA+ Escherichia coli strain TG1 and subsequently transformed into DK1042. Mls resistant colonies were isolated at 37C and plasmid eviction was promoted by overnight growth at room temperature. The resulting mls-sensitive colonies were isolated and confirmed to encode the deletion by PCR-length polymorphism analysis to create DK4564.

The ymfI::tet mutation was transduced into DK4564 to produce DK5170. The transposon from DK4303 was transduced into DK1042 to produce DK4555.

103

∆ynbB – 3610 genomic DNA was PCR amplified using primers 5283/5284 and

5285/5286. The resulting fragments were inserted into the SmaI restriction site of pMiniMAD2 using Gibson assembly to create pKRH70. pKRH70 was passaged through the recA+ Escherichia coli strain TG1 and subsequently transformed into DK1042. Mls resistant colonies were isolated at 37C and plasmid eviction was promoted by overnight growth at room temperature. The resulting mls-sensitive colonies were isolated and confirmed to encode the deletion by PCR-length polymorphism analysis to create DK4604.

The ymfI::tet mutation was transduced into DK4604 to produce DK5172. The transposon from DK4381 was transduced into DK1042 to produce DK4389.

∆ynbA – 3610 genomic DNA was PCR amplified using primers 5287/5288 and

5289/5290. The resulting fragments were inserted into the SmaI restriction site of pMiniMAD2 using Gibson assembly to create pKRH71. pKRH71 was passaged through the recA+ Escherichia coli strain TG1 and subsequently transformed into DK1042. Mls resistant colonies were isolated at 37C and plasmid eviction was promoted by overnight growth at room temperature. The resulting mls-sensitive colonies were isolated and confirmed to encode the deletion by PCR-length polymorphism analysis to create DK4605.

The ymfI::tet mutation was transduced into DK4605 to produce DK5173.

∆gsaB – 3610 genomic DNA was PCR amplified using primers 5291/5292 and

5293/5294. The resulting fragments were inserted into the SmaI restriction site of pMiniMAD2 using Gibson assembly to create pKRH72. pKRH72 was passaged through the recA+ Escherichia coli strain TG1 and subsequently transformed into DK1042. Mls resistant colonies were isolated at 37C and plasmid eviction was promoted by overnight

104 growth at room temperature. The resulting mls-sensitive colonies were isolated and confirmed to encode the deletion by PCR-length polymorphism analysis to create to create

DK4601. The ymfI::tet mutation was transduced into DK4601 to produce DK5171. The transposon from DK4375 was transduced into DK1042 to produce DK4554.

∆ywlG – 3610 genomic DNA was PCR amplified using primers 5295/5296 and

5297/5298. The resulting fragments were inserted into the SmaI restriction site of pMiniMAD2 using Gibson assembly to create pKRH73. pKRH73 was passaged through the recA+ Escherichia coli strain TG1 and subsequently transformed into DK1042. Mls resistant colonies were isolated at 37C and plasmid eviction was promoted by overnight growth at room temperature. The resulting mls-sensitive colonies were isolated and confirmed to encode the deletion by PCR-length polymorphism analysis to create to create

DK4612. The ymfI::tet mutation was transduced into DK4601 to produce DK5174. The transposon from DK4368 was transduced into DK1042 to produce DK4556.

To create the yaaO efp-flag strain, the efp::tet mutation from DS354 was transduced into DK3894 and the amyE::Physpank-efp-flag construct from DK2448 was transduced into the resulting strain to create DK3908. To create the yfkA efp-flag strain, the efp::tet mutation from DS354 was transduced into DK4564 and the amyE::Physpank-efp-flag construct from DK2448 was transduced into the resulting strain to create DK4572. To create the gsaB yaaO efp-flag strain, the efp::tet mutation from DS354 was transduced into

DK4750 and the amyE::Physpank-efp-flag construct from DK2448 was transduced into the resulting strain to create DK4815. The ynbB, gsaB, and ywlG efp-flag strains were

105 created by transducing the transposons from DK4301, DK4304, and DK4368 into DK2448 to create DK4310, DK4313, and DK4573, respectively.

PynbA-ynbB – The ynbA promoter was amplified using p0rimers 5877/5878 and the ynbB open reading frame was amplified using primers 5879/5880 from 3610 genomic

DNA. Fragments were digested with HindIII/NheI and NheI/SphI, respectively, and ligated into the HindIII/SphI sites of pAH25 to create pKRH86. pKRH86 was transformed into

DK4604 to create DK5321 and the ymfI::tet mutation from DS235 was transduced into

DK5309 to create DK5321.

PyfkA-yfkA – The yfkA promoter and open reading frame was amplified using primers

5881/5882 from 3610 genomic DNA. The resulting fragment was digested with

BamHI/NheI and ligated into the BamHI/NheI sites of pAH25 to create pKRH87. pKRH87 was transformed into DK4564 to create DK5165 and the ymfI::tet mutation from DS235 was transduced into DK5165 to create DK5304.

PynbA-ynbB – The ywlF promoter was amplified using primers 5883/5884 and the ywlG open reading frame was amplified using primers 5879/5880 from 3610 genomic

DNA. Fragments were digested with HindIII/NheI and NheI/EcoRI, respectively, and ligated into the HindIII/EcoRI sites of pAH25 to create pKRH88. pKRH88 was transformed into DK4612 to create DK5166 and the ymfI::tet mutation from DS235 was transduced into DK5166 to create DK5326.

PxpcA-yaaO – The xpcA promoter was amplified using primers 5887/5888 and the yaaO open reading frame was amplified using primers 5889/5890 from 3610 genomic

DNA. Fragments were digested with HindIII/NheI and NheI/SphI, respectively, and ligated

106 into the HindIII/SphI sites of pAH25 to create pKRH89. pKRH89 was transformed into

DK3894 to create DK5167 and the ymfI::tet mutation from DS235 was transduced into

DK5167 to create DK5305.

PgsaB-gsaB – The gsaB promoter and open reading frame was amplified using primers 5909/5946 from 3610 genomic DNA. The resulting fragment was digested with

BamHI/NheI and ligated into the BamHI/NheI sites of pAH25 to create pKRH90. pKRH90 was transformed into DK4601 to create DK5308 and the ymfI::tet mutation from DS235 was transduced into DK5308 to create DK5320.

Physpank-yaaO – The yaaO open reading frame was excised from pKRH89 by restriction digestion with NheI/SphI and ligated into the NheI/SphI restriction sites of pDR111 to create pKRH112. pKRH112 was then transformed into DK3894 and DK4077 to create DK5327 and DK5328, respectively.

PPX-GFP Reporter Strains

amyE::Physpank-gfp – pAW92 was transformed into DK1042, DK2050, DK4601,

DK4604, DK4564, DK4612, DK3621, and DK3894 to create AW112, AW114, AW118,

AW120, AW122, AW124, RT01, and RT03 respectively (61).

amyE::Physpank-ppw-gfp – pAW40 was transformed into DK1042, DK2050,

DK4601, DK4604, DK4564, DK4612, DK3621, and DK3894 to create AW113, AW115,

AW119, AW121, AW123, AW125, RT02, and RT04 respectively (61).

amyE::Physpank-ppp-gfp – pAW93 was transformed into DK1042, DK2050,

DK3621, DK3894, DK4601, DK4604, DK4564, and DK4612, to create AW149, AW150,

AW156, AW157, AW158, AW159, AW160, and AW161 respectively (61).

107

amyE::Physpank-ppe-gfp – ppe-gfp was amplified from pAW93 using primers

1657/7514. The resulting fragment was ligated into SphI/NheI digested pDR111 using

Gibson assembly to generate pAW162. pAW162 was transformed into DK1042, DK2050,

DK3621, DK3894, DK4601, DK4604, DK4564, and DK4612, to create AW163, AW164,

AW165, AW166, AW167, AW168, AW169, and AW170 respectively.

amyE::Physpank-fabF – fabF was amplified from Bacillus subtilis 3610 genomic

DNA using primers 6102/6103. The resulting fragment was ligated into SphI/NheI digested pDR111 using Gibson assembly to generate pAW143. pAW143 was transformed into

DK1042 to generate AW146.

amyE::Physpank-fabG – fabG was amplified from Bacillus subtilis 3610 genomic

DNA using primers 6104/6105. The resulting fragment was ligated into SphI/NheI digested pDR111 using Gibson assembly to generate pAW144. pAW144 was transformed into

DK1042 to generate AW147.

amyE::Physpank-fabG – fabG was amplified from Bacillus subtilis 3610 genomic

DNA using primers 6106/6107. The resulting fragment was ligated into SphI/NheI digested pDR111 using Gibson assembly to generate pAW145. pAW145 was transformed into

DK1042 to generate AW148.

3.4.2 YmfI suppressor screen

Transposon delivery vectors for TnYLB and TnHyJump were introduced into the non- swarming, sensitized ymfI mutant background DK3789 by SPP1-mediated phage transduction followed by selection for mls resistance at room temperature. Both delivery vectors contain a transposon encoding a kanamycin resistance cassette as well as a

108 temperature sensitive origin that allows for replication at room temperature but not 42°C in B. subtilis (76, 77). The resulting colonies were used to inoculate 26 separate 3 ml LB cultures and transposon mutagenesis was allowed to occur by incubation at room temperature overnight. Mutants with transposon insertions in the genome were selected by incubating cells at 42°C on LB plates containing kanamycin. Approximately 10,000 of the resulting colonies from each pool were combined and each pool was used to inoculate separate swarming motility agar plates. Following a 7-10 hour incubation at 37°C, swarming proficient mutants emerged from the site of inoculation as a disk of motile cells, colonies were isolated, and the transposon was backcrossed to verify that suppression of ymfI was inseparably linked to the transposon insertion.

To determine the location of the transposon insertion sites, genomic DNA was isolated from each backcrossed suppressor strain, digested with either Sau3A1 or TaqA1 restriction enzymes, and ligated using T4 ligase to create circular fragments. Primer pairs

695/696 and 2567/2818, which anneal to TnYLB and TnHyJump, respectively, and direct polymerization outwards from the transposon were used to PCR amplify the neighboring

DNA. Resulting DNA fragments were subsequently sequenced with 696 and 2567 respectively to determine the transposon insertion site.

3.4.3 Swarming motility assay.

Cells were grown to mid logarithmic phase in LB at 37°C and concentrated to an

OD600 of 10 in PBS pH7.4 (0.8% NaCl, 0.02% KCl, 100 mM Na2HPO4, and 17.5 mM

KH2PO4) plus 0.5% India ink. Cell resuspensions were used to centrally inoculate 0.7% agar LB plates that had been dried for 10 minutes open-faced in a laminar flow hood.

109

Swarm plates were dried an additional 12 minutes after inoculation. Plates were incubated at 37C and swarm radius was monitored along the same axis every 30 minutes for 5.5 hours.

3.4.4 Isoelectric focusing

Strains were grown in 5 ml LB at 37°C with shaking. When cultures had reached mid-log phase, cells were collected and lysed in 25μl lysis buffer (10% glycerol, 25 mM

Tris pH 7.4, 100 mM NaCl, cOmplete mini EDTA-free protease inhibitor tablet (Roche),

1 mg/ml lysozyme, 1.5 units DNase I). Isoelectric focusing gels were prepared as previously described, with an ampholyte range of 4.0-6.5 (Rajkovic et al. 2016). Prior to sample loading, the isoelectric focusing gel was pre-focused at 100V for 45 minutes.

Following sample loading, the bands were resolved at 200V for 1 hour, 300V for 1 hour, and 500V for 30 minutes. Gels were soaked in towbin buffer for 15 minutes, and then transferred to nitrocellulose paper for western blotting. Blotting was completed with a

1:40,000 dilution of anti-EFP polyclonal antisera primary antibody and 1:5,000 dilution of goat anti-Rabbit conjugated to horseradish peroxidase secondary antibody. Blots were developed with BioRad Clarity ECL Substrate.

3.4.5 Semi-native gel electrophoresis

Strains were grown to mid-log phase, concentrated to an OD600 of 10 in lysis buffer

(17.2 mM Tris pH 7.0, 8.6 mM EDTA pH 8.0, 1 mg/mL Lysozyme, 0.1 mg/mL RNaseA,

110 were loaded onto 10% polyacrylamide native (with no added SDS) or 15% polyacrylamide denaturing (with 0.1% SDS) gels. Lysates were resolved at 150 V for 1.25 hours, transferred onto nitrocellulose membranes, and subsequently probed with a 1:40,000 dilution of anti-EF-P or a 1:80,000 dilution anti-SigA polyclonal antiserum. Following incubation with the primary antibodies, nitrocellulose membranes were probed with horseradish peroxidase conjugated goat anti-rabbit immunoglobulin G. Blots were developed using Pierce ECL substrate (Thermo Fisher Scientific).

3.4.6. Mass spectrometry

EF-P-FLAG was purified from DK2448, DK4313, DK4310, DK4572, DK4573, and DK3908. Saturated overnight cultures were back-diluted 1:1000 in 1.5 L LB. When cultures reached mid-log phase, EF-P-FLAG expression was induced with 1mM IPTG for

3 hours. Cells were then collected and lysed in 5 ml lysis buffer (50 mM Tris pH7.4, 150 mM NaCl, 1 mg/ml Lysozyme, cOmplete mini EDTA-free protease inhibitor tablet

(Roche)) for 1 hour at 37°C. EF-P-FLAG was purified with anti-FLAG M2 magnetic beads

(Sigma-Aldrich) at 4°C, following manufacturers instructions with minor alterations. EF-

P was eluted with 100μg/ml FLAG peptide (Sigma-Aldrich), concentrated, and resolved on a 13% SDS-PAGE gel. Bands were visualized with colloidal coomassie stain and excised for in-gel digestion with trypsin.

For DK2448, DK4313, DK4310, DK4572, and DK4573, the generated peptide samples were brought up in 2% acetonitrile in 0.1% formic acid (20 μL) and analyzed (2

μL) by LC/ESI MS/MS with a Thermo Scientific Easy-nLC II (Thermo Scientific,

Waltham, MA) coupled to a hybrid Orbitrap Elite ETD (Thermo Scientific, Waltham, MA)

111 mass spectrometer. In-line de-salting was accomplished using a reversed-phase trap column (100 μm × 20 mm) packed with Magic C18AQ (5-μm 200Å resin; Michrom

Bioresources, Auburn, CA) followed by peptide separations on a reversed-phase column

(75 μm × 250 mm) packed with Magic C18AQ (5-μm 100Å resin; Michrom Bioresources,

Auburn, CA) directly mounted on the electrospray ion source. A 40-minute gradient from

2% to 40% acetonitrile in 0.1% formic acid at a flow rate of 400 nL/minute was used for chromatographic separations. A spray voltage of 2750 V was applied to the electrospray tip and the Orbitrap Elite instrument was operated in the data-dependent mode, switching automatically between MS survey scans in the Orbitrap (AGC target value 1,000,000, resolution 120,000, and injection time 250 milliseconds) with MS/MS spectra detected in the Orbitrap (AGC target value of 50,000, 15,000 resolution and injection time 250 milliseconds). The 10 most intense ions from the Fourier-transform (FT) full scan were selected for fragmentation in the Orbitrap using ETD 100 ms activation time with supplemental CID activation with normalized collision energy of 35%. Selected ions were dynamically excluded for 10 seconds.

For DK2448 and DK3908, The generated peptide samples were brought up in 2% acetonitrile in 0.1% formic acid (20 μL) and analyzed (2 μL) by LC/ESI MS/MS with a

Thermo Scientific Easy-nLC II (Thermo Scientific, Waltham, MA) coupled to a hybrid

Orbitrap Fusion (Thermo Scientific, Waltham, MA) mass spectrometer. Peptides were separated on a reversed-phase column (75 μm × 370 mm) packed with Magic C18AQ (5-

μm 100Å resin; Michrom Bioresources, Auburn, CA) directly mounted on the electrospray ion source. A 40-minute gradient from 2% to 40% acetonitrile in 0.1% formic acid at a

112 flow rate of 300 nL/minute was used for chromatographic separations. A spray voltage of

2100 V was applied to the electrospray tip and the Orbitrap Fusion instrument was operated in the data-dependent mode, switching automatically between MS survey scans in the

Orbitrap (AGC target value 400,000, resolution 120,000, and injection time 50 milliseconds) with MS/MS spectra detected in the Orbitrap (AGC target value of 50,000,

15,000 resolution and injection time 22 milliseconds). The 10 most intense ions from the

Fourier-transform (FT) full scan were selected for fragmentation in the Orbitrap using ETD charge dependent activation time with=al CID activation with normalized collision energy of 35%. Selected ions were dynamically excluded for 10 seconds. Modifications detected when DK2448 was analyzed on the Orbitrap Fusion were the same as modifications detected when analyzed on the Orbitrap Elite ETD. MS/MS data is only presented for

DK2448 analyzed on the Orbitrap Elite ETD, though both data sets are available upon request (Fig 18j-l). For all samples, peptides were mapped using Proteome Discoverer

(Thermo Scientific).

3.4.7. GFP reporter assay.

Saturated overnight cultures were back diluted 1:1000 in LB with 100 μg/ml spectinomycin and grown at 37°C with shaking. When cultures reached mid-log phase,

GFP expression was induced with 1 mM IPTG for 1 hour. Following induction, 1 ml of cells was collected from each culture and washed once in PBS. Fluorescence was then measured with a Horiba Fluorlog spectrofluorometer.

3.4.8. Phenotype microarray

113

Respiration of WT (DK1042) and efp (DK2050) were compared in the Biolog microbial plates PM1-20 according to manufacturer’s instructions. Both strains were grown on BUG+B agar plates overnight, and then subcultured a second time. Cells were collected from plates and suspended 1x IF-0a to an OD of 0.09. Cells were mixed with inoculating fluid additives where appropriate, and aliquoted into wells. Plates were incubated in the OmniLog PM System at 37°C for 24 hours, with readings taken every hour.

3.4.9. Antibiotic sensitivity assay

Overnight saturated cultures were back diluted to an OD600 of 0.01 in LB with 0-5 mg/ml sulfanilamide or 0-20 μg/ml puromycin. Cultures were grown at 37°C with shaking for 18 hours or 8 hours respectively. The optical density of each culture was measured after the indicated length of time.

3.4.10 Phylogenetic analysis

From the GenBank assembly summary file, we obtained faa files of all bacteria known to have their complete genomes sequenced. The removal of duplicate entries yielded a total of 4196 genomes from which we built a blast database using the ncbi-blast-

2.3.0+ software (78). The Bacillus subtilis 168 genes efp, yaaO, ymfI, ynbB, gsaB, yfkA, and ywlG were independently blasted against the database with an arbitrary e-value of

0.001 selected to be the cutoff. Exclusive presence of efp (with a lysine at position 32 or analogous to position 32), yaaO, ymfI, and ynbB were plotted across a taxonomic tree generated using ITOL (79). Access to the taxonomic tree can be found here: http://itol.embl.de/tree/13023848185272891498208300

114

Chapter 4

Regulation of EF-P post-translational modification increases transcript stability in

Bacillus subtilis

4.1 Introduction

Efficient protein synthesis requires a core set of universally conserved translation factors. Disruption of this basic translation machinery can have varied impact on different organisms. For example, the requirement for maintaining translation factor Elongation

Factor P (EF-P) varies across phyla (31). EF-P facilitates efficient translation of proteins containing polyproline motifs, and although the functional role of EF-P is conserved across different species, the physiological effects associated with loss of EF-P vary (4, 5, 7, 8). In many Gram-negative organisms, deletion of efp results in a wide range of pleiotropic phenotypes such as growth defects, aberrant motility, hypersensitivity to antibiotics andloss of virulence (45, 46, 54, 58, 59). In contrast, the Gram-positive organism Bacillus subtilis requires EF-P for very little, with the major phenotype associated with loss of efp being aberrant swarming motility (61).

In order to stimulate translation of polyproline motifs, EF-P continuously binds the ribosome between the P and E sites (6, 12). Upon binding a ribosome paused on a

115 polyproline motif, EF-P will stimulate peptide bond formation, allowing translation to proceed past the pause (7, 8). Post-translational modification (PTM) of EF-P on a conserved residue significantly enhances this activity. Although the position of the modification on EF-P is highly conserved, the structure can vary widely between different organisms. In organisms such as Escherichia coli and Salmonella enterica, EF-P is modified on Lys34 with R-β-lysine (44). This modification has been shown to increase the affinity of EF-P for the ribosome and stabilize the P-site tRNA in a conformation favorable for peptide bond formation (7, 8). In Pseudomonas aeruginosa and Neisseria meningitidis,

EF-P is modified on Arg32 with a cyclic rhamnose moiety (58-60). The way in which rhamnose alters EF-P activity is unclear, but EF-P has been shown to require this modification in vivo. In Bacillus subtilis, EF-P is modified on Lys32 with a 5- aminopentanol moiety that likely impacts EF-P activity through a mechanism similar to R-

β-lysine (61). EF-P appears to be constitutively modified in all examples characterized, and recent work has shown that mutation of the lysine modification residue to an arginine will partially restore swarming motility to a B. subtilis efp mutant, suggesting that these diverse modifications are not necessarily required for EF-P activity. In light of these results, it is unclear why such divergent EF-P modifications have been evolutionarily selected for as opposed to a fully active unmodified EF-P.

In the absence of a fully functional EF-P, ribosomes will pause on polyproline motifs (4, 5). If there is significant ribosomal traffic on the transcript when the lead ribosome pauses, the trailing ribosomes will queue behind the paused ribosome, covering the upstream region of the transcript (4, 5, 17-19). Ribosomal profiling and SILAC data in

116 efp mutants have indicated that while polyproline induced ribosomal queuing generally results in a marked decrease in protein production, there are instances in which such queuing occurs, but net protein levels are unaltered or even elevated (5). This suggests that alternative mechanisms must exist to fine tune gene expression in the absence of EF-P. We hypothesized that in the absence of efp, polyproline induced ribosomal queuing will directly inhibit 5’-3’ exoribonucleolytic degradation and increase the stability of the mRNA which could in turn increase net protein levels. Here, we investigate the relationship between EF-P relieved ribosomal queuing and transcript stability in B. subtilis, and uncover how regulation of EF-P PTM alters this dynamic. Our results indicate that the regulation of EF-P PTM can be used to protect a subset of mRNAs under stress, suggesting that basis for maintaining EF-P modification is to allow for fine-tuned regulation of gene expression.

4.2 Results

4.2.1 RNaseJ1 degraded transcripts display increased stability in the absence of EF-P.

For ribosomal queuing to positively impact transcript stability of any given mRNA, the transcript must be degraded by a 5’-3’ exonuclease that could be inhibited by ribosomal occupancy and also encode a polyproline motif near the 5’ end. In order to identify candidate transcripts that could meet these criteria, we analyzed a list of transcripts that are elevated in response to depletion of 5’-3’ exonuclease RNaseJ1. From this list, we pulled

863 transcripts that display at least a 2-fold increase in abundance in the rnjA knockdown strain. It is likely that these transcripts are degraded by RNaseJ1, though we cannot exclude the possibility that they are upregulated in the RNaseJ1 knockdown through indirect mechanisms. From these targets, we identified 108 transcripts that also encode a

117 polyproline motif within the first 150 AAs (Table 13). Transcripts that encode a polyproline motif this close to the N-terminus are the most likely to induce sufficient ribosomal queuing to block the 5’ end of the mRNA. We then analyzed the steady state mRNA levels in the absence of efp for five test transcripts from this list. Of the transcripts tested, mdr, polA, and bmrC displayed elevated steady state transcript levels in the absence of efp (Fig 25A). To confirm that this increase was due to elevated mRNA stability rather than an increase in gene expression, we inhibited transcription with 500 μg/ml rifampicin and monitored mRNA decay over the course of 12 or 18 minutes. In the absence of efp, each transcript displayed a significant increase in transcript half-life, indicating that each transcript displays increased stability in the absence of efp (Fig 25C-E, Table 14).

118

Gene Length PPX Motif Position amyC 276 PPL 116 araD 229 PPL 111 araE 464 PPA 143 arsB 346 PPL 59 arsC 139 PPH 93 bmr3 512 PPE 130 bmrC 585 PPK, PPF, PPG 36, 106, 389 ccdC 160 PPI 36 comK 192 PPI 79 cotO 227 PPK, PPK, PPA 45, 113, 147 cssR 225 PPV 125 dinB1 414 PPN 78 dltB 395 PPH 111 dnaB 472 PPL 112 ecsA 247 PPL 116 epsJ 344 PPQ 131 epsM 216 PPK 49 ftsW 403 PPV 148 garD 510 PPA, PPL 94, 111 gerM 366 PPQ 37 glcD 470 PPD 130 ktrB 445 PPQ 16 lexA 205 PPS, PPD 25, 113 lmrB 479 PPH 135 lrpC 144 PPS 33 moaD 77 PPV 72 nap 300 PPL 56 nasA 401 PPH 131 nos 363 PPE, PPI 114 obg 428 PPG 90 pdaA 263 PPD, PPR 41, 163 pgcA 581 PPE, PPK, PPQ 148 pnbA 489 PPV, PPE, PPY 32, 44, 392 polA 880 PPE 83 pucB 205 PPF, PPI 70, 144 pucJ 449 PPV 122 pucK 430 PPV, PPT 124, 226 pyrP 435 PPV, PPT 124, 294 qcrA 167 PPL 139 qcrC 255 PPL 65 qdoI 337 PPY 140 rapJ 373 PPF 77 sbcD 391 PPA 55 scpA 251 PPS 128 spmA 196 PPD 83 spoIIP 401 PPP, PPK 141 spoIVFA 264 PPS 36 tepA 245 PPQ 59 thyA2 264 PPC 144 xynC 422 PPS 118 ybfP 295 PPG 99 ycbJ 306 PPS, PPH 122, 207 ycbK 312 PPM 67

Table 13. Transcripts elevated in the absence of RNaseJ1 that encode a polyproline motif in the first 150 amino acids. Continued 119

Table 13. Continued

Gene Length PPX Motif Position ycbR 243 PPV, PPA 109, 226 yceB 331 PPD 147 ydeF 462 PPQ 48 ydeH 148 PPL 77 ydeP 128 PPK 83 ydiP 389 PPW, PPI, PPA 105, 214, 227 yerO 289 PPK 77 yfhA 343 PPW 41 yfhI 397 PPE, PPL 128 yfiK 220 PPE 126 yfjO 466 PPC, PPR 80, 395 ygaD 580 PPV, PPL 76, 497 ygaJ 230 PPS 72 yhcY 379 PPA 65 yisB 100 PPE 83 yitR 97 PPL 77 yjbI 132 PPL 60 yjcD 759 PPE, PPE 51, 179 yjiB 396 PPK 86 yjkB 250 PPE 134 ykyB 154 PPT 86 ylmG 90 PPI 58 yloA 572 PPM, PPA 74, 167 ymdB 264 PPL 122 yocL 110 PPE 83 yojB 78 PPG 52 yonJ 405 PPP 130 yopD 131 PPN 97 ypcP 296 PPV 85 ypjP 203 PPA, PPH, PPN 23, 163, 197 ypuF 174 PPK 34 yqcF 192 PPF 130 yqcK 146 PPS 136 yqeC 297 PPR 59 yqfB 139 PPR 123 yqfO 373 PPI 70 yqjL 253 PPY 94 yrzH 43 PPQ 37 ysmB 146 PPQ 38 ytcQ 498 PPK 54 ytdP 772 PPL, PPQ 79, 350 yuiF 442 PPL 131 yurJ 367 PPK 71 yusU 95 PPL 60 yvbV 305 PPL 35 yveL 227 PPN, PPL 132, 163 yvgM 227 PPT 64 yvnB 1289 PPQ, PPK, PPA 48, 793, 914 yvrN 409 PPV 83 yvrO 229 PPK 139 ywdE 177 PPA 117 ywqO 73 PPV 24 ywrO 175 PPL, PPY 72, 144 yxiI 162 PPD, PPI 36, 155 yybR 125 PPK 79

120

Figure 25. Transcript stability in the absence of efp. Steady state transcript levels (A,B) and mRNA decay (C-H) were measured with qRT-PCR in WT (DK1042) and Δefp (DK2050), mdrP131A (AW194), mdrP131A Δefp (AW185), polAP84A (AW188), polAP84A Δefp (AW197), bmrCP37A (AW190), and bmrCP37A Δefp (AW196). (A,B) Expression was normalized to rpoA as a reference gene. Data were analyzed with a one- sample one-tailed t-test. *p<0.05. (C-H) mRNA decay of mdr (C,F), polA (D,G), and bmrC (E,H) were normalized to rrnB as a reference gene. White circles ○ are strains with efp, and black squares ■ are strains without efp. Data were analyzed with a t-test. *p<0.05.

121

Transcript WT Half Life (min) ∆efp Half Life (min) Fold Change mdr 11.2 +/- 9.1 35.6 +/- 5.2 3.2 polA 3.2 +/- 0.3 4.9 +/- 0.5 1.5 yheI 12.5 +/- 2.1 34.8 +/- 2.9 2.8

Table 14. Transcript half-lives in WT and ∆efp.

122

4.2.2 Increased stability is the result of ribosomal queuing and downregulation of

RNaseJ1.

As B. subtilis maintains over 900 polyproline containing proteins, deletion of efp can impact a wide range of cellular processes. In order to confirm that the increase in transcript stability in the efp mutant directly results from ribosomal queuing, we mutated the encoded polyproline motif in mdr, polA and bmrC so that each gene would encode a

PAX motif rather than a PPX motif. This mutation would eliminate ribosomal queuing on the transcript, even in the efp deletion strain. Although steady state transcript levels were still elevated in the efp deletion after mutation of the polyproline motif, the mRNA half- lives were markedly reduced relative to the wild-type control for polA and mdr (Fig 25F-

H, Table 15). However, we did still observe elevated transcript stability in the efp mutant even in the absence of ribosomal queuing. This indicates that although ribosomal queuing influences transcript stability, it cannot solely account for the differences between wild- type and ∆efp. Additional mechanisms must contribute to the increased transcript stability in the absence of EF-P. As RNaseJ1 contains a PPG motif, we hypothesized that RNaseJ1 requires EF-P for efficient translation and is downregulated in the efp mutant, further enhancing the increased transcript stability observed in the absence of efp. Recent ribosomal profiling data from wild-type and ∆efp B. subtilis indicates that there is an 11- fold increase in ribosomal occupancy at the encoded PPG motif on the rnjA transcript, indicating that RNaseJ1 does require EF-P for efficient translation (80). This data indicates

RNaseJ1 levels are likely decreased in the absence of efp, further exacerbating the increase in stability of RNaseJ1 targets.

123

Transcript WT Half Life (min) ∆efp Half Life (min) Fold Change mdrP131A 2.5 +/- 0.2 6.3 +/- 1.6 2.5 polAP84A 2.9 +/- 0.1 4.2 +/- 0.4 1.4 yheIP37A 8.9 +/- 1.4 32.8 +/- 14.8 3.7

Table 15. Transcript half-lives in WT and ∆efp after mutation of the polyproline motif.

124

4.2.3 Regulation of EF-P PTM correlates with increased transcript stability.

Given that loss of efp increases transcript stability, it is possible that B. subtilis manipulates mRNA turnover through downregulation of EF-P activity. In order to identify a growth condition in which EF-P activity could be downregulated, we analyzed expression data for mdr, polA and bmrC (72). If EF-P activity is negatively regulated under a specific growth condition, then transcript levels for each of these genes will be elevated in response.

Although each transcript displayed a distinct expression pattern across all conditions, all three transcripts were elevated under glucose exhaustion and early sporulation (72). We investigated both the expression and modification state of EF-P in cells grown in M9 minimal media and sporulation media. In early growth in M9 minimal media, B. subtilis expresses fully modified EF-P (Fig 26). However, after 24 hours of growth when glucose levels are depleted, EF-P is still expressed, but migrates as two separate bands on semi- native gel electrophoresis, suggesting a change in PTM state (Fig 26). We were unable to detect EF-P in spores (data not shown).

In order to confirm that the altered EF-P migration observed under glucose exhaustion on semi-native gel electrophoresis was the result of regulation of 5- aminopentanol, an EF-P-FLAG fusion was overexpressed and purified from cells grown for 4 hours or 24 hours in M9 minimal media. Purified protein was gel extracted and analyzed by tandem mass spectrometry. In both growth conditions, unmodified (Fig 27B,

E), 5-aminopentanolylated (Fig27C, F), and hydroxypentatonated (a 5-aminopentanol intermediate, Fig 27D, G) EF-P were observed. We also detected an acetyl group on K29

125 on a small number of peptides that also contained 5-aminopentanol on K32 (data not shown). The acetylation was not observed in the absence of 5-aminopentanol, and recent evidence suggests that EF-P acetylation can occur as an artifact of overexpression (74). No additional modifications were identified in either condition. As has been previously observed, overexpression of EF-P resulted in a significant amount of unmodified EF-P, making precise quantification of the relative levels of 5-aminopentanolylated EF-P at each time point impossible. Nonetheless, these results indicate that the regulation of EF-P under glucose exhaustion is in fact regulation of 5-aminopentanol and not another PTM.

126

Figure 26. EF-P expression and PTM state in M9 minimal media. WT (DK1042) B. subtilis were grown to mid log, 8 hours, or 24 hours in M9 minimal media. Lysates were resolved by SDS-PAGE or semi-native gel electrophoresis, transferred to nitrocellulose paper, and probed with EF-P antisera.

127

Figure 27. MS/MS spectra of EF-P peptide containing Lys32 in exponential or stationary phase in M9 minimal media. EF-P-FLAG was purified from DK2488 and subjected to chymotrypsin in-gel digestion, and analyzed via tandem mass spectrometry as previous described (74). (A) C/Z ions generated from ETD fragmentation of peptide containing Lys32. (B-G) Peptide containing Lys32 from cells in exponential (B-D) or stationary phase (E-G). Spectra are shown for unmodified peptide (B m/z=380.5503, E m/z=570.322), 5- aminopentanolylated peptide (C,F m/z =310.935), or hydroxypentatonated peptide (D, G m/z=310.678). Parent ions are z=2+ (E), z=3+ (B) or z=4+ (C, D, F, G). Continued

128

Figure 27. Continued

129

Given that both modified and unmodified EF-P were observed under glucose exhaustion, it was unclear if each cell within the population was maintaining both forms of the protein or if EF-P modification was stochastically regulated. To test these possibilities, wild-type B. subtilis expressing either a Phyperspank-gfp or Phyperspank-3xpro-gfp reporter construct were used as a read-out of EF-P activity. Cells were grown for 4 or 24 hours in

M9 minimal media and analyzed by flow cytometry. Cells that did not maintain either construct were used to set a cutoff for non-expression (Fig 28E). During exponential phase, nearly 100% of cells were able to express either construct (Fig 28A,B,F) . After 24 hours of growth, only 13% were able to express GFP (Fig 28C). This suggests that a portion of the population has a reduction in translation in this condition, as has previously been observed (81). A significantly higher proportion of cells (63%) were unable to express

3xPro-GFP (Fig 28D). The percentage of cells that were unable to express 3xPro-GFP correlates with the percentage of unmodified EF-P (Fig 26). This suggests that that the downregulation of 5-aminopentanol in a portion of the population inhibits efficient polyproline translation in those cells.

130

Figure 28. Flow cytometry analysis in exponential or stationary phase. B. subtilis expressing Phyperspank- gfp (A,C) or Phyperspank-3xpro-gfp (B,D) were grown for 4 hours (A,B) or 24 hours (C,D) in M9 minimal media. Cells lacking either construct (E) were analyzed to set cutoff for nonexpression (dashed line). (F) Quantification of flow cytometry data. 131

4.3 Discussion

4.3.1 EF-P relieved ribosomal queuing can increase stability of RNaseJ1 targets

Loss of EF-P is known to induce ribosomal queuing on transcripts that encode polyproline motifs (4, 5). If the encoded polyproline motif is near the N-terminus of the protein, this queuing can extend to the 5’ end of the transcript and inhibit translation initiation through occlusion of the start codon (5, 18). Here, we have demonstrated that this

EF-P relieved ribosomal queuing can also prevent transcript turnover through the inhibition of RNaseJ1 (Fig 29A). Although we cannot completely exclude the possibility that queued ribosomes stabilize transcripts through the inhibition of additional nucleases, we find this unlikely. Ribosomal queuing is known to cause an asymmetric distribution of ribosomes along the transcript, resulting in a significant amount of unprotected transcript downstream of the pause (4, 5). This region of the mRNA would then be available for endonucleolytic attack. Several examples suggest that ribosomal queuing enhances endonucleolytic cleavage. Ribosomal pausing has been shown to result in an increase in ribosome associated endonuclease 1 (Rae1) endonucleolytic attack, and deletion of Rae1 suppresses the efp swarming phenotype (82, 83). This suggests that EF-P relieved ribosomal queuing can increase endonucleolytic cleavage of Rae1 degraded transcripts. In organisms that do not maintain a 5’-3’ exonuclease, such as E. coli, ribosomal pausing has been shown to increase transcript turnover (84, 85). As RNaseJ1 is the only known 5’-3’ exonuclease in

B. subtilis, it is likely that the increase in transcript stability in the absence of efp results from inhibition of RNaseJ1. It is also important to note that RNaseJ1 maintains both endo- and exonucleolytic activity (86). From these results, we cannot exclude the possibility that

132 ribosomal queuing inhibits the endonucleolytic activity of this enzyme instead of or in addition to the exonucleolytic activity. However, due to the aforementioned reasons, it is unlikely that the endonucleolytic activity of RNaseJ1 is inhibited by ribosomal queuing.

Together, these examples suggest that the exonucleolytic actvitiy of RNaseJ1 is inhibited by ribosomal queuing (Fig 29A).

Although the relative transcript stability in the efp mutant decreased after mutation of the polyproline motif for mdr and polA, this was not the case for bmrC. It is possible that alternative mechanisms exist to further enhance the stability of this transcript that do not apply to mdr or polA. It is also important to note that bmrC also encodes two other polyproline motifs. It is possible multiple polyproline motifs encoded on a single transcript can compound such stabilization effects, and that mutation of the first PPE motif was not sufficient to alleviate this effect.

The impact that ribosomal queuing has on target transcript stability in the absence of EF-P is further exacerbated by the concomitant downregulation of RNaseJ1 itself (Fig

29A). As RNaseJ1 is required for processing of 16S rRNAs, it is possible that a portion of the ribosomes in the efp mutant strain maintain 16S precursors, though it is also possible that the RNaseJ1 remaining in the efp mutant is sufficient to maintain an appropriate pool of processed 16S rRNA (87). Nonetheless, such precursors resulting from RNaseJ1 depletion do not appear to inhibit ribosome function, as they are found within polysomes and polysome profiles are unaltered by the presence or absence of RNaseJ1 (87). The impact that such ribosomal alteration could have in the absence of efp is unclear.

133

4.3.2 Increase in transcript stability correlates with regulation of EF-P

Our results indicate that deletion of efp increases the stability of mdr, polA and bmrC. Transcript levels of each of these targets are elevated under glucose exhaustion, a condition in which EF-P PTM state is negatively regulated (72). This correlation suggests that regulation of EF-P PTM fine-tunes transcript turnover of these targets under starvation conditions. We have identified 108 transcripts that could fit this model of EF-P regulation.

Of these transcripts, a substantial portion is transport/membrane proteins, sporulation proteins, and transcriptional regulators (Fig 29B). It is possible that under negative EF-P regulation, these target transcripts are protected, allowing for distinct physiological changes such as initiation of sporulation or induction of competence. Future studies will center on understanding how mis-regulation of EF-P PTM alters cellular physiology under starvation conditions.

Six enzymes are known to influence the PTM state of EF-P and could possibly play role in this regulation. While most of these enzymes are still expressed under glucose exhaustion, transcriptomic data indicates that ymfI transcript levels are depleted during this stress (72). Though it is tempting to speculate that regulation of ymfI directly results in the regulation of EF-P PTM, ectopic expression of ymfI under glucose exhaustion did not alter the PTM state of EF-P (data not shown). It is instead likely that the regulation of EF-P

PTM results from depletion of the modification substrate during starvation. Limitation of the substrate would allow for the regulatory response to directly result from the change in environmental conditions. As the initial substrate for 5-aminopentanol synthesis is unknown, this idea remains to be tested.

134

It has previously been shown that polyproline induced ribosomal pausing does not necessarily result in a decrease in protein levels, suggesting that alternative mechanisms must influence net protein production in the absence of EF-P (5). Here, we have demonstrated that depletion of EF-P can increase transcript stability through ribosomal mRNA protection and downregulation or RNaseJ1 (Fig 29A). As many other bacteria do not maintain a 5’-3’ exonuclease, it is unlikely that this specific mode of mRNA protection is universally conserved. It is possible that other bacteria instead use alternative mechanisms to ameliorate the effects of efp depletion. Nonetheless, our results indicate that regulation of EF-P PTM can enhance transcript stability. This regulation suggests that the purpose of the PTM of EF-P is to fine-tune translational control of gene expression.

135

Figure 29. A model for EF-P dependent regulation of transcript stability. (A) Proposed model for ribosomal queuing increasing transcript stability. In the absence of EF-P, RNaseJ1 is downregulated and queued ribosomes inhibit RNaseJ1 activity. (B) Distribution of transcripts that fit this model.

136

4.4 Materials and Methods

4.4.1 Growth conditions and strain construction

Growth conditions and strain construction. Bacteria were grown in either luria broth

(5g NaCl, 5g yeast extract, and 10g tryptone per liter) or M9 minimal media (56.4mM

Na2PO4, 22mM KH2PO4, 8.6 mM NaCl, 18.7 mM NH4Cl, 1mM MgSO4, 100µM CaCl2,

0.2% glucose, and 600µM thiamine) supplemented with 100 μg/ml spectinomycin, 100

μg/ml ampicillin, or 25 µg/ml lincomycin + 1 µg/ml erythromycin (MLS) when appropriate. Strains were constructed as follows. All plasmids, strains, and primers used in this chapter are listed in tables 16, 17, and 18 respectively.

mdrP131A - Two DNA fragments with 500bp homology upstream and downstream of the desired mutation were PCR amplified from B. subtilis 3610 genomic DNA using primer pairs 9983/9984 and 9985/9986. The resulting fragments were ligated into SmaI digested pMiniMad2 with gibson assembly to generate pAW174. The resulting plasmid was transformed into DK1042, and the mutated copy of the gene was integrated through allelic replacement to generate AW194.

mdrP131A∆efp - Two DNA fragments with 500bp homology upstream and downstream of the desired mutation were PCR amplified from B. subtilis 3610 genomic DNA using primer pairs 9983/9984 and 9985/9986. The resulting fragments were ligated into SmaI digested pMiniMad2 with gibson assembly to generate pAW174. The resulting plasmid was

137 transformed into DK2050, and the mutated copy of the gene was integrated through allelic replacement to generate AW185.

polAP84A - Two DNA fragments with 500bp homology upstream and downstream of the desired mutation were PCR amplified from B. subtilis 3610 genomic DNA using primer pairs 9979/9980 and 9981/9982. The resulting fragments were ligated into SmaI digested pMiniMad2 with gibson assembly to generate pAW175. The resulting plasmid was transformed into DK1042, and the mutated copy of the gene was integrated through allelic replacement to generate AW188.

polAP84A ∆efp – pKRH3 (61) was transformed into AW188, and the efp deletion was integrated through allelic replacement to generate AW197.

yheIP37A - Two DNA fragments with 500bp homology upstream and downstream of the desired mutation were PCR amplified from B. subtilis 3610 genomic DNA using primer pairs 9975/9976 and 9977/9978. The resulting fragments were ligated into SmaI digested pMiniMad2 with gibson assembly to generate pAW176. The resulting plasmid was transformed into DK1042, and the mutated copy of the gene was integrated through allelic replacement to generate AW190.

yheIP37A∆efp - pKRH3 (61) was transformed into AW190, and the efp deletion was integrated through allelic replacement to generate AW196.

138

4.4.2 qRT-PCR analysis

Saturated overnight cultures of the indicated B. subtilis strains were back diluted

1:1000 in LB or 1:100 in M9 minimal media and grown to mid-log at 37°C with shaking.

Where indicated, cells were treated with 500 µg/ml rifampicin and harvested every 4 minutes for a total of 12 minutes. Following growth, cells were pelleted and resuspended in 500 µl RNA Later (Invitrogen) and stored overnight at 4°C. RNA was harvested with hot acid phenol chloroform extraction, treated with turbo DNAse (Invitrogen), and reverse transcribed using SuperScript IV reverse transcriptase (Invitrogen). cDNA was diluted

1:100 and quantified with qRT-PCR using SsoAdvanced Universal SYBR Green Supermix

(BioRad). Data was analyzed using ∆∆CT method. Significance was determined using a one-tailed one-sample t-test.

4.4.3 Semi-native and SDS polyacrylamide gel electrophoresis

A saturated overnight culture of DK1042 was back diluted to OD=0.05 in 3 separate tubes containing M9 minimal media. Cultures were grown to mid-log phase, 8 hours, or 24 hours. Following growth, cells were harvested and resuspended to OD=20 in lysis buffer

(10% glycerol, 25mM Tris HCl pH=7.4, 100mM NaCl, 2 units/ml DNaseI), and lysed 30 minutes at 37°C. Following clarification, lysates were resuspended in either SDS-PAGE or semi-native PAGE loading dye and boiled 5 minutes. Lysates were then run on a 13%

SDS-PAGE or 10% semi-native PAGE for 1.25 hours at 150V. The resulting gel was transferred to a nitrocellulose membrane and blotted with a 1:40,000 dilution of anti-EFP polyclonal antisera primary antibody.

139

4.4.4 Mass spectrometry

Saturated overnight cultures were back-diluted 1:100 in 1.5L M9 minimal media.

When cultures had grown to mid-log phase or for 24 hours (stationary phase), they were induced for 3 hours with 1mM IPTG. EF-P-FLAG was purified and analyzed by tandem mass spectrometry as previously described.

4.4.5 Flow cytometry

AW126, AW112 and AW149 were grown to mid-log or stationary phase (24hours) in M9 minimal media. Expression of GFP or 3xPro-GFP was induced for 1 hour with 1mM

IPTG. Cells were then diluted down to 500 cells/μl and resolved on a Guava easyCyte flow cytometer. Populations were analyzed using FlowJo data analysis software, version 10.

140

Plasmid Description Citation pAW174 pMiniMad2-mdrP131A This Publication pAW175 pMiniMad2-polAP84A This Publication pAW176 pMiniMad2-yheIP37A This Publication pKRH3 ΩΔefp mls amp Rajkovic, 2016

Table 16. Plasmids used in chapter 4.

141

Strain Genotype Reference 3610 Wild Type Bacillus subtilis 3610 DK1042 comIQ12L Konkol, 2013 DK2050 comIQ12L ∆efp Rajkovic, 2016

Q12L DK2448 comI ∆efp amyE::Physpank-efp-flag Rajkovic, 2016

Q12L AW112 comI amyE:::Physpank-GFP Rajkovic, 2016 AW126 comIQ12L amyE::spec Rajkovic, 2016

Q12L AW149 comI amyE:::Physpank-3P-GFP Rajkovic, 2016 AW185 comIQ12L ∆efp mdrP131A This Publication AW188 comIQ12L polAP84A This Publication AW190 comIQ12L yheIP37A This Publication AW194 comIQ12L mdrP131A This Publication AW196 comIQ12L ∆efp yheIP37A This Publication AW197 comIQ12L ∆efp polAP84A This Publication

Table 17. Strains used in chapter 4.

142

Number Sequence 9975 GACTCTAGAGGATCCCCAATTAACCTTTTCCTATTCAAGTAATGATTGACAATA AAAG 9976 GAGCTTTGCCGGAAACATTTCAATGACATTGACTG 9977 TTTCCGGCAAAGCTCTTGGGGAACGC 9978 TCGAGCTCGGTACCCGATTCCAGCACTCTGTCATTCAG 9979 GACTCTAGAGGATCCCCCCGCCTAATGTTTACAAAGGTTTAAC 9980 TTCGGAAAGCTCTGCCGGGGTTTTCTGTCTGCC 9981 GCAGAGCTTTCCGAACAAATGCC 9982 TCGAGCTCGGTACCCTCCGTCATAATTGTCGCAAGC 9983 GACTCTAGAGGATCCCCGACAGTCCTCCCTATAAAATCGAGTTTATC 9984 CGTTTTTCTGCCGGAAACAAATCAAAGATAATGG 9985 TCCGGCAGAAAAACGCGGAAAAATGTCC 9986 TCGAGCTCGGTACCCCTGCCGTAGACCGCC

Table 18. Primers used in chapter 4.

143

Chapter 5

Conclusions and Outlook

In all examples characterized to date, EF-P maintains diverse modifications on a highly conserved lysine/arginine residue (31, 32). Structural analyses, in vitro assays and phenotypic characterization of modification mutants have all built a strong case for the functional role of the majority of these modifications (31, 32). Despite this thorough characterization, the basis for the selection of diverse modifications rather than an EF-P that can function independent of modification is unclear. This question is particularly perplexing given that EF-P and its homologs appear to be constitutively modified and are often essential for viability (32). In this work, we have determined that regulation of EF-P

PTM allows for fine-tuned translational control of gene expression and propose that this regulation explains the need for modification rather than constitutive activity.

5.1. The Diversity of EF-P PTMs

Even with the addition of the 5-aminopentanol modification pathway to the list of known EF-P PTM pathways, we can only predict EF-P modification state for roughly 30% of known bacteria. This indicates that either the majority of bacteria do not require EF-P

PTM, or that the full list of modifications is in reality even more divergent than we have already observed. Both possibilities are intriguing, as the former would suggest that in

144 some instances, a constitutively active EF-P has been selected for and regulation is irrelevant, and the latter would suggest an even deeper divergence in potential modifications. Which of these two scenarios is correct can only be answered as more EF-

P modification pathways are uncovered.

When considering all of the known EF-P PTMs, it is evident that although they are required for the same general function, they are diverse in structure (31, 32). It is unusual to see such functional uniformity yet structural diversity. This disparity is even more apparent when considering the diversity in modification pathways and substrates. EF-P modifications can be derived from polyamines, amino acids, cyclic carbohydrates, or even fatty acids, as suggested by the 5-aminopentanol addition pathway discovered in this work

(31, 32). Despite this diversity, all known PTMs seem to enhance EF-P activity. The diversity in EF-P modification pathways introduces the question of whether EF-P simply requires a bulky modification that it can sequester from any free metabolites, or if instead each of these modifications has been specifically selected by evolution and optimized to function in the specific ribosome it evolved with. Differentiation of these two possibilities will require swapping differentially modified EF-P’s both in vitro and in vivo for functional and physiological characterization. Such experiments will reveal if each modification has the same impact on EF-P activity in each organism and what the advantages might be in maintaining one modification over another.

5.2 Regulation of EF-P PTM

Through a forward genetic screen based on the inhibitory nature of one of the 5- aminopentanol synthesis intermediates, we have identified the genes required for EF-P

145 modification in B. subtilis. In all other bacteria characterized to date, the genes required modification are in the same genomic neighborhood as efp (57-59). However, in B. subtilis, the identified modification genes are scattered throughout the genome, and instead sporulation genes surround EF-P. The genomic separation of these genes likely indicates unlike other organisms, B. subtilis does not necessarily co-express EF-P and all of the modification machinery under all conditions. For example, ywlG is upregulated in response to nitrogen limitation through the tnrA regulon, while yaaO expression is enhanced by the

SigW regulon in response to alkaline shock. Though in this work we only observed complete on/off regulation of the full 5-aminopentanol modification, the differential expression of the modification genes suggests that B. subtilis could also allow for more fine-tuned translational control with intermediate modifications under growth conditions that were not assessed in this work.

This is the first example to our knowledge of translational control through regulation of EF-P PTM, and it demonstrates the potential for these diverse PTMs to be used for regulatory purposes. In light of these results, efp genomic neighborhood appears to correlate with regulation. Bacteria that keep the modification genes near efp tend to co- express both and constitutively modify EF-P. Those that use EF-P PTM for regulatory purposes spread modification genes throughout the genome. It is likely that other organisms that do not keep efp in close proximity with its modification genes will also use

PTM for regulatory purposes.

It should also be noted that although EF-P and its homologs appear to be constitutively modified in most organisms, we cannot rule out the possibility that other

146 modes of EF-P PTM regulation exist and have simply not been observed, as it is impractical to assess EF-P modification state in every possible growth condition. For example, the essentiality of eIF5A and its hypusine modification would suggest that an organism must maintain both at all times to survive. However, in Fusarium graminearum, overexpression of DOHH results in attenuated virulence due to an overaccumulation of hypusinated eIF5A, whereas overexpression of DHS results in hypervirulence, presumably through the accumulation of deoxyhypusinated eIF5A (73). This result suggests that although eIF5A is essential and thought to be constitutively modified, hypusine and deoxyhypusine could play differential regulatory roles in eIF5A activity during the F. graminearum life cycle.

While we focused on the PTM of EF-P at the conserved lysine/arginine residue,

EF-P and its homologs can also be regulated through alternative PTMs. For example, eIF5A is phosphorylated in a number of eukaryotic systems (88-92). In Zea maize this phosphorylation is believed to impact subcellular localization, as a phosphomimetic aspartate mutation results in inappropriate nuclear localization of the protein (89, 90). In

Trypanosoma cruzi, eIF5A is phosphorylated under exponential phase and dephosphorylated in stationary phase (91). Phosphomimetic mutations result in an increase in protein synthesis and cell proliferation under exponential phase but cell death in stationary phase (91). This suggests that the regulation of the phosphorylation of eIF5A is important for differential growth, though it is unclear how this phosphorylation also impacts eIF5A activity. The role of alternative modifications on EF-P in bacteria is less well characterized. It has been established that EF-P can be phosphorylated in

Staphylococci after phage infection, but the functional consequences of this

147 phosphorylation are unknown (93). Nonetheless, these studies support the idea that EF-P and its homologs can be regulated by alternative post-translational mechanisms that remain to be discovered.

An important issue that was not addressed in this work is the consequences of misregulation of EF-P. If the evolutionary basis for maintaining an EF-P PTM rather than a fully active unmodified EF-P is truly for regulation, then there must be a condition in which it is detrimental to maintain a fully active EF-P. Based on the results presented here, we anticipate that a constitutively active EF-P would be disadvantageous under glucose exhaustion. However, without fully understanding the basis of EF-P PTM regulation under this condition, we were unable to perturb this regulation and investigate any possible negative effects. Future studies will center on the investigation of the negative consequences of a constitutively active EF-P.

5.3 Conclusion

In this work, we have uncovered a novel EF-P PTM pathway and demonstrated how negative regulation of this pathway can impact mRNA stability. These results demonstrate that regulation of EF-P PTM state can be used as a means to fine tune translational control and gene expression. Future studies will center on uncovering the full diversity of EF-P PTMs and understanding how these divergent modifications can be used for alternative regulation mechanisms.

148

References

1. Rodnina MV, Wintermeyer W. 2016. Protein Elongation, Co-translational Folding and Targeting. J Mol Biol 428:2165-85. 2. Rodnina MV. 2016. The ribosome in action: Tuning of translational efficiency and protein folding. Protein Sci 25:1390-406. 3. Pavlov MY, Watts RE, Tan Z, Cornish VW, Ehrenberg M, Forster AC. 2009. Slow peptide bond formation by proline and other N-alkylamino acids in translation. Proc Natl Acad Sci U S A 106:50-4. 4. Elgamal S, Katz A, Hersch SJ, Newsom D, White P, Navarre WW, Ibba M. 2014. EF-P dependent pauses integrate proximal and distal signals during translation. PLoS Genet 10:e1004553. 5. Woolstenhulme CJ, Guydosh NR, Green R, Buskirk AR. 2015. High-precision analysis of translational pausing by ribosome profiling in bacteria lacking EFP. Cell Rep 11:13-21. 6. Blaha G, Stanley RE, Steitz TA. 2009. Formation of the first peptide bond: the structure of EF-P bound to the 70S ribosome. Science 325:966-70. 7. Ude S, Lassak J, Starosta AL, Kraxenberger T, Wilson DN, Jung K. 2013. Translation elongation factor EF-P alleviates ribosome stalling at polyproline stretches. Science 339:82-5. 8. Doerfel LK, Wohlgemuth I, Kothe C, Peske F, Urlaub H, Rodnina MV. 2013. EF- P is essential for rapid synthesis of proteins containing consecutive proline residues. Science 339:85-8. 9. Doerfel LK, Wohlgemuth I, Kubyshkin V, Starosta AL, Wilson DN, Budisa N, Rodnina MV. 2015. Entropic Contribution of Elongation Factor P to Proline Positioning at the Catalytic Center of the Ribosome. J Am Chem Soc 137:12997-3006. 10. Gutierrez E, Shin BS, Woolstenhulme CJ, Kim JR, Saini P, Buskirk AR, Dever TE. 2013. eIF5A promotes translation of polyproline motifs. Mol Cell 51:35- 45. 11. Huter P, Arenz S, Bock LV, Graf M, Frister JO, Heuer A, Peil L, Starosta AL, Wohlgemuth I, Peske F, Nováček J, Berninghausen O, Grubmüller H, Tenson T, Beckmann R, Rodnina MV, Vaiana AC, Wilson DN. 2017. Structural Basis for Polyproline-Mediated Ribosome Stalling and Rescue by the Translation Elongation Factor EF-P. Mol Cell 68:515-527.e6. 12. Mohapatra S, Choi H, Ge X, Sanyal S, Weisshaar JC. 2017. Spatial Distribution and Ribosome-Binding Dynamics of EF-P in Live Escherichia coli. MBio 8:e00300-17. 149

13. Tollerson R, 2nd, Witzky A, Ibba M. 2017. Elongation Factor P Interactions with the Ribosome Are Independent of Pausing. MBio 8:e01056-17. 14. Katoh T, Wohlgemuth I, Nagano M, Rodnina MV, Suga H. 2016. Essential structural elements in tRNA(Pro) for EF-P-mediated alleviation of translation stalling. Nat Commun 7:11657. 15. Shin BS, Katoh T, Gutierrez E, Kim JR, Suga H, Dever TE. 2017. Amino acid substrates impose polyamine, eIF5A, or hypusine requirement for peptide synthesis. Nucleic Acids Res 45:8392-8402. 16. Buskirk AR, Green R. 2013. Biochemistry. Getting past polyproline pauses. Science 339:38-9. 17. Hersch SJ, Wang M, Zou SB, Moon KM, Foster LJ, Ibba M, Navarre WW. 2013. Divergent protein motifs direct elongation factor P-mediated translational regulation in Salmonella enterica and Escherichia coli. MBio 4:e00180-13. 18. Hersch SJ, Elgamal S, Katz A, Ibba M, Navarre WW. 2014. Translation initiation rate determines the impact of ribosome stalling on bacterial protein synthesis. J Biol Chem 289:28160-71. 19. Peil L, Starosta AL, Lassak J, Atkinson GC, Virumäe K, Spitzer M, Tenson T, Jung K, Remme J, Wilson DN. 2013. Distinct XPPX sequence motifs induce ribosome stalling, which is rescued by the translation elongation factor EF-P. Proc Natl Acad Sci U S A 110:15265-70. 20. McGary K, Nudler E. 2013. RNA polymerase and the ribosome: the close relationship. Curr Opin Microbiol 16:112-7. 21. Elgamal S, Artsimovitch I, Ibba M. 2016. Maintenance of Transcription- Translation Coupling by Elongation Factor P. MBio 7. 22. Schuller AP, Wu CC, Dever TE, Buskirk AR, Green R. 2017. eIF5A Functions Globally in Translation Elongation and Termination. Mol Cell 66:194-205.e5. 23. Pelechano V, Alepuz P. 2017. eIF5A facilitates translation termination globally and promotes the elongation of many non polyproline-specific tripeptide sequences. Nucleic Acids Res 45:7326-7338. 24. Bassani F, Zink IA, Pribasnig T, Wolfinger MT, Romagnoli A, Resch A, Schleper C, Blasi U, La Teana A. 2019. Indications for a moonlighting function of translation factor aIF5A in the crenarchaeum Sulfolobus solfataricus. RNA Biol. 25. Wagner S, Klug G. 2007. An archaeal protein with homology to the eukaryotic translation initiation factor 5A shows ribonucleolytic activity. J Biol Chem 282:13966-76. 26. Xu A, Chen KY. 2001. Hypusine is required for a sequence-specific interaction of eukaryotic initiation factor 5A with postsystematic evolution of ligands by exponential enrichment RNA. J Biol Chem 276:2555-61. 27. Xu A, Jao DL, Chen KY. 2004. Identification of mRNA that binds to eukaryotic initiation factor 5A by affinity co-purification and differential display. Biochem J 384:585-90.

150

28. Schrader R, Young C, Kozian D, Hoffmann R, Lottspeich F. 2006. Temperature-sensitive eIF5A mutant accumulates transcripts targeted to the nonsense-mediated decay pathway. J Biol Chem 281:35336-46. 29. Hoque M, Park JY, Chang YJ, Luchessi AD, Cambiaghi TD, Shamanna R, Hanauske-Abel HM, Holland B, Pe'ery T, Tian B, Mathews MB. 2017. Regulation of gene expression by translation factor eIF5A: Hypusine- modified eIF5A enhances nonsense-mediated mRNA decay in human cells. Translation (Austin) 5:e1366294. 30. Tollerson R, Witzky A, Ibba M. 2018. Elongation factor P is required to maintain proteome homeostasis at high growth rate. Proc Natl Acad Sci U S A 115:11072-11077. 31. Rajkovic A, Ibba M. 2017. Elongation Factor P and the Control of Translation Elongation. Annu Rev Microbiol 71:117–31. 32. Lassak J, Wilson DN, Jung K. 2016. Stall no more at polyproline stretches with the translation elongation factors EF-P and IF-5A. Mol Microbiol 99:219-35. 33. Schmidt C, Becker T, Heuer A, Braunger K, Shanmuganathan V, Pech M, Berninghausen O, Wilson DN, Beckmann R. 2016. Structure of the hypusinylated eukaryotic translation factor eIF-5A bound to the ribosome. Nucleic Acids Res 44:1944-51. 34. Park MH, Cooper HL, Folk JE. 1981. Identification of hypusine, an unusual amino acid, in a protein from human lymphocytes and of spermidine as its biosynthetic precursor. Proc Natl Acad Sci U S A 78:2869-73. 35. Park MH, Cooper HL, Folk JE. 1982. The biosynthesis of protein-bound hypusine (N epsilon -(4-amino-2-hydroxybutyl)lysine). Lysine as the amino acid precursor and the intermediate role of deoxyhypusine (N epsilon -(4- aminobutyl)lysine). J Biol Chem 257:7217-22. 36. Abbruzzese A, Park MH, Folk JE. 1986. Deoxyhypusine hydroxylase from rat testis. Partial purification and characterization. J Biol Chem 261:3085-9. 37. Prunetti L, Graf M, Blaby IK, Peil L, Makkay AM, Starosta AL, Papke RT, Oshima T, Wilson DN, de Crecy-Lagard V. 2016. Deciphering the Translation Initiation Factor 5A Modification Pathway in Halophilic Archaea. Archaea 2016:7316725. 38. Bassani F, Romagnoli A, Cacciamani T, Amici A, Benelli D, Londei P, Martens B, Blasi U, La Teana A. 2018. Modification of translation factor aIF5A from Sulfolobus solfataricus. Extremophiles 22:769-780. 39. Park JH, Wolff EC, Folk JE, Park MH. 2003. Reversal of the deoxyhypusine synthesis reaction. Generation of spermidine or homospermidine from deoxyhypusine by deoxyhypusine synthase. J Biol Chem 278:32683-91. 40. Park JH, Aravind L, Wolff EC, Kaevel J, Kim YS, Park MH. 2006. Molecular cloning, expression, and structural prediction of deoxyhypusine hydroxylase: a HEAT-repeat-containing metalloenzyme. Proc Natl Acad Sci U S A 103:51-6. 41. Fujimura K, Wright T, Strnadel J, Kaushal S, Metildi C, Lowy AM, Bouvet M, Kelber JA, Klemke RL. 2014. A hypusine-eIF5A-PEAK1 switch regulates the pathogenesis of pancreatic cancer. Cancer Res 74:6671-81. 151

42. Imam S, Mirmira RG, Jaume JC. 2014. Eukaryotic translation initiation factor 5A inhibition alters physiopathology and immune responses in a "humanized" transgenic mouse model of type 1 diabetes. Am J Physiol Endocrinol Metab 306:E791-8. 43. Park MH, Mandal A, Mandal S, Wolff EC. 2017. A new non-radioactive deoxyhypusine synthase assay adaptable to high throughput screening. Amino Acids 49:1793-1804. 44. Roy H, Zou SB, Bullwinkle TJ, Wolfe BS, Gilreath MS, Forsyth CJ, Navarre WW, Ibba M. 2011. The tRNA synthetase paralog PoxA modifies elongation factor- P with (R)-β-lysine. Nat Chem Biol 7:667-9. 45. Marman HE, Mey AR, Payne SM. 2014. Elongation factor P and modifying enzyme PoxA are necessary for virulence of Shigella flexneri. Infect Immun 82:3612-21. 46. Zou SB, Roy H, Ibba M, Navarre WW. 2011. Elongation factor P mediates a novel post-transcriptional regulatory pathway critical for bacterial virulence. Virulence 2:147-51. 47. Park JH, Johansson HE, Aoki H, Huang BX, Kim HY, Ganoza MC, Park MH. 2012. Post-translational modification by β-lysylation is required for activity of Escherichia coli elongation factor P (EF-P). J Biol Chem 287:2579-90. 48. Yanagisawa T, Sumida T, Ishii R, Takemoto C, Yokoyama S. 2010. A paralog of lysyl-tRNA synthetase aminoacylates a conserved lysine residue in translation elongation factor P. Nat Struct Mol Biol 17:1136-43. 49. Peil L, Starosta AL, Virumäe K, Atkinson GC, Tenson T, Remme J, Wilson DN. 2012. Lys34 of translation elongation factor EF-P is hydroxylated by YfcM. Nat Chem Biol 8:695-7. 50. Kobayashi K, Katz A, Rajkovic A, Ishii R, Branson OE, Freitas MA, Ishitani R, Ibba M, Nureki O. 2014. The non-canonical hydroxylase structure of YfcM reveals a metal ion-coordination motif required for EF-P hydroxylation. Nucleic Acids Res 42:12295-305. 51. Bullwinkle TJ, Zou SB, Rajkovic A, Hersch SJ, Elgamal S, Robinson N, Smil D, Bolshan Y, Navarre WW, Ibba M. 2013. (R)-β-lysine-modified elongation factor P functions in translation elongation. J Biol Chem 288:4416-23. 52. Peng WT, Banta LM, Charles TC, Nester EW. 2001. The chvH locus of Agrobacterium encodes a homologue of an elongation factor involved in protein synthesis. J Bacteriol 183:36-45. 53. Klee SM, Mostafa I, Chen S, Dufresne C, Lehman BL, Sinn JP, Peter KA, McNellis TW. 2018. An Erwinia amylovora yjeK mutant exhibits reduced virulence, increased chemical sensitivity and numerous environmentally dependent proteomic alterations. Mol Plant Pathol 19:1667-1678. 54. Zou SB, Hersch SJ, Roy H, Wiggers JB, Leung AS, Buranyi S, Xie JL, Dare K, Ibba M, Navarre WW. 2012. Loss of elongation factor P disrupts bacterial outer membrane integrity. J Bacteriol 194:413-25. 55. Bearson SM, Bearson BL, Brunelle BW, Sharma VK, Lee IS. 2011. A mutation in the poxA gene of Salmonella enterica serovar Typhimurium alters protein 152

production, elevates susceptibility to environmental challenges, and decreases swine colonization. Foodborne Pathog Dis 8:725-32. 56. Kaniga K, Compton MS, Curtiss R, 3rd, Sundaram P. 1998. Molecular and functional characterization of Salmonella enterica serovar typhimurium poxA gene: effect on attenuation of virulence and protection. Infect Immun 66:5599-606. 57. Navarre WW, Zou SB, Roy H, Xie JL, Savchenko A, Singer A, Edvokimova E, Prost LR, Kumar R, Ibba M, Fang FC. 2010. PoxA, yjeK, and elongation factor P coordinately modulate virulence and drug resistance in Salmonella enterica. Mol Cell 39:209-21. 58. Lassak J, Keilhauer EC, Fürst M, Wuichet K, Gödeke J, Starosta AL, Chen JM, Søgaard-Andersen L, Rohr J, Wilson DN, Häussler S, Mann M, Jung K. 2015. Arginine-rhamnosylation as new strategy to activate translation elongation factor P. Nat Chem Biol 11:266-70. 59. Rajkovic A, Erickson S, Witzky A, Branson OE, Seo J, Gafken PR, Frietas MA, Whitelegge JP, Faull KF, Navarre W, Darwin AJ, Ibba M. 2015. Cyclic Rhamnosylated Elongation Factor P Establishes Antibiotic Resistance in Pseudomonas aeruginosa. MBio 6:e00823. 60. Yanagisawa T, Takahashi H, Suzuki T, Masuda A, Dohmae N, Yokoyama S. 2016. Neisseria meningitidis Translation Elongation Factor P and Its Active- Site Arginine Residue Are Essential for Cell Viability. PLoS One 11:e0147907. 61. Rajkovic A, Hummels KR, Witzky A, Erickson S, Gafken PR, Whitelegge JP, Faull KF, Kearns DB, Ibba M. 2016. Translation Control of Swarming Proficiency in Bacillus subtilis by 5-Amino-pentanolylated Elongation Factor P. J Biol Chem 291:10976-85. 62. Kearns DB, Chu F, Rudner R, Losick R. 2004. Genes governing swarming in Bacillus subtilis and evidence for a phase variation mechanism controlling surface motility. Mol Microbiol 52:357-69. 63. Hummels KR, Witzky A, Rajkovic A, Tollerson R, Jones LA, Ibba M, Kearns DB. 2017. Carbonyl reduction by YmfI in Bacillus subtilis prevents accumulation of an inhibitory EF-P modification state. Mol Microbiol. 64. Hanawa-Suetsugu K, Sekine S, Sakai H, Hori-Takemoto C, Terada T, Unzai S, Tame JR, Kuramitsu S, Shirouzu M, Yokoyama S. 2004. Crystal structure of elongation factor P from Thermus thermophilus HB8. Proc Natl Acad Sci U S A 101:9595-600. 65. Balibar CJ, Iwanowicz D, Dean CR. 2013. Elongation factor P is dispensable in Escherichia coli and Pseudomonas aeruginosa. Curr Microbiol 67:293-9. 66. Schnier J, Schwelberger HG, Smit-McBride Z, Kang HA, Hershey JW. 1991. Translation initiation factor 5A and its hypusine modification are essential for cell viability in the yeast Saccharomyces cerevisiae. Mol Cell Biol 11:3105- 14. 67. Patel PH, Costa-Mattioli M, Schulze KL, Bellen HJ. 2009. The Drosophila deoxyhypusine hydroxylase homologue nero and its target eIF5A are

153

required for cell growth and the regulation of autophagy. J Cell Biol 185:1181-94. 68. Katz A, Solden L, Zou SB, Navarre WW, Ibba M. 2014. Molecular evolution of protein-RNA mimicry as a mechanism for translational control. Nucleic Acids Res 42:3261-71. 69. Cooper HL, Park MH, Folk JE, Safer B, Braverman R. 1983. Identification of the hypusine-containing protein hy+ as translation initiation factor eIF-4D. Proc Natl Acad Sci U S A 80:1854-7. 70. Sasaki K, Abid MR, Miyazaki M. 1996. Deoxyhypusine synthase gene is essential for cell viability in the yeast Saccharomyces cerevisiae. FEBS Lett 384:151-4. 71. Price AC, Zhang YM, Rock CO, White SW. 2001. Structure of beta-ketoacyl- [acyl carrier protein] reductase from Escherichia coli: negative cooperativity and its structural basis. Biochemistry 40:12772-81. 72. Nicolas P, Mader U, Dervyn E, Rochat T, Leduc A, Pigeonneau N, Bidnenko E, Marchadier E, Hoebeke M, Aymerich S, Becher D, Bisicchia P, Botella E, Delumeau O, Doherty G, Denham EL, Fogg MJ, Fromion V, Goelzer A, Hansen A, Hartig E, Harwood CR, Homuth G, Jarmer H, Jules M, Klipp E, Le Chat L, Lecointe F, Lewis P, Liebermeister W, March A, Mars RA, Nannapaneni P, Noone D, Pohl S, Rinn B, Rugheimer F, Sappa PK, Samson F, Schaffer M, Schwikowski B, Steil L, Stulke J, Wiegert T, Devine KM, Wilkinson AJ, van Dijl JM, Hecker M, Volker U, Bessieres P, et al. 2012. Condition-dependent transcriptome reveals high-level regulatory architecture in Bacillus subtilis. Science 335:1103-6. 73. Martinez-Rocha AL, Woriedh M, Chemnitz J, Willingmann P, Kroger C, Hadeler B, Hauber J, Schafer W. 2016. Posttranslational hypusination of the eukaryotic translation initiation factor-5A regulates Fusarium graminearum virulence. Sci Rep 6:24698. 74. Witzky A, Hummels KR, Tollerson II R, Rajkovic A, Jones LA, Kearns DB, Michael I. 2018. EF-P post-translational modification has variable impact on polyproline translation in Bacillus subtilis. mBio 9:e00306-18. 75. Peters JM, Colavin A, Shi H, Czarny TL, Larson MH, Wong S, Hawkins JS, Lu CHS, Koo BM, Marta E, Shiver AL, Whitehead EH, Weissman JS, Brown ED, Qi LS, Huang KC, Gross CA. 2016. A Comprehensive, CRISPR-based Functional Analysis of Essential Genes in Bacteria. Cell 165:1493-1506. 76. Le Breton Y, Mohapatra NP, Haldenwang WG. 2006. In vivo random mutagenesis of Bacillus subtilis by use of TnYLB-1, a mariner-based transposon. Appl Environ Microbiol 72:327-33. 77. Pozsgai ER, Blair KM, Kearns DB. 2012. Modified mariner transposons for random inducible-expression insertions and transcriptional reporter fusion insertions in Bacillus subtilis. Appl Environ Microbiol 78:778-85. 78. Letunic I, Bork P. 2011. Interactive Tree Of Life v2: online annotation and display of phylogenetic trees made easy. Nucleic Acids Research 39:W475- W478. 154

79. Boratyn GM, Camacho C, Cooper PS, Coulouris G, Fong A, Ma N, Madden TL, Matten WT, McGinnis SD, Merezhuk Y, Raytselis Y, Sayers EW, Tao T, Ye J, Zaretskaya I. 2013. BLAST: a more efficient report with usability improvements. Nucleic Acids Res 41:W29-33. 80. Hummels KR, Kearns DB. 2019. EF-P relieves translational pausing in FliY to promote swarming motility in Bacillus subtilis. Mol Microbiol. 81. Rosenberg A, Sinai L, Smith Y, Ben-Yehuda S. 2012. Dynamic expression of the translational machinery during Bacillus subtilis life cycle at a single cell level. PLoS One 7:e41921. 82. Hummels KR, Kearns DB. 2019. EF-P relieves translational pausing in FliY to promote swarming motility in Bacillus subtilis. Plos Genet. 83. Leroy M, Piton J, Gilet L, Pellegrini O, Proux C, Coppee JY, Figaro S, Condon C. 2017. Rae1/YacP, a new endoribonuclease involved in ribosome-dependent mRNA decay in Bacillus subtilis. Embo j 36:1167-1181. 84. Sunohara T, Jojima K, Tagami H, Inada T, Aiba H. 2004. Ribosome stalling during translation elongation induces cleavage of mRNA being translated in Escherichia coli. J Biol Chem 279:15368-75. 85. Hayes CS, Sauer RT. 2003. Cleavage of the A site mRNA codon during ribosome pausing provides a mechanism for translational quality control. Mol Cell 12:903-11. 86. Mathy N, Benard L, Pellegrini O, Daou R, Wen T, Condon C. 2007. 5'-to-3' exoribonuclease activity in bacteria: role of RNase J1 in rRNA maturation and 5' stability of mRNA. Cell 129:681-92. 87. Britton RA, Wen T, Schaefer L, Pellegrini O, Uicker WC, Mathy N, Tobin C, Daou R, Szyk J, Condon C. 2007. Maturation of the 5' end of Bacillus subtilis 16S rRNA by the essential ribonuclease YkqC/RNase J1. Mol Microbiol 63:127-38. 88. Kang HA, Schwelberger HG, Hershey JW. 1993. Translation initiation factor eIF-5A, the hypusine-containing protein, is phosphorylated on serine in Saccharomyces cerevisiae. J Biol Chem 268:14750-6. 89. Lebska M, Ciesielski A, Szymona L, Godecka L, Lewandowska-Gnatowska E, Szczegielniak J, Muszynska G. 2010. Phosphorylation of maize eukaryotic translation initiation factor 5A (eIF5A) by casein kinase 2: identification of phosphorylated residue and influence on intracellular localization of eIF5A. J Biol Chem 285:6217-26. 90. Lewandowska-Gnatowska E, Szymona L, Lebska M, Szczegielniak J, Muszynska G. 2011. Phosphorylation of maize eukaryotic translation initiation factor on Ser2 by catalytic subunit CK2. Mol Cell Biochem 356:241- 4. 91. Chung J, Rocha AA, Tonelli RR, Castilho BA, Schenkman S. 2013. Eukaryotic initiation factor 5A dephosphorylation is required for translational arrest in stationary phase cells. Biochem J 451:257-67. 92. Carvajal-Gamez BI, Quintas-Granados LI, Arroyo R, Mendoza-Hernandez G, Alvarez-Sanchez ME. 2012. Translation initiation factor eIF-5A, the hypusine- 155

containing protein, is phosphorylated on serine and tyrosine and O- glycosylated in Trichomonas vaginalis. Microb Pathog 52:177-83. 93. Depardieu F, Didier JP, Bernheim A, Sherlock A, Molina H, Duclos B, Bikard D. 2016. A Eukaryotic-like Serine/Threonine Kinase Protects Staphylococci against Phages. Cell Host Microbe 20:471-481.

156