IDENTIFICATION of Trna ELEMENTS IMPORTANT for ANTITERMINATION in the T BOX REGULATORY SYSTEM

IDENTIFICATION OF tRNA ELEMENTS IMPORTANT FOR ANTITERMINATION IN THE T BOX REGULATORY SYSTEM

DISSERTATION

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

Liang-Chun Liu, M.S.

Graduate Program in Microbiology

The Ohio State University

2014

Dissertation Committee:

Professor Tina M. Henkin, Advisor

Professor Irina Artsimovitch

Professor Charles J. Daniels

Professor Michael Ibba

Liang-Chun Liu

2014

ABSTRACT

Expression of many amino acid-related genes in Gram-positive bacteria is controlled by the T box regulatory mechanism. The 5’ untranslated region of the nascent

RNA (the “leader RNA”) of genes in the T box family senses the tRNA charging ratio in the cell through direct interaction with its cognate tRNA. This interaction regulates the expression of the downstream genes by determining whether the leader RNA folds into mutually exclusive terminator or antiterminator structures. Three residues within the leader RNA, termed the Specifier Sequence, resemble a codon that corresponds to the amino acid identity of the downstream genes. Both charged and uncharged tRNAs interact with the Specifier Sequence by codon-anticodon pairing, but only the free acceptor end of uncharged tRNA can form a second interaction with residues in a bulge domain of the antiterminator of the leader RNA; this interaction stabilizes the antiterminator and allows synthesis of the full-length transcript.

Previous mutational analysis revealed that pairings between the Specifier Sequence and the anticodon, and between residues in the antiterminator bulge of the leader RNA and the acceptor end of the tRNA, are necessary but not sufficient for efficient antitermination. The goal of this study is to identify tRNA residues outside of these known positions that contribute to specific interactions with the T box leader RNA. We used the Bacillus subtilis glyQS gene, which encodes glycyl-tRNA synthetase, to

ii characterize such elements in vitro. The Specifier Sequence-anticodon pairing mimics codon-anticodon pairing; therefore, tRNA mutations that result in misreading during translation may have an effect on antitermination. We selected tRNA mutations based on conservation and suppressor tRNA studies, and introduced them into tRNAGly to determine their effects on glyQS antitermination and binding. The results indicate that while some mutations that cause miscoding also affect antitermination, others are tolerated in antitermination. This part of the work provides new insight into the tRNA recognition requirements in antitermination relative to that in translation, and reveals different evolutionary constraints of these processes.

We also used the Clostridium acetobutylicum alaS gene, which encodes alanyl- tRNA synthetase, to show that the cognate tRNAAla directs efficient alaS antitermination, but the heterologous B. subtilis tRNAAla (with the same anticodon and acceptor end sequences) does not. Base substitutions at positions that show sequence variations between these two tRNAs reveal complex intramolecular relationships. These experiments provide the first information about fine-tuning of tRNA to form specific interactions with its cognate leader RNA. The results imply that the T box leader RNA has coevolved with its cognate tRNA in each organism to achieve specificity and efficiency during antitermination.

Because T box riboswitches are found in pathogenic Gram-positive bacteria and regulate several essential amino acid-related genes, this system has become a novel target for antibiotic design. Previous studies have identified lead compounds that bind to the

iii antiterminator model RNA with affinity and specificity. Here, we used in vitro and in vivo assays to test the effect of these compounds on antitermination.

Overall, the roles of T box riboswitches in tRNA evolution and clinical application are revealed.

DEDICATION

This document is dedicated to my family

– Teh-Ho Liu, Ying-Chuan Liao, Chung-Yuan Liu

ACKNOWLEDGEMENTS

I would like to thank my advisor Dr. Tina Henkin for her mentorship throughout these years. It has been a great honor to work with such an outstanding scientist. I am deeply grateful for her help and guidance in developing critical thinking and writing skills. I would also like to thank Dr. Frank Grundy for sharing his expertise and providing his insights into my project. I would like to thank the past and present members in the Henkin lab for their support and help during my graduate life. In particular, I would like to thank Anya Sherwood, Becky Williams-Wagner and Kiel

Kreuzer for their helpful discussions on the T box project, and Chris Woltjen for offering suggestions to all the problems I encountered.

I would like to thank all my committee members Dr. Kurt Fredrick, Dr. Michael

Ibba, Dr. Irina Artsimovitch for their insightful input into my project. I am also thankful for their warm support and encouragement. In particular, I would like to thank Dr.

Fredrick for discussion on the calculation of K1/2 and Kd. I would also like to thank Dr.

Charles Daniels for being on my dissertation committee.

Finally, I would like to thank all the friends I made in Columbus, as well as friends back in Taiwan for their companionship and friendship. I would also like to thank my parents and brother in Taiwan for their endless support and love, which helped me to get through a lot of difficult times.

VITA

October 22, 1981 ...... Born – Taipei, Taiwan

2000...... B.S. Clinical Laboratory Sciences and Medical Biotechnology, National Taiwan University

2004...... M.A. Clinical Laboratory Sciences and Medical Biotechnology, National Taiwan University

2007 to present ...... Graduate Teaching and Research Associate, Department of Microbiology, The Ohio State University

PUBLICATIONS

Gly  Liu, L.C., Grundy, F.J., Henkin, T.M. Conserved elements in tRNA contribute to efficient antitermination of the Bacillus subtilis glyQS T box gene. (Manuscript in preparation)

 Caserta, E., Liu, L.C., Grundy, F.J., Henkin, T.M. Codon-anticodon recognition in the Bacillus subtilis glyQS T box antitermination system. (Manuscript in preparation)

 Liu, L.C., Grundy, F.J., Henkin, T.M. The alaS T box leader RNA in Clostridium acetobutylicum has coevolved with its cognate tRNAAla to achieve specificity and efficiency during antitermination. (Manuscript in preparation)

FIELDS OF STUDY

Major Field: Microbiology vii

TABLE OF CONTENTS

ABSTRACT ...... ii

DEDICATION ...... v

ACKNOWLEDGEMENTS ...... vi

VITA ...... vii

TABLE OF CONTENTS ...... viii

LIST OF TABLES ...... xi

LIST OF FIGURES ...... xii

LIST OF ABBREVIATIONS ...... xiv

CHAPTER 1 GENE REGULATION BY THE T BOX RIBOWITCH ...... 1 1.1 Discovery and Prevalence ...... 3 1.2 Mechanism ...... 4 1.3 Factor-independent transcription antitermination in vitro...... 9 1.4 Kinetic analysis of in vitro transcription antitermination ...... 12 1.5 Composition of T box leader RNAs ...... 14 1.6 tRNA requirements for antitermination and binding ...... 18 1.7. Structural transitions of the leader RNA and tRNA ...... 22 1.7.1 Leader RNA ...... 22 1.7.2 tRNA ...... 26 1.8 The T box system as an antibiotic target ...... 29 1.9 Research goal ...... 33

CHAPTER 2 ROLES OF CONSERVED tRNAGly ELEMENTS IN glyQS ANTITERMINATION ..... 36 viii

2.1 Introduction ...... 36 2.2 Materials and Methods ...... 41 2.2.1 Construction of glyQS and tRNA variants ...... 41 2.2.2 T7 RNAP transcription ...... 41 2.2.3 In vitro transcription antitermination assays ...... 42 2.2.4 Size-exclusion filtration tRNA binding assays ...... 43 2.3 Results ...... 46 2.3.1 U32C ...... 46 2.3.1.1 Effects on antitermination ...... 46 2.3.1.2 Effects on binding ...... 51 2.3.2 C30G-G40C ...... 55 2.3.3 C27G•G43A ...... 58 2.3.4 A26G•G44U ...... 58 2.3.5 U11C, A24G, U11C-A24G ...... 59 2.3.5.1 Effects on antitermination ...... 59 2.3.5.2 Effects on binding ...... 64 2.4 Discussion ...... 66

CHAPTER 3 tRNAAla RECOGNITION BY COGNATE AND NONCOGNATE T BOX LEADER RNAS ...... 74 3.1 Introduction ...... 74 3.2 Materials and Methods ...... 81 3.2.1 Generation of DNA templates ...... 81 3.2.2 In vitro transcription antitermination assays ...... 82 3.2.3 Genetic techniques ...... 84 3.2.4 Bacterial growth conditions and β–galactosidase assays ...... 85 3.3 Results ...... 86 3.3.1 tRNAAla-directed alaS antitermination in vitro ...... 86 3.3.2 tRNAAla requirements for alaS antitermination ...... 90 3.3.3 Expression of alaS-lacZ fusions in B. subtilis ...... 102 3.3.4 tRNA recognition by noncognate leader RNA ...... 104 3.4 Discussion ...... 109

CHAPTER 4 ANALYSIS OF COMPOUNDS DESIGNED TO TARGET THE T BOX ANTITERMINATOR ...... 117 4.1 Introduction ...... 117 4.2 Materials and Methods ...... 120 4.2.1 In vitro transcription antitermination assays ...... 120 4.2.2 Bacterial strains and growth conditions ...... 121 4.2.3 β–galactosidase assays ...... 122 4.3 Results ...... 123 4.3.1 Inhibition of glyQS antitermination ...... 123 4.3.2 Reduction of tyrS-lacZ expression ...... 126 4.4 Discussion ...... 131

CHAPTER 5 CONCLUSIONS AND FUTURE DIRECTIONS...... 137 5.1 Conclusions ...... 137 5.2 Future directions ...... 143

LIST OF REFERENCES ...... 148

APPENDIX A SUPPLEMENTAL MATERALS ...... 162

LIST OF TABLES

Table

2.1. Effect of U32C on binding affinity of the glyQS leader RNA and variants ...... 53

3.1. Oligonucleotide primers for alaS and glyQS leader RNA constructs...... 82

3.2. Oligonucleotide primers for construction of plasmid-borne tRNAAla variants...... 85

3.3. Hybrid tRNAAla ...... 101

3.4. Expression of alaS-lacZ fusions ...... 103

4.1. Effect of compounds on expression of tyrS-lacZ and YP-lacZ...... 129

A.1. Effect of tRNA mutations on antitermination efficiency (K1/2) of the wild-type glyQS leader RNA and leader variants ...... 163

A.2. Effects of tRNA mutations on RTmax of the wild-type glyQS leader RNA and leader variants ...... 164

A.3. Effect of mutations in tRNAGly(UCC)G34U on readthrough (%RT) of constructs with mismatches at position 1, or positions 1 and 3 of the Specifier Sequence ...... 165

A.4. Effect of tRNA mutations on binding affinity (Kd) of the wild-type glyQS leader RNA and leader variants ...... 166

A.5. Results of compound tests...... 168

LIST OF FIGURES

Figure

1.1. Secondary structural model of the B. subtilis tyrS leader RNA ...... 6

1.2. Transcription termination and antitermination of the T box riboswitch ...... 8

1.3. Secondary structural model of the B. subtilis glyQS leader RNA ...... 11

1.4. Model for tRNAGly recognition by the glyQS leader RNA ...... 25

1.5. Structural models of the Specifier Sequence-anticodon helix ...... 27

1.6. Superposition of leader RNA-bound tRNAGly with the structure of free tRNAPhe and tRNAPhe in P/P state ...... 28

1.7. Compound binding to the AM1A model RNA ...... 32

2.1. Structural model of B. subtilis glyQS leader RNA and tRNAGly ...... 40

2.2. In vitro transcription of the wild-type glyQS leader template with increasing amount of wild-type tRNAGly(GCC) ...... 45

2.3. Response of antitermination to mismatches between the Specifier Sequence and anticodon in the presence or absence of U32C ...... 48

2.4. Effect of U32C on binding of constructs with Specifier Sequence-anticodon pairing alterations at positions 1 or 3 ...... 54

2.5. Effect of C30G-G40C, A26G•G44U or C27G•G43A on antitermination...... 57

2.6. Effect of D arm mutations on antitermination ...... 61

2.7. Effect of U11C on binding of constructs with Specifier Sequence-anticodon pairing alterations at position 3 ...... 65

xii

3.1. The predicted structural models of (A) C. acetobutylicum and (B) B. subtilis alaS T box leader RNAs ...... 76

3.2. Cloverleaf diagrams of (A) B. subtilis tRNAAla(GGC), (B) B. subtilis tRNAAla(UGC) and (C) C. acetobutylicum tRNAAla(UGC) ...... 80

3.3. The alaS antitermination in vitro...... 88

3.4. Antitermination of the C. acetobutylicum alaS gene using B. subtilis tRNAAla(UGC) and C. acetobutylicum tRNAAla(UGC) ...... 91

3.5. Effects of tRNA mutations on antitermination ...... 94

Ala 3.6. The RTmax and K1/2 of 32 tRNA (UGC) constructs for C. acetobutylicum alaS antitermination...... 98

3.7. Antitermination of the alaS and glyQS genes with either the wild-type or mutant Specifier Sequence...... 108

4.1. Composition of lead compounds ...... 120

4.2. Effect of compounds on glyQS antitermination in vitro...... 125

A.1. Best-fit curves ...... 162

A.2. Effect of U32C in truncated tRNAGly(GCC)ΔUCCA on binding...... 167

xiii

LIST OF ABBREVIATIONS

%RT percent readthrough

°C degree Celsius

µ micro

A adenosine

Å angstrom aaRS aminoacyl-tRNA synthetase aa-tRNA aminoacyl-tRNA

Ala alanine

ATP adenosine triphosphate bp base pair

BSA bovine serum albumin

C cytidine

DMSO dimethyl sulfoxide

DNA deoxyribonucleic acid

DTT dithiothreitol

E L-glutamic acid

EDTA ethylenediaminetetraacetic acid

EF-Tu elongation factor-Tu

xiv

G guanosine

Gly glycine

GMP guanosine monophosphate

GTP guanosine triphosphate h hour

IPTG isopropyl-β-D-thiogalactopyranoside

K1/2 antitermination constant

KCl potassium chloride

Kd dissociation constant

M molar m milli min minute mRNA messenger RNA n nano na not determined

NMR nuclear magnetic resonance nt nucleotide

PCR polymerase chain reaction

Q L-glutamine

RNA ribonucleic acid

RNAP RNA polymerase

RNase ribonuclease

RT readthrough

RTmax maximum readthrough

SD Shine-Dalgarno sp. species

Tris-HCl tris-(hydroxylmethyl) aminomethane hydrochloride tRNA transfer RNA

U uridine

UTP uridine triphosphate

WT wild-type

xvi

CHAPTER 1

GENE REGULATION BY THE T BOX RIBOWITCH

All living cells have developed a variety of mechanisms to regulate gene expression in response to environmental changes and different physiological conditions. Expression of necessary genes under appropriate growth conditions is essential for cell survival.

Regulation can occur at different stages during gene expression, including transcription initiation (Browning and Busby, 2004), transcription termination (Henkin, 1996; Henkin and Yanofsky, 2002), translation initiation and elongation (Kaberdin and Blasi, 2006;

Shalgi et al., 2013), and the control of mRNA stability (Bernstein et al., 2002; Grunberg-

Manago, 1999; Waters and Storz, 2009). These processes usually utilize regulatory proteins. For example, transcription initiation commonly involves DNA-binding proteins that promote or repress the function of the RNA polymerase (RNAP) at the promoter region. Transcription termination can be regulated by RNA binding proteins that modulate the formation of an intrinsic transcription terminator. Other than regulatory proteins, small untranslated RNAs (sRNAs) act either in cis or in trans to control mRNA stability, the abundance of regulatory proteins, or translation initiation (Storz et al., 2005;

Storz et al., 2011). The discovery of riboswitch RNAs revealed another protein- independent mechanism for gene regulation. 1

Riboswitches are cis-acting nascent RNA transcripts that directly sense environmental signals (usually an effector molecule) and fold into an appropriate structure to control expression of the downstream genes without the aid of trans-acting protein factors (Grundy and Henkin, 2006; Henkin, 2008). Binding of the ligand causes a structural rearrangement of the RNA; this structural change controls the “on” or “off” state of gene expression by modulation of the formation of an intrinsic transcription terminator when regulation occurs at the level of transcription attenuation, or the sequestration of the Shine-Dalgarno (SD) sequence when regulation occurs at the level of translation initiation. Riboswitches are found in all three domains of life where they regulate essential genes in the cell. Recent studies have focused on using riboswitch

RNAs as new antimicrobial drug targets (Winkler and Breaker, 2005).

Riboswitches can be categorized into several different classes and subclasses based on the signal that is monitored (Breaker, 2012; Grundy and Henkin, 2006; Peselis and

Serganov, 2014; Serganov and Nudler, 2013). Temperature-sensing riboswitches, also known as RNA thermosensors, regulate the expression of genes in response to temperature change. Metabolite-sensing riboswitches can be divided into several groups based on their corresponding ligands, which include: 1) coenzymes and derivatives such as adenosylcobalamine (AdoCbi), thiamine pyrophosphate (TPP), flavin mononucleotide

(FMN), tetrahydrofolate (THF), S-adenosylmethionine (SAM-I, II, III), and S- adenosylhomocysteine (SAH); 2) amino acids such as lysine, glycine, and glutamine; and

3) purines and derivatives such as adenine, guanine, ATP, pre-queuosine-1, and cyclic-di-

GMP. Riboswitch RNAs can also sense the abundance of sugars, such as glucosamine-6- 2

phosphate (GlcN6P), and metal ions, such as Mg2+. Riboswitch RNAs usually use a feedback mechanism to monitor the requirement for gene expression (Henkin, 2008).

The accumulated ligand, which is the end-product of the pathway that is regulated, binds to the corresponding riboswitch RNA to induce formation of the alternative structure and results in gene repression. In some rare classes, a ligand that is the precursor for the regulated pathway binds to the riboswitch RNA and induces gene activation (Mandal and

Breaker, 2004). The T box riboswitches represent a special group of riboswitch RNAs in that the effector molecule is a tRNA, and the system measures the charging ratio of a specific tRNA rather than its absolute abundance (Grundy et al., 2005).

1.1 Discovery and Prevalence

The T box riboswitch was discovered by identification of a long 5’ untranslated region (leader RNA) upstream of the Bacillus subtilis tyrS gene, which encodes tyrosyl- tRNA synthetase (Henkin et al., 1992). This leader RNA contains an inverted repeat sequence, which was predicted to be an intrinsic transcription terminator. Upstream of the factor-independent terminator is a 14-nucleotide (nt) conserved sequence found in many other aminoacyl-tRNA synthetase (aaRS) genes and amino acid biosynthesis operons in Bacillus sp. (Condon et al., 1996a; Gendron et al., 1994; Grundy and Henkin,

1993; Luo et al., 1997; Putzer et al., 1995). This conserved sequence element was designated as the “T box”, which was used to name this regulatory system. Finding of the T box sequence upstream of many aaRS genes in Bacillus sp. suggested a common

regulatory mechanism. This is in contrast with the regulation of aaRS genes in

Escherichia coli, which utilizes multiple different regulatory mechanisms.

Genome-wide analyses of T box-regulated genes revealed that this regulatory mechanism is mainly found in Gram-positive bacteria, and rarely found in Gram-negative organisms (Gutierrez-Preciado et al., 2009; Vitreschak et al., 2008), including those in the phyla of Firmicutes, Actinobacteria, δ–Proteobacteria, Deinococcus-Thermus, and

Chloroflexi. These studies revealed that the T box elements regulate amino acid-related genes, including genes that encode aaRSs, proteins for amino acid biosynthesis and transport, and regulatory proteins for amino acid biosynthesis and transport. There is also a small group of T box-regulated genes of unknown function.

1.2 Mechanism

The initial studies used a tyrS-lacZ reporter gene in a tyrosine auxotrophic strain to demonstrate that tyrS expression is regulated at the level of transcription termination rather than transcription initiation, and readthrough of the intrinsic terminator is induced specifically by tyrosine starvation but not by limitation for other amino acids (Henkin et al., 1992). Similar transcription attenuation mechanisms were also reported in the B. subtilis ilv-leu operon (Grandoni et al., 1992), and the thrS and thrZ genes (Putzer et al.,

1992). A subsequent study that analyzed 10 leader RNAs of T box-containing aaRS genes revealed a common structural model (Grundy and Henkin, 1993). The secondary structures of these T box leader RNAs were manually predicted based on possible base pairing interactions and nucleotide covariation analysis. The canonical structure includes 4

Stem I, Stem II and Stem III, and a terminator structure (Fig. 1.1). An alternative antiterminator structure was also predicted. Mutational analysis further demonstrated the formation of a Stem IIA/B pseudoknot structure, which is located at the 3’ end of the

Stem II element (Rollins et al., 1997). Several conserved sequence and structural elements were also identified.

A codon, designated as the “Specifier Sequence,” was identified in the Specifier

Loop of each leader RNA that corresponds to the amino acid identity of the downstream gene (Grundy and Henkin, 1993). For example, the tyrS leader RNA has a UAC tyrosine codon at this position (Fig. 1.1), and the phenylalanyl-tRNA synthetase leader RNA has a

UUC phenylalanine codon. Mutation of the Specifier Sequence from a UAC tyrosine codon to a UUC phenylalanine codon in the context of a tyrS-lacZ transcriptional fusion resulted in loss of response to tyrosine limitation and a switch to a response to phenylalanine limitation, indicating that the amino acid specificity of the construct was altered. The discovery of a codon in the leader RNA that acts as a regulatory element led to the hypothesis that tRNA is the effector molecule for the system (Grundy and Henkin,

1993). Alteration of the tyrosine Specifier Sequence to an amber or ochre nonsense codon resulted in an uninducible phenotype, which suggested that no corresponding effectors in the cell can induce gene expression. A lysyl-tRNA variant with an ochre suppressor anticodon that matches the nonsense codon induced some expression of the amber and ochre nonsense mutants under lysine starvation conditions, indicating that gene expression is induced by the uncharged tRNA construct with matching anticodon.

These experiments demonstrated that tRNA is the effector. 5

Figure 1.1. Secondary structural model of the B. subtilis tyrS leader RNA. The sequence is shown and numbered from the transcription start-site (+1) through the end of the terminator; the alternative antiterminator structure is shown above the terminator. Structural domains, including Stem I, II, IIA/B, III and terminator, are labeled. Residues that form the pseudoknot structure are labeled in purple. The conserved sequence and structural elements, including the GA motif, S-turns, AG bulge, Specifier Loop and T box, are also shown. Asterisks indicate highly conserved nucleotides. Bases at the 5’ end of the terminator (blue) base-pair with part of the T box sequence (red) and stabilize the antiterminator. The UAC tyrosine Specifier Sequence (boxed) that pairs with the GUA anticodon of tRNATyr, and the UGGU sequence in the antiterminator bulge that pairs with the ACCA acceptor end of tRNATyr, are labeled in green. Adapted from (Green et al., 2010). 6

Addition of a U residue 5’ of the UAC tyrosine Specifier Sequence, which was predicted to shift the reading frame of the codon if translation is involved, had no effect on lacZ expression; this indicates that the codon-anticodon interaction in the T box system is independent of translation. Overexpression of a plasmid-borne unchargeable tRNA construct in rich medium resulted in efficient induction of lacZ expression, which indicates that expression of the gene is activated by an uncharged tRNA. Overexpression of a chargeable tRNA construct resulted in reduced induction in the presence of the corresponding uncharged tRNA, which indicates that the charged tRNA is able to repress gene expression. These results suggested that the T box leader RNA monitors both charged and uncharged tRNA, and only the uncharged tRNA induces gene expression

(Grundy et al., 1994).

A model was proposed based on these studies. The T box leader RNA monitors the tRNA charging ratio in the cell and controls the expression of the downstream genes (Fig.

1.2). The charging ratio of a tRNA reflects the cellular requirements for the corresponding aaRS and amino acid. When the appropriate tRNA is highly charged, and the requirement for expression of the regulated genes is low, the leader RNA forms an intrinsic transcription terminator that prevents transcription of the downstream coding sequence. Readthrough of the terminator requires the stabilization of a competing antiterminator structure, which is dependent on an additional interaction with the cognate uncharged tRNA. Regulation at the level of premature transcription termination is found primarily in low G+C Gram-positive bacteria. For high G+C Gram-positive and some

Gram-negative organisms, gene regulation occurs at the level of translation initiation 7

(Gutierrez-Preciado et al., 2009). The T box leader RNAs that regulate gene expression at the level of translation initiation form a helix that sequesters the SD sequence and prevents initiation of translation when tRNA charging ratio is high, and an anti-SD helix that releases the SD sequence and allows translation initiation under low tRNA charging conditions.

Figure 1.2. Transcription termination and antitermination of the T box riboswitch. Both charged and uncharged tRNAs (L-shape in gold color) interact with the T box leader RNA at the Specifier Sequence by codon-anticodon pairing. (A) When the tRNA charging ratio is high, a terminator (T) helix is formed that prevents expression of downstream genes. (B) When the tRNA charging ratio is low, the free 3’ acceptor end of the uncharged tRNA stabilizes the antiterminator (AT) by pairing with the antiterminator bulge, which allows transcription to continue.

Two major pairing interactions between the leader RNA and the tRNA were identified by mutational analysis. The anticodon of the tRNA base-pairs with the

Specifier Sequence of the leader RNA; this interaction resembles the codon-anticodon interaction in translation. Stabilization of the antiterminator requires additional pairings between the acceptor end of the uncharged tRNA and a complementary sequence in the bulge region of the antiterminator (Grundy et al., 1994). Charging of the tRNA prevents the antiterminator-acceptor end interaction through steric hindrance.

Two endonucleases, RNase J1/J2, were identified in B. subtilis that process the readthrough transcript (Condon et al., 1996b; Even et al., 2005). The major cleavage site was reported to be in the terminal loop of the antiterminator. Cleavage of the readthrough transcript is predicted to induce the formation of the terminator structure at the 5’ end of the mRNA and result in increased stability of the RNA transcript. This processing step also allows the release and recycling of the uncharged tRNA from the T box leader RNA.

1.3 Factor-independent transcription antitermination in vitro

Whether uncharged tRNA is sufficient for antitermination, or other cellular factors are required in the process, was tested using a purified in vitro transcription antitermination system. Previous attempts to reconstitute antitermination in vitro using the tyrS leader RNA were unsuccessful (Grundy et al., 2002a). Instead, tRNA-dependent antitermination was demonstrated using the leader RNA of the B. subtilis glyQS gene, which encodes glycyl-tRNA synthetase, and its cognate tRNAGly (Grundy et al., 2002b). 9

The glyQS leader RNA is a natural deletion variant that lacks the Stem II and IIA/B pseudoknot elements (Fig. 1.3). The Specifier Sequence of the glyQS leader RNA is a

GGC glycine codon, which forms stable G-C pairs with the cognate tRNAGly(GCC). The ability to demonstrate tRNA-mediated antitermination in a purified system without any additional cellular factors indicates that tRNA alone is able to interact with the leader

RNA for regulation.

Although tRNA-directed antitermination in vitro was also demonstrated using the leader RNA of the B. subtilis thrS gene (encoding threonyl-tRNA synthetase) (Putzer et al., 2002) and valS gene (encoding valyl-tRNA synthetase) (R.N. Williams-Wagner, F.J.

Grundy, T.M. Henkin, unpublished results), cellular extracts and fully modified tRNA purified from B. subtilis were used. The leader RNAs of both genes contain all the canonical structural features, including the Stem II and IIA/B elements, which suggests that antitermination of genes in this class may have requirements that are different from the natural deletion variants. Cellular extract can be replaced by a high concentration of spermidine in thrS antitermination in vitro; however, a B. subtilis strain that fails to produce spermidine showed normal thrS-lacZ expression under threonine starvation conditions, which suggests that this polyamine is not essential for antitermination in vivo.

Instead, unknown proteins or factors may be required (Putzer et al., 2002).

Figure 1.3. Secondary structural model of the B. subtilis glyQS leader RNA. The sequences are shown from the transcription initiation site through the end of terminator. The alternative antiterminator is shown above the terminator. The strutural domains (Stem I, III) and the conserved structural elements (GA motif, S-turn, AG bulge) are labeled; the Stem II and IIA/B structures are naturally missing in this leader RNA. The conserved T box sequence is labeled in red, and the bases that form pairing interactions with the T box sequence in the antiterminator are labeled in blue. The GGC glycine Specifier Sequence is boxed. Watson-Crick pairs are connected by “-”, and wobble pairs are connected by “•”. Residues are colored as described in Fig. 1.1. Asterisks indicate highly conserved nucleotides.

1.4 Kinetic analysis of in vitro transcription antitermination

The establishment of the glyQS in vitro system facilitated the investigation of the requirements for antitermination of genes in the T box family. Transcription elongation factors NusA (which promotes RNAP pausing) and NusG (which prevents RNAP from isomerization at pause sites) are not required for tRNA-dependent antitermination in vitro, although the addition of NusG resulted in reduced pausing half-life of B. subtilis RNAP

(Grundy and Henkin, 2004; Grundy et al., 2002b). Transcription reactions using E. coli

RNAP resulted in different pausing patterns but a similar level of readthrough as compared with B. subtilis RNAP, indicating that antitermination relies on the properties of the RNA transcript rather than the RNAP. Kinetic analysis using B. subtilis RNAP revealed several transcriptional pauses, including a major one at G138 in the loop region of Stem III (Grundy and Henkin, 2004). Introduction of an EcoRI binding site into this region reduced pausing but had no effect on antitermination. The half-life of this pause was decreased by the addition of high concentrations of GTP, although the level of readthrough was not affected. This indicates that pausing at this site is not necessary for antitermination. However, the addition of high concentrations of NTP, particularly UTP, resulted in increased readthrough in the absence of uncharged tRNA, which indicates that low NTP concentration is required for tRNA-dependent antitermination in vitro.

An EcoRI (E111Q) variant that causes a transcription roadblock was used for further kinetic analysis. This protein recognizes the EcoRI restriction endonuclease binding site on DNA templates but has poor DNA cleavage activity (Grundy et al., 2005).

EcoRI restriction sites were introduced into the glyQS leader DNA template at different 12

locations that had no effect on antitermination. The EcoRI protein was pre-bound to the

DNA template, resulting in paused transcription complexes at designated positions in the absence of uncharged tRNA. Transcription elongation was re-initiated by addition of 0.4

M KCl that released the EcoRI protein from the template and allowed transcription to continue in the presence of tRNAGly. The results showed that addition of uncharged tRNAGly during transcription at any point up to and including the formation of the antiterminator structure directs efficient antitermination, which indicates that the glyQS leader RNA folds into a functional structure in the absence of tRNAGly.

Whether or not both charged and uncharged tRNAs have equal access to the T box leader RNA was tested using a tRNAGlyEX1C charged tRNA mimic (Grundy et al., 2005).

This tRNAGlyEX1C construct has an extra C residue at the 3’ end of the acceptor end to mimic an aminoacylated tRNAGly. Addition of increasing concentrations of tRNAGlyEX1C in the presence of a constant concentration of uncharged tRNAGly resulted in a concentration-dependent inhibition of glyQS antitermination. The EcoRI (E111Q) was used to demonstrate that addition of tRNAGlyEX1C inhibits antitermination at any point throughout the transcription process before the formation of the antiterminator, which indicates that charged and uncharged tRNAs have equal access to the leader RNA transcript until the termination/antitermination decision is made. This result suggests that the T box leader RNA constantly monitors the aminoacylation status of tRNA in response to changes in amino acid abundance, but the system is committed to antitermination once the acceptor end-antiterminator interaction is formed.

1.5 Composition of T box leader RNAs

T box leader RNAs contain several conserved primary sequence and secondary structural elements (Grundy and Henkin, 1993; Rollins et al., 1997). These canonical features were used to identify other T box leader RNAs with the same structural domains, and variants with sequence and structural variations (Gutierrez-Preciado et al., 2009;

Vitreschak et al., 2008). The roles of these elements were studied by mutational analysis

(Grundy et al., 2002a; Rollins et al., 1997; Winkler et al., 2001).

The Specifier Sequence, which represents a codon that specifies the amino acid identity of the downstream genes, is located in the Specifier Loop of the Stem I.

Phylogenetic analysis revealed that the residue at position 3 of the Specifier Sequence in the T box leader RNAs is usually a cytosine residue (the “C rule”), which is independent of the codon usage in each organism (Grundy and Henkin, 1993; Gutierrez-Preciado et al.,

2009; Rollins et al., 1997). The significance of the strong preference for a C at this position is still unclear, but is likely to represent a structural constraint. The Specifier

Sequence of most tRNAAla-regulated T box genes is a GCU alanine codon, which is an exception to the C rule (Gutierrez-Preciado et al., 2009). A recent study revealed that the

Specifier Loop of some genes in the T box family contains two overlapping codons, which suggests that the leader RNA can sense the abundance of two metabolically related amino acids via the corresponding tRNAs (Saad et al., 2013).

The Specifier Loop has a conserved S-turn (also known as loop E) motif, which is composed of conserved 5’-AGUA-3’ residues at the 5’ side of the Specifier Loop, and 5’-

GAA-3’ residues at the 3’ side. Mutations that disrupt the S-turn motif resulted in loss of 14

antitermination (Rollins et al., 1997; N.J. Green, F.J. Grundy and T.M. Henkin, unpublished). This motif is predicted to assist proper presentation of the Specifier

Sequence to allow pairing interactions with the tRNA anticodon (Wang et al., 2010).

Below the Specifier Loop is a GA motif, which introduces a kink-turn to the RNA structure (Green et al., 2010; Klein et al., 2001; Winkler et al., 2001). This motif is composed of two helices separated by a 3-nt bulge. One helix contains A•G and G•A noncanonical pairs, while the other helix is composed of canonical Watson-Crick (W-C) pairs (Schroeder et al., 2010). Mutational analysis showed that alteration of conserved residues in the helical region or changing the bulge size resulted in a severe reduction in antitermination, which indicates that maintaining the canonical structure of this motif is important for its function. This motif is also found in the S box leader RNA and many

RNA elements in eukaryotes and archaea (e.g., U4 snRNAs, box C/D snoRNAs, H/ACA snoRNPs, RPL30, and 23S rRNA) (Liang et al., 2007; Nottrott et al., 1999; Winkler et al.,

2001), and can serve as a binding site for ribosomal L7Ae protein as well as several regulatory proteins (Rozhdestvensky et al., 2003; Turner et al., 2005). It is unclear whether any protein binds this motif in the T box leader RNA.

Above the Specifier Loop is the terminal domain of Stem I, which is composed of an AG bulge and an apical stem-loop structure. Specific residues in the AG bulge covary with specific bases in the apical loop of Stem I (K.D. Kreuzer, N.J. Green, F.J. Grundy,

T.M. Henkin, unpublished results). Mutational analysis demonstrated the importance of the sequence conservation in this region (Rollins et al., 1997). Recent X-ray crystallography studies revealed that the AG bulge and the apical loop of the Stem I form 15

a loop-loop interaction, which stacks against the D-loop/T-loop region of tRNA (Grigg et al., 2013; Zhang and Ferre-D'Amare, 2013). This indicates that the terminal domain of

Stem I forms a functional interaction with the tRNA. In some Actinobacteria, the Stem I terminal domain of leader RNAs of the isoleucyl-tRNA synthetase genes is missing (A.V.

Sherwood, F.J. Grundy, T.M. Henkin, in prep). The absence of the conserved Stem I terminal domain implies that these organisms may have evolved to utilize other structural elements to compensate for the function of the missing domain.

The Stem I domain is followed by the Stem II and IIA/B pseudoknot elements. The

Stem II structure usually has a second S-turn motif within an internal loop, and the Stem

IIA/B element contains a conserved F box sequence (5’-CCGUUA-3’). The role of the

Stem II and IIA/B elements is still unclear; however, it was demonstrated that both the primary sequence and secondary structures are important for tyrS gene expression

(Rollins et al., 1997). These structures are naturally missing in some T box leader RNA variants, such as all of the glycyl genes (e.g., the B. subtilis glyQS leader RNA, Fig. 1.3) and some of the alanyl genes (e.g., the Clostridium acetobutylicum alaS leader RNA, see

Chapter 3). The Staphylococcus aureus ileS gene, which encodes isoleucyl-tRNA synthetase, belongs to another special subclass, in which the Stem II structure is located upstream of the Stem I rather than downstream, and the Stem IIA/B structure is missing

(Grundy et al., 1997a).

Stem II and IIA/B are followed by a linker region and Stem III. Only two unpaired bases (A and G) at the 5’ end of the Stem III helix are conserved in this region. The composition and the size of the Stem III vary among different leader RNAs. Although 16

the functional role of this domain is still unclear, it has been reported that a group of leader RNAs contain a noncognate tRNA embedded within Stem III (Gutierrez-Preciado et al., 2009). This tRNA transcript may be processed after the termination/antitermination decision is made, and may stabilize the RNA transcript after its processing.

Downstream of Stem III is the highly conserved 14-base T box sequence (5’-

AGGGUGGNACCGCG-3’, N = G, A, U, C), which is found in all T box leader RNAs

(Gutierrez-Preciado et al., 2009; Rollins et al., 1997). This sequence forms the 5’ half of the antiterminator structure, and contains bases (5’-UGGN-3’) that pair with the free acceptor end (5’-NCCA-3’, where N is the discriminator base) of the cognate tRNA. The discriminator base in the tRNA and the variable position of the cognate leader RNA exhibit covariance. A mutation at the variable position resulted in reduced antitermination, which was restored by a compensatory mutation at the discriminator base of the tRNA (Grundy et al., 1994; Grundy et al., 2002b).

The antiterminator structure is composed of two helixes (A1 and A2) connected by a 7-nt antiterminator bulge, and a terminal loop on the top of the A2 helix (Grundy et al.,

2002a). The A1 helix at the bottom of the antiterminator is 4-6 bp in length, and shows a strong bias for purine residues at the 5’ side of the helix. The A2 helix is located above the antiterminator bulge, and from 3-12 bp in length. Sequence variations in helix A1 have only a modest effect on antitermination, while variations in the basal region of helix

A2 resulted in a more dramatic reduction in antitermination. This indicates that some alterations in helix A1 can be tolerated, while the sequence in helix A2 is very important 17

for function. The terminal loop on the top of the A2 helix varies in size and sequence, and is not phylogenetically conserved.

The terminator is thermodynamically more stable than the antiterminator (Grundy et al., 2002a), and is formed in the absence of uncharged tRNA or in the presence of charged tRNA (Yousef et al., 2005). The terminator structure has a G+C rich helix that ranges from 7-20 bp in length, and a 3-6 nt terminal loop. Mutations that destabilize the terminator helix resulted in increased antitermination (Grundy et al., 2002a). The terminator helix is followed by a run of 5-7 U residues. Extension of the length of the U run resulted in increased termination (Grundy et al., 2002a). The residue located at the 3’ end of the run of U’s is usually an A residue in most Clostridium sp., an A (60%) or G

(35%) in Bacillus sp., and an equal distribution of a G, A, or C in Staphylococcus,

Streptococcus and Enterococcus sp. (Grundy et al., 2002a).

1.6 tRNA requirements for antitermination and binding

Alteration of the tyrosine Specifier Sequence to a phenylalanine codon in a tyrS- lacZ transcriptional fusion resulted in a switch in response to phenylalanine limitation

(Grundy and Henkin, 1993), which indicates that the effector molecule is changed from tRNATyr to tRNAPhe. However, the efficiency of induction in response to phenylalanine limitation was significantly reduced. A subsequent study that changed the tyrS Specifier

Sequence to several other amino acid codons, with a corresponding change at the variable position of the T box leader RNA to maintain the pairing with the discriminator base of the corresponding tRNA, demonstrated that only certain constructs exhibit a successful 18

switch in amino acid response (Grundy et al., 1997b). Similarly, changing the Specifier

Sequence of the B. subtilis valS-, ilv-leu- and thrS-lacZ transcriptional fusions to other codons did not always result in gene expression under corresponding amino acid starvation in vivo (Luo et al., 1997; Marta et al., 1996; Putzer et al., 1995). None of the mutant constructs resulted in induction comparable to the wild-type, which suggests that other sequence and structural requirements are involved in the leader RNA-tRNA interaction.

Elements in tRNA outside of the anticodon and the discriminator base that contribute to efficient antitermination were identified initially by sequence comparison and mutational analysis. A tyrS-lacZ variant, which contains a threonine Specifier

Sequence and a U222A mutation at the variable position to match the discriminator base of tRNAThr, showed reduced lacZ expression under threonine starvation conditions as compared with the expression of the wild-type construct under tyrosine starvation conditions (Grundy et al., 1997b). Sequence comparison between tRNAThr and tRNATyr, and several other tRNAs that interact with tyrS-lacZ variants with Specifier Sequence mutations, revealed that nucleotides at positions 16, 32, and 60 of tRNAThr differ from the consensus of most other tRNAs. Replacing the purine (A) at position 32 of tRNAThr with a more common pyrimidine (C) enhanced expression, suggesting that C32 could be a determinant for tyrS. However, a C32A mutation in tRNATyr had no effect on expression.

This indicates that C32 in the context of tRNAThr enhances antitermination, but this residue is not a major recognition determinant for tyrS antitermination in the context of tRNATyr. Replacement of the bases at positions 16 and 60 of tRNAThr with consensus 19

residues found in other tRNAs had no effect on antitermination, which indicates that nucleotides at these positions are not major determinants for tyrS antitermination.

In a different study, a series of plasmid-borne tRNATyr variants were constructed to investigate the tRNA requirements for tyrS antitermination. The results showed that deletion of the D and/or T stem-loop of tRNATyr resulted in loss of expression of the tyrS- lacZ gene (Grundy et al., 2000). Experiments that tested a pool of tRNA variants with all possible base combinations at positions 69-72, 18 and 19, and 55 and 56 revealed that only those tRNA constructs with the wild-type sequence directed efficient antitermination.

In addition, only certain combinations of bases at positions 26 and 44, and 57, 59 and 60 were found in tRNA variants that resulted in lacZ expression. Because residues at these positions are involved in tRNA tertiary interactions (Oliva et al., 2006), these results indicate that maintaining the conserved tRNA tertiary structure is required for antitermination.

Although variations at positions important for tertiary tRNA structure are not permitted, some changes in tRNA are allowed. For instance, deletion of the tRNATyr variable arm, or replacing it with the HIV TAR site to increase the size of the stem-loop, was tolerated (Grundy et al., 2000). This suggests that the variable arm of tRNATyr is not important for tyrS antitermination. Base substitutions at the acceptor or anticodon stem of tRNATyr that still maintain the pairing interactions had no effect on antitermination

(Grundy et al., 2000). This indicates that there is certain flexibility for tRNA sequence or structural alteration.

Expression of tRNA variants in vivo may be hindered by cellular tRNA repair systems (Deutscher, 1990), rapid degradation of the mutant tRNA transcript before its maturation (Li et al., 2002), or the lack of proper base modifications that contribute to codon recognition (Agris, 2008; Gustilo et al., 2008). In addition, the presence of endogenous wild-type tRNATyr and other noncognate tRNAs in the cell may also complicate the interpretation of the in vivo results. Therefore, the glyQS in vitro system was used to test additional tRNA variants (Yousef et al., 2003).

Several findings are consistent with the in vivo results. For example, disruption of the conserved tertiary interactions at the D-loop/T-loop (pairings between bases 55 and

18, and 56 and 19) was not tolerated. G-C to C-G changes in helical regions, such as the anticodon stem, T stem or acceptor stem, were tolerated. Also, insertion of 1 bp into the anticodon stem was allowed, whereas insertion of 2 bp greatly reduced antitermination.

In contrast, extension of up to 4 bp at the acceptor stem was tolerated, and extension of 5 and 8 bp abolished antitermination. However, extension of 11 bp, which is a full turn of an RNA helix, completely restored antitermination. This indicates that the acceptor stem of the tRNA exhibits face-of-the-helix dependence. The result also suggests that proper presentation of the tRNA acceptor end to the antiterminator bulge is important for antitermination.

Specific binding between the glyQS leader RNA and tRNAGly was demonstrated by size-exclusion filtration assays (Yousef et al., 2005). Both RNA molecules were generated by T7 RNAP-transcription in vitro. Formation of the leader RNA-tRNA complex resulted in a larger complex, which can be separated from unbound 21

radioactively-labeled tRNA by size. The wild-type glyQS leader RNA, which extends from the +1 transcription initiation site to the 3’ end of the antiterminator (+182), bound efficiently to its cognate uncharged tRNAGly. A mutation at position 1 of the glyQS

Specifier Sequence resulted in loss of tRNA binding, and a mutation at the variable position of the antiterminator bulge resulted in only a moderate decrease in binding. This indicates that efficient binding relies primarily on correct Specifier Sequence-anticodon pairing, and the contribution of the pairing between the acceptor end and the antiterminator bulge to binding is less significant.

In sum, these studies reveal that maintaining the tRNA conserved structure is necessary for efficient antitermination, while some variations in tRNA are tolerated.

Also, matching the pairing between the Specifier Sequence and anticodon promotes efficient tRNA binding to the leader RNA. Whether or not mutations in tRNA outside of the anticodon and acceptor end affects tRNA binding affinity is unclear.

1.7 Structural transitions of the leader RNA and tRNA

1.7.1 Leader RNA

Structural mapping assays were used to monitor the structural transition of the leader RNA upon tRNA binding. Under high Mg2+ concentration and high pH conditions,

RNA residues with a flexible backbone (e.g., unpaired bases) are subject to in-line cleavage, while residues that are structurally constrained are protected from cleavage.

Uncharged tRNAGly induced reductions in cleavage throughout the glyQS leader RNA, including residues in the Specifier Loop, in the linker region between Stem I and III, in 22

the region between Stem III and the antiterminator, and in the antiterminator bulge

(Yousef et al., 2005). The tRNAGlyEX1C charged tRNA mimic resulted in a similar protection pattern in Stem I but failed to cause protection of the residues at the antiterminator bulge and the linker region. This suggests that charged tRNA interacts with the leader RNA at the Specifier Sequence, but is unable to base-pair with the residues at the antiterminator bulge. Oligonucleotide-directed RNase H-cleavage assays, which identify unpaired RNA fragments via formation of DNA-RNA hybrids that can be cleaved by RNase H endonuclease, were used to show the formation of the antiterminator in the presence of uncharged tRNA, and the formation of the terminator in the presence of the charged tRNA mimic or in the absence of uncharged tRNA. These results demonstrated that the leader RNA interacts with both charged and uncharged tRNA, and undergoes a structural rearrangement in response to tRNA binding.

Binding of tRNA resulted in increased cleavage at position 98 of the glyQS leader

RNA in the Specifier Loop (Yousef et al., 2005), which led to the design of a bipartite molecule that resembles the bottom part of the Stem I leader RNA (including the

Specifier Loop and GA motif) with a 2-aminopurine at position 98 (2AP98) (Nelson et al.,

2006). Binding of the cognate tRNAGly anticodon stem-loop resulted in quenching of fluorescence, while binding of the noncognate tRNAPhe anticodon stem-loop did not.

This indicates that structural rearrangement of the Specifier Loop is induced by binding of cognate tRNA but not noncognate tRNA.

A recent X-ray crystal structure study revealed detailed interactions between the

Stem I of the Geobacillus kaustophilus glyQS leader RNA and its cognate tRNAGly 23

(Grigg et al., 2013). In this study, a two-checkpoint model was proposed (Fig. 1.4). The first checkpoint is that the leader RNA utilizes the Specifier Sequence to detect the information content of tRNA via base-pairing with the anticodon. The second checkpoint is that the leader RNA uses the Stem I terminal loop and the AG bulge to measure the length of the tRNA anticodon arm by docking at the elbow (D-loop/T-loop) of tRNA.

Similarly, crystal structure analysis of the Oceanobacillus iheyensis glyQ leader RNA and tRNAGly also demonstrated that the leader RNA and tRNA mutually induced structural rearrangements (Zhang and Ferre-D'Amare, 2013), which is consistent with the results in structural mapping experiments (Yousef et al., 2005). The RNA with only the

Specifier Loop and the GA motif bound weakly to its cognate tRNA, while the transcript with the Stem I terminal loop and the AG bulge improved the binding affinity (Zhang and

Ferre-D'Amare, 2013). This indicates that the terminal region of Stem I contributes to efficient tRNA binding.

Fluorescence-based assays were also used to investigate the structural rearrangement of the antiterminator bulge. An antiterminator model RNA AM1A (see

Section 1.8 for more detail) was labeled with fluorescent residues to monitor specific interaction with different tRNA models (full length tRNA, minihelix, microhelix, and tetranucleotide RNAs) (Means et al., 2007). The results demonstrated that although the pairings between the acceptor end and the antiterminator bulge are important, there are other tRNA structural requirements for specific binding to AM1A. It has also been demonstrated that the A1 helix of the antiterminator coaxially stacks on the acceptor end of uncharged tRNA and stabilizes the antiterminator to allow readthrough of the 24

terminator (Zhang and Ferre-D'Amare, 2014). This result provides an explanation for the strong purine bias at the 5’ end of the A1 helix (Section 1.5).

Figure 1.4. Model for tRNAGly recognition by the glyQS leader RNA. The glyQS leader RNA monitors both the information content (anticodon) and the length of the anticodon stem of tRNA at two checkpoints. The first checkpoint is the pairing between the Specifier Sequence and anticodon (1), and the second checkpoint indicates a docking interaction between the apical loop/AG bulge of the Stem I and the D-loop/T-loop of tRNA (2). Once the two checkpoints are passed, uncharged tRNA forms pairing interactions with the antiterminator bulge (3) and allows expression of the downstream genes. Conserved residues in the GA motif, S-turn and the apical loop of Stem I are labeled in orange and yellow. Modified from (Grigg et al., 2013).

1.7.2 tRNA

Structural mapping experiments also revealed residues in tRNAGly that are protected from in-line cleavage in the presence of the glyQS leader RNA, including those at positions 19, 33 and the anticodon (Yousef et al., 2005). Because the U33 base is highly conserved in all tRNAs, and the T box leader RNA has a semi-conserved residue

(A or G) 3’ of the Specifier Sequence that may form an A-U or G•U pair with U33, it was predicted that there may be a fourth base pair between the Specifier Loop and the anticodon loop. However, mutational analysis revealed that maintaining the base pair at this position is not necessary for antitermination. Instead, mutations at position 33 of tRNAGly can be tolerated as long as the residue located at the 3’ end of the Specifier

Sequence remained a purine (Caserta et al., in prep). A crystal structure study demonstrated that this purine residue stacks underneath the Specifier Sequence and stabilizes the Specifier Sequence-anticodon helix (Zhang and Ferre-D'Amare, 2013). The anticodon loop of tRNAGly undergoes structural rearrangement upon binding to Stem I to avoid steric clash with the loop E motif (Fig. 1.5A). The conformational rearrangement in the anticodon loop of tRNAGly reorients the U33 residue. Both crystal structure and

NMR studies revealed that U33 is not involved in any intermolecular pairing interaction

(Chang and Nikonowicz, 2013; Zhang and Ferre-D'Amare, 2013). This residue may be protected from Mg2+-induced cleavage or chemical probing by the phosphate backbone of the anticodon loop (Fig. 1.5B). Another conserved tRNA residue, A37, which is usually a modified base, stacks on top of the Specifier Sequence-anticodon helix and contributes to additional stability (Chang and Nikonowicz, 2013; Zhang and Ferre- 26

D'Amare, 2013). These studies clearly showed conformational change in the tRNAGly anticodon loop upon binding to the T box leader RNA.

A. B.

Figure 1.5. Structural models of the Specifier Sequence-anticodon helix. (A) Superposition of the free tRNAGly anticodon stem-loop (orange) structure with the structure of tRNAGly in complex with the glyQ leader RNA (green and purple). The X- ray crystallographic structure shows binding-induced distortion in the anticodon loop of tRNAGly. The U33 tRNA residue is looped out from the anticodon loop and is not involved in any pairing interaction with the base in the leader RNA. The Specifier Loop forms the conserved loop E motif. Adapted from (Zhang and Ferre-D'Amare, 2013). (B) The NMR study demonstrates the solution form of the Specifier Sequence-anticodon helix. The phosphate backbone of the anticodon loop (orange) protects U33 of tRNAGly from cleavage by in-line probing (Yousef et al., 2005). The conserved A37 of tRNAGly stacks above the Specifier Sequence-anticodon helix. The Specifier Loop corresponds to the domain of the tyrS leader RNA with GGC glycine Specifier Sequence. Nucleotides in red are tRNA bases, and nucleotides in green are residues in the Specifier Loop. Nucleotides involved in the loop E (S-turn) motif are labeled in blue. Adapted from (Chang and Nikonowicz, 2013).

Upon binding to the glyQ leader RNA, tRNAGly exhibits a 20° bend at positions 26 and 44, and the resulting conformation is similar to the structure of tRNAPhe at the P/P

state during decoding (Fig. 1.6). This distortion shifts the acceptor stem 11 Å away from

Stem I and brings the elbow region (D-/T- loops) of the tRNA closer to the terminal domain of Stem I, which promotes the stacking interaction between the terminal domain of Stem I and the conserved G19-C56 tertiary pair (D-/T-loops) of tRNA (Zhang and

Ferre-D'Amare, 2013). This explains the observation that G19 was protected in the structural mapping experiments, and mutations that alter the conserved G19 in tRNAGly resulted in reduced antitermination in vitro (Yousef et al., 2005).

Figure 1.6. Superposition of leader RNA-bound tRNAGly with the structure of free tRNAPhe and tRNAPhe in P/P state. Binding of tRNAGly (purple and green) to Stem I of the glyQ leader RNA induces a bend near positions 26 and 44, and results in a 20° displacement at the elbow of tRNA (D-loop/T-loop) as compared with free tRNAPhe (orange). The distorted tRNAGly resembles the structure of tRNAPhe in the P/P state in the decoding center of the ribosome (12° displacement at the elbow). Adapted from (Zhang and Ferre-D'Amare, 2013).

1.8 The T box system as an antibiotic target

Bacterial drug resistance has become a serious problem, especially for nosocomial infections. In hospitals, pathogens that cause infections in immunocompromised patients result in higher mortality rates than in healthy individuals. More than half of the bloodstream infections are caused by Gram-positive bacteria, particularly Staphylococcus aureus (Wisplinghoff et al., 2004). Methicillin-resistant S. aureus was first reported in hospitals, and now is also found in the community. In recent years, infections caused by vancomycin-resistance enterococci have increased dramatically (National Nosocomial

Infections Surveillance, 2004). In addition, multidrug-resistant Mycobacterium tuberculosis has emerged and become a global problem (Ormerod, 2005). The need for new effective antimicrobial compounds has therefore increased.

Gram-positive bacteria, including pathogenic Streptococcus pyogenes, S. aureus,

Bacillus anthracis, and M. tuberculosis, have adapted the T box regulatory system to control the expression of many essential amino acid-related genes (Gutierrez-Preciado et al., 2009); therefore, this system has been developed as a target for antibiotic design. The

5’ side of the antiterminator is composed of the highly conserved T box sequence, and stabilization of this structure is necessary for expression of the T box-regulated genes.

Consequently, the antiterminator element was chosen as the target for compound development.

Antibiotics usually inhibit the synthesis of DNA, protein, cell wall and cell membrane in bacteria (Gold and Moellering, 1996). Bacteria develop drug resistance via many ways, including modification of the drug target, inactivation of the drug, and 29

reduction of antibiotic concentration in the cell by an efflux pump (Giedraitiene et al.,

2011). The advantage of using the T box antiterminator as a drug target is that mutations that disrupt the conserved structure would lead to gene repression; therefore, cells are less likely to develop resistance by alteration of this target. Approximately 60% gut of microbiota are Gram-positive bacteria, including Bifidobacterium and Lactobacillus sp.

(Gibson and Roberfroid, 1995), which also utilize the T box system for gene regulation

(Gutierrez-Preciado et al., 2009). Therefore, the disadvantage of using the T box mechanism as a drug target is that compounds may also inhibit the growth of beneficial

Gram-positive microbes in the digestive system. Potential treatment using compounds that target the T box mechanism may require co-administration of Gram-negative microbes (e.g., Bacteroides sp.) to maintain the population of gut microbiota.

We collaborated with Dr. Jennifer Hines at Ohio University (Athens, Ohio) to identify compounds that compete with uncharged tRNA for binding to the antiterminator and disrupt regulation (Fig. 1.7A). An AM1A model RNA, which is composed of two helices (A1 and A2) connected by a 7-base bulge and a terminal loop (Fig. 1.7B), was used to facilitate the initial compound screening process using in vitro biochemical tests.

This 29 nt RNA was initially designed based on the B. subtilis tyrS antiterminator with several modifications (Gerdeman et al., 2002): 1) the variable position (position 9) is an

A instead of a U to prevent the formation of a kissing bulge in vitro due to the palindromic sequence; 2) the terminal loop is a stable UUCG tetraloop rather than a

AAUCA pentaloop; 3) immediately below the tetraloop is a C-G pair that replaces the U-

A and A-U pairs found in the A2 helix of the tyrS antiterminator; and 4) there is an extra 30

G-C pair at the bottom of this model RNA (A1 helix) to promote transcription by the T7

RNA polymerase in vitro. An A73U mutation was introduced into tRNATyr to generate a variant that carries the (5’-UCCA-3’) acceptor end to match the (5’-UGGA-3’) sequence of the AM1A bulge (Gerdeman et al., 2002). An AM1A(C11U) construct was also generated as a specificity control; this variant has a C to U mutation at position 11 of

AM1A (Fig. 1.7B), which corresponds to a C224U mutation in the tyrS leader RNA that resulted in an 18-fold reduction in antitermination in vivo (Rollins et al., 1997). B. subtilis tRNATyr(A73U) binds 3-fold more weakly to the AM1A(C11U) variant than to the AM1A model RNA (Gerdeman et al., 2002).

The solution structure of AM1A revealed an 80° kink between helix A1 and helix

A2 due to the 7-nt intervening bulge (Gerdeman et al., 2003). The four bases at the 5’ end of the antiterminator bulge revealed structural flexibility, which allows interaction with the acceptor end of the tRNA. The three bases at the 3’ end of the bulge showed stacking interactions, which may constrain the flexibility of resides at the 5’ end of the bulge. Ligand binding to AM1A may alter the angle of the kink between the two helices.

Therefore, fluorescence resonance energy transfer (FRET) assays were used to monitor the ligand-induced structural change of AM1A (Means and Hines, 2005).

A. B.

Figure 1.7. Compound binding to the AM1A model RNA. (A) Ligands that compete with the uncharged tRNA for binding to the antiterminator may lead to formation of the terminator and disruption of the T box regulation. Modified from (Jentzsch and Hines, 2012). (B) The AM1A model RNA was constructed based on the tyrS antiterminator with an A instead of a U at the variable position (position 9), the addition of a G-C pair at the base of the model RNA (helix A1), a UUCG tetraloop as the terminal loop, and the replacement of the original U-A/A-U pairs with a C-G pair underneath the tetraloop (helix A2). The C11U mutation that resulted in reduced antitermination in vivo is indicated. Modified from (Means et al., 2006).

Aminoglycosides were first tested due to their RNA binding ability. Binding of aminoglycosides to the fluorophore-labeled AM1A resulted in a change in fluorescence, indicating a conformational change (Means and Hines, 2005). However, in addition to structure-specific binding, aminoglycosides also bind RNA at divalent metal ion-binding sites, which suggests that the interaction is electrostatic-dependent. Due to the abundance 32

of divalent metal ion-binding sites in the cell, aminoglycosides may bind other RNA molecules in addition to the T box antiterminator. Addition of neomycin B (an aminoglycoside) also results in enhanced rather than decreased tRNA binding to AM1A

(Anupam et al., 2008a).

In contrast, oxazolidinones are non-ionic ligands that bind RNAs with less reliance on electrostatic interaction as compared with aminoglycosides; therefore, compounds in this class could confer structure-specific binding to the antiterminator model RNA rather than non-specific binding to divalent metal ion binding sites. A series of 4,5- disubstituted oxazolidinones and 1,4-disubstituted 1,2,3-triazole analogs were generated, and the binding affinities and specificities to antiterminator RNAs were determined by

FRET using AM1A and AM1A(C11U) model RNAs (Acquaah-Harrison et al., 2010;

Anupam et al., 2008b; Maciagiewicz et al., 2011; Means et al., 2006). In addition, fluorescence-based assays were used to test the ability of lead compounds to compete with uncharged tRNA for binding to AM1A, and to monitor ligand-induced structural stability of AM1A (Zhou et al., 2011; Zhou et al., 2012). However, the effects of lead compounds on antitermination were not well-studied. In Chapter 4, we will describe the work on investigation of the effect of lead compounds on antitermination using in vitro and in vivo functional assays.

1.9 Research goal

tRNAs are one of the earliest evolved molecules with the primary function in protein synthesis. Accurate decoding is important for cell viability; therefore, tRNAs 33

have evolved to maintain the fidelity of codon recognition. It has been proposed that residues in tRNA have coevolved with its anticodon to achieve uniform ribosome binding activity (Olejniczak et al., 2005; Saks and Conery, 2007; Yarus, 1982). Besides the canonical role of tRNA in decoding, several non-canonical roles of tRNA in gene regulation, cell wall biosynthesis, antibiotics production, proteolysis, retroviral reverse transcription, cell signaling pathway, etc. have been reported (Dare and Ibba, 2012;

Henkin, 2014; Phizicky and Hopper, 2010; Raina and Ibba, 2014). In B. subtilis and many Gram-positive bacteria, tRNAs serve as effector molecules that mediate expression of genes in the T box family (Green et al., 2010; Henkin, 2014). How each tRNA achieves specific interaction with its cognate T box leader RNA while still maintaining its primary function in translation and other biological processes is not well understood.

This study focused on investigation of elements in tRNA that are important for antitermination and comparisons to their roles in codon recognition in translation.

Studies on suppressor tRNAs and tRNAs that cause misreading have revealed that residues outside of the anticodon contribute to correct codon-anticodon interaction (Hirsh,

1971; Lustig et al., 1989; Schultz and Yarus, 1994a; Smith and Yarus, 1989b; Yarus,

1982). Since the pairing between the Specifier Sequence and anticodon resembles codon-anticodon recognition in translation, it is possible that mutations that reduce decoding accuracy also affect antitermination. In Chapter 2, we designed mutations based on suppressor tRNA and sequence conservation studies, and investigated how these mutations in tRNAGly affect glyQS antitermination and leader RNA binding in vitro.

Because all leader RNAs of glyQS genes are natural deletion variants, while most other T box leader RNAs have all the canonical domains including the Stem II and IIA/B elements, it was unclear whether findings in the glyQS system are completely applicable to antitermination of other genes in the T box family. We found that leader RNAs of alaS genes, which encode alanyl-tRNA synthetases, contain both natural deletion variants and members of the canonical class. In Chapter 3, we established another in vitro system using the C. acetobutylicum alaS leader RNA, which belongs to the class of natural deletion variant, and the B. subtilis alaS leader RNA, which has all the canonical structural domains. tRNA requirements for alaS antitermination were determined and compared with previous findings in glyQS and tyrS antitermination.

We continued the collaboration with Dr. Jennifer Hines and investigated compounds that could target the T box system for therapeutic use. In Chapter 4, we used the glyQS in vitro transcription antitermination assays and tyrS-lacZ transcriptional fusions in B. subtilis to determine the effect of lead compounds on antitermination.

This study provides insights into the importance of conserved tRNA elements in the

T box regulatory system in comparison with that in translation, and the roles of lead compounds in disruption of gene expression by targeting the tRNA-dependent T box antitermination system.

CHAPTER 2

ROLES OF CONSERVED tRNAGly ELEMENTS IN glyQS ANTITERMINATION

2.1 Introduction

The T box riboswitch RNA interacts directly with the cognate tRNA and controls the expression of the downstream amino acid-related genes in response to the charging ratio of the cognate tRNA (Chapter 1). At least two major interactions contribute to the specific interaction between the leader RNA and the tRNA. The anticodon of the tRNA pairs with the Specifier Sequence of the leader RNA and contributes to the first specific interaction. Stabilization of the antiterminator structure requires base-pairing of the free acceptor end of the uncharged tRNA with the complementary sequence in the antiterminator bulge of the leader RNA. Although these two interactions are necessary for antitermination, studies have shown that they are not sufficient for efficient antitermination (Section 1.6). There are other sequence or structural requirements involved in the leader RNA-tRNA interaction (Grundy and Henkin, 1993; Grundy et al.,

1997b; Grundy et al., 2002b). Recent crystal structure studies revealed a docking interaction between the Stem I apical loop/AG bulge of the leader RNA and the elbow region of tRNA (D-/T- loops) (Grigg et al., 2013; Zhang and Ferre-D'Amare, 2013).

Besides these reported interactions, it is still unclear what other elements in tRNA are 36

important for efficient antitermination.

The extended anticodon hypothesis proposes that tRNA elements near the anticodon have evolved to assist efficient and accurate codon-anticodon interaction in translation (Yarus, 1982). A bioinformatics analysis of tRNA sequences from 145 bacterial genomes revealed that residues at positions 32 and 38 (anticodon loop), 27 and

43 (anticodon stem), and 26 and 44 (central core) exhibit sequence conservation depending on the anticodon; this property is termed “anticodon-dependent conservation”

(Saks and Conery, 2007). Bases at positions 32 and 38, which are located in the anticodon loop, were studied extensively. Substitutions at position 32 of tRNAGly, or at positions 32 and 38 of tRNAAla, result in misreading of near-cognate codons (Claesson et al., 1995; Ledoux et al., 2009; Lustig et al., 1993; Olejniczak and Uhlenbeck, 2006).

Base substitutions at positions 27 and 43 of a tRNATrp variant promote misreading of a stop codon (Schultz and Yarus, 1994a). These results suggest that disruption of the conservation at these positions alters decoding accuracy. A mutation at the D arm of tRNATrp (the Hirsh suppressor mutation), which is distant from the anticodon and is not conserved based on the anticodon sequence (Saks and Conery, 2007), also results in increased misreading of stop codons via stabilization of the distortion of tRNA structure during decoding (Cochella and Green, 2005; Hirsh, 1971; Schmeing et al., 2011; Smith and Yarus, 1989a). This result indicates that multiple tRNA residues outside the anticodon have evolved to optimize the codon-anticodon interaction in translation. The possible effects of these elements on tRNA-mediated antitermination are unknown.

The B. subtilis glyQS gene was used previously to investigate the effect of a 37

mismatch at the three different positions of the Specifier Sequence on antitermination in vitro (Caserta et al., in prep). This gene, which encodes glycyl-tRNA synthetase, has a

GGC glycine codon as its Specifier Sequence (Fig. 2.1A). The cognate tRNAGly has an unmodified GCC anticodon (Fig. 2.1B), which avoids potential problems caused by the lack of anticodon modification with in vitro-generated tRNAs. The study demonstrated a unique Specifier Sequence-anticodon recognition pattern in comparison with the codon- anticodon interaction in translation. Similar to codon recognition in translation, a mismatch at the middle position of the Specifier Sequence is more detrimental than at the flanking positions. However, mismatches at either positions 1 or 3 of the Specifier

Sequence result in a similar reduction in antitermination efficiency. This differs from the translational effects of mismatches at position 3 of the codon, which are tolerated in some instances (Crick, 1966; Lagerkvist, 1978), and mismatches at position 1 of the codon, which are usually not tolerated (Gromadski et al., 2006).

In translation, the pairing geometries at the first two positions of the codon are monitored by three 16S ribosomal RNA residues, A1492, A1493 and G530, and the pairing at the third position is less restricted and includes wobble pairs (Crick, 1966; Ogle et al., 2001; Ogle et al., 2002). In contrast, the recent crystal structure of the Stem I of the glyQ leader RNA and the cognate tRNAGly shows that no residues are located nearby to monitor the pairing geometry of the Specifier Sequence and anticodon (Zhang and

Ferre-D'Amare, 2013). Therefore, the codon-anticodon pairing in the T box system depends only on sequence or structural elements in both molecules. This study examines the role of such elements in tRNA outside of the anticodon and the discriminator base. 38

Studies have revealed residues in tRNA outside of the anticodon that affect correct codon recognition in translation; in this study, we tested whether these residues also affect antitermination efficiency.

The Stem I of the leader RNA is capable of specific tRNA binding (Grigg et al.,

2013; Yousef et al., 2005; Zhang and Ferre-D'Amare, 2013); therefore, binding is considered as the first step during antitermination before the RNAP reaches the termination/antitermination decision point. The next critical step involves the acceptor end of the tRNA to be properly presented to the antiterminator bulge (Yousef et al., 2003), and may strongly depend on the interaction between the Specifier Sequence and anticodon. Mutations that alter the conserved elements in tRNA could have effects on either or both steps in antitermination. In the current study, we designed tRNA mutations based on positions that show anticodon-dependent conservation and/or residues that alter codon recognition fidelity in translation. These mutations were introduced into B. subtilis tRNAGly(GCC) or a variant with a UCC (G34U) anticodon (Fig. 2.1B), and their effects on antitermination and binding were evaluated. The results provide new information on the importance of conserved tRNA elements in antitermination in comparison with that in translation.

glyQS mutations Specifier Sequence WT GGC C101A GGA C101U GGU C101G GGG G99A AGC G99U UGC G99C CGC G99A C101A AGA G99U C101A UGA G99C C101A CGA

B. tRNA References mutations G34U (Claesson et al., 1995) U32C (Claesson et al., 1995); (Olejniczak et al., 2005); (Ledoux et al., 2009); (Olejniczak and Uhlenbeck, 2006); (Saks and Conery, 2007) C30G-G40C (Saks and Conery, 2007); (Grundy et al., 2007, unpublished data) C27G•G43A (Saks and Conery, 2007); (Schultz and Yarus, 1994a) A26G•G44U (Grundy et al., 1997); (Saks and Conery, 2007) U11C (Hirsh,1971); (Saks and Conery, A24G 2007); (Smith and Yarus, 1989a) U11C-A24G

Figure 2.1. Structural model of B. subtilis glyQS leader RNA and tRNAGly. (A) Pairing interaction between the glyQS leader RNA and uncharged tRNAGly. The Specifier Sequence (5’-GGC-3’), corresponding to positions 99-101 of the glyQS leader RNA, pairs with the anticodon (5’-GCC-3’) of tRNAGly. The acceptor end (5’-UCCA-3’) of the uncharged tRNA pairs with the antiterminator bulge (5’-UGGA-3’) of the glyQS leader RNA and stabilizes the antiterminator. Substitutions in the Specifier Sequence that were tested in this study are boxed and underlined in the table. (B) Cloverleaf model of tRNAGly(GCC). Positions of mutations in tRNAGly are indicated, and sequences of mutations are listed in the table.

2.2 Materials and Methods

2.2.1 Construction of glyQS and tRNA variants

The wild-type glyQS leader sequence was inserted into plasmid pGEM-4 (Promega) and propagated in E. coli strain DH5α; the U at position +2 (relative to the transcription start-site) was replaced with a C to allow for transcription initiation with an ApC dinucleotide (Sigma) in the in vitro transcription antitermination assays (Grundy et al.,

2005). DNA templates used in this assay were 440 bp PCR products, which extended from 135 bp upstream of the transcription start-site to 85 bp downstream of the termination site at +220. Leader variants with alterations at position(s) 1 and/or 3 of the

Specifier Sequence (Fig. 2.1A) were amplified by oligonucleotide primers carrying the desired substitutions. Templates for T7 RNA polymerase (RNAP) transcription were generated by PCR using an oligonucleotide containing the T7 RNAP promoter fused with the +1 position of the glyQS leader sequence by a tandem G (Yousef et al., 2005). Leader

RNAs for the binding assay (see size-exclusion filtration assays below) extended to +182 at the 3’ end of the antiterminator, or to +150 at the 5’ end of the antiterminator. The tRNA variants were constructed by oligonucleotide cassette-ligation as described previously (Yousef et al., 2003).

2.2.2 T7 RNAP transcription

T7 RNAP transcription of leader RNA and tRNA was carried out by overnight incubation of DNA templates in transcription buffer containing 40 mM Tris-HCl (pH 8),

2 mM spermidine, 22 mM MgCl2, 5 mM DTT, 4 mM NTP (N = G, A, U, C), 4 mM GMP, 41

0.028 μM pyrophosphatase (Roche), 8 U RNase inhibitor (Roche) and 0.45 μM T7

RNAP (Ambion) at 37°C. RNA products were purified on a 6% (w/v) denaturing acrylamide gel, and quantified with an ND1000 Spectrophotometer (NanoDrop

Technologies). For the binding assay, [α-32P] UTP (800 Ci/mmol; 1 Ci = 37 GBq) at a final concentration of 1 μM was added to allow random incorporation into the tRNA during transcription. The resulting tRNAs were folded in the presence of 10 mM MgCl2 by heating to 80°C for 2 min followed by slow cooling over 40 min to room temperature.

2.2.3 In vitro transcription antitermination assays

Single round transcription reactions were carried out as described previously

(Grundy et al. 2002; Caserta et al., in prep). Transcription reactions were initiated by incubation of leader DNA templates at 1 nM in 1X transcription buffer (20 mM Tris-HCl

[pH 7.9], 10 mM MgCl2, 20 mM NaCl, 0.1 mM EDTA), 6 nM His-tagged purified B. subtilis RNAP, 150 μM ApC dinucleotide (Sigma), 2.5 μM ATP and GTP, 0.75 μM of

UTP, and 0.25 μM [α-32P] UTP (800 Ci/mmol; 1 Ci = 37 GBq) at 37°C for 15 min.

Transcription was halted at position +17 of the leader template by omission of CTP.

After the addition of heparin (25 nM, Sigma) to prevent reinitiation of transcription, an

NTP mix (N = G, A, U, C) at 10 μM final concentration was added to allow elongation in the presence of various concentrations of T7 RNAP-transcribed tRNA (final concentration 0, 0.003, 0.01, 0.03, 0.06, 0.1, 0.2 and 0.5 μM). Transcription products were extracted with phenol-chloroform and resolved on a 6% denaturing acrylamide gel.

The results were analyzed on a PhosphorImager (Amersham Biosciences) and quantified 42

by ImageQuant (Fig. 2.2A). The percent readthrough (%RT) was calculated by %RT =

[(paused site transcripts) + (readthrough transcripts)] X 100 / (total transcripts). Data were fit to a one-site binding hyperbola equation to determine the RTmax and K1/2 using

GraphPad Prism version 4.0 (GraphPad Software) (Fig. 2.2B):

K1/2 = [tRNA]input * (RTmax - %RT) / %RT

The RTmax (%) is defined as the maximum amount of percent readthrough estimated by extrapolation of a hyperbolic curve, and the K1/2 (μM) is defined as the tRNA concentration required to reach half of the maximum readthrough. K1/2 is used as a parameter to evaluate the antitermination efficiency. Because the K1/2 is dependent on the RTmax, constructs with the same K1/2 value but lowered RTmax showed decreased antitermination activity. If the K1/2 value was too high to be determined accurately, the percent readthrough (%RT) determined at 0.5 µM tRNA was compared directly. All experiments were repeated at least twice.

2.2.4 Size-exclusion filtration tRNA binding assays

The tRNA binding assay was carried out as described previously (Yousef et al.

2005; Caserta et al., in prep) with minor changes in calculation of Kd (µM). Increasing concentrations of T7 RNAP-generated leader RNAs (final concentration 0.025, 0.05, 0.1,

0.2, 0.4, 0.8, 1.6 μM) were mixed with radioactively labeled tRNAs (0.05 μM) in 1X transcription buffer. The mixture was heated to 65°C for 5 min and cooled to 40°C before applying to a 30K NanoSep filter (Life Science). After 6 washes with 150 μl of

1X transcription buffer, material retained on the filter was mixed with scintillation fluid 43

(Packard BioScience Ultima Gold) and quantified with a scintillation counter (Packard

Tri-Carb 2100TR). The counts were used to calculate the concentration of leader RNA- tRNA complexes formed in the assay. The approximate dissociation binding constant

(Kd) was determined by fitting the amount of the leader RNA-tRNA complex to a quadratic equation using GraphPad Prism version 4.0 (GraphPad Software):

[Complex formed] = (([leader]input + [tRNA]input + Kd) - sqrt(([leader]input + [tRNA]input +

Kd)^2 – 4 * [leader]input * [tRNA]input)) / (2 * 1)

The Kd estimates the tRNA concentration (μM) required to occupy half of the leader RNA binding sites. This Kd was used to determine the Gibbs free energy change

(ΔG°) of the leader RNA-tRNA interaction by ΔG° = -RTlnKd (kcal/mol), where T =

298°K as described previously (Nelson et al., 2006).

Figure 2.2. In vitro transcription of the wild-type glyQS leader template with increasing amount of wild-type tRNAGly(GCC). (A) Single round transcriptions using the wild-type B. subtilis glyQS DNA template and B. subtilis RNAP were carried out in the presence or absence of increasing amounts of T7 RNAP-transcribed wild-type tRNAGly. Terminated (T, 220 nt) and readthrough (RT, 305 nt) transcripts are indicated; a long-lived pause (P, 240 nt) was also observed. Percent readthrough (%RT) is shown at the bottom of each lane. Lane 1, no tRNAGly. Lanes 2-8; 0, 0.003, 0.01, 0.03, 0.06, 0.1, 0.2 and 0.5 μM tRNAGly, respectively. (B) Percent readthrough calculated from (A) was fit to a hyperbolic equation and used to determine the RTmax (94%) and K1/2 (0.031 μM).

2.3 Results

2.3.1 U32C

2.3.1.1 Effects on antitermination

It was reported previously that introduction of a U32C mutation into E. coli

Gly tRNA1 CCC (which specifically recognizes the GGG codon in protein synthesis) allows the tRNA to recognize the three other glycine codons (GGC, GGA, GGU) (Claesson et al., 1995). To investigate whether this mutation would have a similar effect on recognition of the glyQS Specifier Sequence, we introduced a U32C substitution into B. subtilis tRNAGly in the context of a GCC (WT) or UCC (G34U) anticodon, and tested the effect on expression of glyQS variants with mutations at positions 3 and/or 1 that resulted in different Specifier Sequence-anticodon mismatches.

Disruption or alteration of the pairing at position 3 of the Specifier Sequence resulted in a 1.9- to 7.4-fold increase in the K1/2 (Fig. 2.3A, Table A.1), indicative of decreased antitermination efficiency. Antitermination efficiency decreased according to the identity of the pairing at position 3 as follows: C-G > U•G > A-U > G•U > G•G >

A•G > C•U > U•U. The U32C substitution in the wild-type tRNAGly(GCC) resulted in a

2.1-fold decrease in the K1/2 of the wild-type glyQS leader RNA (C-G at position 3); however, this mutation also reduced the RTmax from 94% to 75% (Table A.2). In comparison with the wild-type tRNAGly, U32C led to enhanced antitermination at tRNA concentrations below 0.06 μM, and reduced antitermination at tRNA concentrations at and above 0.06 μM (Fig. A.1A). The differential response suggests that the

antitermination process involves at least two steps, which could be affected differentially by U32C.

The lowered RTmax was not observed in other constructs with pairing alterations at position 3, which allows for direct comparison of antitermination efficiency using the K1/2 value. U32C resulted in a 1.7-fold increase in antitermination efficiency of the construct with an A-U W-C pair at position 3, and had no effect on antitermination of constructs with a G•U or U▪G wobble pair (Fig. 2.3A). A 1.6- to 2.2-fold increase in antitermination efficiency was also observed in constructs with purine-purine (R•R) mismatches (G•G and A•G) or pyrimidine-pyrimidine (Y•Y) mismatches (C•U and U•U) at position 3 of the Specifier Sequence. This indicates that the presence of U32C reduces the variation in antitermination efficiency between the construct with wild-type Specifier Sequence- anticodon interaction and constructs with mismatches at position 3. This result suggests that U32C results in misreading of other glycine Specifier Sequences (GGA, GGU, GGG) in addition to GGC, which is similar to the effect on codon recognition in translation.

Figure 2.3. Response of antitermination to mismatches between the Specifier Sequence and anticodon in the presence or absence of U32C. Point mutations were introduced into the Specifier Sequence and anticodon (dark boxes) to investigate the effects on antitermination of mismatches at position 3 (A), position 1 (B and C) or both positions (D) of the Specifier Sequence. Position 32 of tRNAGly was changed from a U to a C (gray box, U32C). Antitermination activity (K1/2 in A and B, and %RT in C and D) of each construct in the absence (dark bars) or presence (gray bars) of U32C was determined as described in Materials and Methods. A dot “•” between nucleotides represents a mismatch interaction and a dash “-” represents a perfect match. Substitutions in the Specifier Sequence and anticodon are underlined in each bar graph. See Table A.1 for K1/2 values, and Table A.3 for %RT values at 0.5 μM tRNA concentration. Error bars represent the standard error of the mean (SEM).

Figure 2.3

A mismatch at position 1 of the Specifier Sequence resulted in a 3.1- to 10-fold decrease in antitermination (Fig. 2.3B, Table A.1). The U32C substitution in tRNAGly(GCC) resulted in a 1.7- to 2.5-fold increase in antitermination efficiency, which is comparable to the effect of this mutation on antitermination of constructs with a mismatch at position 3 of the Specifier Sequence. This suggests that the effect of U32C is not limited to constructs with position 3 mismatches. However, the RTmax of the construct with a C•C mismatch at position 1 was decreased from 100% to 76% (Table

A.2), which suggests that introduction of U32C may constrain the antitermination activity.

The fitted curves of the construct with a C•C mismatch at position 1 showed that U32C resulted in enhanced antitermination below 0.2 μM tRNA, and reduced antitermination above 0.2 μM tRNA relative to the wild-type tRNA (Fig. A.1B). This suggests that the effect of U32C on the RTmax may be specific to particular pairing interactions between the

Specifier Sequence and the anticodon.

We also tested the effect of U32C on antitermination of constructs with an A-U pair at position 3 and a mismatch at position 1 (A•C, U•C or C•C), and constructs with double mismatches at positions 1 and 3 of the Specifier Sequence (Figs. 2.3C and 2.3D).

Antitermination activities of these constructs were too low to determine the K1/2 accurately; therefore, the percent readthrough (%RT) determined under the highest tRNA concentration (0.5 μM) tested in the assay was used to evaluate the effect of the mutation on antitermination. The A-U pair at position 3 in the construct with a wild-type G-C pair at position 1 showed up to 80% readthrough, while a mismatch at position 1 resulted in

6.7-25% readthrough (Fig. 2.3C). A similar detrimental effect was observed in constructs 50

with a C•U mismatch at position 3 and a mismatch at position 1 (Fig. 2.3D). This suggests that the A-U (W-C) pair at position 3 is as detrimental as the C•U (Y•Y) mismatch in constructs with mismatches at position 1. The U32C mutation resulted in enhanced antitermination of two constructs with detrimental pairing interactions between the Specifier Sequence and the anticodon. This included the construct with an A-U pair at position 3 and a U•C mismatch at position 1 (readthrough increased from 9.1% to 22%,

Table A.3), and the construct with a C•U mismatch at position 3 and an A•C mismatch at position 1 (readthrough increased from 8.5% to 16%). The U32C mutation had no effect on other constructs with pairing alterations at both positions 1 and 3, which suggests that the effect may be specific to particular mismatch interactions between the Specifier

Sequence and anticodon.

2.3.1.2 Effects on binding

Although the U32C substitution in wild-type tRNAGly resulted in a 2.1-fold increase in antitermination (decrease in K1/2) of the wild-type glyQS leader RNA, the

RTmax was also decreased (Section 2.3.1.1). The fitted curves revealed a differential effect of U32C on antitermination below and above 0.06 μM. Antitermination is predicted to be a multi-step process, which involves at least two steps (tRNA binding to the leader RNA and stabilization of the antiterminator structure). Therefore, we hypothesized that the U32C mutation enhances tRNA binding to the leader RNA via an enhanced pairing interaction between the Specifier Sequence and anticodon; however, an enhanced pairing interaction between the wild-type GGC Specifier Sequence and wild- 51

type GCC anticodon may constrain the proper positioning of the acceptor end relative to the antiterminator bulge, and thus limit the maximum readthrough of this construct.

We tested whether U32C enhanced tRNA binding to the leader RNA using size-exclusion filtration tRNA binding assays. The antitermination K1/2 is calculated by values determined after a dynamic multi-step transcription process. In contrast, the apparent dissociation constant Kd determined in the binding assays reflects the leader RNA-tRNA binding affinity at approximately equilibrium conditions. Therefore, the binding assay can provide information about one of the steps in antitermination.

T7 RNAP-transcribed wild-type glyQS leader RNA that extends from the transcription start-site to the 3’ end of antiterminator structure (1-182) was tested first.

Gly The wild-type tRNA (GCC) bound the 1-182 leader RNA with a Kd of 0.092 ± 0.018

µM (free energy change ΔG° = -9.6 ± 0.1 kcal/mol at 25°C), and the U32C substitution had no effect on binding affinity and ΔG° (Table 2.1). Our previous binding experiments revealed that deletion of the glyQS antiterminator structure results in reduced binding

(Grundy et al. 2005), which suggests that the interaction at the antiterminator bulge contributes to stable binding. Therefore, we also tested a truncated leader RNA that extends only to the 5’ end of the antiterminator (1-150) without the bulge sequence.

Although the wild-type tRNAGly bound this truncated leader RNA (1-150) with a 36-fold decrease in affinity relative to the 1-182 leader RNA, the U32C substitution resulted in a

2.4-fold increase in binding affinity and a small difference in ΔG° (-7.5 kcal/mol vs. -8.0 kcal/mol). This indicates that the effect of U32C is more prominent without the interaction at the antiterminator bulge. This result supports the first part of our prediction 52

that enhanced antitermination (at tRNA below 0.06 µM) is due to enhanced tRNA binding. Whether or not the enhanced tRNA binding to the wild-type leader RNA affects proper positioning of the acceptor arm of tRNA relative to the antiterminator bulge remains to be determined.

Table 2.1. Effect of U32C on binding affinity of the glyQS leader RNA and variants The glyQS leader RNA 1-182 The glyQS leader RNA 1-150 a b Kd ΔG° Kd ΔG° (µM) (kcal/mol) (µM) (kcal/mol) WT tRNAGly(GCC) 0.092 ± 0.018 -9.6 ± 0.1 3.3 ± 0.5 -7.5 ± 0.1 tRNAGly(GCC)U32C 0.063 ± 0.011 -9.8 ± 0.1 1.4 ± 0.2 -8.0 ± 0.1 a Kd was determined by size-exclusion filtration tRNA binding assays. b ΔG° was determined by ΔG° = -RTlnKd where T = 298°K.

We also tested whether the U32C substitution enhanced antitermination of constructs with a mismatch at positions 1 or 3 of the Specifier Sequence as a result of enhanced binding. Leader RNA variants (1-182) that carry corresponding mutations in the Specifier Sequence were used. Without U32C, a mismatch or pairing alteration at positions 1 or 3 resulted in increased Kd by 2.0- to 8.9-fold (Fig. 2.4, Table A.4), indicative of decreased binding affinity. Introduction of the U32C substitution in tRNAGly resulted in a 1.4- to 2.5-fold increase in binding affinity of constructs with W-C, wobble, or a mismatch at position 3. This suggests that the effects of U32C on binding and antitermination are correlated in these constructs. However, the effects on binding and antitermination were not correlated for position 1 changes. U32C resulted in reduced 53

binding affinity (by 4.2- and 2.3-fold) in constructs with an A•C or C•C mismatch at position 1 of the Specifier Sequence, respectively, and had no effect on binding of the construct with a U•C mismatch at position 1 (Fig. 2.4); all of these constructs exhibited enhanced antitermination efficiency when U32C was introduced. This indicates that

U32C has differential effects on antitermination and binding in constructs with a mismatch at position 1. This could be a mismatch-specific result.

Figure 2.4. Effect of U32C on binding of constructs with Specifier Sequence- anticodon pairing alterations at positions 1 or 3. Response of binding to mismatches between the Specifier Sequence and anticodon in the absence (WT, dark bars) or presence of U32C is shown. Dissociation constant (Kd) was determined using size-exclusion filtration assays as described in the Materials and Methods. Symbols and base substitutions are denoted as described in Fig. 2.3. Error bars represent the standard error of the mean (SEM).

We further generated truncated tRNAGly variants without the UCCA acceptor end sequence that is unable to form the interaction at the antiterminator bulge, and tested the

effect of U32C on tRNA binding to leader RNA variants (1-182) with mutations at position 1. We observed that although tRNAGly(GCC)ΔUCCA bound to the wild-type glyQS leader RNA with an 18-fold reduction in binding affinity relative to tRNAGly containing the UCCA acceptor end, the presence of the U32C mutation resulted in only a

8.5-fold increase in affinity (Fig. A.2A). Similarly, although a mismatch at position 1 of the Specifier Sequence greatly reduced binding affinity in the absence of the interaction at the antiterminator bulge (>180-fold) (Figs. A.2B-2D), introduction of U32C into the truncated tRNAGly resulted in enhanced binding affinity (by 2.0- to 3.5-fold). This result indicates that U32C enhances binding in constructs with a mismatch at position 1, which is consistent with the effect of U32C on binding to the constructs with a mismatch at position 3. This also suggests that the interaction at the antiterminator bulge may be disrupted in constructs with a mismatch at position 1 in the presence of U32C.

2.3.2 C30G-G40C

Residues at positions 30 and 40 in tRNAGly isoacceptors exhibit anticodon- dependent sequence conservation; a C30-G40 pair is conserved in tRNAGly with a GCC anticodon, and a G30-C40 pair is conserved in tRNAGly with a UCC anticodon (Saks and

Conery, 2007; F.J. Grundy, unpublished data). To determine whether this pattern of conservation contributes to antitermination, C30G-G40C substitutions were introduced

Gly into tRNA (GCC) (Fig. 2.1B). The mutations reduced the RTmax of the wild-type leader

RNA from 94% to 80% without an effect on the K1/2 value (Tables A.1 and A.2). The fitted curves demonstrated decreased antitermination throughout the tested tRNA 55

concentrations (Fig. A.1C), indicating that the conservation at positions 30 and 40 of tRNAGly(GCC) contributes to efficient interaction between the wild-type leader RNA and tRNA. These substitutions had no effect on other leader RNA-tRNA mutant constructs

(Fig. 2.5, Tables A.1 and A.3), suggesting that pairing alterations between the Specifier

Sequence and the anticodon may counterbalance the effect of the sequence conservation at positions 30 and 40. Whether these residues are recognized directly by the leader RNA or contribute to structural recognition is unclear.

Figure 2.5. Effect of C30G-G40C, A26G•G44U or C27G•G43A on antitermination. Antitermination efficiency of constructs with a mismatch at position 3 (A) or position 1 (B) of the Specifier Sequence was determined in the absence (dark bars, WT) or presence of C30G-G40C (dark gray bars), C27G•G43A (light gray bars) or A26G•G44U (white bars). Symbols and mutations are denoted as described in Fig. 2.3. * indicates not determined. Error bars represent the standard error of the mean (SEM). 57

2.3.3 C27G•G43A

Nucleotides at positions 27 and 43 in tRNAs generally covary with the anticodon sequence (Saks and Conery, 2007). Previous translational studies showed that

C27G•G43A substitutions in the Su7 G36 suppressor tRNA (a derivative of E. coli tRNATrp with a CUG anticodon) resulted in efficient misreading of CAA and UAG codons, which form an A•C mismatch at the third position and a U•G wobble at the first position of the codon, respectively (Schultz and Yarus, 1994a; Schultz and Yarus, 1994b).

To test whether the same substitutions in tRNAGly affect glyQS antitermination, the C-G pair at positions 27 and 43 of tRNAGly was replaced with a G•A mismatch in the context of a GCC or UCC (G34U) anticodon (Fig. 2.1B). The C27G•G43A substitutions had no effect on K1/2 or RTmax of all tested constructs (Fig. 2.5, Tables A.1-A.3), which suggests that weakening of the pairing interaction at these positions of tRNAGly was tolerated.

This result also indicates that a G•A mismatch at positions 27 and 43 did not alter the overall tertiary structure of tRNAGly that can be efficiently recognized by the glyQS leader RNA. Alteration of the conserved residues at these positions in Su7 G36 resulted in misreading, while the same mutations in tRNAGly had no major effect on antitermination, which revealed that the role of conservation at these positions are more important for codon recognition than for antitermination.

2.3.4 A26G•G44U

Nucleotides at positions 26 and 44, which are involved in tertiary interactions at the tRNA central core, also showed anticodon-dependent conservation among various tRNA 58

species (Saks and Conery, 2007). Bases at these positions were changed from the native

A•G cis W-C pair to a G•U wobble pair (the sequence found in tRNATyr, which is the regulatory tRNA for the tyrS gene) in the context of a GCC or UCC (G34U) anticodon

(Fig. 2.1B). Similar to the C27G•G43A mutations, the A26G•G44U substitutions had no effect on any of the tested constructs, with the exception of the construct with a U•G wobble pair at position 3, which showed a 1.6-fold increase in K1/2, indicating decreased antitermination activity (Fig. 2.5, Tables A.1-A.3). This indicates that this alteration at the conserved central core sequence is generally tolerated.

2.3.5 U11C, A24G, U11C-A24G

2.3.5.1 Effects on antitermination

The Hirsh suppressor mutation (G24A), which is located at the D stem of E. coli tRNATrp, changes the original U11•G24 wobble pair to a more stable U11-A24 W-C pair.

This mutation results in a suppressor tRNA with a CCA anticodon that recognizes both the cognate UGG tryptophan codon and a UGA stop codon, which forms an A•C mismatch at position 3 of the codon (Hirsh, 1971). Mutational analysis showed that a

C11•A24 mismatch in Su7 (a suppressor tRNA derived from tRNATrp with a CUA anticodon) results in the most efficient suppression of a UAA stop codon (also forming an A•C mismatch at position 3) (Smith and Yarus, 1989a). Although there is no anticodon-dependent sequence conservation at positions 11 and 24 (Saks and Conery,

2007), the ability of a G24A mutation to alter the decoding accuracy led us to hypothesize that substitutions at the D arm of tRNAGly might also affect antitermination. 59

Wild-type tRNAGly has a U11-A24 pair, which is equivalent to the interaction formed by the Hirsh suppressor mutation in tRNATrp. Therefore, a U11C mutation that forms a

C11•A24 pair at the D arm (Fig. 2.6A) was introduced into tRNAGly with a GCC or UCC

(G34U) anticodon. This C11•A24 mismatch, which caused efficient misreading in Su7, was predicted to cause increased antitermination of the glyQS leader variants.

We observed that the effect of U11C on antitermination was dependent on the sequence of the anticodon. In the context of the native GCC anticodon, U11C resulted in a 2.6-fold reduction in antitermination efficiency of the wild-type leader RNA, and a 2.9- to 3.5-fold reduction for leader variants with a base substitution at position 3 of the

Specifier Sequence (Fig. 2.6B and Table A.1). This is different from the effect of

C11•A24 in Su7, that causes enhanced translational misreading. An A24G mutation in tRNAGly(GCC), which forms a more stable U11•G24 wobble pair than the C11•A24 mismatch (Fig. 2.6A), resulted in a 1.8- to 2.6-fold reduction in antitermination efficiency of the same leader variants (Fig. 2.6B, Table A.1). Antitermination efficiency was improved by the U11C-A24G substitutions that fully stabilized the pairing at the D arm

(Figs. 2.6A and 2.6B). This indicates that destabilization of the pairing at positions 11 and 24 in the context of the GCC anticodon resulted in decreased antitermination efficiency.

Figure 2.6. Effect of D arm mutations on antitermination. Base substitutions in tRNAGly at positions 11 and/or 24 are boxed in the D stem structure, and underlined (A). Antitermination efficiency (B-D) or percent readthrough (E and F) of constructs with pairing alterations at positions 3 (B and C), 1 (D), or both (E and F) in the absence (WT, dark bars) or presence of U11C (dark gray bars), A24G (light gray bars) or U11C-A24G (white bars) was determined. Symbols and base substitutions are denoted as described in Fig. 2.3. Error bars represent the standard error of the mean (SEM).

Figure 2.6.

In contrast, a U11C or A24G substitution in the context of a UCC (G34U) anticodon had no effect on antitermination of constructs with a stable A-U or a G•U pair at position 3 of the Specifier Sequence (Fig. 2.6C). However, U11C resulted in a 2-fold increase in antitermination efficiency for constructs with a C•U or U•U (Y•Y) mismatch at position 3, while A24G and U11C-A24G had no effect on these constructs (Fig. 2.6C).

This indicates that the same D arm mutation(s) in different anticodon contexts had distinct effects on antitermination of the glyQS leader RNA and its variants. It is also possible that the effect of D arm mutations is dependent on the context of the Specifier

Sequence-anticodon interaction. These results imply that the interaction between the

Specifier Sequence and anticodon may have an effect on how mutations in the D arm affect antitermination.

The effect of D arm mutations on antitermination was also tested in constructs with a mismatch at position 1 of the Specifier Sequence. The U11C or A24G substitution in the context of a GCC anticodon resulted in a 1.5- to 2.3-fold reduction in antitermination efficiency of constructs with a mismatch at position 1 (Fig. 2.6D, Table A.1). Although the K1/2 value of the construct with a C•C mismatch at position 1 was not affected by

A24G, the RTmax was lowered from 98% to 75% (Table A.2) and the fitted curves showed reduced antitermination throughout the tested concentrations relative to the antitermination by the wild-type tRNAGly (Fig. A.1D), which indicates reduced activity.

The U11C-A24G substitutions had no effect on antitermination of these constructs.

In contrast, U11C in the context of the UCC (G34U) anticodon resulted in a small increase in antitermination of constructs with an A•C (from 25% to 30%) or U•C (from 63

9.1% to 14%) mismatch at position 1 (Fig. 2.6E, Table A.3), and a greater increase in readthrough of constructs with an A•C or U•C mismatch at position 1 and a C•U mismatch at position 3 (Fig. 2.6F, Table A.3). The A24G and U11C-A24G substitutions in the context of the UCC anticodon had no effect on antitermination of constructs with a mismatch at position 1 (Fig. 2.6E), and constructs with double mismatches at positions 1 and 3 (Fig. 2.6F). These results indicate that the effects of D arm mutations are dependent on the context of the anticodon sequence rather than the position of the mismatch.

2.3.5.2 Effects on binding

Since the U11C substitution in tRNAGly affected antitermination depending on the anticodon sequence, we tested whether this mutation affected tRNA binding in an anticodon-dependent manner. The binding affinity of tRNAGly(GCC) or UCC (G34U) in the presence or absence of U11C to the wild-type leader RNA (with GGC Specifier

Sequence) and the variant with a GGA (C101A) Specifier Sequence was determined.

These constructs formed four different pairing interactions (C-G, A-U, A•G and C•U) at position 3 of the Specifier Sequence. All leader RNAs extended from +1 to +182

(including the antiterminator structure). Relative to the wild-type tRNA, tRNAGly(GCC)U11C resulted in a 2.1-fold and 2.3-fold reduction in binding affinity for constructs with a C-G pair and an A-G mismatch at position 3, respectively (Fig. 2.7,

Table A.4). Antitermination efficiency of these constructs was reduced in the presence of

U11C (section 2.3.5.1), which suggests that the effect of U11C on binding in the context 64

of a GCC anticodon correlated well with that on antitermination. However, U11C in the context of a UCC (G34U) anticodon resulted in a 2.7-fold decrease in binding affinity for the construct with an A-U pair at position 3, and had no effect on binding for the construct with a C•U mismatch at position 3. In contrast, antitermination efficiency of both constructs was enhanced by U11C (Section 2.3.5.1). This indicates that U11C in the context of a UCC anticodon reduces tRNA binding to the leader RNA variants, and the effect is not correlated with the anticodon sequence. The results showed that effect of

U11C on binding is not anticodon-dependent, and imply that U11C may affect other steps in antitermination in addition to binding.

Figure 2.7. Effect of U11C on binding of constructs with Specifier Sequence- anticodon pairing alterations at position 3. Response of binding to mismatches between the Specifier Sequence and anticodon in the absence (WT, dark bars) or presence of U11C (gray bars) is shown. Dissociation constant (Kd) was determined using size- exclusion filtration assays as described in the Materials and Methods. Symbols and base substitutions are denoted as described in Fig. 2.3. Error bars represent the standard error of the mean (SEM).

2.4 Discussion

The goal of this study was to identify elements in tRNA outside of the anticodon and the acceptor end that are important for efficient antitermination of the glyQS T box riboswitch. Mutations were chosen based on previous anticodon-dependent conservation and suppressor tRNA studies. We found that a U32C mutation, which results in increased misreading in translation, also enhanced antitermination efficiency of leader

RNA variants with mutations at positions 1 or 3. This indicates that the effect of this mutation on antitermination is similar to that on decoding. We also found that mutations at positions 11 and 24 of tRNAGly alter antitermination efficiency depending on the anticodon sequence, which is different from the finding for the Hirsh suppressor mutation in which a mutation in the D arm of tRNATyr affects decoding accuracy independent of the codon-anticodon mismatch. Other tRNA mutations in the central core or anticodon stem failed to cause a major effect on antitermination, which indicates that sequence conservation at these positions are less important for antitermination than for translation.

Transcription antitermination of genes in the T box family is a dynamic, multi-step process. Because the Stem I of the leader RNA is capable of specific tRNA binding

(Grigg et al., 2013; Yousef et al., 2005; Zhang and Ferre-D'Amare, 2013), binding is considered as the first step during antitermination before the decision between termination or antitermination is made. The second critical step requires proper presentation of the acceptor end of the tRNA to the antiterminator bulge to stabilize the antiterminator structure (Yousef et al., 2003). Mutations in the tRNA could have effects on either or both steps in antitermination. 66

The U32C mutation in the wild-type tRNAGly, which forms a GGC-GCC Specifier

Sequence-anticodon interaction with the wild-type glyQS leader RNA, resulted in enhanced antitermination when tRNA is limited, and decreased antitermination when tRNA is abundant (Fig. A.1A). We hypothesize that this result is due to differential effects of U32C on tRNA binding and stabilization of the antiterminator. At low concentrations of tRNA, when binding is the rate-limiting step, introduction of U32C may enhance tRNA binding affinity and result in increased antitermination. At high concentrations of tRNA, the rate-limiting step is switched from efficient binding to stabilization of the antiterminator structure, enhanced interaction at the Specifier

Sequence may constrain the proper presentation of the acceptor end to the antiterminator bulge, resulting in reduced antitermination (RTmax). We used a truncated wild-type glyQS leader RNA (1-150) without the antiterminator structure to demonstrate that U32C in tRNAGly enhanced binding, which suggests that this mutation enhances the interaction between the Specifier Sequence and the anticodon. However, whether the positioning of the acceptor end to the antiterminator bulge is affected by enhanced tRNA binding remains to be determined.

We also demonstrated that the presence of U32C enhanced tRNA binding to leader

RNAs variants (1-182) with a mutation at position 3 of the Specifier Sequence. However, the results of the binding assays did not correlate with the results of the antitermination assays in constructs with a mismatch at position 1. Preliminary binding assays using truncated tRNA variants without the acceptor end (ΔUCCA) showed that U32C enhanced tRNA binding in constructs with a position 1 mismatch, which correlated well with the 67

effect on antitermination. This indicates U32C may enhance binding interaction at the

Specifier Sequence of these constructs, but the interaction at the antiterminator bulge may be disrupted due to particular structural effect caused by the Specifier Sequence- anticodon mismatch at position 1, and therefore resulted in reduced binding of the full- length leader RNA constructs.

Gly A previous study showed that a U32C substitution in tRNA1 CCC results in increased tRNA binding to mispaired codons in the A site of the decoding center via reduced dissociation rate (Olejniczak and Uhlenbeck, 2006). Because the residue at position 32 usually forms a bifurcated hydrogen bond with the residue at position 38

(Auffinger and Westhof, 1999), and both nucleotides exhibit strong sequence conservation with the anticodon (Saks and Conery, 2007), alterations at these positions may result in a conformational change in the anticodon loop structure. It was proposed that the effect of U32C in tRNAGly on misreading is due to modulation of the anticodon loop geometry by the 32-38 base-pair (Ledoux et al., 2009; Olejniczak and Uhlenbeck,

2006). A recent X-ray crystal structure study revealed an anticodon loop rearrangement of tRNAGly upon binding to the glyQ T box leader RNA (Zhang and Ferre-D'Amare,

2013). The wild-type U32-A38 pair in tRNAGly forms only a single hydrogen bond in complex with the leader RNA, and the reshaped loop geometry superimposes well with the structure of free yeast tRNAPhe. Since tRNAPhe has a C32-A38 pair in the anticodon loop, it is possible that the U32C mutation in tRNAGly (which also forms a C32-A38 pair) pre-organizes the anticodon loop structure and facilitates binding.

We also observed that destabilization of the pairing at the D arm of tRNAGly(GCC) reduced antitermination (Figs. 2.5B and 2.5D), which indicates that antitermination efficiency is dependent on the pairing stability at positions 11 and 24. This effect was observed in constructs with a W-C, wobble, or mismatch at positions 1 or 3 of the

Specifier Sequence, which suggests that this is an anticodon-dependent effect rather than a pairing-specific effect. This is in contrast with the result of mutational analysis at positions 11 and 24 of tRNA Su7 (a UAG suppressor tRNA derived from tRNATrp), which showed that there is no direct correlation between the misreading efficiency and the pairing stability of the D arm (Smith and Yarus, 1989a). Instead, miscoding is promoted by a newly formed hydrogen bond between A24 and G44 that stabilizes the deformed tRNA structure required during the initial codon recognition process

(Schmeing et al., 2011). The different effect of D arm mutations on antitermination and decoding could be attributed to different tRNA conformational changes during these processes.

One major tRNA structural distortion during the initial decoding step (the A/T state) is a “bend and twist” located at the junction between the D arm and the anticodon arm

(Ogle et al., 2001; Valle et al., 2003). A crystal structure study of the 70S ribosome bound with G24A Trp-tRNATrp-EF-Tu revealed that stabilization of this deformed tRNA facilitates misreading (Schmeing et al., 2011). However, due to the severe deformation of tRNA structure, any displacement caused by a codon-anticodon mismatch shifts the phosphodiester backbone of the tRNA up to position 31 and does not pass beyond this residue. Therefore, the effect of the Hirsh suppressor mutation is independent of the 69

mismatch interaction. In contrast, tRNA binding to the glyQ leader RNA results in a deformed structure similar to that in the P/P state (Zhang and Ferre-D'Amare, 2013). The bend is less prominent in the P/P state than in the A/T state (Dunkle et al., 2011).

Therefore, the effect of the pairing interaction between the Specifier Sequence and the anticodon may be transmitted to the D arm of tRNA and has an impact on antitermination.

Different from the effect of D arm mutations in the context of the GCC anticodon, the stability of the pairings at the D arm in the context of the UCC anticodon did not correlate with antitermination efficiency (Fig. 2.5C). In particular, the U11C mutation, which resulted in severe reduction in antitermination in the context of the GCC anticodon, enhanced antitermination efficiency of constructs with a mismatch at either position 1 or

3 in the context of the UCC anticodon. Mutations at positions 11 and 24 should have the same structural effect on tRNAGly regardless of the anticodon sequence. In the context of the native GCC anticodon, which could provide stable stacking interactions in the

Specifier Sequence-anticodon helix, destabilization of the pairing at the D arm may disrupt the proper presentation of the acceptor end to the antiterminator bulge and result in decreased antitermination. However, replacing the G (R) with a U (Y) at the first position of the anticodon may reduce the stacking interaction and introduce some flexibility to the Specifier Sequence-anticodon helix. This flexibility may compensate for the effect of the destabilized pairing at the D stem and promote antitermination efficiency.

This could explain why the tRNAGly(UCC)U11C directed enhanced antitermination particularly in the constructs with mismatches between the Specifier Sequence and anticodon. 70

Although the U11C mutation showed an anticodon-dependent effect on antitermination, the effect on tRNA binding was not anticodon-dependent. This mutation resulted in reduced tRNA binding in the context of either the GCC or UCC anticodon.

Based on the X-ray crystal structures (Grigg et al., 2013; Zhang and Ferre-D'Amare,

2013), we predicted that destabilization of the pairing at the D arm may disrupt the docking interaction of the Stem I apical loop/AG bulge of the leader RNA on the D-/T- loops of tRNAGly (see Chapter 1). If the Stem I of the leader RNA is unable to properly cradle the tRNA and direct the acceptor end to interact with the antiterminator bulge, it is possible that the leader RNA cannot form a stable complex with the tRNA under equilibrium conditions, which therefore results in reduced binding affinity. However, because transcription antitermination is a more rapid and dynamic process than binding, and the decision between termination or antitermination has to be made in a short time frame, it is possible that the mismatch between the Specifier Sequence and anticodon results in a flexible helix that could compensate the effect of U11C, and therefore promote antitermination efficiency.

Mutations that disrupt the anticodon-dependent conservation at positions 30-40, 27-

43 and 26-44 had no major effect on antitermination, which suggests that sequence alterations at these positions are tolerated. This is consistent with the previous findings

(Grundy et al., 2000; Yousef et al., 2003) that showed there is some flexibility for tRNA sequence variation in antitermination. In addition to the anticodon and the acceptor end sequences, the T box leader RNA recognizes the overall tertiary structure of the cognate tRNA rather than specific nucleotides (Grundy et al., 2000; Yousef et al., 2003, 2005). 71

Nucleotide conservation at the anticodon stem or central core may be important in other tRNA-related processes such as translation and aminoacylation. For example, mutations at positions 30 and 40 of tRNAAla result in an increased tRNA dissociation rate from the

P site of the decoding center (Olejniczak et al., 2005). In addition, the conserved G30-

C40 nucleotides in the initiator tRNAfMet directly contact 16S rRNA residues G1338 and

A1339, and are necessary for translation initiation (Lancaster and Noller, 2005; Samhita et al., 2012). Mutations at positions 27 and 43 in tRNATrp or tRNAThr affect misreading efficiency, the level of tRNA aminoacylation, or tRNA abundance in the cell (Komine and Inokuchi, 1990; Schultz and Yarus, 1994a; Schultz and Yarus, 1994b). Although

A26G•G44U had almost no effect on antitermination, a distortion at these positions was observed in tRNAGly upon binding to the glyQ leader RNA (Zhang and Ferre-D'Amare,

2013). The A26G•G44U substitutions may still allow the conformational distortion at the central core, and therefore were tolerated during antitermination. The structural alteration at the central core is also observed during the initial decoding step and in the aminoacylation process (Caulfield et al., 2011; Peterson and Uhlenbeck, 1992; Robertus et al., 1974), which indicates that a flexible core region is important for tRNA function.

Studies on the tRNA sequence conservation revealed that some conserved elements are responsible for formation of the canonical L-shape of tRNA or recognition by its cognate aminoacyl-tRNA synthetase (Oliva et al., 2006; Saks and Conery, 2007;

Zagryadskaya et al., 2004), and some anticodon-dependent conserved elements are important for tuning tRNA for accurate decoding (Olejniczak et al., 2005; Olejniczak and

Uhlenbeck, 2006; Saks and Conery, 2007; Shepotinovskaya and Uhlenbeck, 2013). 72

Because tRNAs are also involved in the regulation of expression of amino acid-related genes in the T box family in Gram-positive organisms, they may have been subject to selective pressures from the T box leader RNA in addition to those from the translation machinery. This study reveals that some residues, either near (at position 32) or distant

(at the D arm) from the anticodon, contribute to the interaction between tRNAGly and the glyQS leader RNA; these residues were reported to affect miscoding in translation, which suggests that the Specifier Sequence-anticodon recognition in antitermination shares some properties with codon-anticodon recognition in translation. The T box leader RNA may have evolved to recognize its cognate tRNA in a way that is similar to tRNA recognition by the ribosome. However, some mutations at positions where the residues show anticodon-dependent conservation were tolerated in antitermination, which suggests that the selective pressures imposed by the T box leader RNA are not as strong as those from the translation machinery. Indeed, errors in protein synthesis may have more detrimental effects on cell survival than errors in gene regulation. Therefore, the different stringency in tRNA recognition during antitermination and translation may be due to different evolutionary constraints. This study provides new insights into the role of conserved tRNA elements in antitermination relative to the role in translation.

Whether the findings here are specific for antitermination of the glyQS gene, or can be applied to antitermination of other genes in the T box family, will be addressed in the next chapter.

CHAPTER 3

tRNAAla RECOGNITION BY COGNATE AND NONCOGNATE

T BOX LEADER RNAS

3.1 Introduction

In the previous chapter, we reported the identification of tRNA mutations that affect antitermination of the glyQS gene. However, as described in Chapter 1, the glyQS leader RNA is a natural deletion variant that lacks the Stem II and IIA/B pseudoknot structures, and represents a special group of leader RNAs in the T box family. The majority of T box leader RNAs, such as tyrS, have all the canonical structural elements

(Green et al., 2010; Rollins et al., 1997). Therefore, it was unclear whether or not the findings using the glyQS in vitro system are completely applicable to antitermination of other T box leader RNAs, or are specialized for the glyQS gene. It therefore is important to analyze another natural deletion variant that belongs to a different class of aaRS gene for comparison purposes.

All glyQS T box leader RNAs are natural deletion variants (Grundy et al., 2002b); however, the leader RNAs of alaS genes (which encode alanyl-tRNA synthetases) include both natural deletion variants and members of the canonical class (F.J. Grundy and T.M. Henkin, unpublished results). In this chapter, we compared the C. 74

acetobutylicum alaS leader RNA, which is a natural deletion variant (Fig. 3.1A), and the

B. subtilis alaS leader RNA, which has all the canonical structural elements (Fig. 3.1B), and tested for their ability to show tRNA-dependent antitermination in vitro.

Figure 3.1. The predicted structural models of (A) C. acetobutylicum and (B) B. subtilis alaS T box leader RNAs. Numbers start at the predicted transcription start-site for each gene. The RNAs are shown in the terminator form, and the alternative antiterminator structures are shown to the right of the terminator structures. The alanine Specifier Sequence (GCU) is boxed. The canonical leader structures, including Stem I, Stem II and IIA/B pseudoknot (missing in the C. acetobutylicum alaS leader RNA) and Stem III, are labeled. The conserved T box sequence is labeled in red, and the nucleotides that pair with the T box sequence in the antiterminator structure are denoted in blue. Bases are colored as described in Fig. 1.1. Asterisks indicate conserved residues.

Figure 3.1.

Continued

Figure 3.1. Continued

There are two tRNAAla isoacceptors in B. subtilis, with either GGC or UGC anticodons (Figs. 3.2A and 3.2B). B. subtilis tRNAAla(UGC) has a 5-methoxyuridine

(mo5U) at position 1 of the anticodon, while tRNAAla(GGC) has no anticodon modification; both tRNAAla isoacceptors can recognize a GCU codon (Kanaya et al.,

1999; Yamada et al., 2005). In contrast, there is only one tRNAAla isoacceptor in C. acetobutylicum, which has a UGC anticodon (Fig. 3.2C); this suggests that tRNAAla(UGC) is the effector molecule for the alaS gene in this organism. Although B. subtilis tRNAAla(UGC) and C. acetobutylicum tRNAAla(UGC) have the same anticodon and discriminator base, sequence comparisons between these two tRNAs revealed variations at five regions (Figs. 3.2B and 3.2C). Two of the five regions are located at the anticodon loop (position 32) and the central core (positions 26 and 44), which showed anticodon-dependent sequence conservation (Saks and Conery, 2007) and their importance for glyQS antitermination has been studied in Chapter 2. The goal of this study was to determine whether the sequence variations in tRNAAla(UGC) play a role in alaS antitermination.

A. B. C.

Figure 3.2. Cloverleaf diagrams of (A) B. subtilis tRNAAla(GGC), (B) B. subtilis tRNAAla(UGC) and (C) C. acetobutylicum tRNAAla(UGC). The sequence differences between the two tRNAAla(UGC) molecules are boxed and assigned a letter (A/a, B/b, C/c, D/d and E/e); lower case letters in red represent sequences found in B. subtilis tRNAAla(UGC) (abcde), and upper case letters in green represent sequences found in C. acetobutylicum tRNAAla(UGC) (ABCDE). Modifications in tRNAAla are not shown. Numbers were assigned based on tRNA numbering rules (Sprinzl et al., 1998). Note that all three tRNAAla molecules have the G3•U70 identity determinant recognized by AlaRS (Hou and Schimmel, 1988; McClain et al., 1988).

As described in previous chapters, although correct Specifier Sequence-anticodon and acceptor end-antiterminator bulge pairing are necessary for antitermination, tRNA elements outside of the two major sites may also contribute to efficient interaction with the cognate leader RNA. This chapter continues to investigate important tRNA determinants for efficient antitermination using the T box leader RNA of the alaS gene.

The results were compared with the previous findings in glyQS and tyrS antitermination, and used to determine general features of tRNA recognition by the T box leader RNAs.

This part of the work also provided insights into the importance of tRNA sequence variation among different organisms in antitermination relative to the conserved tRNA residues.

3.2 Materials and Methods

3.2.1 Generation of DNA templates

A C. acetobutylicum alaS-lacZ transcriptional fusion, which has the C. acetobutylicum alaS promoter region and leader sequence fused to a lacZ reporter gene in plasmid pFG328 (Grundy et al., 1993; J. Williams, unpublished data), was used to generate PCR templates for in vitro transcription (primers are listed in Table 3.1). The resulting product was a 374 bp DNA fragment that extends from 119 bp upstream of the transcription start-site (+1) to 53 bp downstream of the predicted termination site (+202).

A similar B. subtilis alaS-lacZ transcriptional fusion was constructed in pFG328 (see below) and used to generate a 510 bp template that extends from 120 bp upstream of the

+1 site to 134 bp downstream of the predicted termination site (+265).

A C. acetobutylicum alaS leader variant with a glycine (GGC) Specifier Sequence, denoted alaS(Gly), and a B. subtilis glyQS leader variant with an alanine (GCU) Specifier

Sequence, denoted glyQS(Ala), were generated using oligonucleotide-directed mutagenesis as described in Chapter 2. Oligonucleotides carrying the Specifier Sequence mutations (Table 3.1) were purchased from Integrated DNA Technologies (Coralville, IA) and used in PCR to generate mutant plasmids containing the desired changes in the

Specifier Sequence. The resulting products were treated with DpnI (New England 81

BioLabs) to allow cleavage of the template plasmid DNA and then propagated in E. coli strain DH5α. The Specifier Sequence mutations were confirmed by DNA sequencing

(Genewiz, Inc.).

Templates for T7 RNAP transcription of tRNA contained the T7 RNAP promoter sequence positioned at the first nt of the tRNA, and were constructed by oligonucleotide cassette ligation and amplified by PCR as described in Chapter 2. The T7 RNAP transcription protocol described in Chapter 2 was used to generate unmodified tRNA molecules.

Table 3.1. Oligonucleotide primers for alaS and glyQS leader RNA constructs. Construct Primer Sequence (5’-3’) Cac.a GlyCaUSXba GGGGAATGATACTTTCTAGACCTGTAGGAGTC alaS-lacZ GlyCAaDSH3 CATTTAATCCCAAGCTTTTCATAATCCAATCTCCTCC Bsu.b BsAlaSUSXba GGTATCGTCTAGATGATTCTGGCTG alaS-lacZ BsAlaSDSH3RC TAATGACAAGCTTGGCTCTACCGCA CaAlaS(Gly)GGC_F CTTCCTATGTAATGATTTGAAGGCATCTCTGAACTTTTTTAAGAG alaS(Gly) CaAlaS(Gly)GGC_RC CTCTTAAAAAAGTTCAGAGATGCCTTCAAATCATTACATAGGAAG GlyQS(ala)GCU_F GGATGAGCACGCAACGAAAGCTATTCTTGAGCAATTTTAAAAAAG glyQS(Ala) GlyQS(ala)GCU_RC CTTTTTTAAAATTGCTCAAGAATAGCTTTCGTTGCGTGCTCATCC a Cac. indicates C. acetobutylicum. b Bsu. indicates B. subtilis.

3.2.2 In vitro transcription antitermination assays

Single round transcription reactions using the alaS template were carried out as described in Chapter 2 with some modifications. Transcription reactions were initiated in the presence of 1 nM PCR template, 140 μM GpU dinucleotide (Sigma), 6 nM His-

tagged purified B. subtilis RNAP, 2.5 μM ATP and GTP, 0.75 μM UTP and 0.25 μM [α-

32P] UTP (800 Ci/mmol; 1 Ci = 37 GBq) in 1X transcription buffer (20 mM Tris-HCl, pH

7.9, 10 mM MgCl2, 20 mM NaCl, 0.1 mM EDTA) at 37°C for 15 min. Omission of CTP resulted in a halt at position +39 of the C. acetobutylicum alaS leader template, and at position +9 of the B. subtilis alaS template. Heparin (Sigma) was added at 25 nM to prevent reinitiation. Transcription elongation was restarted by addition of NTPs at 10

μM final concentration in the presence of increasing amounts of tRNAAla (final concentration 0, 0.25, 0.5, 1, 5, 10 μM). After 15 min incubation at 37°C, reactions were stopped by addition of 17 μl 3X stop buffer (7 M urea, 0.1 M EDTA, 4.8% glycerol, xylene cyanol and bromophenol blue) and resolved on 6% denaturing gels. The results were analyzed on a PhosphorImager (Amersham Biosciences) and quantified by

ImageQuant 5.2 software. The percent readthrough (%RT) was calculated by dividing the amount of readthrough transcript by that of the total transcript (termination and readthrough products) and multiplying by 100. The %RT determined without tRNA was subtracted from each result to eliminate the tRNA-independent readthrough, and the resulting values (corrected %RT) were used to calculate the K1/2 and RTmax as described in Chapter 2.

In vitro transcription of the mutant alaS(Gly) and glyQS(Ala) constructs was carried out using the protocols for the alaS gene (described above) and the glyQS gene

(described in Chapter 2), respectively. The T7 RNAP-transcribed tRNAs were added at

0.025 and 1 μM. Data were normalized to the result for wild-type leader RNA and wild- type tRNA at 1 μM. Each experiment was repeated at least twice. 83

3.2.3 Genetic techniques

The B. subtilis alaS-lacZ transcriptional fusion was constructed using primers

(Table 3.1) that hybridize to the region 97-120 bp upstream of the alaS transcription start- site and 110-124 bp downstream of the termination site, respectively. These primers carry XbaI and HindIII restriction enzyme sites to permit insertion into plasmid pFG328

(Grundy et al., 1993). The fusion construct was then introduced into B. subtilis strain

ZB307A (SPβc2del2::Tn917::pSK10Δ6) by transformation to allow integration into the

SPβ prophage through homologous recombination, and transformants were selected using medium containing 5 µg/ml chloramphenicol (Zuber and Losick, 1987). Phage carrying the alaS-lacZ fusion were purified by passage through B. subtilis strain ZB449 (trpC2 pheA1 abrB703, SPβ cured) (Nakano and Zuber, 1989) and introduced into B. subtilis strain 1A434 (ala-1 leuB8 metA5 pur thrC5 trpC) from the Bacillus Genetic Stock Center

(Ohio State University) by transduction. Strain 1A434 did not grow in minimal medium without L-alanine after 24 hours of incubation at 37°C.

To construct plasmid-borne tRNA variants, we amplified the tRNA sequence by

PCR using primers (Table 3.2) that contain HindIII and SalI restriction enzyme sites to allow insertion of the tRNA gene into plasmid pDG148 (Stragier et al., 1988). The resulting constructs were introduced into E. coli strain DH5α by transformation. After propagation and purification, the plasmid was introduced into B. subtilis strain 1A434 containing the alaS-lacZ fusion construct by transformation and selected with 5 µg/ml of both chloramphenicol and neomycin. Expression of the plasmid-borne tRNA was

controlled by the Pspac promoter, and was induced by the addition of 1 mM isopropyl-β-

D-thiogalactopyranoside (IPTG).

Table 3.2. Oligonucleotide primers for construction of plasmid-borne tRNAAla variants. tRNA Primer name Sequence (5’-3’) construct a Cac Ca_AlatRNAUSHindIII CCTACTGGAAGCTTGGGGGATTAGCTCAGCTGGGA tRNAAla(UGC) U70C Ca_AlatRNADSU70CSalI TAGGAGAGGTCGACTGGTGGGGAATAAGAGATTCG b Bsu Bs_AlatRNAUSHindIII CCTACTGGAAGCTTGGGGCCTTAGCTCAGCTGGGAG tRNAAla(UGC) U70C Bs_AlatRNADSU70CSalI TAGGAGAGGTCGACTGGTGGGGCCTAGCGGGATCGAACCGC Bsu. Bs_AlatRNAGGCUSHindIII CCTACTGGAAGCTTGGGGCCATAGCTCAGCTGGG tRNAAla(GGC) U70C Bs_AlatRNAGGCDSU70CSalI TAGGAGAGGTCGACTGGTGGGGCCAAGGGGGCTCGAA aCac. indicates C. acetobutylicum. bBsu. indicates B. subtilis.

3.2.4 Bacterial growth conditions and β–galactosidase assays

For alanine starvation experiments, B. subtilis strain 1A434 carrying an alaS-lacZ transcriptional fusion was grown to mid-exponential phase in minimal medium

(Anagnostopoulos and Spizizen, 1961) with 5 µg/ml chloramphenicol and 50 µg/ml of the appropriate amino acids (L-alanine, L-leucine, L-methionine, L-threonine, L- tryptophan) and adenine, harvested and split equally into fresh medium with or without

50 µg/ml L-alanine. Cells were collected at 1 h intervals for β–galactosidase assays. For

IPTG induction assays, strain 1A434 with an alaS-lacZ fusion construct and plasmid pDG148 carrying a tRNA construct was grown in minimal medium as described above with the addition of 5 µg/ml neomycin. Expression of the tRNA construct was induced

by adding 1 mM IPTG. β–galactosidase activities were measured using toluene- permeabilized cells as described by Miller (Miller, 1972). Each experiment was repeated at least twice.

3.3 Results

3.3.1 tRNAAla-directed alaS antitermination in vitro

We predicted structural models for the C. acetobutylicum and the B. subtilis alaS leader RNAs (Fig. 3.1) based on canonical features of T box leader RNAs. The C. acetobutylicum alaS leader RNA, which is similar to the glyQS leader RNA, belongs to the group of natural deletion variants that lack the Stem II and IIA/B elements; other structural elements, including Stem I, III and the mutually exclusive terminator and antiterminator structures, match the canonical pattern. Although the B. subtilis alaS leader RNA, which is similar to the tyrS leader RNA, has the Stem II and IIA/B elements, the conserved S-turn in Stem II is missing.

Also, this RNA has a non-canonical GA motif below the Specifier Loop. The canonical GA motif is composed of two short helices connected by a 3-nt bulge; there are two noncanonical pairs (G•A and A•G) in one helix, and three canonical W-C pairs on the other. This motif forms a kink-turn structure and is found at the base of the Stem I of most T box leader RNAs (Winkler et al., 2001). The GA motif in the B. subtilis alaS leader RNA is composed of a 4-nt bulge, and the helix at bottom of the bulge has a U•U mismatch that separates the two C-G pairs. Noncanonical kink-turn motifs have been reported previously (Lescoute et al., 2005; Schroeder et al., 2010; Winkler et al., 2001), 86

with variations in the bulge region or the pairing interaction at the helix. Therefore, it is possible that the kink-turn motif in the Stem I of the B. subtilis alaS leader RNA is still formed, but the structural arrangement may be different from the canonical motif found in other T box leader RNAs. The Specifier Sequence of both alaS leader RNAs is a GCU codon, which represents an exception from the conserved C residue at position 3 of the

Specifier Sequence in most T box leader RNAs and is commonly found in tRNAAla- regulated T box genes (Gutierrez-Preciado et al., 2009).

We tested the ability of B. subtilis tRNAAla isoacceptors and C. acetobutylicum tRNAAla to mediate antitermination of both alaS genes in vitro. In Chapter 2, we showed that antitermination of the wild-type glyQS gene reaches saturation at 0.5 µM tRNAGly.

Here, we chose to use 1 µM and 10 µM tRNAAla for the initial test for alaS antitermination. When 1 µM tRNAAla was included in the assay, no antitermination of the B. subtilis alaS gene was observed using either B. subtilis tRNAAla isoacceptor or C. acetobutylicum tRNAAla(UGC) (data not shown). When 10 µM B. subtilis tRNAAla(GGC) was tested, readthrough of the cognate B. subtilis alaS terminator was still <5% (Fig.

3.3A), while B. subtilis tRNAAla(UGC) resulted in a 2-fold increase in antitermination relative to tRNAAla(GGC). This indicates that B. subtilis tRNAAla(UGC) directed more efficient antitermination of its cognate alaS gene than tRNAAla(GGC), even though the

UGC anticodon forms a less stable U•U mismatch at position 3 of the GCU Specifier

Sequence than the U•G wobble pair formed by the GGC anticodon. Antitermination of the B. subtilis alaS gene using C. acetobutylicum tRNAAla(UGC) was <5%, which is similar to the result using B. subtilis tRNAAla(GGC). It appears that antitermination of 87

the B. subtilis alaS gene is not very efficient under these conditions, even in the presence of the cognate tRNAAla. This is consistent with the general observation that leader RNAs with the complete canonical structures (including Stems II and IIA/B) generally exhibit low antitermination efficiency in vitro (Grundy et al., 2002a; Putzer et al., 2002; R.N.

Williams-Wagner, F.J. Grundy, T.M. Henkin, unpublished results). Certain unknown cellular factors or the use of the fully modified tRNA may be required for efficient antitermination of leader RNAs in the class with all the canonical structural elements in the in vitro conditions.

A . B . s u b tilis a la S B . C . a c e to b u ty lic u m a la S

3 5 3 5

3 0 3 0

)

d e

e 2 5 2 5

e r

r 2 0 2 0

o c

c 1 5 1 5

(

T T

R 1 0

% % 5 5

0 0 ) ) ) ) ) ) C C C C C C G G G G G G G U U G U U la ( la ( la ( la ( la ( la ( A A A A A A A A A A A A N N N N N N R R R R tR t t tR t t ...... u u c u u c s s a s s a B B C B B C

Figure 3.3. The alaS antitermination in vitro. Antitermination of the (A) B. subtilis and (B) C. acetobutylicum alaS genes in the presence of 10 µM T7 RNAP-transcribed tRNAAla was determined by in vitro transcription assays. The tRNA-independent readthrough was subtracted from each result. Error bars represent the standard error of the mean (SEM). Bsu. represents B. subtilis; Cac. represents C. acetobutylicum. 88

When the C. acetobutylicum alaS gene was tested, the cognate C. acetobutylicum tRNAAla(UGC) at 1 µM resulted in 23% readthrough (data not shown), while this tRNA at 10 µM resulted in 28% readthrough (Fig. 3.3B). This indicates that antitermination is very efficient in the C. acetobutylicum alaS leader RNA, which is a natural deletion variant. The result is consistent with our previous in vitro tests that showed tRNA- mediated antitermination using the B. subtilis glyQS leader RNA (a natural deletion variant) (Grundy et al., 2002b).

B. subtilis tRNAAla(GGC) at 10 µM resulted in <10% readthrough of the C. acetobutylicum alaS terminator, although the activity was 2-fold greater than B. subtilis tRNAAla(UGC) (Fig. 3.3B). This is opposite to the findings in antitermination of the B. subtilis alaS gene, where B. subtilis tRNAAla(UGC) directed 2-fold more efficient antitermination than tRNAAla(GGC). The result suggests that the tRNA requirements for antitermination of the C. acetobutylicum alaS gene may be different from that of the B. subtilis alaS gene. Although the U•G wobble pair at position 3 of the Specifier Sequence may allow more efficient antitermination than the U•U mismatch, the cognate C. acetobutylicum tRNAAla(UGC) also forms a U•U mismatch at position 3 as the B. subtilis tRNAAla(UGC). This suggests that there are other requirements for efficient antitermination in addition to the pairing stability between the Specifier Sequence and anticodon. B. subtilis tRNAAla(UGC) and C. acetobutylicum tRNAAla(UGC) share 80% sequence identity, including conservation of the anticodon and discriminator base that form pairing interactions with the alaS Specifier Sequence and the variable position of the antiterminator bulge, respectively. However, B. subtilis tRNAAla(UGC) failed to 89

direct efficient antitermination of the C. acetobutylicum alaS gene in vitro, suggesting that tRNA elements outside of the anticodon and discriminator base may contribute to efficient C. acetobutylicum alaS antitermination.

3.3.2 tRNAAla requirements for alaS antitermination

tRNAAla-dependent antitermination of the wild-type C. acetobutylicum alaS gene was more efficient using its tRNAAla(UGC) than the heterologous B. subtilis tRNAAla(UGC) (Fig. 3.4). We compared the sequences of B. subtilis tRNAAla(UGC) and

C. acetobutylicum tRNAAla(UGC) to identify elements that may contribute to the efficiency. Variations at five regions, including positions 32 (anticodon loop), 26 and 44

(central core), 59 and 60 (T loop), and pairings at 50-64, 51-63, 52-62 (T stem), and 5-68,

6-67 (acceptor stem), were identified (Figs. 3.2B and 3.2C). Only the base at position 32 and the pair 52-62 at the T stem showed moderate anticodon-dependent conservation among heterologous tRNAAla(UGC) found in both Gram-positive and -negative bacteria

(Saks and Conery, 2007); residues at other positions exhibit no conservation in tRNAAla(UGC). Upper-case (C. acetobutylicum tRNAAla(UGC)) and lower-case (B. subtilis tRNAAla(UGC)) letters were assigned to represent the nucleotide identities at each position.

Figure 3.4. Antitermination of the C. acetobutylicum alaS gene using B. subtilis tRNAAla(UGC) and C. acetobutylicum tRNAAla(UGC). Single round transcription reactions of the C. acetobutylicum alaS leader template were performed in the presence of increasing amounts of tRNAAla at 0, 0.25, 0.5, 1.0, 5.0, 10 μM final concentrations. C. acetobutylicum tRNAAla(UGC) (green) directed efficient antitermination, but B. subtilis tRNAAla(UGC) (red) did not. tRNA-independent readthrough was subtracted from each data point, and the resulting values were fit to a hyperbolic equation to determine the K1/2 and RTmax. All experiments were repeated at least twice.

Elements “abcde” found in B. subtilis tRNAAla(UGC) were replaced individually by elements “ABCDE” found in C. acetobutylicum tRNAAla(UGC) to identify elements that promote efficient C. acetobutylicum alaS antitermination. An anticodon loop U32C mutation (aA) in B. subtilis tRNAAla(UGC) had no effect on antitermination of the C. acetobutylicum alaS gene (Abcde, Fig. 3.5A). This result is different from the effect of a

U32C substitution in tRNAGly that enhanced glyQS antitermination of a construct with a

GGU•UCC Specifier Sequence-anticodon pair, which also has a U•U mismatch at position 3 of the Specifier Sequence (Chapter 2). The tRNA variant with G26A•A44G

substitutions (bB) directed 2-fold less efficient antitermination (aBcde, Fig. 3.5A) as compared with C. acetobutylicum tRNAAla(UGC), and the addition of U32C resulted in a

2-fold increase in antitermination efficiency (ABcde, Fig. 3.5B) relative to the construct with only G26A•A44G substitutions. This indicates that the B. subtilis tRNAAla(UGC) variant with base substitutions at the central core promotes C. acetobutylicum alaS antitermination with reduced activity relative to the cognate C. acetobutylicum tRNAAla, and the addition of the anticodon loop substitution resulted in activity comparable to that of the cognate C. acetobutylicum tRNAAla. This result showed that the U32C enhanced the activity of the construct that exhibited reduced antitermination efficiency relative to the wild-type tRNAAla, which is similar to the role of U32C in glyQS antitermination.

A B. subtilis tRNAAla(UGC) variant with nucleotides transplanted from the T stem of C. acetobutylicum tRNAAla(UGC) (G50A-C64U, C51G-G63C, G52A-C62U, dD) directed antitermination (abcDe, Fig. 3.5A) equivalent to that of C. acetobutylicum tRNAAla(UGC), and the introduction of U32C resulted in a 3-fold increase in antitermination (AbcDe, Fig. 3.5B) relative to the construct with T stem substitutions alone. These substitutions reduced the pairing stability at the T stem (G-C vs. A-U), which indicates that destabilization of the pairings at the T stem of B. subtilis tRNAAla(UGC) is sufficient for the tRNA variant to direct C. acetobutylicum alaS antitermination as well as the cognate C. acetobutylicum tRNAAla(UGC), and the addition of the anticodon loop substitution that enhances the interaction at the Specifier Sequence further promotes the efficiency. The T stem substitutions may assist proper presentation of the acceptor end to the antiterminator bulge, and promote antitermination. 92

A B. subtilis tRNAAla(UGC) variant with acceptor stem substitutions (C5G-G68U,

C6A-G67U, eE) resulted in a 3-fold reduction in antitermination (abcdE, Fig. 3.5A) relative to C. acetobutylicum tRNAAla(UGC). The C-G to G•U and A-U substitutions reduced the pairing stability at the acceptor stem, which indicates that destabilization of the pairing at the acceptor stem of B. subtilis tRNAAla(UGC) promotes C. acetobutylicum alaS antitermination, although the efficiency is not comparable to that of the cognate tRNAAla. Addition of the U32C mutation in this tRNA variant resulted in a 7-fold reduction in the K1/2 and a reduction in the RTmax from 35% to 21% (AbcdE, Fig. 3.5B).

A similar effect was observed when the U32C mutation was introduced into tRNAGly in the context of the wild-type glyQS leader RNA-tRNAGly(GCC) interaction (Chapter 2).

The result indicates that the anticodon loop and acceptor stem substitutions in tRNAAla may have an effect on alaS antitermination similar to that of the U32C mutation in tRNAGly on glyQS antitermination (Chapter 2).

Substitutions at the T loop (U59A C60U, cC) of B. subtilis tRNAAla(UGC) completely abolished antitermination (abCde, Fig. 3.5A). However, the addition of

U32C in this tRNA construct resulted in efficient antitermination (AbCde, Fig. 3.5B), similar to that of C. acetobutylicum tRNAAla(UGC). Because neither the T loop nor anticodon loop substitutions in B. subtilis tRNAAla(UGC) affected antitermination, the fact that the combination of both substitutions enhanced antitermination suggests that the structural effects of substitutions at both positions have synergistic effects on antitermination.

Figure 3.5. Effects of tRNA mutations on antitermination. (A) Cloverleaf model of B. subtilis tRNAAla(UGC) with sequences shown at five regions (abcde) where there are differences from C. acetobutylicum tRNAAla. Elements found in C. acetobutylicum tRNAAla(UGC) (ABCDE) were introduced individually. The resulting construct was tested with the C. acetobutylicum alaS template in in vitro transcription assays to determine the RTmax and K1/2, and the results are listed next to the tested substitutions. (B) Effects of substitutions in B. subtilis tRNAAla(UGC)U32C (Abcde) on antitermination. (C) Effects of substitutions in B. subtilis tRNAAla(UGC)U59A C60U (abCde) on antitermination. (D) Effects of substitutions in B. subtilis tRNAAla(UGC)U32C U59A C60U (AbCde) on antitermination. * indicates that the RTmax and K1/2 were calculated without the readthrough values determined at the highest two tRNA concentrations.

In sum, we showed that base substitutions in B. subtilis tRNAAla(UGC) at the central core, the T stem or the acceptor stem with nucleotides found in C. acetobutylicum tRNAAla(UGC) enhanced C. acetobutylicum alaS antitermination. This suggests that residues at these positions are not directly recognized by the alaS leader RNA. Instead, they may have some structural effects (e.g., destabilization of the pairing interaction) on tRNAAla, which allows efficient alaS antitermination. B. subtilis tRNAAla(UGC) variants with either the U32C or U59A C60U mutations failed to direct efficient C. acetobutylicum alaS antitermination; however, the combination of both substitutions showed a synergistic effect on antitermination. Addition of U32C in tRNAAla constructs with substitutions at the central core or T stem enhanced antitermination efficiency, while addition of U32C in the construct with acceptor stem substitutions resulted in differential effects on antitermination. Based on the effect of U32C in tRNAGly on glyQS antitermination, we predicted that U32C in tRNAAla may also enhance the interaction between the Specifier Sequence and anticodon. However, this effect was not sufficient for B. subtilis tRNAAla(UGC) to promote efficient alaS antitermination, which suggests that other interactions in antitermination (e.g., acceptor end-antiterminator bulge) may be required.

We further investigated the relationship between the T loop substitutions and substitutions at the central core, T stem or acceptor stem. Base substitutions at the T loop of B. subtilis tRNAAla(UGC) with those found in C. acetobutylicum tRNAAla(UGC)

(cC, U59A C60U) generally reduced antitermination and antagonized the effect of other effective tRNA substitutions (Fig. 3.5C); this antagonistic effect of U59A C60U 95

was partially reduced by the addition of U32C (Fig. 3.5D). The B. subtilis tRNAAla(UGC) variant with the G26A•A44G substitutions (aBcde, Fig. 3.5A) directed 2-fold less efficient antitermination as compared with the cognate C. acetobutylicum tRNAAla, and the activity was completely inhibited by U59A C60U (aBCde, Fig. 3.5C). The addition of U32C in this construct (ABCde, Fig. 3.5D) resulted in increased RTmax relative to the tRNA variant with U59A C60U and G26A•A44G substitutions, although the activity was still very low. This indicates that the structural effects caused by the combination of

U59A C60U and G26A•A44G are very detrimental. U59A C60U antagonized the effect of G26A•A44G in B. subtilis tRNAAla(UGC) on alaS antitermination, although the antagonistic effect can be partly reduced by enhanced interaction at the Specifier

Sequence.

The B. subtilis tRNAAla(UGC) variant with less stable pairings at the acceptor stem

(abcdE, Fig. 3.5A) directed 3-fold less efficient antitermination than the cognate C. acetobutylicum tRNAAla, while the addition of U59A C60U in this tRNA variant resulted in a 2-fold decrease in antitermination efficiency and RTmax (abCdE, Fig. 3.5C). This indicates that tRNA variant with less stable U59A C60U than the variant with the central core substitutions. The addition of U32C (AbCdE, Fig. 3.5D) resulted in a 2-fold increase in antitermination efficiency relative to the variant with U59A C60U and the acceptor stem substitutions, which indicates that enhanced interaction at the Specifier

Sequence reduces the effect of U59A C60U.

The B. subtilis tRNAAla(UGC) variant with the less stable pairings at the T stem

(abcDe, Fig. 3.5A) directed antitermination that is comparable to the cognate C. 96

acetobutylicum tRNAAla, and the activity was also inhibited by the addition of U59A

C60U (abCDe, Fig. 3.5C). However, the addition of U32C (AbCDe, Fig. 3.5.D) had no effect in the activity, indicating that the U59A C60U substitutions is very detrimental to the variant with T stem substitutions. Together, this study revealed that the U59A C60U substitutions disrupt the function effect of the substitutions at the central core, T stem and acceptor stem in B. subtilis tRNAAla(UGC) to a different degree, and the U32C mutation could partially reduce the antagonistic effect of U59A C60U. This also indicates that the identity at positions 59 and 60 are important for efficient antitermination, which is consistent with the previous findings that showed only certain combinations of bases at these positions in tRNATyr directs efficient tyrS antitermination in vivo (Grundy et al.,

2000).

We tested all possible combinations of base substitutions at these five positions, and the results are listed in Table 3.3 and summarized in Fig. 3.6. Constructs that exhibited high RTmax and low K1/2 values indicate efficient antitermination, while constructs that exhibited low RTmax and high K1/2 values indicate low antitermination activity. 14 of the 32 tRNAAla constructs, including B. subtilis tRNAAla(UGC), showed indeterminable high K1/2 values (n/a, Fig. 3.6). The other constructs revealed various levels of antitermination activity relative to the wild-type C. acetobutylicum tRNAAla(UGC).

Ala Figure 3.6. The RTmax and K1/2 of 32 tRNA (UGC) constructs for C. acetobutylicum alaS antitermination. The RTmax and K1/2 of each construct were determined by in vitro transcription antitermination assays. The C. acetobutylicum tRNAAla (ABCDE) and B. Ala subtilis tRNA (abcde) are circled. n/a indicates that the K1/2 value was too high to be determined accurately.

Among the 14 tRNAAla constructs that showed inefficient C. acetobutylicum alaS antitermination, 11 of them have the element “e”, which represents the stable C5-G68 and C6-G67 pairs found at the acceptor stem of B. subtilis tRNAAla(UGC). These constructs include the one with (ABCDe) elements, which represents a C. acetobutylicum tRNAAla(UGC) variant with base substitutions at the acceptor stem from the B. subtilis tRNAAla(UGC) (Ee). Base substitutions in C. acetobutylicum tRNAAla(UGC) at the anticodon loop (Aa, aBCDE) and the T loop (Cc, ABcDE) resulted in reduced

activity but still maintained the ability to direct alaS antitermination, while substitutions at the central core (Bb, AbCDE) or the T stem (Dd, ABCdE) resulted in reduced

RTmax from 32% to 24%, but a 6.3- and 3.6-fold increase in antitermination efficiency, respectively. This indicates that stabilization of the pairing at the acceptor arm of C. acetobutylicum tRNAAla(UGC) are detrimental to antitermination. The C5G-G68U,

C6A-G67U substitutions (eE) that destabilized the pairing in most tRNA constructs were sufficient to promote antitermination (e.g., abcdeabcdE). This suggests that a less rigid acceptor stem structure is important for efficient antitermination of the C. acetobutylicum alaS gene. There are 5 tRNA constructs that contain stable pairing at the acceptor stem and still directed efficient antitermination (aBcde, ABcde, AbCde, abcDe,

AbcDe). This could be due to combination of specific elements at other positions of tRNAAla have an effect on tRNA structure that could compensate the rigidity at the acceptor stem.

We found that the U32C substitution (aA) enhanced the activity in most of the tRNA constructs that showed effective antitermination, which indicates that enhanced interaction at the Specifier Sequence promotes antitermination. Replacement of the bases found in B. subtilis tRNAAla(UGC) with bases found in C. acetobutylicum tRNAAla(UGC) at the central core (bB), T loop (cC) or T stem (dD) revealed that these substitutions resulted in increased or decreased antitermination depending on the context of each tRNA construct. For example, B. subtilis tRNAAla(UGC) variants with substitutions at either the central core (aBcde) or the T stem (abcDe) directed efficient alaS antitermination, suggesting that these substitutions have a positive effect on 99

antitermination. However, combination of both the central core and the T stem substitutions (aBcDe) resulted in a complete loss of antitermination, which indicates that these substitutions antagonized the effect of each other. Antitermination directed by the tRNA variant with elements (AbcDE) was not very efficient; however, the addition of the

T loop substitutions (AbCDE), the central core substitutions (ABcDE), or both (ABCDE) resulted in efficient antitermination. The results show that bases at these positions may collaborate to achieve efficient antitermination, or antagonize the effect of each other depending on their context.

Overall, we showed that sequence variations between B. subtilis tRNAAla(UGC) and C. acetobutylicum tRNAAla(UGC) at the anticodon loop, central core, T loop, T stem and acceptor stem contribute to the efficiency in antitermination of the C. acetobutylicum alaS gene. Residues at these positions may not be recognized directly by the alaS leader

RNA. Instead, they may contribute to antitermination by causing a structural effect, which could affect the interaction between the Specifier Sequence and anticodon, or the presentation of the acceptor end to the antiterminator bulge. Our work revealed that C. acetobutylicum tRNAAla(UGC) is finely tuned by these sequence elements to form an optimal interaction with its cognate T box leader RNA.

100

3.3. Hybrid tRNAAla

101

3.3.3 Expression of alaS-lacZ fusions in B. subtilis

In Section 3.3.1, we showed that tRNA-dependent antitermination can be demonstrated in vitro using the C. acetobutylicum alaS leader RNA and its cognate tRNAAla(UGC), while antitermination was greatly reduced using both B. subtilis tRNAAla isoacceptors. Because B. subtilis tRNAAla isoacceptors also resulted in limited antitermination of the cognate alaS gene in vitro, it was unclear whether this result is due to the lack of proper tRNA modification or due to the structural complexity of the B. subtilis alaS leader RNA. To test the ability of the fully modified B. subtilis tRNAAla to direct antitermination of the C. acetobutylicum alaS gene, we introduced a C. acetobutylicum alaS-lacZ transcriptional fusion into B. subtilis strain 1A434 (an alanine auxotroph) and monitored its expression under alanine starvation conditions. A B. subtilis alaS-lacZ transcriptional fusion was used as a control to demonstrate that B. subtilis tRNAAla is capable of inducing expression under alanine starvation conditions.

We observed a 2-fold increase in expression of the B. subtilis alaS-lacZ fusion when cells were starved for alanine, and no significant induction of the C. acetobutylicum alaS-lacZ fusion (Table 3.4). This indicates that tRNAAla in B. subtilis directs antitermination of the

B. subtilis alaS gene, but not C. acetobutylicum alaS.

102

Table 3.4. Expression of alaS-lacZ fusions. β-galactosidase activity (Miller units) Induction ratio uninduced induced B. subtilis alaS-lacZ 11 ± 1 22 ± 1 2.0 C. acetobutylicum alaS-lacZ 5.9 ± 1.7 6.4 ± 1.3 1.1

To further test if the C. acetobutylicum alaS-lacZ gene can be expressed in B. subtilis, we introduced a plasmid-borne IPTG-inducible C. acetobutylicum tRNAAla(UGC) construct and monitored the expression of the lacZ gene in the presence and absence of

IPTG. The G3•U70 pair is a highly conserved, major tRNAAla recognition determinant for AlaRS (Hou and Schimmel, 1988; McClain and Foss, 1988). A U70C mutation was introduced into the plasmid-borne tRNAAla to allow accumulation of uncharged tRNAAla construct after induction. This U70C mutation was tested in the in vitro transcription antitermination assays and resulted in enhanced antitermination efficiency (RTmax = 48%

Ala ± 1; K1/2 = 0.33 μM ± 0.03) relative to the wild-type C. acetobutylicum tRNA (RTmax =

32% ± 1; K1/2 = 0.44 μM ± 0.06), which indicates that the plasmid-borne tRNA should function in antitermination in vivo.

A plasmid-borne C. acetobutylicum tRNAAla(UGC)U70C construct was introduced into B. subtilis strains carrying either the C. acetobutylicum alaS-lacZ fusion or the B. subtilis alaS-lacZ fusion. However, no expression of the reporter gene was observed in either construct under inducing conditions (data not shown). To ensure that the plasmid- borne tRNA variants can be expressed properly in B. subtilis, we constructed plasmid- borne B. subtilis tRNAAla(UGC)U70C and introduced it into the strain carrying the B. 103

subtilis alaS-lacZ gene to test if the expression of the lacZ gene could be induced. We also constructed a plasmid-borne B. subtilis tRNAAla(GGC)U70C variant to test if this tRNAAla isoacceptor is the effector molecule for the B. subtilis alaS gene. However, we did not observe any significant induction of expression of the reporter gene. Therefore, it is still unclear whether C. acetobutylicum tRNAAla(UGC) forms a specific interaction only with its cognate alaS leader RNA in a B. subtilis host, or whether the tRNAAla(UGC) in B. subtilis could also interact with C. acetobutylicum alaS. Expression of the plasmid- borne tRNA constructs in vivo remains to be determined.

3.3.4 tRNA recognition by noncognate leader RNA

To compare the tRNA requirements for alaS and glyQS antitermination, we swapped the Specifier Sequence of the glyQS and alaS leader RNAs, and determined the antitermination activities using tRNAGly and tRNAAla with corresponding changes in the anticodon and discriminator base to match the interactions at the Specifier Sequence and antiterminator bulge. Four leader RNAs and eight tRNA constructs were generated, and the antitermination activities were determined using 0.025 and 1 µM tRNA in in vitro transcription antitermination assays. We predicted that the wild-type leader RNA-tRNA interaction (glyQS-tRNAGly or alaS-tRNAAla) shows the best activity relative to other leader RNA-tRNA variants.

In the context of the alanine Specifier Sequence-anticodon interaction (GCU•UGC),

0.025 µM tRNAAla and tRNAGly variant resulted in equivalent antitermination of the alaS leader RNA, while 1 µM tRNAGly resulted in a 25% increase in antitermination as 104

compared with tRNAAla (125% vs. 100%, Fig. 3.7A). The GCU•UGC to GGC-GCC change (AlaGly) in the context of alaS leader RNA-tRNAAla resulted in a 1.9-fold increase in antitermination at 0.025 µM relative to the wild-type construct, and no major effect on antitermination at 1 µM. This GGC-GCC pair in the context of tRNAGly-alaS leader RNA also resulted in a 1.6-fold increase in antitermination at both 0.025 and 1 µM relative to the original GCU•UGC pair. This indicates that stabilization of the Specifier

Sequence-anticodon pairing enhances alaS antitermination. The tRNAGly variant at 0.025

µM directed antitermination similar to that of tRNAAla, and a 1.9-fold enhanced antitermination relative to tRNAAla at 1 µM. This result indicates that elements in tRNAGly promote alaS antitermination when the tRNA concentration is close to saturation. The effect may be related to proper interaction between the acceptor end and the antiterminator bulge.

In the context of the glycine Specifier Sequence-anticodon interaction (GGC-GCC), the wild-type tRNAGly at either 0.025 or 1 µM promoted the most efficient antitermination of the glyQS leader RNA relative to other tRNA variants (Fig. 3.7B).

The tRNAAla variant with a glycine anticodon resulted in a 5-fold decrease in antitermination at 0.025 µM and a 1.6-fold decrease in antitermination at 1 µM relative to the wild-type tRNAGly, which indicates that tRNAAla is less effective in directing glyQS antitermination than tRNAGly, especially when tRNA is limited. This implies that tRNAAla has less binding affinity to the glyQS leader RNA as compared to tRNAGly.

The Specifier Sequence-anticodon change from GGC-GCC to GCU•UGC

(GlyAla) resulted in a 17-fold decrease in antitermination in the context of tRNAGly at 105

0.025 µM, and had limited effects on antitermination at 1 µM. This suggests that the

U•U mismatch at position 3 of the Specifier Sequence has detrimental effects on antitermination when tRNAGly is limited, and the defect could be compensated by high abundance of tRNAGly. In contrast, the GlyAla Specifier Sequence-anticodon swap in the context of tRNAAla resulted in a 2-fold reduction in antitermination at 0.025 µM, and a small increase in antitermination from 62% to 73% at 1 µM. This indicates that the

U•U mismatch at position 3 of the Specifier Sequence in the context of the glyQS leader

RNA reduces antitermination when tRNAAla is limited, and resulted in greater antitermination when tRNAAla is saturated as compared with the C-G match at position 3.

This effect could be explain by the two-step model, in which a mismatch between the

Specifier Sequence and anticodon affect tRNA binding step when the tRNA concentration is low, and the positioning step when tRNA concentration is high (see

Chapter 2).

In sum, these experiments showed that tRNAGly resulted in greater antitermination of either its cognate glyQS gene and the alaS gene than tRNAAla, especially at 1 µM under the same pairing interactions between the Specifier Sequence and anticodon, and between the antiterminator bulge and acceptor end. This indicates that elements in tRNAGly outside of the anticodon and discriminator base may assist efficient antitermination, and elements in tRNAAla outside of the anticodon and discriminator base may constrain efficient antitermination. In most cases, the construct with a more stable

GGC-GCC Specifier Sequence-anticodon pair resulted in greater antitermination as compared with the less stable GCU•UGC pair. The only exception is that in the context 106

of the glyQS leader RNA and tRNAAla, the GCU•UGC pair resulted in a small increase in antitermination relative to that of the GGC-GCC pair at 1 µM tRNAAla. This suggests that the less stable pairing at position 3 of the Specifier Sequence may allow tRNAAla to promote efficient antitermination.

107

Figure 3.7. Antitermination of the alaS and glyQS genes with either the wild-type or mutant Specifier Sequence. (A) Antitermination of the wild-type C. acetobutylicum alaS(Ala) gene and the alaS(Gly) variant (with a GGC glycine Specifier Sequence) was determined using 0.025 and 1.0 µM wild-type tRNAAla(Ala)_ACCA, tRNAAla(Gly)_ACCA (with a glycine anticodon), tRNAGly(Ala)_ACCA (with an alanine anticodon and an ACCA acceptor end) and tRNAGly(Gly)_ACCA (with an ACCA acceptor end) in in vitro transcription assays. (B) Antitermination of the wild-type B. subtilis glyQS(Gly) gene and the glyQS(Ala) variant (with a GCU alanine Specifier Sequence) using 0.025 and 1.0 µM wild-type tRNAGly(Gly)_UCCA, tRNAAla(Gly)_UCCA (with a glycine anticodon and a UCCA acceptor end), tRNAGly(Ala)_UCCA (with an alanine anticodon) and tRNAAla(Ala)_UCCA (with a UCCA acceptor end). All results were normalized to the readthrough values determined using the wild-type leader RNA and 1.0 µM wild-type tRNA. Each experiment was repeated at least twice. 108

3.4 Discussion

Previous studies on antitermination of the glyQS and tyrS genes have revealed tRNA requirements for efficient antitermination (Grundy et al., 2000; Grundy et al.,

1997b; Yousef et al., 2003). However, the results were hard to compare due to differences between in vitro and in vivo experimental conditions, and also due to use of genes that belong to different aaRS families. The requirements for antitermination of leader RNAs in the class of natural deletion variants may be different from those with the canonical structures. In this chapter, we utilized the leader RNA of the C. acetobutylicum alaS gene to establish another in vitro system and investigated tRNAAla requirements for efficient alaS antitermination; the results were compared with the previous findings in glyQS and tyrS antitermination and used to determine the general features for efficient antitermination.

In the process of testing alaS leader RNAs for in vitro transcription antitermination, we observed that antitermination of the B. subtilis alaS leader RNA (with canonical structural elements) is not very efficient in vitro, while antitermination of the C. acetobutylicum alaS leader RNA (without Stem II and IIA/B elements) is more efficient.

This result is similar to our previous observations in the tyrS and glyQS in vitro analyses

(Grundy et al., 2002a; Grundy et al., 2002b), where tRNA-dependent antitermination could be demonstrated using the glyQS natural deletion variant, but not the tyrS leader

RNA with all the canonical structural elements. tRNA-dependent antitermination of leader RNAs with the Stem II and IIA/B elements was previously demonstrated in vitro using fully modified tRNA extracted from the cell and cellular extracts, or modified 109

tRNA and high concentration of polyamine (Putzer et al., 2002; R. Williams-Wagner, F.J.

Grundy and T.M. Henkin, unpublished results), suggesting different requirements for antitermination of leader RNAs in this class as compared with those in the class of natural deletion variants.

In this study, we showed that C. acetobutylicum tRNAAla(UGC) directs efficient C. acetobutylicum alaS antitermination in vitro, and less efficient B. subtilis alaS antitermination. Although neither B. subtilis tRNAAla isoacceptors directed efficient antitermination of the C. acetobutylicum alaS gene in vitro and in vivo, expression of B. subtilis alaS-lacZ in vivo was observed in B. subtilis under alanine starvation conditions, which indicates that tRNAAla in B. subtilis can interact with the cognate alaS leader RNA.

Together, these results imply that each T box leader RNA may have coevolved with its cognate tRNA in each organism to achieve interaction specificity.

Sequence comparison between B. subtilis tRNAAla(UGC) and C. acetobutylicum tRNAAla(UGC) revealed nucleotide variations at five regions. Although B. subtilis tRNAAla(UGC) directed inefficient antitermination of the C. acetobutylicum alaS gene in vitro, base substitutions in B. subtilis tRNAAla(UGC) at any one of three of these regions

(the central core, the T arm, or the acceptor arm) with the corresponding residues from C. acetobutylicum tRNAAla was sufficient to promote efficient antitermination. This indicates that the residues at these positions are not specific determinants that are recognized by the leader RNA. Instead, substitutions may tune the tRNA conformation to allow proper interaction with the alaS leader RNA, which implies that the leader RNA recognizes the overall tRNA structure rather than individual bases at these positions. This 110

finding is similar to our previous in vivo and in vitro findings (Grundy et al., 2000;

Yousef et al., 2005), and reveals that residues in tRNAAla at these five positions contribute to specific interaction with the cognate leader RNA.

An anticodon loop mutation (U32C) substitution was of particular interest. As discussed in Chapter 2, this mutation in tRNAGly may reorganize the anticodon loop conformation, which results in an enhanced pairing interaction at the Specifier Sequence.

However, this substitution was not sufficient for B. subtilis tRNAAla(UGC) to direct efficient C. acetobutylicum alaS antitermination. The U32C mutation was effective only in combination with other base substitutions, which suggests that there are additional requirements (e.g., proper presentation of the acceptor end of tRNA to the antiterminator bulge, Chapter 2) for alaS antitermination. It is also possible that the C38 residue in tRNAAla results in different anticodon loop geometry as compared with the A38 base in tRNAGly. Therefore, the effect of U32C in tRNAAla and tRNAGly may be different.

The combination of U32C (aA) and the acceptor stem substitutions (eE) in B.

Ala subtilis tRNA (UGC) (AbcdE) resulted in reduced K1/2 and RTmax as compared with the tRNA variant with only the acceptor stem substitutions (abcdE). A similar result was observed in antitermination of the glyQS gene when a U32C mutation was introduced into the wild-type B. subtilis tRNAGly(GCC) (Chapter 2). It was predicted that U32C results in enhanced interaction between the glycine Specifier Sequence and anticodon.

However, a tightened interaction at the Specifier Sequence may constrain the positioning of the tRNA acceptor end relative to the antiterminator bulge, and therefore reduce the

RTmax. The results of the current study indicate that U32C may also enhance the 111

interaction between the alanine Specifier Sequence and anticodon, and this effect may restrict the proper presentation of the acceptor arm of the tRNAAla variant with acceptor stem mutations to the antiterminator bulge. Whether the U32C mutation in tRNAAla results in enhanced binding to the alaS leader RNA remains to be determined.

We also observed that the U59A C60U mutations played both positive and negative roles in alaS antitermination, depending on the context of the tRNA. The unpaired nucleotides 59 and 60 are located at the juxtaposition of the two coaxially-stacked helixes

(D/anticodon stem and T/acceptor stem) in the L-shape tRNA structure (Zagryadskaya et al., 2004). Base alterations at positions 59 and 60 may alter the angle of the two stacking helices in tRNA, and may impair proper interaction between the leader RNA and tRNA.

This could explain why U59A C60U antagonized the effect of some substitutions that resulted in efficient antitermination in the context of B. subtilis tRNAAla(UGC). However, the antagonistic effect could be partly compensated by the addition of U32C, which showed a synergistic effect on antitermination with U59A C60U. U59A C60U may reduce the interaction stability between the C. acetobutylicum alaS leader RNA and tRNAAla variant, while U32C may enhance the stability and allow enhanced antitermination. Previous mutational analysis of tRNA requirements for expression of a tyrS-lacZ fusion revealed that only certain combinations of bases at positions 59 and 60 were found in tRNATyr constructs active in induction of lacZ expression, indicating that the nucleotide identities at these positions are important for antitermination (Grundy et al.,

2000). The results of the current study show that base substitutions at positions 59 and

112

60 alter the antitermination efficiency, which is consistent with the findings in the tyrS in vivo system, and emphasizes the importance of these nucleotides in antitermination.

Substitutions at the central core, the T stem or the acceptor stem of B. subtilis tRNAAla(UGC) may also affect the presentation of the acceptor end. As described in

Chapter 1, the X-ray crystal structure revealed that there is a distortion at positions 26 and

44 of tRNAGly upon binding to the Stem I of the glyQ leader RNA (Zhang and Ferre-

D'Amare, 2013). The resulting tRNA structure mimics a bent tRNAPhe structure in the

P/P state during translation. Thus, the G26A•A44G substitutions in B. subtilis tRNAAla(UGC) may reorient the central core conformation, leading to proper presentation of the acceptor end sequence and efficient antitermination. Because base pairs found at the T arm and the acceptor arm of the B. subtilis tRNAAla(UGC) are composed of more stable pairings than those found in C. acetobutylicum tRNAAla(UGC), nucleotide changes at the T stem (G50A-C64U, C51G-G63C, G52A-C62U, dD) and the acceptor stem

(C5G-G68U, C6A-G67U, eE) of B. subtilis tRNAAla(UGC) may introduce some flexibility into the structurally more rigid region, which assists in proper interaction between the acceptor end and the antiterminator bulge.

We also showed that stabilization of the pairing at positions 3 and 70 of C. acetobutylicum tRNAAla(UGC) resulted in enhanced C. acetobutylicum alaS antitermination, while stabilization of the pairings 2 bp away (5-68, 6-67) from the 3-70 pair resulted in loss of antitermination. Because the acceptor stem adopts an A-form

RNA conformation, which is composed of 11 bp per turn (Ramos and Varani, 1997), stabilization of the pairings at 5-68 and 6-67 in C. acetobutylicum tRNAAla(UGC) may 113

shift the acceptor end and disrupt the pairing with the antiterminator bulge, while stabilization of the 3-70 pair may promote the pairing interaction due to a shift at the acceptor end in an opposite direction. This implies that proper presentation of the tRNA acceptor end to the antiterminator bulge relies on the balance between tRNA structural rigidity and flexibility.

The flexibility of tRNA structure has been revealed to be important during decoding and aminoacylation (Guigou and Mirande, 2005; Schmeing et al., 2011; Schultz and Yarus, 1994a; Schultz and Yarus, 1994b; Smith and Yarus, 1989a, b; Valle et al.,

2002). Rather than acting as a passive adaptor, tRNA plays an active role in translation because its conformational change during decoding contributes to high-fidelity tRNA selection (Cochella and Green, 2005). Our results support the idea that tRNA is a molecular spring (Yarus et al., 2003), which plays an active role not only in translation, but also in tRNA-mediated antitermination of the T box regulatory system.

We demonstrated that tRNAAla in B. subtilis directed antitermination only of the B. subtilis alaS gene, and not the C. acetobutylicum alaS gene. However, we were unable to show expression of C. acetobutylicum alaS-lacZ in vivo by induction of expression of the plasmid-borne C. acetobutylicum tRNAAla. One possible explanation is that the plasmid- borne tRNA construct was not properly expressed in vivo, and therefore failed to induce the expression of the fusion. Another possibility is that B. subtilis tRNAAla isoacceptors in the cell may serve as competitive inhibitors that interfere with the interaction between

C. acetobutylicum tRNAAla construct and the C. acetobutylicum alaS leader RNA. This problem can be avoided by the use of an amber suppression system (Grundy and Henkin, 114

1994; Grundy et al., 1997b). Future experiments will focus on construction of an alaS leader variant with a UAG nonsense codon and a suppressor tRNAAla variant with a CUA anticodon to avoid possible competition by endogenous tRNAAla.

It was surprising to observe that a B. subtilis tRNAGly variant with an alanine UGC anticodon and an ACCA acceptor end resulted in antitermination of the C. acetobutylicum alaS gene equivalent to that of the cognate C. acetobutylicum tRNAAla(UGC), especially when the wild-type heterologous B. subtilis tRNAAla(UGC) failed to direct efficient C. acetobutylicum alaS antitermination in vitro (Fig. 3.4). It is possible that elements in the tRNAGly body contribute to certain properties that fulfill the requirements for efficient alaS antitermination. In contrast, antitermination of the B. subtilis glyQS construct was less efficient using the tRNAAla variant with a GCC anticodon and a UCCA acceptor end as compared to the wild-type tRNAGly, suggesting that elements in the tRNAAla body may restrict proper interaction with the non-cognate leader RNA. The switch from GGC-GCC to GCU-UGC in the context of the glyQS leader RNA may introduce flexibility into the Specifier Sequence-anticodon helix, which allows tRNAAla at a saturated concentration to promote antitermination. In addition to the elements in tRNA that may cause the differences in antitermination, elements in the leader RNAs may also contribute to discrimination of noncognate tRNAs. It is possible that the glyQS leader RNA has evolved to better discriminate against non-cognate tRNA than the alaS leader RNA.

Protein synthesis is essential for cell survival. Therefore, tRNAs should have been subject to primary selective pressures from the translation machinery. In addition to the 115

conserved tRNA elements that maintain the canonical L-shape structure (Oliva et al.,

2006; Rich and RajBhandary, 1976; Zagryadskaya et al., 2004), some elements in tRNA have coevolved with the anticodon sequence to tune the idiosyncratic tRNAs for uniform decoding activity during translation (Ledoux et al., 2009; Olejniczak et al., 2005; Saks and Conery, 2007). We tested the importance of conserved elements in glyQS antitermination in Chapter 2; the conservation of those elements is found in the majority of tRNAs, including isoacceptors and different tRNA species from organisms that has no

T box regulation. In the current work, we found that sequence variations between C. acetobutylicum tRNAAla(UGC) and B. subtilis tRNAAla(UGC) are located at positions that show less anticodon-dependent conservation among heterologous tRNAAla(UGC) from different organisms, including those that do not utilize the T box system to regulate amino acid-related genes (Saks and Conery, 2007; L.C. Liu, F.J. Grundy, T.M. Henkin, unpublished). The results from this chapter suggest that the T box leader RNAs may have imposed selective pressures onto their cognate tRNAs to achieve specific interactions. This specificity is accomplished by sequence variations at positions that showed less nucleotide conservation so that the tRNA is still able to maintain the L-shape structure, uniform decoding efficiency during translation, recognition by AlaRS, RNase P and EF-Tu, and proper interaction with other molecules in the cell. The sequence variations between heterologous tRNAs could be the result of different selective pressures imposed by leader RNAs that belong to the class of natural deletion variants or the class with canonical structures to fulfill their structural requirements.

116

CHAPTER 4

ANALYSIS OF COMPOUNDS DESIGNED TO TARGET

THE T BOX ANTITERMINATOR

4.1 Introduction

The emergence of multi-drug resistance pathogens is a serious problem in the medical field. One approach to this problem is to search for new targets for antimicrobial agents. Riboswitches are ideal candidates due to their roles in regulating expression of essential genes (Anupam et al., 2008b; Green et al., 2010; Mehdizadeh Aghdam et al.,

2014; Trausch and Batey, 2014). Because many pathogenic Gram-positive bacteria, such as Streptococcus pyogenes, Staphylococcus aureus, Bacillus anthracis, Clostridium difficile, etc., have adapted the T box system to regulate important amino acid-related genes (Gutierrez-Preciado et al., 2009), chemical compounds that specifically target the T box antitermination mechanism could simultaneously inhibit the expression of multiple essential genes in one organism, resulting in efficient growth inhibition.

In collaboration with Dr. Jennifer Hines at Ohio University (Athens, Ohio), we developed compounds that compete with tRNA for binding to the T box antiterminator

(Chapter 1). Binding of uncharged tRNA stabilizes the antiterminator structure, and switches the T box-regulated gene expression from the default “off” state to the “on” 117

state. Compounds that disrupt tRNA binding to the antiterminator could prevent the expression of essential genes and inhibit bacterial growth. Because the 5’ half of the antiterminator is composed of the highly conserved T box sequence, any mutation in this conserved region that disrupts the formation of antiterminator could be detrimental to cell survival. As a result, bacteria are less likely to develop drug resistance via alterations in the T box sequence. Therefore, the T box antiterminator was selected as a potential target for antibiotics.

Previous biochemical studies utilized an AM1A T box antiterminator model RNA and a reduced-function variant AM1A(C11U) to screen for compounds that specifically target the T box antiterminator (Chapter 1). A series of 4,5-disubstituted oxazolidinones was generated and tested for binding to the model RNAs using a fluorescence resonance energy transfer (FRET) assay (Means et al., 2006). Two oxazolidinones, ANB-22 and

40, that bound AM1A efficiently were tested in the glyQS in vitro system to determine their effect on antitermination (Anupam et al., 2008b). ANB-22 resulted in reduced tRNAGly-dependent antitermination, while ANB-40 led to increased readthrough of the terminator regardless of the presence or absence of tRNAGly; this indicates that compound binding to the antiterminator structure may reduce or enhance antitermination.

In a subsequent study, a series of 1,4,-disubstituted 1,2,3-triazole analogs was generated based on the composition and chemical properties of ANB-22 and 40. These triazole analogs were synthesized by replacing the oxazolidinone rings with 1,2,3-triazole rings in order to improve water solubility, and to eliminate the stereocenter of oxazolidinones in order to prevent the synthesis of enantiomers that may have reduced 118

biological activities (Acquaah-Harrison et al., 2010; Brooks et al., 2011). Binding of triazole analogs to the antiterminator model RNAs was monitored by FRET binding assays (Acquaah-Harrison et al., 2010), and the ability of analogs to compete with tRNA for binding to AM1A was determined by fluorescence anisotropy binding assays (Zhou et al., 2011). A fluorescence-based thermal denaturation assay was also used to monitor the ligand-induced structural stability of AM1A, and to determine whether compounds stabilize or disrupt the antiterminator structure (Zhou et al., 2012). Together, these studies identified lead compounds that bind AM1A with specificity, alter the structural stability of AM1A, and efficiently disrupt tRNA binding to AM1A (Table A.5).

The goal of this work was to determine the effects of these lead compounds on antitermination of the T box-regulated genes using double-blind functional assays. Two oxazolidinones, IMB-16 and 60 (Fig. 4.1A), and eight 1,4-disubstituted 1,2,3-triazole analogs, GHB-7, 23, 54, 56, 60, 76, 134, 146 (Fig. 4.1B), were selected based on biochemical tests performed by the laboratory of Dr. Jennifer Hines (results listed in

Table A.5). We tested these compounds in in vitro transcription antitermination assays using the wild-type glyQS leader template and tRNAGly(GCC). Compounds that resulted in inhibition of tRNAGly-dependent glyQS antitermination were further tested in vivo using a tyrS-lacZ reporter system. The results from this study allowed us to correlate the in vitro observations with in vivo results, and to understand the functional roles of lead compounds in the T box antitermination system.

119

Figure 4.1. Composition of lead compounds. (A) Oxazolidinones and (B) 1,4- disubstituted 1,2,3-triazole analogs (J.V. Hines, unpublished results; Maciagiewicz et al., 2011) .

4.2 Materials and Methods

4.2.1 In vitro transcription antitermination assays

In vitro transcription of the wild-type glyQS leader DNA was carried out as described in Chapter 2. Compound stocks (50 mM) were diluted with 100% dimethyl sulfoxide (DMSO) and added at final concentrations of 1.4 and 2.9 mM after the halted transcription complex was formed; an equal volume of 100% DMSO was added as a solvent control. Transcription elongation was reinitiated by adding NTPs at 10 μM in the presence or absence of wild-type tRNAGly(GCC) at 0.11 µM final concentration to test the effect of the compounds on glyQS antitermination. The resulting transcripts were

120

resolved on a 6% denaturing gel and the %RT was determined as described in Chapter 2.

Data were normalized to the %RT of the solvent control in the presence of tRNAGly.

4.2.2 Bacterial strains and growth conditions

A tyrS-lacZ transcriptional fusion, which has the promoter sequence and the leader region of the B. subtilis tyrS gene fused to the lacZ reporter gene, was introduced into B. subtilis strain BR151 (lys-3 metB10 trpC2) in single copy (Henkin et al., 1992).

Expression of tyrS-lacZ was induced by a tyrosine analog (4-amino-L-phenylalanine hydrochloride, Sigma) to mimic tyrosine starvation conditions (Grundy et al., 2002a). A

YP-lacZ transcriptional fusion, which has only the promoter region of the tyrS gene fused to the lacZ gene without the leader sequence, was also introduced into strain BR151 (F.J.

Grundy and T.M. Henkin, unpublished data). Expression of YP-lacZ is constitutive and independent of the T box regulatory mechanism.

The minimal inhibitory concentrations (MICs) of IMB-16, GHB-7, 23, 54, 56, 60,

76, 134, 146 for S. aureus, B. subtilis, E. faecalis and E. coli were determined by the laboratory of Dr. Nigel Priestley (University of Montana, Missoula, Montana) using broth dilution assays (Table A.5, N. Priestley, unpublished results). Among these compounds,

GHB-54, 56, and 76 specifically inhibited the growth of Gram-positive organisms (S. aureus, B. subtilis, E. faecalis) that have adapted the T box system to regulate gene expression, and had no effect on E. coli, a Gram-negative bacterium that does not use the

T box system for gene regulation. IMB-16 and GHB-7 also inhibited the growth of S. aureus and B. subtilis, although the effects on E. faecalis and E. coli were small (>1000 121

µM MIC). GHB-54 and GHB-56 had the lowest MIC (16 µM) for B. subtilis, and IMB-

16 and GHB-76 had intermediate MICs (31 and 125 µM, respectively); other compounds that had MICs over 250 µM were not included in this in vivo test. The effect of compounds on antitermination was assayed at sub-inhibitory concentrations (0.25, 0.5 and 0.75 MICs).

For induction assays, cells were grown until mid-log phase in minimal medium

(Anagnostopoulos and Spizizen, 1961) containing 0.5% (w/v) glucose, 50 µg/mL of the appropriate amino acids (L-lysine, L-methionine and L-tryptophan), and 5 µg/mL chloramphenicol. Cells were collected by centrifugation and resuspended in fresh medium in the presence or absence of 2 µg/mL 4-amino-L-phenylalanine hydrochloride, as inducer. Inhibitor compounds were added immediately after induction, and cells were collected at designated time points (T1, T2, T3 or T4, where T1 = 1 hour after induction).

All conditions were tested at least twice.

4.2.3 β–galactosidase assays

β–galactosidase assays were performed as described in Chapter 3. Data were normalized to the Miller units of the DMSO control under inducing conditions. Each permeabilized cell sample was assayed at least twice.

122

4.3 Results

4.3.1 Inhibition of glyQS antitermination

We used the B. subtilis glyQS gene to test the effects of lead compounds on antitermination in vitro. When compounds were tested at 1.4 mM (Fig. 4.2A), IMB-16 resulted in a 17% decrease in tRNAGly-directed antitermination, and had no effect on tRNA-independent readthrough. This suggests that IMB-16 is an antagonist that competes with tRNAGly for binding to the glyQS antiterminator and reduces antitermination. Although GHB-54 and 56 resulted in a 69% and 54% decrease in tRNAGly-directed antitermination respectively, both compounds also caused a 2-fold increase in tRNAGly-independent readthrough. These compounds therefore act as partial agonists, because they act as an antagonist in the presence of the tRNAGly agonist, and serves as a less effective agonist on their own (Bolonna and Kerwin, 2005). All other compounds tested had no effect on antitermination at this concentration.

When the compounds were added at 2.9 mM, IMB-16 and GHB-54 had no further effect on antitermination as compared with their effects at 1.4 mM (Fig. 4.2B). GHB-7,

23 and 76 led to a 7%-17% decrease in tRNAGly-mediated antitermination, and had no effect on readthrough of the glyQS terminator in the absence of tRNAGly, which suggests that these compounds are antagonists. Both GHB-56 and IMB-60 completely inhibited the transcription reaction when added at 2.9 mM, indicating that these compounds have adverse effects on overall transcription at this concentration. GHB-60 resulted in a 12% reduction in tRNAGly-dependent antitermination and a 1.7-fold increase in readthrough in the absence of tRNAGly, which indicates that GHB-60 is a partial agonist. GHB-134 123

reduced tRNAGly-dependent antitermination by 28%, and resulted in a 1.6-fold increase in readthrough of the terminator in the absence of tRNAGly. As a result, GHB-134 is also considered as a partial agonist. GHB-146 had no effect on tRNAGly-directed antitermination, but resulted in a 4.5-fold increase in readthrough of the terminator in the absence of tRNAGly, suggesting that this compound is an agonist. In sum, IMB-16, GHB-

7, 23, 54, 56, 76, 134 are compounds that inhibited tRNAGly-dependent antitermination, and GHB-54, 56, 60 and 134 also increased tRNAGly-independent readthrough. GHB-

146 only increased readthrough of the glyQS terminator in the absence of tRNAGly.

GHB-56 and IMB-60 at 2.9 mM had detrimental effects on the overall transcription reaction.

124

Figure 4.2. Effect of compounds on glyQS antitermination in vitro. Antitermination of the B. subtilis wild-type glyQS gene was determined using in vitro transcription assays, in the presence (gray bars) or absence (white bars) of 0.11 µM wild-type tRNAGly(GCC) with either (A) 1.4 mM or (B) 2.9 mM compounds; the same volume of 100% DMSO was included as a no ligand control. All data were normalized to the DMSO control +tRNA as 100%. Asterisks (*) indicate that the transcription reaction was inhibited by the compound.

125

4.3.2 Reduction of tyrS-lacZ expression

We next investigated the effect of compounds on antitermination in B. subtilis, which is a model organism for Gram-positive bacteria and contains 19 T box-regulated genes, using a tyrS-lacZ transcriptional fusion construct. The minimal inhibitory concentration (MIC) of each compound for B. subtilis was determined by the laboratory of Dr. Priestley (Table A.5), and the effect of compounds on antitermination was determined at sub-inhibitory concentrations (0.25, 0.5 and 0.75 MICs). We focused on

GHB-54, 56, 76 and IMB-16 because of the low MIC values for B. subtilis among all lead compounds. The effect of compounds on antitermination in vivo was monitored by changes in induction of the expression of a tyrS-lacZ transcriptional fusion construct.

Expression of the reporter gene was induced by a tyrosine analog that mimics tyrosine starvation conditions, and compounds that specifically target the T box leader RNA were tested for their effect on induction. Any effect on gene expression that is independent of the T box regulation was monitored by the expression of a YP-lacZ construct, which lacks the tyrS leader sequence and is constitutively expressed. Cells were grown to mid- exponential phase, induced with the tyrosine analog, and incubated with the inhibitor compound for 1-4 h (T1-T4) before collection.

In the absence of any inhibitor compound, expression of tyrS-lacZ was induced by the addition of the tyrosine analog, with induction ratios ranged from 3.0 to 9.2 at T1-T4

(Table 4.1). The induction ratios of YP-lacZ expression were between 0.89 and 1.2, indicating that the tyrosine analog had no effect on expression of this construct, as expected. When cells were incubated with IMB-16 at 0.25 MIC for 3 h (cells collected at 126

T3), a 23% decrease in expression of tyrS-lacZ was observed under induced conditions, and there was no effect on expression of tyrS-lacZ under uninduced conditions.

However, addition of this compound also decreased the expression of YP-lacZ by 22% under induced conditions, and 23% under uninduced conditions. The induction ratio of both constructs remained the same as that of the no ligand control, which indicates that the effect of IMB-16 under these conditions is not specific to the T box regulatory system. When cells were collected at T4, induction of tyrS-lacZ expression was reduced from 5.3- to 3.7-fold, while induction of YP-lacZ expression was the same as the no ligand control. This implies that IMB-16 may have a more specific effect on the T box- regulated gene than on general gene expression under these conditions. Addition of

IMB-16 at 0.5 MIC had no effect on induction of both tyrS-lacZ and YP-lacZ expressions at either T3 or T4, which indicates that increasing the concentration of IMB-16 did not enhance the effect of the compound on expression of the reporter gene.

Addition of GHB-54, 56 and 76 at 0.25 and 0.5 MICs had no effect on induction of tyrS-lacZ and YP-lacZ expression at T3 and T4. Although addition of GHB-54 at 0.75

MIC decreased the expression of tyrS-lacZ by 15% under induced conditions at T3, and had no major effect under uninduced conditions, the induction ratio was comparable to the no ligand control. Also, addition of this compound reduced the expression of YP-lacZ by 10% under induced conditions, and by 35% under uninduced conditions, while the induction ratio remained the same as the no ligand control. This indicates that the effect of GHB-54 was not specific to the T box regulation under this condition. In contrast, addition of GHB-54 at 0.75 MIC had no effect on expression of tyrS-lacZ under induced 127

conditions at T4, and resulted in a decrease from 17% to 13% under uninduced conditions; this reduction led to an increase in the induction ratio from 5.9 to 7.8.

Because the induction ratio of YP-lacZ was not affected at T4, GHB-54 may have a more specific effect on expression of the T box-regulated gene than general gene expression under this condition. Addition of GHB-56 at 0.75 MIC had no effect on expression of tyrS-lacZ at T3 and T4 under induced and uninduced conditions. Although the induction of YP-lacZ expression was not affected by GHB-56 at 0.75 MIC, this compound reduced the expression of YP-lacZ by 12% and 17% under induced and uninduced conditions at

T3, and by 11% and 4% under induced and uninduced conditions at T4, respectively. This indicates that GHB-56 has a more effect on general gene expression than the expression of the T box-regulated gene under this condition.

128

Table 4.1. Effect of compounds on expression of tyrS-lacZ and YP-lacZ. tyrS-lacZ YP-lacZ induced uninduced ratio induced uninduced ratio

No ligand 100a 16 ± 2 6.3 ± 0.8 100 110 ± 10 0.92 ± 0.10 IMB-16 77 ± 1 14 ± 1 5.5 ± 0.1 78 ± 3 87 ± 9 0.91 ± 0.13 T3 GHB-54 100 ± 10 16 ± 3 6.6 ± 1.9 97 ± 1 120 ± 10 0.81 ± 0.08 GHB-56 97 ± 3 16 ± 2 6.2 ± 1.0 97 ± 1 110 ± 20 0.91 ± 0.18 GHB-76 97 ± 3 15 ± 2 6.6 ± 1.1 100 ± 10 110 ± 10 0.93 ± 0.18

5 MIC No ligand 100 19 ± 2 5.3 ± 0.6 100 120 ± 30 0.89 ± 0.22

0. IMB-16 82 ± 4 22 ± 1 3.7 ± 0.4 77 ± 16 97 ± 23 0.88 ± 0.37 GHB-54 100 ± 10 19 ± 3 5.5 ± 1.4 94 ± 7 130 ± 30 0.78 ± 0.23 T4 GHB-56 100 ± 10 21 ± 1 4.8 ± 0.7 100 ± 20 110 ± 30 1.0 ± 0.5 GHB-76 100 ± 10 18 ± 2 5.7 ± 1.2 98 ± 10 120 ± 30 0.89 ± 0.3 No ligand 100 12 ± 1 8.4 ± 0.7 100 94 ± 6 1.1 ± 0.1 IMB-16 80 ± 5 8.7 ± 0.6 9.3 ± 1.2 79 ± 9 74 ± 13 1.1 ± 0.3

T3 GHB-54 95 ± 4 11 ± 1 8.7 ± 1.1 96 ± 15 88 ± 13 1.1 ± 0.3 GHB-56 92 ± 5 11 ± 1 8.5 ± 1.2 100 ± 10 92 ± 8 1.1 ± 0.2 GHB-76 96 ± 2 9.4 ± 0.1 10 ± 1 93 ± 18 88 ± 10 1.1 ± 0.3 No ligand 100 11 ± 1 9.2 ± 0.8 100 96 ± 4 1.0 ± 0.1 0.5 MIC IMB-16 83 ± 5 9.9 ± 0.2 8.4 ± 0.7 80 ± 9 73 ± 14 1.2 ± 0.3

T4 GHB-54 97 ± 2 9.0 ± 0.5 11 ± 1 99 ± 11 94 ± 6 1.1 ± 0.2 GHB-56 100 ± 10 9.4 ± 0.3 11 ± 1 100 ± 10 94 ± 6 1.1 ± 0.2 GHB-76 94 ± 1 9.4 ± 0.3 10 ± 1 96 ± 14 90 ± 8 1.1 ± 0.3 No ligand 100 15 ± 1 6.7 ± 0.4 100 110 ± 10 0.92 ± 0.08

T3 GHB-54 85 ± 4 11 ± 1 7.8 ± 1.1 90 ± 2 75 ± 11 1.2 ± 0.2 GHB-56 94 ± 4 13 ± 1 7.3 ± 0.9 88 ± 5 83 ± 5 1.1 ± 0.1

5 MIC

7 No ligand 100 17 ± 1 5.9 ± 0.3 100 93 ± 5 1.1 ± 0.1

0. T4 GHB-54 98 ± 3 13 ± 2 7.8 ± 1.4 90 ± 1 79 ± 8 1.2 ± 0.1 GHB-56 98 ± 3 17 ± 2 5.6 ± 0.9 89 ± 3 89 ± 4 1.0 ± 0.1 No ligand 100 34 ± 2 3.0 ± 0.2 100 96 ± 4 1.0 ± 0.1

T1 GHB-54 110 ± 10 29 ± 4 3.9 ± 0.9 81 ± 1 86 ± 3 0.9 ± 0.1 GHB-56 96 ± 4 31 ± 3 3.1 ± 0.4 82 ± 2 87 ± 1 0.9 ± 0.1 No ligand 100 18 ± 2 5.6 ± 0.6 100 99 ± 1 1.0 ± 0.1

0.5 MIC T2 GHB-54 86 ± 2 14 ± 1 6.2 ± 0.6 87 ± 7 90 ± 1 0.97 ± 0.09 GHB-56 92 ± 4 16 ± 1 5.8 ± 0.6 89 ± 4 96 ± 2 0.93 ± 0.06 No ligand 100 29 ± 3 3.5 ± 0.4 100 87 ± 2 1.2 ± 0.1

T1 GHB-54 71 ± 3 25 ± 5 3.0 ± 0.7 71 ± 10 57 ± 6 1.3 ± 0.3 GHB-56 93 ± 7 26 ± 4 3.7 ± 0.8 85 ± 3 72 ± 1 1.2 ± 0.1

5 MIC

7 No ligand 100 14 ± 1 7.2 ± 0.5 100 96 ± 2 1.0 ± 0.1

0. T2 GHB-54 58 ± 4 10 ± 1 5.9 ± 1.0 73 ± 11 68 ± 10 1.1 ± 0.3 GHB-56 71 ± 4 10 ± 1 7.2 ± 1.1 87 ± 3 82 ± 7 1.1 ± 0.1 aAll data were normalized to the no ligand control at inducing condition as 100%.

129

One potential problem with collecting cells at the T3 and T4 time points is that the effects of compounds on antitermination may be masked when induction reaches maximum. Therefore, we also collected cells at the T1 and T2 time points, a time-frame when there would be approximately 50% of the maximal induction. When GHB-54 was tested at 0.5 MIC, expression of tyrS-lacZ under induced conditions decreased 14% at T2, relative to the no ligand control; however, this reduction was not observed at T1. This suggests that 1 h of incubation may not be sufficient for GHB-54 to be delivered into the cells and have its effect, and 2 h of incubation is more appropriate. However, this compound reduced the expression of YP-lacZ by 19% (induced) and 10% (uninduced) at

T1, and by 13% (induced) and 9% (uninduced) at T2; the induction ratio was unchanged.

This suggests that the effect of GHB-54 at T2 is not completely specific to the expression of the T box-regulated gene. The construct with YP-lacZ is more sensitive to the repression of GHB-54 at T1 than the construct with tyrS-lacZ, which suggests that GHB-

54 had more direct impact on general transcription than the T box regulation. Addition of

GHB-54 at 0.75 MIC reduced the expression of tyrS-lacZ by 29% under induced conditions at T1, and had no effect under uninduced conditions. The reduction in expression of tyrS-lacZ was more severe at T2 (42% under induced and 4% under uninduced conditions). This suggests that the effect of GHB-54 on expression of tyrS- lacZ is more prominent at 0.75 MIC than at 0.5 MIC, and the effect was more prominent at T2 than T1. However, the expression of YP-lacZ also decreased at both T1 and T2 under induced and uninduced conditions (the reduction ranged from 27% to 30%), while the induction ratios were unchanged. This indicates that GHB-54 at 0.75 MIC has a 130

similar effect on expression of YP-lacZ and tyrS-lacZ at T1, and may have a more specific effect on expression of the T box-regulated gene at T2.

GHB-56 at 0.5 MIC had no effect on expression of tyrS-lacZ at both T1 and T2; however, this compound decreased the expression of YP-lacZ under induced (18%) and uninduced conditions (13%) at T1, and under induced conditions (11%) but not under uninduced conditions at T2. This is similar to what was observed with GHB-54, which showed reduced expression of YP-lacZ but not tyrS-lacZ at T1. GHB-56 at 0.75 MIC had no effect on expression of tyrS-lacZ at T1, and resulted in a reduction in expression at T2 under induced (29%) and limited effect under uninduced conditions. This compound also reduced the expression of YP-lacZ by 13-15% at T1 and T2 under both induced and uninduced conditions. These results showed that GHB-56 at 0.75 MIC has similar effect on general transcription at T1 and T2, and has a greater effect on expression of the T box- regulated gene at T2.

4.4 Discussion

T box riboswitches are novel targets for antibiotics because they regulate the expression of multiple amino acid-related genes in most pathogenic Gram-positive organisms. Compounds that compete with tRNA for binding to the antiterminator structure could target several genes in one organism; therefore, mutations in the T box leader RNA that alter the antiterminator structure of a single gene may not be sufficient to confer drug resistance. This system regulates many essential aminoacyl-tRNA synthetase genes, and the default state of the gene expression is “off”; as a result, 131

mutations that disrupt the formation of the antiterminator would be detrimental to cell survival. In addition, compounds that inhibit the expression of several nonessential T box-regulated genes, or partially inhibit expression of several essential genes, could lead to a synthetic lethal phenotype. Therefore, the use of T box riboswitches as antibiotic targets could allow for effective growth inhibition and avoid the development of drug resistance by modification of the drug target. To continue the work on finding compounds that have the potential to disrupt the T box regulatory system via competing with uncharged tRNA for binding to the antiterminator, we investigated the effect of lead compounds on antitermination using in vitro and in vivo functional assays.

We used in vitro transcription antitermination assays to identify antagonist compounds that specifically decreased tRNAGly-dependent glyQS antitermination, as well as agonists that stimulate antitermination in the absence of tRNA. Out of the ten compounds tested, four compounds (IMB-16, GHB-7, 23, and 76) were identified as antagonists, and one compound (GHB-146) was considered as an agonist. IMB-60 inhibited the overall transcription reaction when used at 2.9 mM, and the effect on antitermination of the T box-regulated gene was therefore unclear. We also observed that

GHB-54, 56, 60 and 134 acted as partial agonists that not only led to decreased tRNAGly- directed antitermination, but also increased readthrough of the terminator in the absence of tRNAGly. Thermal denaturation assays revealed that GHB-54 and 56 resulted in an increased Tm of AM1A RNA (Table A.5, ΔTm = 5.2 and 4.2 °C, respectively), indicating that these compounds stabilize the antiterminator model RNA. This result is consistent with the finding that GHB-54 and 56 are partial agonists. However, GHB-60 and 134 132

had no effect on the structural stability of AM1A model RNA (Table A.5, ΔTm = -0.3 and

0.1 °C, respectively), although they enhanced tRNA-independent readthrough in vitro.

This indicates that GHB-60 and 134 may interact with the leader RNA in a way that is different from GHB-54 and 56. Although GHB-146 showed the feature of an agonist that resulted in increased tRNA-independent readthrough, this compound had no effect on the structural stability of the AM1A model RNA (Table A.5, ΔTm = -0.4 °C). This indicates that GHB-146 may affect readthrough of the glyQS gene in a way that is different from stabilization of the antiterminator structure (e.g., destabilization of the terminator structure). GHB-7, 23 and 76 at 2.9 mM decreased tRNA-dependent antitermination, which is consistent with the observations in the fluorescence anisotropy experiments that these compounds disrupt tRNA binding to the AM1A RNA (Table A.5, ∆Anisotropy = -

79, -58 and -71%, respectively; Zhou et al., 2011). However, GHB-7 stabilized the structure of the AM1A model RNA (ΔTm = 3.2 °C) while GHB-23 and 76 had no major effect on the structural stability of AM1A (Table A.5, ΔTm = -0.7 and 0.3 °C; Zhou et al.,

2012). This suggests that ligand-induced stabilization of the antiterminator structure may not always result in enhanced antitermination.

One limitation of using the in vitro assays is that the results do not always resemble what occurs in living cells. Therefore, a cell-based tyrS-lacZ reporter system was applied to test whether the lead compounds had similar effects on antitermination in vivo as compared with that in vitro. We observed GHB-56 and IMB-60 at 2.9 mM completely inhibited transcription in vitro, it was therefore important to test whether the compounds affected general gene expression rather than a specific effect on expression of the T box- 133

regulated gene. A YP-lacZ transcriptional fusion was therefore used as a control; expression of this construct is constitutive and independent of the T box regulatory mechanism. We observed decreased expression of tyrS-lacZ in the presence of IMB-16, which is consistent with the in vitro result that this compound reduced tRNAGly-directed glyQS antitermination. However, IMB-16 also resulted in a similar level of reduction in expression of YP-lacZ under induced conditions, which indicates that the effect of IMB-

16 is not specific to the T box-regulated gene. GHB-76 had no effect on expression of either tyrS-lacZ or YP-lacZ in vivo, although there was a 7% reduction in tRNA- dependent antitermination in vitro. It is possible that the conditions used for testing

GHB-76 in vivo are not ideal. Alternatively, the effect of GHB-76 may be more prominent in the purified in vitro system than in vivo.

We observed that GHB-54 and 56 resulted in a more severe reduction of expression of tyrS-lacZ at earlier time points (T1 and/or T2) than later time points (T3 and T4), which indicates that prolonged incubation may mask the effect of compounds on T box- regulated genes in vivo. However, GHB-54 and 56 had a greater effect on expression of tyrS-lacZ at T2 than at T1, and had a similar effect on expression of YP-lacZ at both T1 and T2. This suggests that the effect of GHB-54 and 56 on expression of YP-lacZ may have reached maximum at T1 and T2, while the effect on expression of tyrS-lacZ appears in a time-dependent manner. The result also indicates that the effects of GHB-54 and 56 are not completely specific to the T box regulatory mechanism. This is consistent with the findings for GHB-56 in vitro, where this compound at 1.4 mM inhibited tRNAGly- dependent glyQS antitermination, and inhibited overall transcription reactions at 2.9 mM. 134

Although GHB-54 at 1.4 and 2.9 mM did not completely inhibit transcription reactions in vitro, it is possible that the inhibitory effect requires higher concentrations of this compound. Additionally, although both GHB-54 and 56 resulted in increased tRNAGly- independent readthrough in vitro and were considered as partial agonists, we did not observe any increase in tyrS-lacZ expression under uninduced conditions, indicating that the roles of these compounds in vivo may be different from the roles in vitro. By definition, a partial agonist acts as an effective antagonist in the presence of agonists, and a less effective agonist on its own. Because in the T box system, the leader RNA monitors the charging ratio of the cognate tRNA in the cell, we could not completely eliminate the uncharged tRNA (endogenous agonist) in vivo. Therefore, a leader variant with a nonsense codon Specifier Sequence would be required to demonstrate the partial agonist property in vivo.

Despite the fact that the in vivo environment is much more complicated than in vitro, it is also possible that these compounds have differential effects on antitermination of different classes of aaRS genes. A previous study revealed that T box leader RNAs from different classes of aaRS genes may have different potential to accommodate the ligand-induced stabilization of the antiterminator structure (Jentzsch and Hines, 2012).

In this case, ligand binding to the antiterminator of the glycyl leader RNA may have more dramatic effects on antitermination than binding to the antiterminator of the tyrosyl leader

RNA. Therefore, it is necessary to understand whether the lead compounds have any aaRS-specific effect on antitermination. The alaS in vitro and in vivo systems described in Chapter 3 would be good options to test. 135

Finally, it is also possible that the optimal condition to investigate lead compounds in vivo has not been well established. After the addition of the tyrosine analog, uncharged tRNATyr accumulates and induces the expression of tyrS-lacZ. Compounds were added as potential competitive inhibitors that prevent binding of tRNATyr to the antiterminator and repress induction. Sufficient time should be provided for compounds to be delivered into the cells; however, prolonged incubation may mask the effect of compounds. At later time points, the compound pool may be reduced after binding to the

T box antiterminator while the uncharged tRNA pool increases after induction. As a result, transcription at later time points may result in higher expression due to the accumulation of uncharged tRNA, and may be less sensitive to repression by lead compounds. This could explain why we observed reduced expression of the reporter gene in the presence of compounds at early time points such as T1 or T2, and no effect on expression at later time points such as T3 and T4. On the other hand, the concentration of the ligand should be high enough to disrupt the T box antitermination mechanism, but low enough to avoid inhibition of overall cell growth. Therefore, collection of cells at different time points after induction, with different concentrations of ligand, would be necessary to understand the kinetics of drug delivery and efficacy. Future experiments could focus on adding compounds to cells at early-exponential phase, using different concentrations of compounds, and collecting cells at shorter time intervals.

136

CHAPTER 5

CONCLUSIONS AND FUTURE DIRECTIONS

5.1 Conclusions

The “RNA world hypothesis” proposes that cells developed from a world based on

RNAs due to their self-replication property, enzyme-like catalytic activity and the ability to carry genetic information (Gilbert, 1986). The discovery of riboswitch RNAs supports this hypothesis because they sense the requirements for gene expression in the cell, undergo structural rearrangements and control the expression of necessary genes without the aid of additional protein factors (Breaker, 2012; Serganov and Nudler, 2013). The T box leader RNAs represent a special class of riboswitches that monitor the requirements for expression of amino acid-related genes via direct interaction with the cognate tRNA

(Green et al., 2010; Grundy and Henkin, 1993). This system uses tRNA as a regulatory molecule, which reveals a special role for tRNA in gene regulation (Banerjee et al., 2010;

Green et al., 2010; Phizicky and Hopper, 2010). The anticodon and the acceptor end of tRNA are two major positions that participate in major functional interactions during translation, aminoacylation and antitermination of genes in the T box family (Crick,

1966; Grundy et al., 1994; Sprinzl and Cramer, 1975). However, other elements are also important to assist the function of tRNAs in these processes (Fahlman et al., 2004; Giege 137

et al., 1998; Grundy et al., 2000; Grundy et al., 1997b; Ibba and Soll, 2000; Ledoux et al.,

2009; Ledoux and Uhlenbeck, 2008; Saks and Conery, 2007; Shepotinovskaya and

Uhlenbeck, 2013). This study focused on identification of elements in tRNA other than the two major positions that are important for efficient antitermination.

In Chapter 2, we used the B. subtilis glyQS in vitro system to study tRNA requirements for antitermination. We found that an anticodon loop mutation (U32C) in tRNAGly promoted glyQS antitermination and tRNA binding in vitro, especially in constructs with a mismatch at position 3 of the Specifier Sequence. These results are consistent with that of a U32C substitution in tRNAGly that caused misreading of other near-cognate glycine codons in translation due to reduced dissociation rate in the A site of the decoding center (Olejniczak and Uhlenbeck, 2006), and suggest that the Specifier

Sequence-anticodon interaction in antitermination shares some similarities with the codon-anticodon interaction in translation. However, D arm mutations in tRNAGly resulted in effects on glyQS antitermination dependent on the anticodon sequence, which is different from the effect of the Hirsh suppressor mutation (a mutation at the D arm of tRNATrp) on misreading (Schmeing et al., 2011). This suggests that tRNA recognition by the T box leader RNA is different from that by the translation machinery. In addition, mutations at positions 30-40, 27-43 and 26-44, which usually contain anticodon- dependent conserved residues (Saks and Conery, 2007), were generally tolerated in antitermination, while residues at these positions affect translation efficiency or decoding accuracy (Komine and Inokuchi, 1990; Olejniczak et al., 2005; Robertus et al., 1974;

138

Schultz and Yarus, 1994a). The conservation at these positions of tRNA may be attributed to strong selective pressures imposed by the translation machinery.

Accurate decoding is critical for cell survival; therefore, tRNAs have been subject to major selective pressures from the translation machinery. Although other molecules, such as the T box leader RNAs, may also impose selective pressures onto tRNAs, the effect may be weaker. Misreading in translation can cause more detrimental effects on cell viability than mispairing between the Specifier Sequence and anticodon in antitermination. Therefore, it is possible that the selective pressures from the translation machinery dominate tRNA evolution. Residues in tRNA that ensure codon recognition fidelity may have more of an evolutionary advantage than those that favor other regulatory processes.

In Chapter 3, we used the C. acetobutylicum alaS gene to establish another in vitro system, and demonstrated that only C. acetobutylicum tRNAAla(UGC) directs efficient antitermination; the heterologous B. subtilis tRNAAla(UGC), which has the same anticodon and acceptor end sequence as C. acetobutylicum tRNAAla(UGC), directs less efficient antitermination. We tested combinations of residues that differ between B. subtilis tRNAAla(UGC) and C. acetobutylicum tRNAAla(UGC) to show that some residues have synergistic effects on antitermination, while some antagonize the effect of other elements. The results revealed that tRNAAla is tuned by residues at these positions to achieve optimal interaction with its cognate T box leader RNA. Bases at these positions are not highly conserved in tRNAAla(UGC) among different organisms (Saks and Conery,

2007), including those without the T box regulatory mechanism. Our results suggest that 139

the T box leader RNA may utilize regions that are less likely to affect the primary function of tRNA to achieve specific and efficient interactions with its cognate tRNA.

We observed that mutations in tRNA may affect antitermination by alteration of the pairing stability between the Specifier Sequence and the anticodon. We found that the

U32C mutation enhanced tRNAGly binding to the wild-type glyQS leader RNA presumably due to a pre-formed anticodon loop structure that may enhance the pairing interaction between the Specifier Sequence and anticodon (Chapter 2). Although enhanced pairing at the Specifier Sequence may increase antitermination efficiency

(reduced K1/2), this effect may also constrain the positioning of the tRNA acceptor end to the antiterminator bulge, and therefore limit the maximal antitermination activity

(lowered RTmax). A similar effect was observed in C. acetobutylicum alaS antitermination when introducing a U32C mutation into a B. subtilis tRNAAla(UGC) variant with acceptor arm mutations (Chapter 3). These results imply that tightening the pairing between the Specifier Sequence and the anticodon may have an impact on positioning of the acceptor end.

Another example is the U11C mutation in tRNAGly. We showed that U11C resulted in reduced antitermination in the context of a GCC anticodon, and enhanced antitermination of the construct with a mismatch at positions 1 or 3 of the Specifier

Sequence in the context of a UCC anticodon (Chapter 2). We predicted that the change from a G to a U at the first position of the anticodon may reduce the stacking interaction in the Specifier Sequence-anticodon helix and introduce flexibility to the helix. This flexibility is more prominent in the presence of a mismatch, which could compensate the 140

effect of U11C and promote antitermination by proper presentation of the acceptor end to the antiterminator bulge. Therefore, the effect of U11C depends on the interaction in the

Specifier Sequence-anticodon helix. These studies revealed the versatility of the RNA-

RNA interaction, and pointed out that efficient antitermination is attributed to a balance between flexibility and rigidity of the interaction between the leader RNA and tRNA.

In addition to the interaction flexibility between the leader RNA and tRNA, the flexibility of tRNA structure may also contribute to antitermination efficiency. We demonstrated that destabilization of the pairing at the D arm of tRNAGly(GCC) resulted in reduced glyQS antitermination (Chapter 2). We also showed that stabilization of the pairing at positions 3 and 70 of C. acetobutylicum tRNAAla(UGC) resulted in enhanced alaS antitermination, while stabilization of the pairings at 5-68 and 6-67 completely inhibited antitermination (Chapter 3). These results imply that the flexibility and rigidity of the tRNA structure determine antitermination efficiency. However, due to the cellular quality control mechanism (Li et al., 2002), mutant tRNA constructs with pairing alterations at the helical regions may be subject to degradation in vivo. Therefore, the correlation between tRNA structural flexibility and antitermination may be studied only in vitro.

It has been reported that each tRNA is tuned by multiple anticodon-dependent elements to achieve a uniform decoding rate during translation (Olejniczak et al., 2005;

Saks and Conery, 2007). tRNA isoacceptors must maintain the identity determinants as well as antideterminants to interact specifically with their cognate aaRS. Each aa-tRNA is tuned by the three base pairs in the T arm (49-65, 50-64, 51-63) to compensate for the 141

different thermodynamic contributions of amino acids that have different chemical properties, and to achieve equivalent EF-Tu binding affinity (Schrader et al., 2011;

Schrader and Uhlenbeck, 2011). Pre-tRNA tuning for RNase P recognition has also been reported (Sun et al., 2006). Our study reveals an additional role for tRNA residues, which participate in tuning tRNA for optimal interaction with the cognate T box leader

RNA.

Due to the high prevalence of this regulation in Gram-positive pathogens, the T box regulatory system has been selected as a novel target for antibiotics. The second focus of this study was to investigate compounds that target the T box antitermination mechanism for potential therapeutic use (Chapter 4). Compounds that target this system can simultaneously inhibit the expression of multiple essential and nonessential genes, which could avoid the development of antibiotic resistance by alteration of a single gene target.

Inhibition of multiple nonessential genes may result in synthetic lethality, which further reduces the possibility of development of drug resistance. Compounds that compete with tRNA for binding to the antiterminator disrupt gene expression, rather than serving as amino acid analogues that target translation. Therefore, cells are less likely to develop resistance via indirect effects that increase particular intracellular amino acid pools as reported previously in the L box riboswitch study (Ataide et al., 2007). We identified several inhibitor compounds that resulted in inhibition of glyQS antitermination in vitro.

However, the inhibitory effect on tyrS-lacZ expression in vivo was not completely specific to the T box regulatory mechanism. Additional work on testing different

142

experimental conditions and developing compounds with improved binding specificity are required.

5.2 Future directions

Several puzzles are still unsolved. Although the X-ray crystallography and NMR studies have revealed the interactions between the Stem I of the glyQ(S) T box leader

RNA and uncharged tRNAGly (Chang and Nikonowicz, 2013; Grigg et al., 2013; Zhang and Ferre-D'Amare, 2013), information about specific interactions between the full length leader RNA and tRNA at the molecular level is still missing. It is particularly important to understand the dynamic interaction between the leader RNA and tRNA because our data suggest that the interaction between the Specifier Sequence and anticodon may impact the presentation of the tRNA acceptor end to the antiterminator bulge (Chapters 2 and 3). The structural flexibility of tRNA may also contribute to efficient interaction with the leader RNA.

It was proposed that antitermination is a multi-step process, which involves at least two steps: 1) tRNA binding to the leader RNA; and 2) stabilization of the antiterminator structure (Chapters 1-3). Each step has been studied individually by binding and fluorescence experiments (Grigg et al., 2013; Means et al., 2009; Means et al., 2007;

Nelson et al., 2006; Yousef et al., 2005; Zhang and Ferre-D'Amare, 2013, 2014).

However, it has not been shown if antitermination happens in a step-wise manner. To address this question, we could combine the glyQS in vitro transcription antitermination assays and the fluorescence quenching assays to separate the two-step antitermination 143

process. EcoRI (E111Q), an EcoRI mutant that recognizes the restriction site but is unable to cleave the DNA template (Grundy et al., 2005), can halt the transcription elongation complex at the terminator loop of the glyQS leader RNA before the formation of a terminator structure (e.g., using Eco191, a glyQS construct with an EcoRI site at positions 191-196) (Grundy et al., 2005). A tRNAGly variant, which has a 2-aminopurine in place of the A76 residue at the acceptor end (5’-UCC-2-AP76-3’) (Ling et al., 2009), can be used to pre-bind the Eco191 leader construct at the Specifier Sequence and monitor fluorescence quenching due to interaction at the antiterminator bulge. After isolation of the pre-bound leader RNA-tRNA complex, the EcoRI protein can be removed by addition of 0.4 M KCl, which allows continuation of transcription elongation. By monitoring the quenching of fluorescence, we could determine the binding affinity between the acceptor end and the antiterminator bulge under the condition where the Specifier Sequence-anticodon interaction is already formed. This experiment would allow us to separate the interaction at the antiterminator bulge from that at the Specifier Sequence, and study whether mutations in tRNA affect the stabilization of the antiterminator structure.

It is unclear whether or not formation of the antiterminator structure is reversible, since this structure is thermodynamically less stable than the terminator. A molecular beacon (Chinnappan et al., 2013; Liu et al., 2009; Navani and Li, 2006; Zhou et al.,

2012), which is composed of a quencher dye and a reporter dye connected by a stem-loop structure carrying the sequence complementary to the 3’ side of the antiterminator, can be used in the in vitro transcription reaction to monitor the stability of the antiterminator 144

structure. When the antiterminator structure is formed, the molecular beacon is unable to bind the leader RNA transcript; therefore, fluorescence is quenched. When the target sequence is released from pairing with the T box sequence in the antiterminator and is available for binding by the molecular beacon, fluorescence is detected. This part of the study will allow us to determine the dynamics between the mutually exclusive terminator and antiterminator structures.

Whether or not tRNA modification plays an important role in antitermination is not yet well understood. Modifications in tRNA have been reported to affect codon recognition fidelity (Gustilo et al., 2008; Kothe and Rodnina, 2007), anticodon loop architecture (Agris, 2004; Denmon et al., 2011; Murphy et al., 2004; Yarian et al., 2002), folding of tRNA (Bhaskaran et al., 2012; Nobles et al., 2002) and the dynamics of tRNA structure (Agris, 2008; Vermeulen et al., 2005). It was demonstrated previously that fully modified tRNATyr resulted in increased antitermination in vitro as compared with the unmodified tRNATyr using a glyQS leader RNA variant with a tyrosine Specifier

Sequence and a change at the variable position to match the discriminator base of tRNATyr (F.J. Grundy, T.M. Henkin, unpublished results). This result suggests that tRNA modifications may be necessary for efficient antitermination, at least for some tRNA species. In B. subtilis, both tRNAAla isoacceptors recognize the GCU alanine codon; however, tRNAAla(UGC) has a mo5U modification at the first position of the anticodon, and tRNAAla(GGC) has no anticodon loop modification (Kanaya et al., 1999). It is unclear whether either or both tRNAs are responsible for alaS antitermination in vivo. In vitro transcription antitermination assays and tRNA binding experiments that use fully 145

modified tRNAAla isolated from B. subtilis would allow better understanding of the roles of tRNA modification in alaS antitermination.

It is still unclear whether EF-Tu plays a role in antitermination. Once tRNA is charged in the cell, it is quickly bound with EF-Tu and forms a ternary complex (EF-

Tu·GTP·aa-tRNA). Although it has been demonstrated that an EX1C charged tRNA mimic is able to inhibit antitermination in the in vitro transcription antitermination assays

(Grundy et al., 2005), this experiment does not completely resemble what happens in the cell. Similarly, although it has been demonstrated that in vitro-generated charged tRNAGly is capable of inhibiting antitermination in the absence of EF-Tu (Zhang and

Ferre-D'Amare, 2014), it is unknown whether the charged tRNA has equivalent affinity for the T box leader RNA and EF-Tu, or if aa-tRNA forms a ternary complex with EF-

Tu·GTP and serves as a better competitive inhibitor than the charged tRNA alone.

Isothermal titration calorimetry experiments can be used to determine the affinity of charged tRNA for EF-Tu relative to the T box leader RNA, and to obtain several thermodynamic parameters under equilibrium conditions (Feig, 2007; Zhang and Ferre-

D'Amare, 2013). In vitro transcription antitermination assays that use the ternary complex as a competitive inhibitor in the presence of uncharged tRNA could provide information about the role of the ternary complex in antitermination relative to that in translation.

It is necessary to continue the investigation of antimicrobial agents that disrupt the

T box antitermination system. Lead compounds should be tested in other amino acid classes of T box riboswitches to determine if the recognition of the antiterminator is 146

general to all antiterminator elements, or if there is any amino acid-specific effect

(Chapter 4). It is also important to continue searching for compounds that bind the antiterminator structure with improved affinity and specificity. Future studies on investigation of potential treatments for infections by multi-drug resistant Gram-positive pathogens could focus on combining other antibiotics (e.g., efflux pump inhibitors) with the compounds that target the T box regulatory system to control different drug resistance mechanisms.

147

LIST OF REFERENCES

Acquaah-Harrison, G., Zhou, S., Hines, J.V., and Bergmeier, S.C. (2010). Library of 1,4- disubstituted 1,2,3-triazole analogs of oxazolidinone RNA-binding agents. J Comb Chem 12, 491-496.

Agris, P.F. (2004). Decoding the genome: a modified view. Nucl Acids Res 32, 223-238.

Agris, P.F. (2008). Bringing order to translation: the contributions of transfer RNA anticodon-domain modifications. EMBO Rep 9, 629-635.

Anagnostopoulos, C., and Spizizen, J. (1961). Requirements for transformation in Bacillus subtilis. J Bacteriol 81, 741-746.

Anupam, R., Denapoli, L., Muchenditsi, A., and Hines, J.V. (2008a). Identification of neomycin B-binding site in T box antiterminator model RNA. Bioorg Med Chem 16, 4466-4470.

Anupam, R., Nayek, A., Green, N.J., Grundy, F.J., Henkin, T.M., Means, J.A., Bergmeier, S.C., and Hines, J.V. (2008b). 4,5-Disubstituted oxazolidinones: high affinity molecular effectors of RNA function. Bioorg Med Chem Lett 18, 3541-3544.

Ataide, S.F., Wilson, S.N., Dang, S., Rogers, T.E., Roy, B., Banerjee, R., Henkin, T.M., and Ibba, M. (2007). Mechanisms of resistance to an amino acid antibiotic that targets translation. ACS Chem Biol 2, 819-827.

Auffinger, P., and Westhof, E. (1999). Singly and bifurcated hydrogen-bonded base-pairs in tRNA anticodon hairpins and ribozymes. J Mol Biol 292, 467-483.

Banerjee, R., Chen, S., Dare, K., Gilreath, M., Praetorius-Ibba, M., Raina, M., Reynolds, N.M., Rogers, T., Roy, H., Yadavalli, S.S., et al. (2010). tRNAs: cellular barcodes for amino acids. FEBS Lett 584, 387-395.

Bernstein, J.A., Khodursky, A.B., Lin, P.-H., Lin-Chao, S., and Cohen, S.N. (2002). Global analysis of mRNA decay and abundance in Escherichia coli at single-gene resolution using two-color fluorescent DNA microarrays. Proc Natl Acad Sci USA 99, 9697-9702. 148

Bhaskaran, H., Rodriguez-Hernandez, A., and Perona, J.J. (2012). Kinetics of tRNA folding monitored by aminoacylation. RNA 18, 569-580.

Bolonna, A.A., and Kerwin, R.W. (2005). Partial agonism and schizophrenia. Br J Psychiatry 186, 7-10.

Breaker, R.R. (2012). Riboswitches and the RNA world. Cold Spring Harb Perspect Biol 4, a003566.

Brooks, W.H., Guida, W.C., and Daniel, K.G. (2011). The significance of chirality in drug design and development. Curr Topics Med Chem 11, 760-770.

Browning, D.F., and Busby, S.J. (2004). The regulation of bacterial transcription initiation. Nat Rev Microbiol 2, 57-65.

Caulfield, T.R., Devkota, B., and Rollins, G.C. (2011). Examinations of tRNA range of motion using simulations of cryo-EM microscopy and X-ray data. J Biophys 2011, 219515.

Chang, A.T., and Nikonowicz, E.P. (2013). Solution NMR determination of hydrogen bonding and base pairing between the glyQS T box riboswitch Specifier domain and the anticodon loop of tRNAGly. FEBS Lett 587, 3495-3499.

Chinnappan, R., Dubé, A., Lemay, J.-F., and Lafontaine, D.A. (2013). Fluorescence monitoring of riboswitch transcription regulation using a dual molecular beacon assay. Nucl Acids Res 41, e106.

Claesson, C., Lustig, F., Boren, T., Simonsson, C., Barciszewska, M., and Lagerkvist, U. (1995). Glycine codon discrimination and the nucleotide in position 32 of the anticodon loop. J Mol Biol 247, 191-196.

Cochella, L., and Green, R. (2005). An active role for tRNA in decoding beyond codon:anticodon pairing. Science 308, 1178-1180.

Condon, C., Grunberg-Manago, M., and Putzer, H. (1996a). Aminoacyl-tRNA synthetase gene regulation in Bacillus subtilis. Biochimie 78, 381-389.

Condon, C., Putzer, H., and Grunberg-Manago, M. (1996b). Processing of the leader mRNA plays a major role in the induction of thrS expression following threonine starvation in Bacillus subtilis. Proc Natl Acad Sci USA 93, 6992-6997.

Crick, F.H. (1966). Codon--anticodon pairing: the wobble hypothesis. J Mol Biol 19, 548-555. 149

Dare, K., and Ibba, M. (2012). Roles of tRNA in cell wall biosynthesis. Wiley Interdiscip Rev RNA 3, 247-264.

Denmon, A.P., Wang, J., and Nikonowicz, E.P. (2011). Conformation effects of base modification on the anticodon stem-loop of Bacillus subtilis tRNATyr. J Mol Biol 412, 285-303.

Deutscher, M.P. (1990). Ribonucleases, tRNA nucleotidyltransferase, and the 3' processing of tRNA. Prog Nucleic Acid Res Mol Biol 39, 209-240.

Dunkle, J.A., Wang, L., Feldman, M.B., Pulk, A., Chen, V.B., Kapral, G.J., Noeske, J., Richardson, J.S., Blanchard, S.C., and Cate, J.H. (2011). Structures of the bacterial ribosome in classical and hybrid states of tRNA binding. Science 332, 981-984.

Even, S., Pellegrini, O., Zig, L., Labas, V., Vinh, J., Brechemmier-Baey, D., and Putzer, H. (2005). Ribonucleases J1 and J2: two novel endoribonucleases in B. subtilis with functional homology to E. coli RNase E. Nucl Acids Res 33, 2141-2152.

Fahlman, R.P., Dale, T., and Uhlenbeck, O.C. (2004). Uniform binding of aminoacylated transfer RNAs to the ribosomal A and P sites. Mol Cell 16, 799-805.

Feig, A.L. (2007). Applications of isothermal titration calorimetry in RNA biochemistry and biophysics. Biopolymers 87, 293-301.

Gendron, N., Putzer, H., and Grunberg-Manago, M. (1994). Expression of both Bacillus subtilis threonyl-tRNA synthetase genes is autogenously regulated. J Bacteriol 176, 486- 494.

Gerdeman, M.S., Henkin, T.M., and Hines, J.V. (2002). In vitro structure-function studies of the Bacillus subtilis tyrS mRNA antiterminator: evidence for factor- independent tRNA acceptor stem binding specificity. Nucl Acids Res 30, 1065-1072.

Gerdeman, M.S., Henkin, T.M., and Hines, J.V. (2003). Solution structure of the Bacillus subtilis T-box antiterminator RNA: seven nucleotide bulge characterized by stacking and flexibility. J Mol Biol 326, 189-201.

Gibson, G.R., and Roberfroid, M.B. (1995). Dietary modulation of the human colonic microbiota: introducing the concept of prebiotics. J Nutr 125, 1401-1412.

Giedraitiene, A., Vitkauskiene, A., Naginiene, R., and Pavilonis, A. (2011). Antibiotic resistance mechanisms of clinically important bacteria. Medicina (Kaunas) 47, 137-146.

150

Giege, R., Sissler, M., and Florentz, C. (1998). Universal rules and idiosyncratic features in tRNA identity. Nucl Acids Res 26, 5017-5035.

Gilbert, W. (1986). Origin of life: the RNA world. Nature 319, 618.

Gold, H.S., and Moellering, R.C. (1996). Antimicrobial-drug resistance. N Engl J Med 335, 1445-1453.

Grandoni, J.A., Zahler, S.A., and Calvo, J.M. (1992). Transcriptional regulation of the ilv-leu operon of Bacillus subtilis. J Bacteriol 174, 3212-3219.

Green, N.J., Grundy, F.J., and Henkin, T.M. (2010). The T box mechanism: tRNA as a regulatory molecule. FEBS Lett 584, 318-324.

Grigg, J.C., Chen, Y., Grundy, F.J., Henkin, T.M., Pollack, L., and Ke, A. (2013). T box RNA decodes both the information content and geometry of tRNA to affect gene expression. Proc Natl Acad Sci USA 110, 7240-7245.

Gromadski, K.B., Daviter, T., and Rodnina, M.V. (2006). A uniform response to mismatches in codon-anticodon complexes ensures ribosomal fidelity. Mol Cell 21, 369- 377.

Grunberg-Manago, M. (1999). Messenger RNA stability and its role in control of gene expression in bacteria and phages. Annu Rev Genet 33, 193-227.

Grundy, F.J., Collins, J.A., Rollins, S.M., and Henkin, T.M. (2000). tRNA determinants for transcription antitermination of the Bacillus subtilis tyrS gene. RNA 6, 1131-1141.

Grundy, F.J., Haldeman, M.T., Hornblow, G.M., Ward, J.M., Chalker, A.F., and Henkin, T.M. (1997a). The Staphylococcus aureus ileS gene, encoding isoleucyl-tRNA synthetase, is a member of the T-box family. J Bacteriol 179, 3767-3772.

Grundy, F.J., and Henkin, T.M. (1993). tRNA as a positive regulator of transcription antitermination in B. subtilis. Cell 74, 475-482.

Grundy, F.J., and Henkin, T.M. (1994). Inducible amber suppressor for Bacillus subtilis. J Bacteriol 176, 2108-2110.

Grundy, F.J., and Henkin, T.M. (2004). Kinetic analysis of tRNA-directed transcription antitermination of the Bacillus subtilis glyQS gene in vitro. J Bacteriol 186, 5392-5399.

151

Grundy, F.J., and Henkin, T.M. (2006). From ribosome to riboswitch: control of gene expression in bacteria by RNA structural rearrangements. Crit Rev Biochem Mol Biol 41, 329-338.

Grundy, F.J., Hodil, S.E., Rollins, S.M., and Henkin, T.M. (1997b). Specificity of tRNA- mRNA interactions in Bacillus subtilis tyrS antitermination. J Bacteriol 179, 2587-2594.

Grundy, F.J., Moir, T.R., Haldeman, M.T., and Henkin, T.M. (2002a). Sequence requirements for terminators and antiterminators in the T box transcription antitermination system: disparity between conservation and functional requirements. Nucl Acids Res 30, 1646-1655.

Grundy, F.J., Rollins, S.M., and Henkin, T.M. (1994). Interaction between the acceptor end of tRNA and the T box stimulates antitermination in the Bacillus subtilis tyrS gene: a new role for the discriminator base. J Bacteriol 176, 4518-4526.

Grundy, F.J., Waters, D.A., Allen, S.H., and Henkin, T.M. (1993). Regulation of the Bacillus subtilis acetate kinase gene by CcpA. J Bacteriol 175, 7348-7355.

Grundy, F.J., Winkler, W.C., and Henkin, T.M. (2002b). tRNA-mediated transcription antitermination in vitro: codon-anticodon pairing independent of the ribosome. Proc Natl Acad Sci USA 99, 11121-11126.

Grundy, F.J., Yousef, M.R., and Henkin, T.M. (2005). Monitoring uncharged tRNA during transcription of the Bacillus subtilis glyQS gene. J Mol Biol 346, 73-81.

Guigou, L., and Mirande, M. (2005). Determinants in tRNA for activation of arginyl- tRNA synthetase: evidence that tRNA flexibility is required for the induced-fit mechanism. Biochemistry 44, 16540-16548.

Gustilo, E.M., Vendeix, F.A., and Agris, P.F. (2008). tRNA's modifications bring order to gene expression. Curr Opin Microbiol 11, 134-140.

Gutierrez-Preciado, A., Henkin, T.M., Grundy, F.J., Yanofsky, C., and Merino, E. (2009). Biochemical features and functional implications of the RNA-based T-box regulatory mechanism. Microbiol Mol Biol Rev 73, 36-61.

Henkin, T.M. (1996). Control of transcription termination in prokaryotes. Annu Rev Genet 30, 35-57.

Henkin, T.M. (2008). Riboswitch RNAs: using RNA to sense cellular metabolism. Genes Dev 22, 3383-3390.

152

Henkin, T.M. (2014). The T box riboswitch: A novel regulatory RNA that utilizes tRNA as its ligand. Biochim Biophys Acta, In Press.

Henkin, T.M., Glass, B.L., and Grundy, F.J. (1992). Analysis of the Bacillus subtilis tyrS gene: conservation of a regulatory sequence in multiple tRNA synthetase genes. J Bacteriol 174, 1299-1306.

Henkin, T.M., and Yanofsky, C. (2002). Regulation by transcription attenuation in bacteria: how RNA provides instructions for transcription termination/antitermination decisions. Bioessays 24, 700-707.

Hirsh, D. (1971). Tryptophan transfer RNA as the UGA suppressor. J Mol Biol 58, 439- 458.

Hou, Y.M., and Schimmel, P. (1988). A simple structural feature is a major determinant of the identity of a transfer RNA. Nature 333, 140-145.

Ibba, M., and Soll, D. (2000). Aminoacyl-tRNA synthesis. Annu Rev Biochem 69, 617- 650.

Jentzsch, F., and Hines, J.V. (2012). Interfacing medicinal chemistry with structural bioinformatics: implications for T box riboswitch RNA drug discovery. BMC Bioinformatics 13 Suppl 2, S5.

Kaberdin, V.R., and Blasi, U. (2006). Translation initiation and the fate of bacterial mRNAs. FEMS Microbiol Rev 30, 967-979.

Kanaya, S., Yamada, Y., Kudo, Y., and Ikemura, T. (1999). Studies of codon usage and tRNA genes of 18 unicellular organisms and quantification of Bacillus subtilis tRNAs: gene expression level and species-specific diversity of codon usage based on multivariate analysis. Gene 238, 143-155.

Klein, D.J., Schmeing, T.M., Moore, P.B., and Steitz, T.A. (2001). The kink-turn: a new RNA secondary structure motif. EMBO J 20, 4214-4221.

Komine, Y., and Inokuchi, H. (1990). Importance of the G27-A43 mismatch at the anticodon stem of Escherichia coli tRNAThr2. FEBS Lett 272, 55-57.

Kothe, U., and Rodnina, M.V. (2007). Codon reading by tRNAAla with modified uridine in the wobble position. Mol Cell 25, 167-174.

Lagerkvist, U. (1978). "Two out of three": an alternative method for codon reading. Proc Natl Acad Sci USA 75, 1759-1762. 153

Lancaster, L., and Noller, H.F. (2005). Involvement of 16S rRNA nucleotides G1338 and A1339 in discrimination of initiator tRNA. Mol Cell 20, 623-632.

Ledoux, S., Olejniczak, M., and Uhlenbeck, O.C. (2009). A sequence element that tunes Escherichia coli tRNAAlaGGC to ensure accurate decoding. Nat Struct Mol Biol 16, 359- 364.

Ledoux, S., and Uhlenbeck, O.C. (2008). Different aa-tRNAs are selected uniformly on the ribosome. Mol Cell 31, 114-123.

Lescoute, A., Leontis, N.B., Massire, C., and Westhof, E. (2005). Recurrent structural RNA motifs, Isostericity Matrices and sequence alignments. Nucl Acids Res 33, 2395- 2409.

Li, Z., Reimers, S., Pandit, S., and Deutscher, M.P. (2002). RNA quality control: degradation of defective transfer RNA. EMBO J 21, 1132-1138.

Liang, B., Xue, S., Terns, R.M., Terns, M.P., and Li, H. (2007). Substrate RNA positioning in the archaeal H/ACA ribonucleoprotein complex. Nat Struct Mol Biol 14, 1189-1195.

Ling, J., So, B.R., Yadavalli, S.S., Roy, H., Shoji, S., Fredrick, K., Musier-Forsyth, K., and Ibba, M. (2009). Resampling and editing of mischarged tRNA prior to translation elongation. Mol Cell 33, 654-660.

Liu, J., Cao, Z., and Lu, Y. (2009). Functional nucleic acid sensors. Chem Rev 109, 1948-1998.

Luo, D., Leautey, J., Grunberg-Manago, M., and Putzer, H. (1997). Structure and regulation of expression of the Bacillus subtilis valyl-tRNA synthetase gene. J Bacteriol 179, 2472-2478.

Lustig, F., Boren, T., Claesson, C., Simonsson, C., Barciszewska, M., and Lagerkvist, U. (1993). The nucleotide in position 32 of the tRNA anticodon loop determines ability of anticodon UCC to discriminate among glycine codons. Proc Natl Acad Sci USA 90, 3343-3347.

Lustig, F., Boren, T., Guindy, Y.S., Elias, P., Samuelsson, T., Gehrke, C.W., Kuo, K.C., and Lagerkvist, U. (1989). Codon discrimination and anticodon structural context. Proc Natl Acad Sci USA 86, 6873-6877.

154

Maciagiewicz, I., Zhou, S., Bergmeier, S.C., and Hines, J.V. (2011). Structure-activity studies of RNA-binding oxazolidinone derivatives. Bioorg Med Chem Lett 21, 4524- 4527.

Mandal, M., and Breaker, R.R. (2004). Adenine riboswitches and gene activation by disruption of a transcription terminator. Nat Struct Mol Biol 11, 29-35.

Marta, P.T., Ladner, R.D., and Grandoni, J.A. (1996). A CUC triplet confers leucine- dependent regulation of the Bacillus subtilis ilv-leu operon. J Bacteriol 178, 2150-2153.

McClain, W.H., Chen, Y.M., Foss, K., and Schneider, J. (1988). Association of transfer RNA acceptor identity with a helical irregularity. Science 242, 1681-1684.

McClain, W.H., and Foss, K. (1988). Changing the identity of a tRNA by introducing a G-U wobble pair near the 3' acceptor end. Science 240, 793-796.

Means, J., Katz, S., Nayek, A., Anupam, R., Hines, J.V., and Bergmeier, S.C. (2006). Structure-activity studies of oxazolidinone analogs as RNA-binding agents. Bioorg Med Chem Lett 16, 3600-3604.

Means, J.A., and Hines, J.V. (2005). Fluorescence resonance energy transfer studies of aminoglycoside binding to a T box antiterminator RNA. Bioorg Med Chem Lett 15, 2169-2172.

Means, J.A., Simson, C.M., Zhou, S., Rachford, A.A., Rack, J.J., and Hines, J.V. (2009). Fluorescence probing of T box antiterminator RNA: Insights into riboswitch discernment of the tRNA discriminator base. Biochem Biophys Res Commun 389, 616-621.

Means, J.A., Wolf, S., Agyeman, A., Burton, J.S., Simson, C.M., and Hines, J.V. (2007). T box riboswitch antiterminator affinity modulated by tRNA structural elements. Chem Biol Drug Des 69, 139-145.

Mehdizadeh Aghdam, E., Barzegar, A., and Hejazi, M.S. (2014). Evolutionary origin and conserved structural building blocks of riboswitches and ribosomal RNAs: riboswitches as probable target sites for aminoglycosides interaction. Adv Pharm Bull 4, 225-235.

Miller, J.H. (1972). Experiments in molecular genetics. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory.

Murphy, F.V., Ramakrishnan, V., Malkiewicz, A., and Agris, P.F. (2004). The role of modifications in codon discrimination by tRNALysUUU. Nat Struct Mol Biol 11, 1186- 1191.

155

Nakano, M.M., and Zuber, P. (1989). Cloning and characterization of srfB, a regulatory gene involved in surfactin production and competence in Bacillus subtilis. J Bacteriol 171, 5347-5353.

National Nosocomial Infections Surveillance System. (2004). National Nosocomial Infections Surveillance (NNIS) System Report, data summary from January 1992 through June 2004, issued October 2004. Am J Infect Control 32, 470-485.

Navani, N.K., and Li, Y. (2006). Nucleic acid aptamers and enzymes as sensors. Curr Opin Chem Biol 10, 272-281.

Nelson, A.R., Henkin, T.M., and Agris, P.F. (2006). tRNA regulation of gene expression: interactions of an mRNA 5'-UTR with a regulatory tRNA. RNA 12, 1254-1261.

Nobles, K.N., Yarian, C.S., Liu, G., Guenther, R.H., and Agris, P.F. (2002). Highly conserved modified nucleosides influence Mg2+-dependent tRNA folding. Nucl Acids Res 30, 4751-4760.

Nottrott, S., Hartmuth, K., Fabrizio, P., Urlaub, H., Vidovic, I., Ficner, R., and Luhrmann, R. (1999). Functional interaction of a novel 15.5kD [U4/U6.U5] tri-snRNP protein with the 5' stem-loop of U4 snRNA. EMBO J 18, 6119-6133.

Ogle, J.M., Brodersen, D.E., Clemons, W.M., Tarry, M.J., Carter, A.P., and Ramakrishnan, V. (2001). Recognition of cognate transfer RNA by the 30S ribosomal subunit. Science 292, 897-902.

Ogle, J.M., Murphy, F.V., Tarry, M.J., and Ramakrishnan, V. (2002). Selection of tRNA by the ribosome requires a transition from an open to a closed form. Cell 111, 721-732.

Olejniczak, M., Dale, T., Fahlman, R.P., and Uhlenbeck, O.C. (2005). Idiosyncratic tuning of tRNAs to achieve uniform ribosome binding. Nat Struct Mol Biol 12, 788-793.

Olejniczak, M., and Uhlenbeck, O.C. (2006). tRNA residues that have coevolved with their anticodon to ensure uniform and accurate codon recognition. Biochimie 88, 943-950.

Oliva, R., Cavallo, L., and Tramontano, A. (2006). Accurate energies of hydrogen bonded nucleic acid base pairs and triplets in tRNA tertiary interactions. Nucl Acids Res 34, 865-879.

Ormerod, L.P. (2005). Multidrug-resistant tuberculosis (MDR-TB): epidemiology, prevention and treatment. Br Med Bull 73-74, 17-24.

156

Peselis, A., and Serganov, A. (2014). Themes and variations in riboswitch structure and function. Biochim Biophys Acta, In Press.

Peterson, E.T., and Uhlenbeck, O.C. (1992). Determination of recognition nucleotides for Escherichia coli phenylalanyl-tRNA synthetase. Biochemistry 31, 10380-10389.

Phizicky, E.M., and Hopper, A.K. (2010). tRNA biology charges to the front. Genes Dev 24, 1832-1860.

Putzer, H., Condon, C., Brechemier-Baey, D., Brito, R., and Grunberg-Manago, M. (2002). Transfer RNA-mediated antitermination in vitro. Nucl Acids Res 30, 3026-3033.

Putzer, H., Gendron, N., and Grunberg-Manago, M. (1992). Co-ordinate expression of the two threonyl-tRNA synthetase genes in Bacillus subtilis: control by transcriptional antitermination involving a conserved regulatory sequence. EMBO J 11, 3117-3127.

Putzer, H., Laalami, S., Brakhage, A.A., Condon, C., and Grunberg-Manago, M. (1995). Aminoacyl-tRNA synthetase gene regulation in Bacillus subtilis: induction, repression and growth-rate regulation. Mol Microbiol 16, 709-718.

Raina, M., and Ibba, M. (2014). tRNAs as regulators of biological processes. Front Genet 5, 171.

Ramos, A., and Varani, G. (1997). Structure of the acceptor stem of Escherichia coli tRNAAla: role of the G3·U70 base pair in synthetase recognition. Nucl Acids Res 25, 2083-2090.

Rich, A., and RajBhandary, U.L. (1976). Transfer RNA: molecular structure, sequence, and properties. Annu Rev Biochem 45, 805-860.

Robertus, J.D., Ladner, J.E., Finch, J.T., Rhodes, D., Brown, R.S., Clark, B.F., and Klug, A. (1974). Structure of yeast phenylalanine tRNA at 3 Å resolution. Nature 250, 546-551.

Rollins, S.M., Grundy, F.J., and Henkin, T.M. (1997). Analysis of cis-acting sequence and structural elements required for antitermination of the Bacillus subtilis tyrS gene. Mol Microbiol 25, 411-421.

Rozhdestvensky, T.S., Tang, T.H., Tchirkova, I.V., Brosius, J., Bachellerie, J.P., and Huttenhofer, A. (2003). Binding of L7Ae protein to the K-turn of archaeal snoRNAs: a shared RNA binding motif for C/D and H/ACA box snoRNAs in Archaea. Nucl Acids Res 31, 869-877.

157

Saad, N.Y., Stamatopoulou, V., Braye, M., Drainas, D., Stathopoulos, C., and Becker, H.D. (2013). Two-codon T-box riboswitch binding two tRNAs. Proc Natl Acad Sci USA 110, 12756-12761.

Saks, M.E., and Conery, J.S. (2007). Anticodon-dependent conservation of bacterial tRNA gene sequences. RNA 13, 651-660.

Samhita, L., Shetty, S., and Varshney, U. (2012). Unconventional initiator tRNAs sustain Escherichia coli. Proc Natl Acad Sci USA 109, 13058-13063.

Schmeing, T.M., Voorhees, R.M., Kelley, A.C., and Ramakrishnan, V. (2011). How mutations in tRNA distant from the anticodon affect the fidelity of decoding. Nat Struct Mol Biol 18, 432-436.

Schrader, J.M., Chapman, S.J., and Uhlenbeck, O.C. (2011). Tuning the affinity of aminoacyl-tRNA to elongation factor Tu for optimal decoding. Proc Natl Acad Sci USA 108, 5215-5220.

Schrader, J.M., and Uhlenbeck, O.C. (2011). Is the sequence-specific binding of aminoacyl-tRNAs by EF-Tu universal among bacteria? Nucl Acids Res 39, 9746-9758.

Schroeder, K.T., McPhee, S.A., Ouellet, J., and Lilley, D.M. (2010). A structural database for k-turn motifs in RNA. RNA 16, 1463-1468.

Schultz, D.W., and Yarus, M. (1994a). tRNA structure and ribosomal function : I. tRNA nucleotide 27-43 mutations enhance first position wobble. J Mol Biol 235, 1381-1394.

Schultz, D.W., and Yarus, M. (1994b). tRNA structure and ribosomal function. II. Interaction between anticodon helix and other tRNA mutations. J Mol Biol 235, 1395- 1405.

Serganov, A., and Nudler, E. (2013). A decade of riboswitches. Cell 152, 17-24.

Shalgi, R., Hurt, J.A., Krykbaeva, I., Taipale, M., Lindquist, S., and Burge, C.B. (2013). Widespread regulation of translation by elongation pausing in heat shock. Mol Cell 49, 439-452.

Shepotinovskaya, I., and Uhlenbeck, O.C. (2013). tRNA residues evolved to promote translational accuracy. RNA 19, 510-516.

Smith, D., and Yarus, M. (1989a). Transfer RNA structure and coding specificity. I. Evidence that a D-arm mutation reduces tRNA dissociation from the ribosome. J Mol Biol 206, 489-501. 158

Smith, D., and Yarus, M. (1989b). Transfer RNA structure and coding specificity. II. A D-arm tertiary interaction that restricts coding range. J Mol Biol 206, 503-511.

Sprinzl, M., and Cramer, F. (1975). Site of aminoacylation of tRNAs from Escherichia coli with respect to the 2'- or 3'-hydroxyl group of the terminal adenosine. Proc Natl Acad Sci USA 72, 3049-3053.

Sprinzl, M., Horn, C., Brown, M., Ioudovitch, A., and Steinberg, S. (1998). Compilation of tRNA sequences and sequences of tRNA genes. Nucl Acids Res 26, 148-153.

Storz, G., Altuvia, S., and Wassarman, K.M. (2005). An abundance of RNA regulators. Annu Rev Biochem 74, 199-217.

Storz, G., Vogel, J., and Wassarman, K.M. (2011). Regulation by small RNAs in bacteria: expanding frontiers. Mol Cell 43, 880-891.

Stragier, P., Bonamy, C., and Karmazyn-Campelli, C. (1988). Processing of a sporulation sigma factor in Bacillus subtilis: how morphological structure could control gene expression. Cell 52, 697-704.

Sun, L., Campbell, F.E., Zahler, N.H., and Harris, M.E. (2006). Evidence that substrate‐specific effects of C5 protein lead to uniformity in binding and catalysis by RNase P. EMBO J 25, 3998-4007.

Trausch, J.J., and Batey, R.T. (2014). A disconnect between high-affinity binding and efficient regulation by antifolates and purines in the tetrahydrofolate riboswitch. Chem Biol 21, 205-216.

Turner, B., Melcher, S.E., Wilson, T.J., Norman, D.G., and Lilley, D.M. (2005). Induced fit of RNA on binding the L7Ae protein to the kink-turn motif. RNA 11, 1192-1200.

Valle, M., Sengupta, J., Swami, N.K., Grassucci, R.A., Burkhardt, N., Nierhaus, K.H., Agrawal, R.K., and Frank, J. (2002). Cryo-EM reveals an active role for aminoacyl- tRNA in the accommodation process. EMBO J 21, 3557-3567.

Valle, M., Zavialov, A., Li, W., Stagg, S.M., Sengupta, J., Nielsen, R.C., Nissen, P., Harvey, S.C., Ehrenberg, M., and Frank, J. (2003). Incorporation of aminoacyl-tRNA into the ribosome as seen by cryo-electron microscopy. Nat Struct Mol Biol 10, 899-906.

Vermeulen, A., McCallum, S.A., and Pardi, A. (2005). Comparison of the global structure and dynamics of native and unmodified tRNAVal. Biochemistry 44, 6024-6033.

159

Vitreschak, A.G., Mironov, A.A., Lyubetsky, V.A., and Gelfand, M.S. (2008). Comparative genomic analysis of T-box regulatory systems in bacteria. RNA 14, 717- 735.

Wang, J., Henkin, T.M., and Nikonowicz, E.P. (2010). NMR structure and dynamics of the Specifier Loop domain from the Bacillus subtilis tyrS T box leader RNA. Nucl Acids Res 38, 3388-3398.

Waters, L.S., and Storz, G. (2009). Regulatory RNAs in bacteria. Cell 136, 615-628.

Winkler, W.C., and Breaker, R.R. (2005). Regulation of bacterial gene expression by riboswitches. Annu Rev Microbiol 59, 487-517.

Winkler, W.C., Grundy, F.J., Murphy, B.A., and Henkin, T.M. (2001). The GA motif: an RNA element common to bacterial antitermination systems, rRNA, and eukaryotic RNAs. RNA 7, 1165-1172.

Wisplinghoff, H., Bischoff, T., Tallent, S.M., Seifert, H., Wenzel, R.P., and Edmond, M.B. (2004). Nosocomial bloodstream infections in US hospitals: analysis of 24,179 cases from a prospective nationwide surveillance study. Clin Infect Dis 39, 309-317.

Yamada, Y., Matsugi, J., Ishikura, H., and Murao, K. (2005). Bacillus subtilis tRNAPro with the anticodon mo5UGG can recognize the codon CCC. Biochim Biophys Acta 1728, 143-149.

Yarian, C., Townsend, H., Czestkowski, W., Sochacka, E., Malkiewicz, A.J., Guenther, R., Miskiewicz, A., and Agris, P.F. (2002). Accurate translation of the genetic code depends on tRNA modified nucleosides. J Biol Chem 277, 16391-16395.

Yarus, M. (1982). Translational efficiency of transfer RNA's: uses of an extended anticodon. Science 218, 646-652.

Yarus, M., Valle, M., and Frank, J. (2003). A twisted tRNA intermediate sets the threshold for decoding. RNA 9, 384-385.

Yousef, M.R., Grundy, F.J., and Henkin, T.M. (2003). tRNA requirements for glyQS antitermination: a new twist on tRNA. RNA 9, 1148-1156.

Yousef, M.R., Grundy, F.J., and Henkin, T.M. (2005). Structural transitions induced by the interaction between tRNAGly and the Bacillus subtilis glyQS T box leader RNA. J Mol Biol 349, 273-287.

160

Zagryadskaya, E.I., Kotlova, N., and Steinberg, S.V. (2004). Key elements in maintenance of the tRNA L-shape. J Mol Biol 340, 435-444.

Zhang, J., and Ferre-D'Amare, A.R. (2013). Co-crystal structure of a T-box riboswitch stem I domain in complex with its cognate tRNA. Nature 500, 363-366.

Zhang, J., and Ferre-D'Amare, A.R. (2014). Direct evaluation of tRNA aminoacylation status by the T-box riboswitch using tRNA-mRNA stacking and steric readout. Mol Cell 55, 148-155.

Zhou, S., Acquaah-Harrison, G., Bergmeier, S.C., and Hines, J.V. (2011). Anisotropy studies of tRNA-T box antiterminator RNA complex in the presence of 1,4-disubstituted 1,2,3-triazoles. Bioorg Med Chem Lett 21, 7059-7063.

Zhou, S., Acquaah-Harrison, G., Jack, K.D., Bergmeier, S.C., and Hines, J.V. (2012). Ligand-induced changes in T box antiterminator RNA stability. Chem Biol Drug Des 79, 202-208.

Zuber, P., and Losick, R. (1987). Role of AbrB in Spo0A- and Spo0B-dependent utilization of a sporulation promoter in Bacillus subtilis. J Bacteriol 169, 2223-2230.

161

APPENDIX A

SUPPLEMENTAL MATERALS

A. B.

C. D.

Figure A.1. Best-fit curves. The best-fit curves were generated as described in Fig. 2.2. The effects of U32C (A, B), C30G-G40C (C) and A24G (D) relative to the wild-type tRNAGly on antitermination of the glyQS wild-type leader template (A, C, D) or its variant with a G99C (B) substitution are shown.

162

Table A.1. Effect of tRNA mutations on antitermination efficiency (K1/2) of the wild-type glyQS leader RNA and leader variants. Pairing alterations at position 3 Pairing alterations at position 1

Specifier b. 5’-GGC-3’ GGG GGU GGG GGA GGC GGU AGC UGC CGC Sequence GGA b. anticodon 3’-CCG-5’ CCU CCG CCG CCG CCU CCU CCG CCG CCG CCU pair WT A-U G•U U•G G•G A•G C•U U•U A•C U•C C•C

a. 0.031 ± 0.007 0.081 ± 0.017 0.086 ± 0.023 0.060 ± 0.009 0.11 ± 0.03 0.13 ± 0.02 0.20 ± 0.06 0.23 ± 0.07 0.095 ± 0.018 0.22 ± 0.06 0.32 ± 0.09 WT c. (1.0) (2.6) (2.8) (1.9) (3.5) (4.2) (6.5) (7.4) (3.1) (7.1) (10)

0.015 ± 0.004 0.047 ± 0.010 0.067 ± 0.012 0.045 ± 0.009 0.067 ± 0.012 0.078 ± 0.016 0.090 ± 0.017 0.13 ± 0.03 0.055 ± 0.014 0.087 ± 0.017 0.17 ± 0.02 U32C d. (0.48) (0.58) (0.78) (0.75) (0.61) (0.6) (0.45) (0.57) (0.58) (0.40) (0.53)

0.037 ± 0.010 e. 0.073 ± 0.017 0.13 ± 0.03 0.15 ± 0.03 0.12 ± 0.03 0.30 ± 0.08 0.28 ± 0.05 C30G-G40C n/a n/a n/a (1.2) n/a (1.2) (1.2) (1.2) (1.3) (1.4) (0.88)

0.034 ± 0.008 0.071 ± 0.014 0.087 ± 0.019 0.070 ± 0.007 0.14 ± 0.03 0.16 ± 0.04 0.18 ± 0.03 0.19 ± 0.03 0.13 ± 0.02 0.31 ± 0.05 0.36 ± 0.09 C27G•G43A (1.1) (0.89) (1.0) (1.2) (1.3) (1.2) (0.9) (0.83) (1.4) (1.4) (1.1) 0.039 ± 0.010 0.075 ± 0.018 0.10 ± 0.03 0.098 ± 0.020 0.14 ± 0.03 0.12 ± 0.02 0.24 ± 0.07 0.20 ± 0.04 0.11 ± 0.04 0.30 ± 0.06 0.26 ± 0.03 A26G•G44U (1.3) (0.93) (1.2) (1.6) (1.3) (0.92) (1.2) (0.87) (1.2) (1.4) (0.81)

tRNA mutations 0.081 ± 0.022 0.053 ± 0.011 0.075 ± 0.020 0.21 ± 0.06 0.34 ± 0.07 0.38 ± 0.10 0.094 ± 0.018 0.13 ± 0.02 0.18 ± 0.03 0.34 ± 0.06 0.75 ± 0.28 U11C

163 (2.6) (0.65) (0.93) (3.5) (3.1) (2.9) (0.47) (0.57) (1.9) (1.5) (2.3) 0.057 ± 0.011 0.11 ± 0.02 0.13 ± 0.02 0.15 ± 0.02 0.25 ± 0.04 0.34 ± 0.06 0.21 ± 0.05 0.28 ± 0.03 0.15 ± 0.02 0.37 ± 0.08 0.38 ± 0.11 A24G (1.8) (1.4) (1.5) (2.5) (2.1) (2.6) (1.1) (1.2) (1.6) (1.7) (1.2) 0.027 ± 0.006 0.066 ± 0.097 0.086 ± 0.010 0.090 ± 0.012 0.16 ± 0.02 0.14 ± 0.02 0.21 ± 0.02 0.30 ± 0.03 0.077 ± 0.008 0.18 ± 0.02 0.24 ± 0.04 U11C-A24G (0.87) (0.81) (1.0) (1.5) (1.5) (1.1) (1.1) (1.3) (0.81) (0.82) (0.75) a. K values (µM) in Mean ± SEM. 1/2 b. Specifier Sequence and anticodon mutations were underlined. c. Fold loss in antitermination efficiency due to pairing alterations between the Specifier Sequence and anticodon relative to the wild-type. K (mutant/0.031). 1/2 d. Fold loss in antitermination in the presence of tRNA mutations relative to the absence of tRNA mutations under the same Specifier Sequence-anticodon interaction. K (+tRNA 1/2 mutation/-tRNA mutation). e. n/a: not applicable.

163

Table A.2. Effects of tRNA mutations on RTmax of the wild-type glyQS leader RNA and leader variants. RT (%) max Interactions at position 3 Interactions at position 1

a. A-U G•U U•G G•G A•G C•U U•U b. U•C C•C C-G A•C WT 94 ± 5 100 ± 10 100 ± 10 99 ± 5 100 ± 10 100 ± 10 100 ± 20 100 ± 10 100 ± 10 100 ± 10 100 ± 10

U32C 75 ± 4 95 ± 6 100 ± 10 100 ± 10 99 ± 6 100 ± 10 100 ± 10 94 ± 8 86 ± 7 99 ± 7 76 ± 5

C30G-G40C 80 ± 6 c. n/a 98 ± 8 100 ± 10 100 ± 10 n/a n/a 100 ± 10 100 ± 10 99 ± 10 n/a C27G•G43A 93 ± 6 100 ± 10 100 ± 10 99 ± 3 100 ± 10 100 ± 10 100 ± 10 100 ± 10 96 ± 5 95 ± 8 91 ± 13

A26G•G44U 88 ± 6 97 ± 8 100 ± 10 100 ± 10 100 ± 10 93 ± 5 100 ± 10 96 ± 8 100 ± 10 100 ± 10 100 ± 20

U11C 100 ± 10 93 ± 6 97 ± 9 100 ± 10 100 ± 10 100 ± 20 97 ± 7 100 ± 10 100 ± 10 98 ± 9 100 ± 30

tRNA tRNA mutations A24G 89 ± 5 100 ± 10 100 ± 10 98 ± 5 96 ± 7 100 ± 10 88 ± 10 83 ± 5 98 ± 6 100 ± 10 75 ± 13

U11C-A24G 95 ± 6 100 ± 10 99 ± 4 100 ± 10 100 ± 10 100 ± 10 99 ± 5 94 ± 5 100 ± 10 100 ± 10 92 ± 7 a. The first letter represents the nucleotide at position 3 of the Specifier Sequence, and the second letter represents the nucleotide at position 1 of the anticodon; mutations are underlined. b. 164 The first letter represents the nucleotide at position 1 of the Specifier Sequence, and the second letter represents the nucleotide at position 3 of the anticodon. These constructs have a C-G perfect match at position 3.

c. n/a: not applicable

164

Table A.3. Effect of mutations in tRNAGly(UCC)G34U on readthrough (%RT) of constructs with mismatches at position 1, or positions 1 and 3 of the Specifier Sequence. %RT at [tRNA] = 0.5 (μM) a. Position 3 A-U C•U b. Position 1 G-C A•C U•C C•C G-C A•C U•C C•C

WT 80 ± 1 25 ± 2 9.1 ± 0.8 6.7 ± 0.4 73 ± 2 8.5 ± 0.4 6.6 ± 0.8 4.9 ± 0.2

U32C 78 ± 1 27 ± 2 22 ± 1 6.4 ± 0.2 78 ± 1 16 ± 1 9.6 ± 2.4 4.5 ± 0.1

C27G•G43A 84 ± 1 26 ± 2 7.8 ± 1.1 5.9 ± 0.1 72 ± 1 8.6 ± 1.5 5.5 ± 0.7 6.8 ± 2.3

A26G•G44U 77 ± 2 22 ± 2 8.7 ± 0.8 5.2 ± 1.1 66 ± 7 13 ± 2 5.0 ± 2.3 4.0 ± 1.2

U11C 76 ± 2 30 ± 2 14 ± 1 6.7 ± 1.0 75 ± 5 18 ± 4 14 ± 2 5.0 ± 2.5

tRNA tRNA mutations A24G 76 ± 2 21 ± 6 12 ± 2 8.2 ± 0.1 57 ± 1 6.4 ± 0.5 9.1 ± 3.8 6.5 ± 2.7

U11C-A24G 85 ± 1 25 ± 1 10 ± 1 6.8 ± 0.2 67 ± 3 7.2 ± 0.3 5.7 ± 0.1 4.8 ± 0.2 a. The first letter represents the nucleotide at position 3 of the Specifier Sequence, and the second letter represents the nucleotide at position 1 of the anticodon; mutations are underlined. b. The first letter represents the nucleotide at position 1 of the Specifier Sequence, and the second letter represents the nucleotide at position 3 of the anticodon.

165

Table A.4. Effect of tRNA mutations on binding affinity (Kd) of the wild-type glyQS leader RNA and leader variants. Pairing alterations at position 3 Pairing alterations at position 1

Specifier b. 5’-GGC-3’ GGG GGU GGG GGA GGC GGU AGC UGC CGC Sequence GGA anticodon 3’-CCG-5’ CCU CCU CCG CCG CCG CCU CCU CCG CCG CCG

pair WT A-U G•U U•G G•G A•G C•U U•U A•C U•C C•C

a. 0.092 ± 0.018 0.18 ± 0.03 0.25 ± 0.03 0.29 ± 0.04 0.37 ± 0.03 0.22 ± 0.03 0.45 ± 0.06 0.45 ± 0.05 0.21 ± 0.03 0.82 ± 0.05 0.47 ± 0.05 WT c. (1) (2.0) (2.7) (3.2) (4.0) (2.4) (4.9) (4.9) (2.3) (8.9) (5.1) 0.063 ± 0.011 0.082 ± 0.014 0.18 ± 0.03 0.19 ± 0.02 0.23 ± 0.03 0.15 ± 0.03 0.20 ± 0.03 0.18 ± 0.02 0.89 ± 0.08 0.69 ± 0.05 1.1 ± 0.1 U32C d. (0.68) (0.46) (0.72) (0.66) (0.62) (0.68) (0.44) (0.4) (4.2) (0.84) (2.3) 0.19 ± 0.02 0.49 ± 0.06 e. 0.50 ± 0.06 0.52 ± 0.05 U11C n/a n/a n/a n/a n/a n/a tRNA tRNA mutations (2.1) (2.7) n/a (2.3) (1.2) a. K values (µM) in Mean ± SEM. 166 d b. Specifier Sequence and anticodon mutations are underlined.

c. Fold loss in binding affinity due to pairing alterations between the Specifier Sequence and anticodon relative to the wild-type. K (mutant/0.092). d d. Fold loss in binding affinity in the presence of tRNA mutations relative to the absence of tRNA mutations under the same Specifier Sequence-anticodon interaction. K (+tRNA mutation/-tRNA mutation). d e. n/a: not applicable.

166

Figure A.2. Effect of U32C in truncated tRNAGly(GCC)ΔUCCA on binding. The 5’- UCCA-3’ end of wild-type tRNAGly(GCC) and a U32C variant was deleted, and the binding affinity for the wild-type leader RNA (A) and variants with a G99A (B), G99U (C) or G99C (D) mutation at position 1 of the Specifier Sequence was determined by size-exclusion filtration assays as described in Materials and Methods.

167

Table A.5. Results of compound tests (provided by Dr. Jennifer Hines). ∆FRET ∆Anisotropy ∆Tm MIC a. b. c. d. AM1A (%) (%) (°C) (µM) Compound S. aureus B. subtilis E. faecalis E. coli IMB-16 9 -11 0.3 31 31 n/ae. n/a GHB-7 10 -79 3.2 500 250 n/a n/a GHB-23 8 -58 -0.7 n/a n/a n/a n/a GHB-54 -31 -53 5.2 16 16 62 1000 GHB-56 -13 -58 4.2 31 16 62 - GHB-60 12 -59 -0.3 n/a 500 n/a n/a GHB-76 12 -71 0.3 250 125 250 n/a GHB-134 23 -100 0.1 1000 500 n/a - GHB-146 19 -61 -0.4 n/a n/a - n/a a. ∆FRETAM1A (%) was determined by [F/F0-1] * 100, where “F” is the fluorescence detected in the presence of compound and “F0” is the fluorescence detected in the absence of compound (Maciagiewicz et al., 2011). b. ∆Anisotropy (%) was determined by [(r-r+)/(r+-r-)] * 100, where “r” is the anisotropy of the AM1A/tRNA complex in the presence of compound, “r+” is the anisotropy of the AM1A/tRNA complex, and “r-” is the anisotropy of AM1A alone (Zhou et al., 2011). c. ∆Tm (°C) was determined by Tm-Tm0, where “Tm” is the melting temperature of AM1A in the presence of compound, and “Tm0” is the melting temperature of AM1A in the absence of compound (Zhou et al., 2012). d.MIC (µM) was determined by broth dilution assays (S. Bergmeier, N. Priestley, unpublished data). e.na indicates MIC ≥ 1000 µM.

168