Translational Regulation of Mrna by G-Quadruplex Structures

Home , Structural motif, Translational regulation

Translational Regulation of mRNA by G-Quadruplex Structures

A dissertation submitted to the Department of Chemistry at Kent State University in partial fulfillment of the requirements for the degree of Doctor of Philosophy

Mark J. Morris

August 2012

Dissertation written by Mark Morris B.S., Union College, 2005 Ph.D., Kent State University, 2012

Approved by

______, Chair, Doctoral Dissertation Committee

______, Advisor, Doctoral Dissertation Committee Soumitra Basu , Ph.D.

______, Member, Doctoral Dissertation Committee William Merrick, Ph.D.

______, Member, Doctoral Dissertation Committee Hanbin Mao, Ph.D.

______, Member, Doctoral Dissertation Committee Roger Gregory, Ph.D.

Accepted by

______, Chair, Dept. Chemistry and Biochemistry Michael Tubergen, Ph.D.

______, Dean, College of Arts and Sciences John R.D. Stalvey, Ph.D.

Table of Contents

Page

LIST OF FIGURES ……………………………………………………………….……..… vi

LIST OF TABLES ………………………………………………………………….……...... xii

ACKNOWLEDGMENTS …………………………………………………………...………xiii

CHAPTER 1 Introduction and Background …………………………………..………… 1

1.1 Discovery of G-quadruplexes ………………………………………………….... 1

1.2 G-quadruplex structures in the 5'-UTR of mRNAs………………………………..5

1.3 Modulation of quadruplex structure and function by small molecules…………….7

CHAPTER 2 An extremely stable G-quadruplex within 5'-UTR of the MT3 matrix metalloproteinase mRNA represses translation in eukaryotic cells...... 10 2.1 Introduction ………………………………………………………………………………..10

2.2 Materials and Methods …………………………………………………………………….11

2.3 Results……………………………………………………………………………………….15

2.4 Discussion…………………………………………………………………………………..26

2.5 Conclusion………………………………………………………………………………….30

CHAPTER 3 The porphyrin TmPyP4 unfolds the extremely stable G-quadruplex in MT3-MMP mRNA and alleviates its repressive effect to enhance translation in eukaryotic cells ………………………………………………………………...31

iii

3.1 Introduction ………………………………………………………………………………...31

3.2 Materials and Methods ………………………………………………………………….....32

3.3 Results and Discussion……………………………………………………………………...36

3.4 Conclusion…………………………………………………………………………………..54

CHAPTER 4 An RNA G-quadruplex is essential for cap-independent translation initiation in human VEGF IRES ………………………………………………………….....55

4.1 Introduction ………………………………………………………………………………..55

4.2 Materials and Methods…………………………………………………………………….57

4.3 Results ………………………………………………………………………………….....62

4.4 Discussion………………………………………………………………………………….77

4.5 Conclusion……………………………………………………………………………..…..82

CHAPTER 5 RNA DOMAIN SWAPPING SHOW CONTEXT DEPENDENT EFFECT

OF G-QUADRUPLEXES ON TRANSLATION …………………………………………83

5.1 Introduction ………………………………………………………………………..….….83

5.2 Materials and Methods ……………………………………………………………….…..85

5.3 Results and Discussion…………………………………………………………………….88

5.4 Conclusion……………………………………………………………………………...….95

CONCLUDING REMARKS……………………………………………………………...…96

REFERENCES………………………………………………………………………………..97

List of Figures

Page

CHAPTER 1 Introduction and Background

Figure 1.1. Structure and H-bond formation of a G-tetrad ………………………………..... 1

Figure 1.2. Structure of a G-quadruplex ………………………………………………...….. 2

Figure 1.3 General arrangements of G-quadruplexes based on orientation of the

strand……………………………………………………………………………..3

CHAPTER 2 An unusually stable G-quadruplex within 5'-UTR of the MT3 matrix

metalloproteinase mRNA represses translation in eukaryotic cell

Figure 2.1 Schematic representation of MT3-MMP mRNA …………………………..……11

Figure 2.2 Circular dichroism spectra of M3Q and mut-M3Q in the presence of various

concentrations of KCl. …………………………………………………………..16

Figure 2.3 Circular dichroism melting curves of M3Q RNA ……………………………….17

Figure 2.4 Circular dichroism cooling curve and first derivative plot of M3Q

RNA………………………………………………………………….…………..18

Figure 2.5 Plot of Tm values for M3Q RNA at various strand concentrations ……..…...... 20

Figure 2.6 RNase T1 footprinting of M3Q and mut-M3Q. …………………………………21

Figure 2.7 Schematic representation of the plasmids used to investigate the effect

of the 5'-UTR of MT3-MMP on translation. …………………….……….…….23

Figure 2.8 Histogram representing the ratio of Renilla/firefly luciferase activities

in HeLa cells………………………………………………………………….…23

Figure 2.9 Schematic representation of the plasmids used to investigate the effect of the

M3Q on translation……………………………………………………………..24

Figure 2.10 Histogram representing the ratio of Renilla/firefly luciferase activities in

HeLa cells …………...………………………………………………………...25

Figure 2.11 Circular dichroism first derivative plot of M3Q RNA in the presence

of 5 mM KCl…………………………………………………………………..26

CHAPTER 3 The porphyrin TmPyP4 unfolds the extremely stable G-quadruplex in MT3-

MMP mRNA and alleviates its repressive effect to enhance translation in eukaryotic cells

Figure 3.1 CD spectra of 4 µM prefolded (in 100 mM KCl) M3Q in the absence

and presence of increasing concentrations of TmPyP4………………………..37

Figure 3.2 CD spectra of TmPyP4 in the absence and presence of folded M3Q………….38

Figure 3.3 Plot of calculated fraction folded vs. TmPyP4 concentration………………….39

Figure 3.4 CD spectra of 4 µM M3Q RNA (in 100 mM KCl) folded in the absence

and presence of 0, 2, 5, 10, and 20 µM TmPyP4……………………………...40

Figure 3.5 Native gel shift assay of M3Q (final concentration of about 250 nM) in

the absence and presence of increasing concentrations of TmPyP4…………..41

Figure 3.6 NMR spectra of 0.42 mM M3Q titrated with TmPyP4………………………..43

Figure 3.7 Visible absorbtion spectra of TmPyP4 in the absence and presence

vii

of increasing concentration of prefolded M3Q………………………………...45

Figure 3.8 Visible absorbtion spectra of TmPyP4 in the absence and presence

of mut-M3Q……………………………………………………………………46

Figure 3.9 A) Schematic of dual luciferase bi-cistronic constructs. B) Histogram

showing % activity of the translation of the Renilla gene as a function

of TmPyP4 concentration…………………………………………………..…48

Figure 3.10 Histogram representing the ratio of Renilla to firefly mRNA CT

values in HeLa cells determined by qRT-PCR………………………………50

Figure 3.11 CD-melting spectrum of the DNA version of M3Q (4 µM) in the

presence of 100 mM KCl…………………………………………………….52

Figure 3.12 Plot of fraction folded vs. TmPyP4 concentration of 4 µM

prefolded (in 100 mM KCl) DNA version of M3Q in the absence

and presence of increasing concentrations of TmPyP4……………………...53

CHAPTER 4 An RNA G-quadruplex is essential for cap-independent translation initiation in human VEGF IRES

Figure 4.1. Primary nucleotide sequence of the human VEGF IRES-A ……...... 56

Figure 4.2. RNase T1 footprinting in the presence of 150 mM K+, 150 mM Li+ and 1 mM

MgCl2 …………………………………………………………………………62

viii

Figure 4.3 DMS footprinting in the presence of 150 mM K+ or 150 mM

Li+……………………………………………………………………………..63

Figure 4.4 Schematic of a subset of G-quadruplex structures that shows the different G-

stretches……………………………………………………………………….64

Figure 4.5 Circular dichroism (CD) spectra of an oligoribonucleotide encompassing the

protected region in hVEGF IRES A…………………………………………..66

Figure 4.6 Schematic of various dual luciferase bi-cistronic constructs………………....68

Figure 4.7 Histogram showing % activity of the mutant constructs normalized

to the wild type construct……………………………………………………..69

Figure 4.8 Histogram representing the ratio of Renilla to firefly luciferase

activities in HeLa cells……………………………………………………...…70

Figure 4.9 Scanned images of gels showing RNase T1 footprinting of the

293 nt human VEGF IRES-A and its various mutants……………………….71

Figure 4.10 Scanned images of gels showing DMS footprinting of the

293 nt human VEGF IRES-A and its various mutants………………..…….73

Figure 4.11 Histogram showing % activity of the dual luciferase

rescue mutant construct (Rescue Quad) normalized to its parent

construct (G774, 789U)………………………………………………………76

Figure 4.12 Scanned image of a gel showing RNase T1 footprinting

of the mutant Rescue Quad version of the transcribable

293 nt human VEGF IRES-A in the presence of K+ and Li+………………..77

Figure 4.13 Histograms representing quantitation of the gel shown

in Figure 4.12………………………………………………………………..79

CHAPTER 5 Effect of G-quadruplex domain exchange on translation

Figure 5.1 Schematic of RNA G-quadruplex swapping………………...……………….83

Figure 5.2 Dual luciferase reporter assay results demonstrating the roles

in translation of endogenous VEGF G-quadruplex, the 4MVF

sequence, the M3Q G-quadruplex forming sequence , and the

NRAS quadruplex forming sequence all within the 5' UTR

of hVEGF IRES-A…………………………………………………………...89

Figure 5.3 Scanned image of a gel showing results of RNase T1 footprinting

of the wild-type, quadrupole mutant, M3Q, and the NRAS

quadruplex forming sequence inserted in the 5'-UTR of the

hVEGF IRES-A………………………………………………………………92

Figure 5.4 Dual luciferase reporter assay results demonstrating the roles

x in translation of endogenous M3Q G-quadruplex, the mutated sequence and the QVEG G-quadruplex forming sequence all within the 5' UTR of MT3-MMP…………………………………………….94

List of Tables

Page

CHAPTER 2 An unusually stable G-quadruplex within 5'-UTR of the MT3 matrix metalloproteinase mRNA represses translation in eukaryotic cell

Table 2.1 Tm values for M3Q RNA in the presence of various cations………………………19

Table 2.2 Thermodynamic parameters for the folding of M3Q……………………………....20

CHAPTER 5 Effect of G-quadruplex domain exchange on translation

Table 5.1 Primer sequences used for various RNA G-quadruplex

swapping experiments…………………………………………………..………….86

xii

ACKNOWLEDGEMENTS

I would like to thank my advisor Dr, Soumitra Basu for all his patience and allowing me to have academic freedom when pursuing new ideas and projects. I truly wouldn’t change anything about the way I have been trained as a biochemist by him. Through his guidance I honestly feel that I have become a self-sufficient biochemist capable of formulating and executing ideas to the point of a successful completion. Without his help and encouragement I wouldn’t have been as successful in the field of RNA

G-quadruplexes as I am today. However, he still owes me lunch for not thinking that the RNA structure swapping project wouldn’t work. I would also like to thank all previous and current members of the Basu lab. I thank J.D., Eric, and Abby for putting up with my goofiness and not so comical jokes (but seriously

I am hilarious guys). These three coworkers have been also good friends that I have had the pleasure to get to know on a personal as well as professional level. I would like to thank Mohammed, Deb and

Gayan who all have been a pleasure to work with especially Mohammed for laughing at all my jokes when I make fun of Eric (see Eric he will tell you I’m hilarious). I would like to thank my former undergraduate students that I have had the pleasure to work with; Katy P., Katherine W., and Jag. I wish you all much success in whatever career you decide to pursue. I am sure that all of you will have much success. I would like to thank all the members of my committee; Dr. Gregory, Dr. Mao, Dr. Merrick and

X for taking time out of their busy schedules to sit on my committee. I would also like to thank previous teachers and professors that helped mold and shape me into a creative and hard worker. I thank Dr.

McEvoy for all her positive, influential and motivating comments that she would always make to me.

She always had faith in me and she was one of the main reasons that I decided to travel down this crazy career path. I thank Martha Asselin for her positive influence and help during my undergraduate degree.

I would also like to thank my highschool algebra teacher, Mr. Dillion. He was the first and only teacher that I ever had in highschool that had high expectations and hopes for me. I would also like to thank people in my personal life that helped me become the man I am today. First, my mother, her unconditional love. I love you so much and you taught me everything about life and I can’t thank you

xiii enough. My sister and brother, they are the strongest people in my life. They have helped me out all my life and I hope one day that I could even come close to repaying them with the love they have given me. I would also like to thank Stacy Grant for being an extremely positive motivator throughout my career and being a shoulder to lean on for all these years. You will always be my best friend. Lastly, I would like to thank Stephan, Dan, DRB, Fiester, Bade, Dave and Stu. These guys have also been there for me both professionally and personally. They are strong and kind-heated good people. I’ll miss all of you and I hope we keep in touch for all the years to come. I feel that so many people had a hand in my education and the development of which I have become as a man today; I would like to thank them all deeply.

xiv

CHAPTER ONE

INTRODUCTION AND BACKGROUND

1.1 DISCOVERY OF G-QUADRUPLEXES

Since the discovery of the structure of the DNA double helix by Watson and Crick

(1), DNA has largely been thought to exist as a right handed helix. While the discovery of the canonical DNA double helix structure comprising of Watson-Crick base pairing has provided the basis for our understanding of the genetic code, it is becoming increasingly evident that non-Watson-Crick interactions between bases and non-canonical nucleic acids structures have importance in biology. Over fifty years ago Gellert and coworkers

Figure 1.1 Structure and H-bond formation of a G-tetrad

collected X-ray fiber diffraction data that revealed the structural basis for the formation of regular hydrogen-bonded helices based on the assembly of tetrameric units, now known as guanine (G)-quartets (2).

Four guanines can form a square and co-planar tetrad in which each guanine serves as both a hydrogen bond donor and acceptor (Figure 1.1). The pairing of the N1 on the first guanine with the O6 on the second guanine along with the pairing of N2 on the first guanine with the N7 on the second guanine results in eight hydrogen bonds per G-tetrad. These G- tetrads are further stabilized by the presence of a monovalent cation in the center of the tetrad, which coordinates to the O6 carbonyl of the guanines. The cation is typically Na+ or

K+ and reduces repulsion of the oxygen atoms located in the center of the tetrad. The result is increased hydrogen bond strength and more stable stacking of tetrads (3). Two or more of these G-tetrads stack upon each other to form a G-quadruplex structure.

Figure 1.2 Structure of a G-quadruplex

These quadruplex structures can adopt uni-, bi-, or tetra-molecular forms consisting of one, two, or four strands respectively (3). Unimolecular G-quadruplexes result in intramolecular structures, but if the molecularity is greater than one an intermolecular structure will form. For an intramolecular G-quadruplex to form at least four stretches of guanosines containing a minimum of two G residues in tandem is required. Typical gap

Figure 1.3 General arrangements of typical G-quadruplexes based on the 5′ to 3′ orientation of the DNA strand. between two consecutive G-stretches can range from one to seven nucleotides (3). Because four G-stretches are needed for an intramolecular G-quadrupelx formation it will encompass three gaps, however the size of the gaps can be uniform or they can vary. These gaps turn into the loops when the sequence adopts a G-quadruplex structure. The number and identity of nucleotides of the gaps have been linked to the stability of the G-quadruplex structures (3).

The orientation of these strands determines whether the quadruplex is termed parallel or antiparallel. A quadruplex is termed parallel if the polarity of all the strands are oriented in the same direction with respect to one another (3). In contrast, if each strand has an opposite polarity with respect to the two adjacent strands, the quadruplex is termed anti-parallel

(Figure 1.3) (3).

Guanine-rich sequences of both RNA and DNA can form G-quadruplex structures via hydrogen bonding between Watson-Crick and major-groove base edges (3). DNA G-rich sequences have been reported to be prevalent throughout the entire human genome (4-8), especially in some of the key growth regulatory genes and oncogenes (9-13). Although,

DNA G-quadruplexes have been studied extensively, formation of RNA G-quadruplexes, particularly their regulatory roles in translational control are just beginning to emerge. There is no apparent physicochemical barrier towards RNA G-quadruplex formation (14, 15). It is logical to assume that the formation of RNA G-quadruplexes is more facile than formation of their DNA counterparts because RNA G-quadruplexes have been found to be more stable in the folded form (14, 16, 17) and do not have to compete with a complementary strand. In fact, it has been suggested that the 2'-OH of RNA may have stabilizing effects on RNA G- quadruplexes, possibly through hydrogen bonding interactions (16, 18). In the 1990s there were early reports of RNA G-quadruplex formation from a 19-nt oligonucleotide from E.

Coli 5S RNA (15). Based on NMR and molecular dynamics data, the UG4U sequence motif

was proposed to form a parallel stranded tetraplex structure containing four stacked G-tetrads and at least one U quartet (14). The NMR data showed that the UG4U aggregate is so stable

that it takes days for the imino protons of 3 of its 4 G tetrads to exchange with solvent at 40

°C. A computational survey found G-quadruplex forming sequences to be enriched within

mRNA processing sites (19) and in the 5'-UTRs of mRNAs of genes related to cancer (17).

G-quadruplex motifs have been characterized in several naturally occurring RNAs (17, 20-

27) and have been shown to have an inhibitory effect on translation (17, 23, 25, 27, 28) as

well as to be essential for translation.

1.2 G-QUADRUPLEX STRUCTURES IN THE 5'-UTR OF mRNAs

While DNA strands spend very little time in the single-stranded form, as they quickly

pair up with the complimentary strand, transcribed RNA molecules are relatively

unconstrained because of the lack of the complimentary strand and can fold intramolecularly to adopt a wide variety of structures. Common secondary structures include: base-pairs,

hairpins, bulges, and internal loops. For example, the secondary structure of transfer RNA

(tRNA) is organized into three hairpin structures that evolve from a multibranch loop. The

formation of secondary and tertiary structures within the 5' untranslated regions (UTRs) of

mRNAs have been shown to play important roles in the regulation of gene expression (29,

30). Historically, a majority of the reports have been based on RNA structures that depend

on canonical purine/pyrimidine base pairs. However, various other types of interactions can

exist between the nucleotides that constitute the fold of an RNA structure, including the one

that lead to formation of G-quadruplexes. The first report of a naturally occurring RNA G-

quadruplex located in the 5'-UTR of an mRNA was in 2007 when Balasubramanian and

coworkers discovered one such structure in the 5'-UTR of the proto-oncogene NRAS (17).

The authors discovered an 18 nucleotide (nt.) motif highly conserved both in position (near the 5'-cap) and sequence among various species (17). This sequence was found to fold into a very stable quadruplex sequence in vitro upon analysis via circular dichroism and UV- melting experiments. A dual luciferase reporter assay showed that there was a 3.7 fold

enhancement in activity of both a mutant and the deletion plasmid over the wild-type

indicating that this particular structure had an inhibitory effect on translation. On the basis of

this study, the same authors reasoned that analogous RNA G-quadruplexes may exist within

the 5′-UTR of other genes. To investigate this, bioinformatics studies were conducted in which they reported the presence of 2,922 putative RNA G-quadruplex motifs in the 5′-UTRs of human genes. This was the first study that reported the presence of a RNA G-quadruplex structure as being inhibitory toward translation in the context of its entire 5′-UTR in a cell- free assay. Later, the same group chose to investigate how the position and stability (in the context of the entire 5′-UTR) of this structure effect the translational repression of the gene.

They reported that the G-quadruplex forming sequence repressed translation when situated relatively close to the 5'-end of the mRNA (31). However, it had no significant effect on translation if located comparatively away from the 5'-end. Additionally, they demonstrated that the greater the thermodynamic stability of the RNA G-quadruplex at its native position the greater its repressive effect on translation. Following this, Wieland and Hartig published a study, which investigated the influence on quadruplex-forming sequences and the stability and how it relates to translational repression when place before the Shine-Dalgarno (SD) sequence (28). They reported that the amount of translational repression is both dependent on the number of stacks of the G-tetrads combined with the number of short loops between the G-stretches flanking the SD sequence. The shorter the loops between the G-stretches the more repressive the structure is toward translation. Additionally the higher the number of stacks of G-tetrads that are in the quadruplex structure then the more repressive the effect on translation.

In 2008, Maiti and coworkers reported the inhibition of translation in living eukaryotic cells by an RNA G-quadruplex motif. They reported that an evolutionary conserved G-quadruplex motif from the 5′-UTR of the mRNA from the zinc-finger protein of

the cerebellum 1 (Zic-1) drastically repressed translation without affecting the mRNA in

HeLa cells (23). The authors used similar biophysical methods to determine that the 27 base

RNA sequence forms a stable RNA G-quadruplex structure in physiological relevant ionic

conditions (100 mM K+). Using a dual luciferase construct, where the entire 5′-UTR of the zic1 mRNA was placed before a renilla luciferase gene, they reported that the RNA G- quadruplex structure repressed translation by 80%. This report was the first to show that

RNA G-quadruplexes not only forms in vitro but may also form intracellularly. Subsequent to these reports, additional studies have shown the presence of G-quadruplex forming sequences within the 5′-UTRs of mRNA of several human genes (17, 20-27). A similar strategy was used where the biophysical analysis of the G-rich sequence, mutagenesis and

reporter gene-based expression assays were utilized to confirm that theses sequences indeed

form RNA G-quadruplex structures in vitro and can modulate translation in eukaryotic cells.

The genes studied included the matrix metalloproteinase MT3-MMP (25), human vascular endothelial growth factor (hVEGF) (26), the estrogen receptor ESR1 (32), the apoptotic regulator BCL2 (33), and the telomere shelterin protein TRF2 (27).

1.3 MODULATION OF QUADRUPLEX STRUCTURE AND FUNCTION BY

SMALL MOLECULES

Small molecules that bind to structural elements within the 5′-UTRs of mRNAs have been explored with a view to interfering with gene expression at the translation level (34).

The special structural features of G-quadruplexes and the discovery of RNA G-quadruplex

forming sequences in the 5′-UTRs of numerous mRNAs, including several from proto-

oncogenes, led to the proposal that such RNA motifs could be suitable targets for small

molecules. Despite the growing number of RNA quadruplexes being discovered, reports on

their interactions with small molecules are few (35-38), although other types of RNA secondary structures have been extensively studied in this regard for the purpose of drug development (39).

The majority of the studies have focused on the small molecule interactions with

DNA G-quadruplexes. Several small molecule ligands have been reported to interact with

DNA quadruplexes, on which they exert either a stabilizing or a destabilizing effect (8, 40).

However, the few studies conducted on clearly established RNA G-quadruplexes reported

small molecule mediated stabilization of the quadruplex structure. Balasubramanian and

coworkers have shown the stabilization and repression of translation in a eukaryotic cell free

system by targeting an RNA quadruplex located in the 5'-UTR of NRAS with a pyridine-2,6-

bis-quinolino-dicarboxamide derivative (36). The compound reduced the translation

efficiency of the NRAS gene by ~ 50% at 1.25 μM concentration. They also tested variants

of this molecule and found that a para-fluorophenyl compound improved G-quadruplex

selectivity. Its inhibitory effect was also retained in the presence of large excess of double-

stranded DNA or hairpin RNA competitors, and compared favorably with studies based on in

vitro selected RNA aptamers (41). Neidle and coworkers have reported the interactions

between tetrasubstituted naphthalene diimides among other ligands, and the telomeric RNA

quadruplex forming sequence (37). Recently, Hartig and coworkers also reported

stabilization of designed RNA quadruplexes in the presence of bisquinolinium compounds

(38). Thus there are a few examples of small molecule RNA interactions, but interestingly in all of the cases the RNA quadruplex was stabilized in presence of the small molecule.

CHAPTER TWO

AN EXTREMELY STABLE G-QUADRUPLEX WITHIN 5'-UTR OF THE MT3

MATRIX METALLOPROTEINASE mRNA ACTS AS A TRANSLATION

REPRESSOR IN EUKARYOTIC CELLS

2.1 INTRODUCTION

Matrix metalloproteinases (MMPs) are zinc-dependent endopeptidases, which are capable of degrading extracellular matrix proteins (42, 43). In particular, it has been shown that upregulation of MT3-MMP mRNA and the increase in MMP protein expression is associated with the invasiveness of cancer cells (44), as in gastric cancer (45), prostate cancer

(46, 47), or renal carcinoma (48) and with pathological processes such as angiogenesis in tumors (49, 50). Despite the growing knowledge on the mechanism of MT3-MMP expression, its regulation at the translational level is not well delineated. The presence of a

20 nucleotide highly G-rich sequence (M3Q) comprised exclusively of purines and located upstream of the initiation codon (Figure 2.1) of MT3-MMP mRNA can be a potential regulator of its translation. Such G-rich sequences with multiple patches of contiguous

guanosines are known to adopt three dimensional folded forms (17, 22, 28).

Figure 2.1 Schematic representation of MT3-MMP mRNA. The G-tracts of the G-quadruplex forming sequence (M3Q) are shown in bold and the translation start site is shown in blue.

In this study, various biochemical and biophysical studies were performed to

elucidate M3Q’s ability to form a stable structure. Data obtained from CD spectroscopy,

temperature dependent melting, and enzymatic footprinting strongly suggest that this 20

nucleotide RNA sequence adopts a G-quadruplex structure with extremely high stability. A functional assay based upon measurement of dual luciferase activity performed in eukaryotic cells confirms the ability of the M3Q sequence in the context of the entire 282 nucleotide wild type 5'-UTR of MT3-MMP to repress the expression of the reporter gene at the

translational level.

2.2 MATERIALS AND METHODS

2.2.1 Preparation of oligonucleotide sequences. The RNA sequence 5'-

rGAGGGAGGGAGGGAGAGGGA-3' (M3Q) was purchased from Dharmacon, Inc. DNA

templates used for in vitro transcription (51) of the RNA sequence 5'-

rGAGAUAAGUGAGUGAGAGAGA-3' (mut-M3Q) were purchased from Integrated DNA

Technologies (IDT). RNA products were purified via denaturing 17% polyacrylamide gel

electrophoresis (PAGE). Full length products were visualized by UV shadowing and excised

from the gel. The RNA was harvested via the crush and soak method by tumbling the gel

slice overnight at 4 °C in a solution of 300 mM NaCl, 10 mM Tris-HCl and 0.1 mM EDTA

(pH 7.5). Salt was removed by ethanol precipitation of the RNA twice, with two 70%

ethanol washes in between each precipitation. The RNA pellet was dissolved in 10 mM Tris-

HCl and 0.1 mM EDTA (pH 7.5). RNA concentrations were determined based on their absorbance value at 260 nm and extinction coefficients calculated using nearest neighbor parameters (52).

2.2.2 5'-Labeling of RNA Oligonucleotides. The 5'-end phosphates of transcribed RNA were removed using Calf-intestinal alkaline phosphatase (New England BioLabs Inc.). CIP was removed by phenol-extraction and subsequent ethanol precipitation using 20 μg of glycogen.

The RNA was 5'-end labeled by treatment with T4 polynucleotide kinase (Promega), [γ-32P]

ATP (Perkin Elmer), and incubated for 45 min at 37 °C. The reaction was stopped by the

addition of an equal volume of stop buffer (7 M urea, 10 mM Tris-HCl and 0.1 mM EDTA,

pH 7.4). The radiolabeled full length RNA was isolated by 17% denaturing PAGE. The

RNA was extracted from the gel via the crush and soak method as described above.

2.2.3 Circular Dichroism (CD) studies. All measurements were recorded at room

temperature. RNA was folded by heating the samples at various concentrations of KCl in 10

mM Tris-HCl and 0.1 mM EDTA (pH 7.5) at 97 °C for 5 min, followed by slow cooling to

room temperature over a 90 min period. The circular dichroism (CD) spectra were recorded

using a Jasco J-810 spectrophotometer with a 0.1 cm cell at a scan speed of 50 nm/min with a

response time of 1 s. The spectra were averaged over five scans. For each sample, a buffer

baseline was obtained in the same cuvette and subtracted from the average scan.

2.2.4 CD-Melting experiments. CD-melt spectra were recorded using a 0.1 cm path-length

cell. Samples were prepared by heating oligonucleotides in 10 mM Tris-HCl and 0.1 mM

EDTA (pH 7.5) in the presence of various salt and oligonucleotide concentrations at 97 °C

for 5 min and then gradually cooled to room temperature at a gradient of 15 °C/hour.

Mineral oil was placed on top of the sample to prevent evaporation. The melting curves were

obtained by monitoring a 263 nm CD peak. Thermodynamic parameters and Tm values were

calculated using the van’t Hoff method (53, 54).

2.2.5 Footprinting by RNase T1. The 5'-end labeled RNA was folded by heating the samples in the presence of various concentrations of KCl in 10 mM Tris-HCl and 0.1 mM EDTA (pH

7.5) at 97 °C for 5 min and then slow cooled to room temperature. Once reactions attained

room temperature, the RNA was digested with 0.5 units of RNase T1 (Ambion) for 5 min at

room temperature. The reaction was terminated by using an equal volume of stop buffer as

described previously.

2.2.6 Plasmid construction. p-M3Q and p-mut-M3Q. p-M3Q and p-mut-M3Q were inserted

at the NheI site within the psiCHECK-2 vector (Promega) using the following

oligonucleotides: M3Q-sense (5'-CTAGCGAGGGAGGGAGGGAGAGGGAA-3'), M3Q- antisense (5'-CTAGTTCCCTCTCCCTCCCTCCCTCG-3'), mut-M3Q-sense (5'-

CTAGCGAAAAAGGGAGGGAGAGGGAA-3'), mut-M3Q-antisense (5'-

CTAGTTCCCTCTCCCTCCCTTTTTCG-3'). Ligated sequences were transformed into

JM109 cells and the plasmids were isolated. The insertion of the correct sequence was verified by sequencing performed at Ohio State University’s Plant-Microbe Genomics

Facility.

2.2.7 wt-UTR and mut-UTR. The entire 282 nucleotide sequence of MT3-MMP (wt-UTR) was obtained from GenScript Corporation, and the sequence was verified by the vendor.

NheI sites were incorporated at the flanks of the segment for subsequent cutting and ligation into the psiCHECK-2 vector. A mutant (mut-UTR) was also purchased with the following base substitutions: G214A, G215U, G218U, G222U, and G228A. Ligated sequences were transformed and the correctness of the sequence was verified as described above.

2.2.8 Cell Culture. HeLa cells were grown in 96-well plates in Dulbecco’s modified Eagles medium (DMEM) supplemented with 10% fetal bovine serum and the antibiotics streptomycin and penicillin at 37 °C in 5% CO2 in a humidified incubator.

2.2.9 Dual Luciferase and quantitative RT-PCR assays. HeLa cells were transfected with plasmids described above using Lipofectamine™ 2000 (Invitrogen) according to the manufacturer’s protocol. Twenty-four hours after transfection, renilla (RL) and firefly (FL) luciferase activities were measured using a Dual-Glo™ Luciferase assay system (Promega)

as per the manufacture’s supplied protocol on a Synergy 2 microplate reader (BioTek

Instruments). Total RNA was extracted from transfected HeLa cells using a NucleoSpin®

RNA II kit (Clontech). Prior to reverse transcription DNA was removed from each RNA

sample upon treatment with RQ1 RNase-Free DNase (Promega). Renilla and firefly mRNAs

were reverse transcribed using AMV-RT (New England BioLabs) and cDNA was subjected to quantitative real-time PCR using a SYBR® Green PCR Master Mix kit (Applied

Biosystems) on an ABI PRISM 7000 Sequence Detection System in presence of the appropriate set of primers. Forward primers RL 5'-(GTAACGCTGCCTCCAGCTAC)-3', FL

5'-(TTCGCTAAGAGCACCCTGAT)-3' and reverse primers RL 5'-

(GTAGGCAGCGAACTCCTCAG)-3' and FL 5'-(GCTGCAGCAGGATAGACTCC)-3'.

2.3 RESULTS

2.3.1 The M3Q sequence forms a parallel G-quadruplex structure

To determine whether the G-rich M3Q RNA forms a structure we used circular dichroism (CD) spectroscopy. A CD spectrum with a peak at 263 nm and a trough at 240 nm indicated that M3Q adopts a parallel G-quadruplex structure (3). Interestingly, M3Q generated a similar spectrum even in the absence of any added salt, suggesting an inherently strong propensity of the sequence to form a G-quadruplex, which have been previously reported (Figure 2.2) (17). However, addition of 1 mM KCl increased the intensity of the

CD spectrum at 263 nm, most likely due to enhanced G-quadruplex formation (55) and not because of any change in absorbance. It has been observed previously that the change in absorbance at 260 nm due to quadruplex formation is very small (~4%) (56, 57) and thus the

change in CD spectrum observed in Figure 2.2 can be ascribed to further stabilization of such

structure and possibly more G-quadruplex formation. It is well established that at least four

consecutive G-rich stretches are required for formation of an intramolecular G-quadruplex

Figure 2.2 Circular dichroism spectra of M3Q and mut-M3Q at a 5 μM strand concentration in 0.1 mM EDTA, 10 mM Tris-HCl, pH 7.5, and in the presence of various concentrations of KCl.

(3). Changes to any of the G stretches should interfere with M3Q’s ability to form an intramolecular G-quadruplex. A mutant sequence (mut-M3Q) in which base substitutions

G214A, G215U, G218U, G222U, and G228A was designed with the expectation that such

mutations would disrupt intramolecular G-quadruplex formation (23). Figure 2.2 shows the

CD spectrum of mut-M3Q in the presence of physiologically relevant K+ concentrations (100

mM KCl), and as expected there were no noticeable features that can be attributed to G- quadruplex formation. Furthermore, the observed decrease in the intensity of the CD

spectrum suggested a lack of helicity which suggests the presence of unstructured, single- stranded RNA (58).

2.3.2 A G-quadruplex of extreme stability.

To determine the stability of the structure formed by the M3Q RNA, CD-melting experiments were performed by monitoring changes in CD intensity at 263 nm. Studies in the presence of 100 mM KCl showed no sign of melting of the structure even at temperatures

+ above 95 °C, which forced us to measure the Tms at lower K concentrations. Figure 2.3 shows the CD-melting curve of M3Q in the presence and absence of salt. In the absence of

Figure 2.3 Circular dichroism melting curves of M3Q RNA at a 4 μM strand concentration in 0.1 mM EDTA, 10 mM Tris-HCl, pH 7.5, and in the presence of 0 and 1 mM KCl. added salt, the structure begins to unfold just above the ambient temperature resulting in a Tm

of 38±1 °C. However, addition of 1 mM KCl increased the Tm of the G-quadruplex by more

than 30 °C (72±1 °C). The Tm values determined by van't Hoff plots and by first derivative

analyses agreed with each other quite well (Figure 2.4).

Figure 2.4 Circular dichroism cooling curve and first derivative plot of M3Q RNA at a 4 μM strand concentration in 0.1 mM EDTA, 10 mM Tris-HCl, pH 7.5, and in the presence of 1 mM KCl.

The annealing and melting curves showed reversible characteristics, which suggests that the

molecules were at thermodynamic equilibrium (54). In the presence of various monovalent

cations, M3Q showed the expected trend for G-quadruplex melting with Tms following the

order K+ > Na+ > Li+ (Table 2.1) (59).

Table 2.1. Tm values for M3Q RNA in the presence of various cations.

*Cation (1 mM ) Tm (°C)

K+ 72

Na+ 45

Li+ 38

*For all experiments the buffer was 0.1 mM EDTA, 10 mM Tris-HCl, pH 7.5 in the presence of 1 mM of the indicated cation. Tm values are within ±1 °C.

The difference between Na+ and K+ is significant and most likely suggests a G-quadruplex

structure.

2.3.3 Molecularity and thermodynamic properties of the G-quadruplex.

To determine the molecularity of the formed G-quadruplex, melting experiments were performed at various strand concentrations but under identical salt concentrations. The

Tm remained unchanged as the M3Q concentration was varied from 1 to 40 µM (Figure 2.5),

which indicates intramolecular G-quadruplex formation (17, 57, 60). Thermodynamic parameters from the melting curves were calculated based on a two-state model (folded → unfolded) (53, 54). In the presence of 1 mM KCl, the Gibbs free energy (ΔG°) at 37 °C increased 20-fold compared to the value in the absence of added salt (Table 2.2).

Figure 2.5 Plot of Tm values for M3Q RNA at various strand concentrations. All experiments were conducted in 0.1 mM EDTA, 10 mM Tris-HCl, pH 7.5 and 1 mM KCl.

Table 2.2. Thermodynamic parameters for the folding of M3Q.

KCl ΔH° ΔS° ΔG°37°C [mM] (KJ/mol) (KJ/molK) (KJ/mol)

0 -163 ± 7 -0.52 ± 0.03 -1.2 ± 0.6

1 -213 ± 20 -0.61 ± 0.07 -21.5 ± 2.7

The calculated ΔH° and ΔS° values (Table 2.2) compare well with previously reported values

(61, 62).

Figure 2.6 RNase T1 footprinting of M3Q and mut-M3Q. Structures were pre-formed in 0 or 100 mM KCl, then incubated for 5 min. at room temperature in the presence of 0.5 mU of RNase T1. Lanes: base hydrolysis ladders for M3Q (lane 1) and mut- M3Q (lane 2). M3Q (lane 3) and mut-M3Q (lane 4) in the absence of KCl. M3Q (lane 5) and mut-M3Q (lane 6) in the presence of 100 mM KCl.

2.3.4 Footprinting on M3Q shows protection from RNase T1.

To further probe the formation of G-quadruplex structure, footprinting experiments on M3Q and mut-M3Q were conducted with ribonuclease T1 (RNase T1). RNase T1 catalyzes cleavage at guanosines in single-stranded RNA, but not guanosines that are present

in the context of secondary or tertiary structures. Thus, guanosines that directly participate in

quartet formation should be more protected from RNase T1 mediated cleavage compared to

guanosines that are in the loop or the unstructured regions, and indeed such patterns have

been reported before (22, 24, 63, 64). Results from the footprinting experiment on M3Q showed that in the presence of 100 mM KCl, all guanosines except G1, a dangling end, and

G15, which is apparently located within a single-stranded loop region were protected (Figure

2.6). In the absence of added salt, modifications were observed at all of the guanosines

within the M3Q sequence at a slightly higher level compared to that observed in the presence

of 100 mM KCl. This suggests that in the absence of added salt either a smaller fraction of

M3Q were folded into a G-quadruplex structure or the overall population formed a less rigid

structure. The observation is consistent with the CD spectrum in the absence of any added

salt. However, RNase T1 footprinting on mut-M3Q in presence of 100 mM KCl resulted in

no protection. All of the guanosine sites underwent cleavage under the experimental

conditions, presumably due to the lack of structure formation. The data suggest that the

M3Q RNA has a strong tendency to form an intramolecular G-quadruplex structure.

2.3.5 The M3Q motif inhibits translation of a reporter gene in eukaryotic cells.

To determine that the M3Q motif was responsible for translation inhibition in the

context of the full length wild type 5'-untranslated region of MT3-MMP, the entire 282

nucleotide 5'-UTR or a mutated version (each G-run of M3Q strategically mutated, see 2.2

Figure 2.7 Schematic representation of the plasmids used to investigate the effect of the 5'-UTR of MT3-MMP on translation.

for details) was inserted upstream of the renilla luciferase coding sequence in the

psiCHECK-2 vector (wt-UTR and mut-UTR) and the resultant vectors were transfected into

Figure 2.8 Histogram representing the ratio of Renilla/firefly luciferase activities in HeLa cells (black bars). The ratio of Renilla/firefly luciferase of the mut-UTR plasmid was set to one and all other values were normalized accordingly. The ratio of CT values for Renilla/firefly determined by qRT-PCR (white bars). All experiments were conducted in triplicate at a minimum.

HeLa cells. When present in the context of the whole 5'-UTR, the M3Q sequence

demonstrates inhibition of the reporter gene by 55% (Figure 2.7 and 2.8). However, the

mutated version of the 5'-UTR (mut-UTR) did not affect translation at all. The mRNA levels

of the reporter genes remained unchanged as confirmed by qRT-PCR experiments. The results demonstrate that the M3Q G-quadruplex motif has the ability to inhibit gene expression by repressing translation of its mRNA.

To ensure that the G-quadruplex located in the 5'-UTR of MT3-MMP has an effect on

translation, the M3Q or mutated M3Q sequence was also inserted upstream and in close

proximity of the Renilla luciferase coding sequence in the psiCHECK-2 vector (p-M3Q) and

transfected into HeLa cells (Figure 2.9 and 2.10). Identical mutations in the G-quadruplex

forming motif in NRAS mRNA has reported to disrupt G-quadruplex formation (17).

Quantitation of the dual luciferase in HeLa cells showed inhibition of Renilla luciferase activity by 60%, while the mutated M3Q showed full activity. To ensure the loss of activity

Figure 2.9 Schematic representation of the plasmids used to investigate the effect of the M3Q on translation.

is due to translation repression we performed qRT-PCR experiments that showed no change in mRNA levels of either of the reporter genes.

Figure 2.10 Histogram representing the ratio of Renilla/firefly luciferase activities in HeLa cells (black bars). The ratio of Renilla/firefly luciferase of the psiCHECK-2 plasmid was set to one and all other values were normalized accordingly. The ratio of CT values for Renilla/firefly determined by qRT-PCR (white bars). All experiments were conducted in triplicate at a minimum.

The ability of only the 20 nucleotide M3Q sequence and the M3Q sequence present in

the context of the wild type entire 5'-UTR to repress translation of a reporter gene by a similar amount, suggests that this G-quadruplex motif is responsible for repression of translation.

2.4 DISCUSSION

Although many naturally occurring sequences have been predicted to form RNA G- quadruplexes (17, 19), very few such structures and their functional relevance have actually been confirmed experimentally. Our data strongly suggest that a 20 nucleotide G-rich region in the 5'-UTR of MT3-MMP mRNA forms an extremely stable, intramolecular, parallel, G- quadruplex structure that can repress translation of a reporter mRNA. Formation of a parallel structure in an RNA G-quadruplex is expected, since an antiparallel CD spectrum would be

Figure 2.11 Circular dichroism first derivative plot of M3Q RNA at a 4 μM strand concentration in 0.1 mM EDTA, 10 mM Tris-HCl, pH 7.5, and in the presence of 5 mM KCl.

unlikely due to the more favored C3'-endo conformation of the ribose ring (65), although a combination of C2'-endo and C3'-endo have also been observed (66). CD and enzymatic footprinting show that selected mutations (mut-M3Q) disrupt the formation of a G-

quadruplex (Figure 2.2 and 2.6) (23). Such a disruption would be anticipated as it is well

known that the presence of four G-tracts in a single stand is an essential requirement for

intramolecular G-quadruplex formation (3).

The M3Q sequence demonstrated remarkable stability and showed an increase in the

Tm value by greater than 30 °C upon the addition of 1 mM KCl, which is rather unexpected

based upon the Tm values reported for biologically relevant RNA G-quadruplexes. In fact, at

5 mM KCl, the Tm increased by another ~10 °C (Figure 2.11), suggesting that at

physiological K+ concentration (~130 mM), unprecedented stability can be attributed to the

M3Q RNA G-quadruplex structure. Based upon the melting profile, thermodynamic parameters were calculated that provided an excellent basis for comparison of the stability of different structures. Reported values for biologically relevant RNA G-quadruplexes range

from -39 to -50 kcal/mol, -2.7 (at 37 °C) to -3.6 (at 25 °C) kcal/mol, and -100 to -130 cal/mol for ΔH°, ΔG°, and ΔS° respectively. These values calculated for M3Q in the presence of 1 mM KCl are in close agreement to those reported previously, but with a significant difference. The considerable difference that exists is between the conditions used to obtain the published values and conditions at which our experiments were performed, as our values were calculated from experiments performed at ≥100-fold lower K+ concentration (67-69).

The ability of M3Q RNA to attain a thermodynamic stability at par with the reported RNA

G-quadruplexes, despite being in 2-orders of magnitude less salt concentration, points

towards its extreme stability. Such unusual stability can be rationalized due to some rather

unique features present in the proposed G-quadruplex structure of M3Q RNA. To accommodate the single nucleotide loops at positions 6 and 10, the putative folded structure must be compact and may also be thought to be not feasible due to strain in conformation.

However, the presence of single nucleotide loops in intramolecular G-quadruplexes has been observed in numerous previous studies (17, 28, 61, 62, 70, 71). For example, in the C-MYC promoter DNA sequence, there is an inherent propensity for the selection of the G-tracts participating in the tetrad core to favor the formation of single nucleotide loops, which suggested that such loops are the most stable “double-chain-reversal loops” that bridge three tetrads (72). G-quadruplex stability is also linked to the actual number of short loops; it has been shown that increasing the number of such loops within a structure resulted in enhanced stability (73). The extreme stability of M3Q in comparison to other reported biologically relevant intramolecular RNA G-quadruplexes (17, 21, 22, 24) may be the result of an

additional single nucleotide loop present in the M3Q structure as compared to other reported

motifs. The adenine residues at positions 2 and 20 flanking the G-quadruplex core of M3Q

may also contribute to its stability, as it has been found that dangling ends containing

adenines stack particularly well to cap the G-quadruplex structures (70, 74, 75). For example, in a recent NMR study of the human telomere, it was found that adenine residues located in the loops and ends stacked and stabilized a G-quadruplex structure (75). Thus, the unique sequence of M3Q most likely contributes to its unprecedented stability.

Stable secondary RNA structures in the 5'-UTRs of mRNAs have been shown to regulate translation (17, 76-79). In fact, RNA G-quadruplexes containing short or one- nucleotide loops have been found to be extremely stable and can inhibit translation (17, 28,

60). For example, an RNA G-quadruplex containing a (G3U)n motif directly upstream of the

Shine-Dalgarno sequence showed a very strong inhibition of translation of a reporter gene in

bacteria (28). Furthermore, it has been shown in an in vitro assay that a G-quadruplex present in the human NRAS 5'-UTR, when inserted upstream of a reporter mRNA, reduced its translational efficiency (17). In eukaryotic cells, compelling evidence was found for the inhibitory role of a G-quadruplex forming sequence present within the Zic-1 5'-UTR (23). In our study, the unusually stable G-quadruplex structure adopted by the M3Q sequence, present in the 5'-UTR of the MT3-MMP mRNA, showed a strong inhibitory role in translation, which is consistent with reports described above. However, most of the previous reports did not show the functional role of a G-quadruplex forming sequence in the context of the native 5'-UTR in eukaryotic cells.

To investigate whether any secondary structure within the entire 5'-UTR can effectively compete with the formation of M3Q, we examined the sequence with QGRS

Mapper (19) and mFOLD (80). QGRS Mapper revealed no putative G-quadruplex motifs within 100 nucleotides upstream of M3Q and two extremely weak ones beyond that point.

Additionally, secondary structural analysis of the 40 flanking nucleotides on either side of

M3Q using mFOLD resulted in a weak secondary structure with a predicted ΔG that is substantially higher than that calculated for M3Q folding (Table 2.2). From the investigation above it seems highly unlikely that any G-quadruplex or other secondary structure could compete with M3Q G-quadruplex formation. Thus, the translational repression most likely can be attributed to the M3Q structure.

Since, computational studies have predicted the presence of thousands of putative G- quadruplex forming motifs in the 5'-UTR region of mRNAs (17), it is not far-fetched to expect structural and functional diversity in such motifs. To better understand the

significance of these studies it is imperative that such motifs are investigated to define their

role in the regulation of translation. This work adds to our nascent knowledge on the role of

G-quadruplex forming RNA motifs in translation control. In light of the present findings,

particularly the demonstrated inhibitory role of the M3Q G-quadruplex, it may be postulated

that yet to be identified cellular factors would have to unwind this very stable structure, thus

facilitating translation of mRNAs with G-quadruplex motifs located in their 5'-UTR. Such a

proposition may be more significant in case of certain types of cancers in which the MT3-

MMP expression is upregulated.

2.5 CONCLUSION

Herein, we have shown that M3Q forms an unusually stable intramolecular G- quadruplex that inhibits translation in eukaryotic cells. To our knowledge, no intramolecular

RNA G-quadruplex structure has been shown to be as stable as the M3Q structure at such low salt concentrations. The unique purine-only sequence of M3Q is likely stabilized via short nucleotide loops, the potential stacking of the flanking adenines, and the lack of competing intramolecular Watson-Crick base pairing. In general, this study supports the notion that RNA G-quadruplex motifs in the 5'-UTR can control translation and in particular begins to provide a molecular basis for translational regulation of MT3-MMP.

CHAPTER 3

THE PORPHYRIN TmPyP4 UNFOLDS THE EXTREMELY STABLE G-

QUADRUPLEX IN MT3-MMP mRNA AND ALLEVIATES ITS REPRESSIVE

EFFECT TO ENHANCE TRANSLATION IN EUKARYOTIC CELLS

3.1 INTRODUCTION

Matrix metalloproteinases (MMPs) are zinc-dependent endopeptidases, which are

capable of degrading extracellular matrix proteins (42, 43). We reported that a 20 nucleotide

all purine sequence (M3Q) located in the 5′-UTR of the MT3-MMP mRNA forms an

extremely stable intramolecular G-quadruplex that inhibits translation in eukaryotic cells

(Figures 2.3 and 2.8) (25). G-quadruplex motifs have been characterized in several naturally

occurring RNAs (17, 20-27) and have been shown to have an inhibitory effect on translation

(17, 23, 25, 27, 28). In particular, it has been shown that an RNA quadruplex motif (M3Q)

located in the 5'-UTR of the MT3-MMP mRNA forms an extremely stable G-quadruplex

structure and inhibits translation in eukaryotic cells when analyzed alone as well as in the

context of the entire 5'-UTR (Figure 2.8) (25). Despite the growing number of RNA quadruplexes being discovered, reports on their interactions with small molecules are few

(35-38) even though other types of RNA secondary structures have been extensively studied in this regard for the purpose of drug development (39). Several small-molecule ligands have been reported to interact with DNA quadruplexes, on which they exert either a stabilizing or a destabilizing effect (8, 40). Many of these studies employed the cationic

porphyrin, 5,10,15,20-tetra (N-methyl-4-pyridyl) porphyrin (TmPyP4), which was usually found to stabilize DNA quadruplexes in vitro as well as in vivo (9, 12, 81-84). Besides the quadruplex structures, TmPyP4 can also bind to several duplex DNA sequences with comparable affinities (85, 86). It was also reported that this ligand can unfold a bimolecular

DNA quadruplex (87) but its effect on the stability of RNA quadruplexes is poorly understood at this time. In this chapter, CD spectrophotometry, 1D 1H NMR spectroscopy and gel electrophoresis convincingly demonstrate that TmPyP4 unfolds the extremely stable,

20 nucleotide RNA G-quadruplex forming sequence located in the 5'-UTR of the MT3-MMP mRNA. We also report the in vivo effect of TmPyP4 on translation of the MT3-MMP

mRNA.

3.2 MATERIALS AND METHODS

3.3.1 RNA Purification. The RNA sequences 5'-rGAGGGAGGGAGGGAGAGGGA-3'

(M3Q) and 5'-rGAGAUAGUGAGUGAGAGAGA-3' (mut-M3Q) were purchased from

Dharmacon, Inc. RNA products were purified via denaturing 17% polyacrylamide gel

electrophoresis (PAGE). Full-length products were visualized by UV shadowing and excised

from the gel. The RNA was harvested via the crush and soak method by tumbling the

crushed gel slices overnight at 4 °C in a solution of 300 mM NaCl, 10 mM Tris-HCl and 0.1

mM EDTA (pH 7.5). The RNA was isolated by ethanol precipitation followed by two 70%

ethanol washes of the precipitate. The final RNA pellet was dissolved in 10 mM Tris-HCl

and 0.1 mM EDTA (pH 7.5). RNA concentrations were determined on the basis of their

absorbance value at 260 nm and appropriate extinction coefficients (52). The RNAs were

folded in the presence of 100 mM KCl, 10 mM Tris-HCl and 0.1 mM EDTA (pH 7.5) by heating for 5 min at 95 °C followed by cooling to room temperature over a 90 min period.

5'-Labeling of RNA Oligonucleotides. The RNA was 5'-end-labeled by treating with T4 polynucleotide kinase (Promega) and [γ-32P] ATP (Perkin Elmer). The reaction was stopped by the addition of an equal volume of stop buffer [7 M urea, 10 mM Tris-HCl and 0.1 mM

EDTA (pH 7.5)]. The radiolabeled full-length RNA was isolated by 17% denaturing PAGE.

The RNA was extracted from the gel via the crush and soak method as described above.

3.2.2 Native gel electrophoresis. 5'-end-labeled M3Q RNA was folded as described above.

After cooling, samples were incubated with various concentrations of TmPyP4 (Calbiochem) for 15 min at room temperature in the dark. The samples were then loaded on a native 10% gel and electrophoresed at 4 °C, dried on Whatman paper and exposed to a phosphorimager screen. The gel images were visualized by scanning the screen on a Typhoon

Phosphorimager 8600 (Molecular Dynamics).

3.2.3 Circular Dichroism (CD). All measurements were recorded at room temperature using folded RNA samples (4 µM). Spectra were recorded using a Jasco J-810 spectrophotometer with a 0.1 cm cell at a scan speed of 50 nm/min with a response time of 1 s. The spectra were averaged over three scans and the spectrum from a blank sample containing only buffer was subtracted from the averaged data. After the collection of the initial spectrum, 1 µL aliquots of TmPyP4 in the same buffer as the RNA were added to the sample in the cuvette and was mixed several times before obtaining a new spectrum.

3.2.4 UV-Visible Spectroscopic Analysis. A concentrated sample of either M3Q or mut-

M3Q folded in the presence of 100 mM KCl 10 mM Tris-HCl and 0.1 mM EDTA (pH 7.5)

was titrated into a 4 µM sample of TmPyP4 (100 µL). The absorption spectra at different

sample addition points were collected with a Cary 300 scan UV-vis spectrophotometer

(Varian) in the wavelength range from 350 to 500 nm using a 1 cm path length microcuvette.

The concentration of free TmPyP4 was determined using an extinction coefficient of 2.26 x

105 M-1 cm-1 and absorbance values at 424 nm (88). The titration was stopped when three successive additions of the M3Q RNA resulted in no further shift of the Soret band. All values were corrected for dilution effect. The fraction (α) of TmPyP4 bound was determined as follows:

α = (Abs TmPyP4free − Absmixture) / (Abs TmPyP4free − Absbound), where Abs TmPyP4free is

the absorbance of the free TmPyP4 in the absence of any added RNA, Absmixture is

absorbance at any point after the beginning of the addition of the RNA, and Absbound is the

absorbance of fully bound TmPyP4 measured at 424 nm (the Soret maxima for TmPyP4).

Concentrations of free TmPyP4 ([TmPyP4]free) and the concentration of bound TmPyP4

([TmPyP4]bound) were calculated as follows:

[TmPyP4]free = [TmPyP4]corrected(1 − α)

[TmPyP4]bound = ([TmPyP4]corrected − [TmPyP4]free), where [TmPyP4]corrected represents the

concentration of TmPyP4 corrected for the change in volume that occurred due to the titrated

RNA. The percent hypochromicity of the Soret band is calculated as follows:

% hypochromicity = [(εfree − εbound) / εfree] × 100, where

εbound = Abs TmPyP4bound / [TmPyP4]bound (89)

3.2.5 NMR. 1D 1H NMR spectra of M3Q were recorded at 37 °C using a 750 MHz Varian

spectrometer. A 300 µL sample of 430 µM M3Q was folded in 20 mM potassium phosphate

buffer, pH 7.0 and 8% D2O as described above. Aliquots of 55 mM TmPyP4 were added in

0.2 molar equivalents and allowed to equilibrate for approximately 1 minute between

acquisitions. Imino proton NMR spectra were collected at 37 °C utilizing the Watergate 3-9-

19 sequence (90) with sweep width of 20 ppm. Recycle delays of two seconds were

employed between each of 128 scans. NMR spectra were processed and superimposed with

iNMR (http://www.inmr.net) using the following parameters: a 40 Hz solvent filter was

applied to clean up the residual water signal, the fid was apodized with a cosine2 function, linear predicted from 2k to 3k points, and zero filled to 4k for Fourier transformation and automatic baseline correction in the frequency domain. During spectral comparison and plotting, the same scale was maintained for each spectrum to permit accurate comparisons of signal intensity.

3.2.6 Plasmid Construction. Plasmids p-M3Q and wt-UTR were constructed as previously described (25). The plasmid del-M3Q was contructed by using a QuikChange site-directed mutagenesis kit (Stratagene) where nucleotides 213-229 were deleted from the wt-UTR plasmid.

3.2.7 Cell Culture. HeLa cells were grown in 96-well plates in Dulbecco’s modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum and the antibiotics streptomycin and penicillin at 37 °C in 5% CO2 in a humidified incubator.

3.2.8 Dual Luciferase Assays. HeLa cells were transfected with plasmids described above using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s protocol. Following incubation at 37 °C for 2 h, each plate was supplemented with a final concentration of 0, 50 or 100 µM TmPyP4. Twenty-four hours after transfection, Renilla (RL) and firefly (FL) luciferase activities were measured using a Dual-Glo Luciferase assay system (Promega) as per the manufacturer’s supplied protocol using a Synergy 2 microplate reader (BioTek

Instruments). The ratio of Renilla to firefly luciferase activities was calculated for each plasmid. In each case, the value of the ratio obtained in the absence of TmPyP4 was normalized to 100%.

3.3 RESULTS & DISCUSSION

3.3.1 TmPyP4 unfolds an RNA G-quadruplex (M3Q) structure and inhibits its formation.

It has been previously shown that a G-quadruplex located in the 5'-UTR of MT3-

MMP (M3Q) mRNA forms an extremely stable quadruplex structure and inhibits translation of a reporter gene in eukaryotic cells (25). The physiological processes related to upregulation of MT3-MMP make it a target to find small molecules that could selectively interact with the M3Q motif and affect translation. To determine the effect of TmPyP4 on the folded M3Q G-quadruplex several biochemical and biophysical techniques were utilized.

As shown in Figure 3.1, the CD spectrum of folded M3Q in 100 mM KCl shows a peak at

263 nm and a trough at 240 nm, which is characteristic of a parallel RNA quadruplex (3, 25).

Figure 3.1 CD spectra of 4 µM prefolded (in 100 mM KCl) M3Q in the absence and presence of increasing concentrations of TmPyP4. The chemical structure of TmPyP4 is shown in the inset. The arrow defines the decrease in CD signal as a function of increasing TmPyP4 concentration.

Titrating increasing concentrations of TmPyP4 resulted in a decrease in the CD signal at 263

nm, reflecting a disappearance of the quadruplex structure presumably due to its unfolding.

According to previously published results, at this concentration of K+ M3Q was unable to be

unfolded even at a temperature as high as 95 ºC (25). Despite the extreme stability of this

quadruplex structure under these conditions it was remarkable that TmPyP4 was able to

unfold M3Q. The possibility that the reduction in the CD signal could result from non-

specific associations of the added ligand to the RNA and not due to a destabilization of the

quaduplex was considered. To ensure that this was not the case and to distinguish between

the two possibilities, CD spectra were taken at different wavelengths to check if there was a

negative induced circular dichoism (ICD), but there was no evidence of such phenomenon

indicating that the reduction at 263 nm was a direct effect of the quadruplex being

destabilized (Figure 3.2) (91, 92).

Figure 3.2 CD spectra of TmPyP4 in the absence and presence of folded M3Q. The initial concentration of TmPyP4 was 4 µM to which a final concentration of 20 µM (5 equivalents) of folded M3Q RNA was added. The experiment was performed at 100 mM KCl, 0.1 mM EDTA and 10 mM Tris-HCl (pH. 7.5).

The results from Figure 3.1 at different concentrations of TmPyP4 were normalized and

graphed against CD signal by assuming that the signal at 263 nm for M3Q in the absence of

TmPyP4 corresponds to 100% folded while that at 100 µM TmPyP4 represented 0% folded.

As shown in Figure 3.3, the data fit well with an exponential function which permitted

interpolation for a value of TmPyP4 corresponding to that for 50% unfolding (11 μM).

Figure 3.3 Plot of calculated fraction folded vs. TmPyP4 concentration.

Additionally, the presence of 20 µM TmPyP4 prevents M3Q quadruplex formation when

present during the folding of the G-quadruplex, as indicated by the complete lack of any

structural features in the CD spectrum (Figure 3.4) (91, 92).

Figure 3.4 CD spectra of 4 µM M3Q RNA (in 100 mM KCl) folded in the absence and presence of 0, 2, 5, 10, and 20 µM TmPyP4.

A possible explanation for the lack of spectral features of a quadruplex structure when

TmPyP4 is added before folding of the M3Q motif might be that TmPyP4 binds to the

unfolded form of M3Q RNA (93) thus preventing the formation of the quadruplex structure.

To determine whether TmPyP4 can interact and influence the folding and subsequent

migration patterns of M3Q RNA, native gel electrophoresis was performed. As shown in the

first two lanes of Figure 3.5A, M3Q has a greater mobility than a mutated version of this

sequence (mut-M3Q) which has been previously shown not to form any structure (25).

Figure 3.5 Native gel shift assay of M3Q (final concentration of about 250 nM) in the absence and presence of increasing concentrations of TmPyP4. M denotes mut-M3Q. Concentrations used are 0, 1, 5, 10, 20 and 50 µM TmPyP4. B) Native gel shift assay of mut-M3Q in the absence and presence of increasing concentrations of TmPyP4. M denotes the single stranded RNA marker mut-M3Q. Concentrations of TmPyP4 are the same as in A.

Since the two oligonucleotides have the same charge, the greater mobility of M3Q can be

ascribed to formation of a quadruplex structure (94). Therefore, if TmPyP4 unfolds the M3Q quadruplex, then a species with lower mobility should be observed. At total concentrations of TmPyP4 from 5 to 50 μM, an RNA species emerges that has a lower mobility than mut-

M3Q which may represent a complex of TmPyP4 with unfolded M3Q. It has been previously shown by the Fry group that the mobility of an RNA sequence that presumably adopts a quadruplex structure can be retarded by TmPyP4 (94). The mut-M3Q sequence

(Figure 3.5B) shows no change in mobility in the presence of up to 50 µM of TmPyP4 indicating the specificity of the interaction between TmPyP4 and the M3Q sequence. The

lower mobility of the M3Q band resulting in the unfolding of the sequence can be

corroborated with CD experiments (Figure 3.1) which indicate that the M3Q quadruplex is almost completely unfolded in the presence of 50 μM TmPyP4. In addition, the binding of

TmPyP4 with unfolded M3Q was evidenced by spectral shifts of its Soret absorption band

(see below, Figure 3.7). Interestingly, at concentrations of the ligand from 1 to 10 μM, a band with intermediate mobility was observed, which may reflect the existence of a transient, intermediate complex of TmPyP4 with partially unfolded M3Q. While this may be an intermediate structure, its exact identity cannot be established at this point. It should also be noted that the different levels of retardation in mobility of the M3Q band observed may also be due to binding of the cationic TmPyP4 molecule resulting in a diminished net negative charge on the RNA which would cause its lower mobility. Considering the prevailing dynamic, non-equilibrium experimental condition where RNA anions and TmPyP4 cations are being separated by electrophoresis, the cause of different lower mobilities cannot be

determined unambiguously. Overall, the CD and native gel electrophoresis data suggest a

change in nature of the M3Q quadruplex structure in the presence of TmPyP4 presumably

due to unfolding of M3Q.

3.3.2 Proton NMR reveals unfolding of the M3Q RNA quadruplex.

The disruption of the folded M3Q quadruplex structure in the presence of TmPyP4 was further established by 1H NMR. It has been shown that the imino protons from the

Hoogsteen hydrogen bonding between the guanines of the tetrads in the quadruplex show distinct peaks from 10-12 ppm (3). Figure 3.6 shows the imino peaks from the tetrads in the

M3Q quadruplex structure that are visible even at 37 °C which still resulted in sharp imino

peaks attesting to the extreme stability of the quadruplex structure; most RNA imino protons

from bases engaged in Watson-Crick pairs, exchange broaden at about room temperature.

Thus, the NMR data suggest the presence of a thermodynamically hyper-stable quadruplex structure at a biologically relevant temperature.

Figure 3.6 NMR spectra of 0.42 mM M3Q titrated with TmPyP4. 1D 3-9-19 WATERGATE spectra were collected at 37 °C. PAGE purified and dialyzed M3Q was prefolded in NMR buffer (20 mM potassium phosphate, pH 7.0 and 8% D2O) and titrated with 0.2 molar equivalent increments of 55 mM TmPyP4. Each experiment was collected for 5 minutes with 128 scans and 2-second recycle delays. The left portion of the spectrum shows the imino region, while the right shows a portion of the anomeric, i.e. H1' region. data collected by Dr. Tom Leeper, University of Akron

TmPyP4 was titrated into the preformed M3Q RNA G-quadruplex in 0.2 equivalent portions and the spectra were taken immediately. Titration of this RNA with the TmPyP4 molecule resulted in accelerated imino exchange (Figure 3.6) consistent with G-quadruplex structure melting. As the concentration of TmPyP4 is increased there is a reduction in intensity of the imino peaks until eventually there is a complete lack of signal while new signals appear

upfield of the main cluster of iminos at substoichiometric levels. The new iminos’ appearance and disappearance of other imino signals is consistent with a ligand dissociation that is slower than the NMR time-scale. Such signal behavior is often seen with modest to higher affinity complexes with Kd’s typically less than 20 µM (95, 96). These new peaks probably represent ring-current shifted iminos resulting from bound but not fully unfolded molecules (97). Because these resonances are upfield shifted, we would hypothesize that this ligand bound and partially unfolded state represents stacking of the TmPyP4 upon the G- tetrad (98, 99), although additional structure probing would be required to validate this notion. Several of the resonances in the anomeric proton region are present even at the highest concentration of TmPyP4. No visible precipitate was present at this highest concentration; this and the persistent signal from non-exchangeable protons suggest that the disappearance of imino resonances is not due to non-specific aggregation.

3.3.3 UV-Vis Spectroscopy shows binding of M3Q to TmPyP4.

The binding of M3Q and TmPyP4 was also monitored via UV-Vis spectroscopy.

Folded M3Q was titrated into a solution of TmPyP4 and the Soret band was monitored as a function of M3Q concentration. As the concentration of M3Q was increased there was substantial hypochromicity (maximum 75 %) as well as a rather unusually large bathochromic shift of 22 nm (from λmax 424 nm to 446 nm), which is indicative of binding of

TmPyP4 to the M3Q RNA G-quadruplex (Figure 3.7) (100).

Figure 3.7 Visible absorption spectra of TmPyP4 in the absence and presence of increasing concentration of prefolded M3Q. The initial concentration of TmPyP4 was 4 µM (100 µL) to which 0.21 nmoles of the prefolded M3Q RNA quadruplex was added in 0.5 µL increments. The experiment was performed at 100 mM KCl, 0.1 mM EDTA and 10 mM Tris-HCl (pH. 7.5). The arrow pointing downward indicates the decrease in the λmax 424 nm of TmPyP4 with the addition of M3Q and the arrow pointing upward indicates the shifted Soret band.

A sharp isosbestic point was observed at 436 nm. When mut-M3Q was titrated, no shift or hypochromicity was apparent even after addition of up to five equivalents of TmPyP4, which

suggests specificity of TmPyP4 interaction with the M3Q RNA (Figure 3.8).

Figure 3.8 Visible absorption spectra of TmPyP4 in the absence and presence of mut-M3Q. The initial concentration of TmPyP4 was 4 µM to which 20 µM of the mut-M3Q RNA was added. The experiment was performed at 100 mM KCl, 0.1 mM EDTA and 10 mM Tris-HCl (pH. 7.5).

These results correlate well with the native gel electrophoresis data on the specificity of the

interaction between M3Q and TmPyP4.

Hypochromicity and a bathochromic shift of the Soret band reported previously were

observed in cases of stabilization of the G-quadruplex structure (100, 101). An important

aspect in case of destabilization is the fate of the Soret band after presumed disruption of the

M3Q G-quadruplex structure. We observed that the Soret band shift remained unchanged

even after destabilization of the M3Q quadruplex. Porphyrins are known to bind to single-

stranded nucleic acids resulting in shift of the Soret band in the same amount and direction as

the change when interacting with a nucleic acid secondary structure (93, 102, 103), which

may explain the lack of change in shift of the Soret band even after the quadruplex structure

was destabilized. Presumably the ability of TmPyP4 to prevent the M3Q quadruplex formation is due to the fact that it can bind to single stranded RNA.

The mode of interaction between TmPyP4 and RNA quadruplexes is unknown, while the mechanism of interaction between DNA G-quadruplexes and the TmPyP4 is a point of debate in the literature (35, 103, 104). We observed a large bathochromic shift of the Soret band accompanied by a high % hypochromicity by UV spectroscopy when the pre-folded

M3Q RNA was titrated into a TmPyP4 solution. The large shift in the Soret band resulted from the transfer of the π-electrons from the purine bases to the pyrrole rings (105), suggesting stacking on the G-quadruplex or perhaps intercalation between the tetrads, both of which have been proposed in the literature (35, 103, 104).

It is important to note that the mechanisms of binding are for reported stabilization of the G-quadruplex structures due to TmPyP4 binding (35, 103, 104). However, here we report unfolding of an usually stable RNA quadruplex. Although the mode of recognition of this stable RNA quadruplex by TmPyP4 is unclear at this point, the unfolding detected by the various biophysical methods described above indicates that the interaction between TmPyP4 and unfolded and non-quadruplex form of M3Q is thermodynamically favorable. Previous reports have shown that TmPyP4 can bind to non-quadruplex nucleic acid forms. For example, it has been reported that TmPyP4 binds to various double stranded DNA sequences with affinities that are similar to quadruplex DNA (85, 86), which raises the question on the selectivity of its binding. Additionally, the TmPyP4 can bind to single stranded DNA with

preference for sequences with stretches of guanines; however the particular mechanism of such interaction is still unknown.

3.3.4 TmPyP4 enhances translation activity of a reporter gene in live eukaryotic cells.

The possibility that TmPyP4 could unfold the M3Q motif in eukaryotic cells and affect translation efficiency was addressed by using a dual luciferase reporter construct in

Figure 3.9 A) Schematic of dual luciferase bi-cistronic constructs. B) Histogram showing % activity of the translation of the Renilla luciferase as a function of TmPyP4 concentration. which the M3Q sequence was placed just before a Renilla luciferase gene, while the upstream firefly luciferase was under the control of a HSV-TK promoter (p-M3Q, Figure

3.9A).

It has been shown previously that the quadruplex has an inhibitory effect on the translation of

the Renilla mRNA in this particular plasmid (25). This plasmid was transfected in HeLa

cells and incubated with 0, 50 or 100 µM TmPyP4 respectively for 24 hours. The cells

where inspected prior to the luciferase assay and no detectable change in the cell number or

morphology in any of the treatment groups was observed. As can be seen in Figure 3.9B, 50

µM TmPyP4 had a moderate effect on translation with an increase of 15 ± 4%. However,

when the concentration of TmPyP4 was increased to 100 µM it had an increase of 35 ± 2% in

translation. In order to increase translation of renilla mRNA, the TmPyP4 would have to relieve the repressive effect of the quadruplex, presumably by destabilizing the quadruplex structure. It was also of interest to compare these results with those found from having the

M3Q sequence in the context of the entire 282 nucleotide 5'-UTR of MT3-MMP mRNA (wt-

UTR). A similar increase in activity would suggest that TmPyP4 is binding to the quadruplex embedded within the entire 5'-UTR and is modulating its activity. The entire 5'-

UTR was placed in front of the renilla gene and again assayed after the cells where incubated in the presence of 0, 50 or 100 µM TmPyP4 respectively for 24 hours. As shown in Figure

3.9B the data for the two constructs correlate well with wt-UTR increasing activity 22 ± 4% in the presence of 50 µM of TmPyP4 and 37 ± 5% when the cells were treated with 100 µM

TmPyP4. While evaluating the enhancement of translation by TmPyP4 it should be taken into account that the translation was repressed by the G-quadruplex alone (no ligand) by

55%. We performed qRT-PCR experiments that showed no change in mRNA levels (Figure

3.10) of the wt-UTR construct at 0 and 100 µM concentration of TmPyP4 which shows that

the gain in activity was at the translational level.

Figure 3.10 Histogram representing the ratio of Renilla to firefly mRNA CT values in HeLa cells determined by qRT-PCR. The ratio of Renilla to firefly luciferase of the wt-UTR plasmid with no TmPyP4 was set to one, and all other values were normalized accordingly

We then tested the effect of TmPyP4 on a mutant of wt-UTR in which the M3Q sequence

was deleted (del-UTR). The strategy of using the deletion mutant as a control was based

upon two previous independent reports (36, 94) where the quadruplex forming segment

deleted to test the functional consequence of small molecule G-quadruplex interactions.

Figure 3.9B shows that at concentrations up to 100 µM there is no increase in renilla

luciferase activity. The lack of any change in the level of translation of the deletion mutant

can be explained by presuming that TmPyP4 is specifically disrupting the M3Q quadruplex

in vivo to significantly mitigate its repressive effect on translation.

3.3.5 The destabilization of the M3Q sequence by TmPyP4 is unprecedented.

Until recently, the majority of reports have focused on the small molecule interactions

with DNA G-quadruplexes, however, the interactions between small molecules and RNA G- quadruplexes are barely understood. The studies on clearly established RNA G-quadruplexes

interacting with small molecules reported stabilization of the quadruplex structure.

Balasubramanian and coworkers have shown the stabilization and repression of translation in

a eukaryotic cell-free system by targeting an RNA quadruplex located in the 5'-UTR of

NRAS with a pyridine-2,6-bis-quinolino-dicarboxamide derivative (36). Neidle and

coworkers have reported the interactions between tetrasubstituted naphthalene diimides

among other ligands, and the telomeric RNA quadruplex forming sequence (37). Recently,

Hartig and coworkers also reported stabilization of designed RNA quadruplexes in the

presence of bisquinolinium compounds (38).

In the current report the interaction between the previously well-characterized quadruplex forming sequence, M3Q, and the known DNA stabilizing cationic porphyrin

TmPyP4 is described. A variety of biochemical and biophysical data strongly suggest the destabilization of the very stable M3Q RNA G-quadruplex structure by TmPyP4, unlike most of the previous studies that reported stabilization of DNA quadruplexes in the presence of this small-molecule (35). Interestingly, when we analyzed a DNA version of the M3Q sequence in the presence of TmPyP4 the quadruplex (Tm = 71 ± 1 °C at 100 mM KCl, Figure

3.11) was also destabilized.

Figure 3.11 CD-melting spectrum of the DNA version of M3Q (4 µM) in the presence of 100 mM KCl, 0.1 mM EDTA and 10 mM Tris-HCl (pH. 7.5).

However, it required a higher concentration to unfold the DNA quadruplex compared to the

M3Q RNA version (Figure 3.12) although the RNA quadruplex could not be melted at 100 mM KCl (Tm = 72 ± 1 at 1 mM KCl,). Thus, the DNA version of M3Q is unfolded by

TmPyP4 but not as efficiently as it unfolds the RNA version of M3Q.

Destabilization of G-quadruplex structures by small-molecules have been shown in

only a few reports and they mostly involve DNA quadruplexes (92, 94). The

Balasubramanian group has described a triarylpyridine molecule that induced unfolding of

the quadruplex formed by the DNA sequence found within the c-kit promoter, a structure that

based upon the reported Tm is significantly less stable than M3Q (92, 106).

Figure 3.12 Plot of fraction folded vs. TmPyP4 concentration of 4 µM prefolded (in 100 mM KCl) DNA version of M3Q in the absence and presence of increasing concentrations of TmPyP4.

Destabilization of a d(CGG) repeat in presence of TmPyP4 was reported by the Fry group.(87) Additionally, Fry and coworkers have shown that the interactions of an RNA quadruplex forming motif r(CGG)33 were destabilized by TmPyP4, however, according to the authors the nature of the secondary structure that this sequence adopts is still rather unclear (94).

3.4 CONCLUSION

The functional consequence of the M3Q G-quadruplex destabilization would result in increased translation of the reporter mRNA. In vitro biochemical and biophysical data

presented in this report indicated destabilization of the M3Q G-quadruplex which may be

valuable for other G-quadruplexes that regulate translation. The findings of this report can

be useful in cases where up regulation of a genes may have a therapeutic benefit.

CHAPTER 4

AN RNA G-QUADRUPLEX IS ESSENTIAL FOR CAP-INDEPENDENT

TRANSLATION INITIATION IN HUMAN VEGF IRES

4.1 INTRODUCTION

Structures present within the 5'-UTR of mRNAs, including cellular mRNAs, have been shown to regulate translation.(76) Many mRNAs have been reported to initiate translation in a 5' cap-independent manner (107-111). Such cap-independent translation is initiated by internal ribosomal entry sites (IRESs) found within the 5'-UTR of these RNAs.

Originally, IRES elements were reported in viruses (110-114) but later these elements were discovered in a wide variety of cellular RNAs occurring in organisms ranging from fruit flies to humans. A list of reported cellular IRESs show that they can be found in mRNAs from proto-oncogenes, growth factors, transcription factors, and translation factors to cell cycle genes, among others (115-120). Besides the IRES, other secondary structural elements are known to regulate translation of mRNA. For example, 5'-UTR of bacterial mRNAs contains

RNA motifs termed 'riboswitches', which in response to metabolite or protein factor binding undergo a conformational switch to regulate translation (121-125).

Human vascular endothelial growth factor (hVEGF) is a key angiogenic growth factor whose overexpression is critical for tumor outgrowth, metastasis and prognosis of

several cancers (126). The 5'-UTR region of VEGF mRNA is unusually long (1038 nt), GC-

rich and can initiate translation via a cap-independent mechanism (115, 116, 127). This alternative translation mode can be particularly critical for VEGF expression under hypoxia, a condition often encountered during neovascularization, which is essential to tumor outgrowth and metastasis (128, 129). Interestingly, this region harbors two separate internal ribosomal entry sites, which are capable of independently initiating translation without the help of the 5'-cap. One of the sequences is a 463 nucleotide fragment (nt 91 to 554, IRES B) that initiates translation at the CUG 499,(130, 131) whereas the other, a 293 nucleotide portion, (nt 745 to 1038, IRES A) has been shown to initiate translation at the canonical

AUG codon (116, 127) (Figure 4.1).

Figure 4.1 Primary nucleotide sequence of the human VEGF IRES-A

We focused our studies on IRES-A as it is known to maintain VEGF translation under

hypoxia (132). Also, it is known that the protein product synthesized from initiation at the

AUG codon is a secreted form of VEGF, and thus important for the autocrine and paracrine

functions of the VEGF ligand (131).

Although scores of cellular IRESes have been identified thus far, the effects of RNA

structures on the cellular IRES function are not well defined. Based upon the accumulated

evidence it has been suggested that RNA structures in this region are beneficial to the

translation process through binding with ribosome subunits and initiation factors (112, 133,

134). However, the molecular details on the secondary and tertiary structures of cellular

IRES and how they regulate IRES function are not well understood, and there is paucity of

experimental data on hVEGF IRES structures in particular.

Herein, we show that a G-quadruplex in the 5'-UTR of hVEGF IRES-A alone is essential for its translation initiation activity. A 17 nt sequence identified by footprinting can potentially adopt multiple G-quadruplex structures, which is a deviation from the RNA quadruplex sequences reported in the literature and such multiple structures are possible because of the redundancy in the number of G-stretches. Mutational analyses of the hVEGF

IRES-A present in the context of a bicistronic dual luciferase reporter system show that a

‘switchable’ sequence containing greater than four G-stretches (nts 774-790) provides enough redundancy to ensure the formation of RNA G-quadruplex structures that are critical for the cap-independent translation initiation. Furthermore, a sequence designed by modifying a crippled mutant sequence was able to partially rescue IRES-A activity. Our data suggest that by choosing different combinations of four G-stretches, a conformational diversity can be generated within the previously defined segment, which can potentially modulate efficiency of cap-independent translation initiation in VEGF IRES-A, while pointing towards a novel mechanism of control of gene expression at the translation level.

4.2 MATERIALS AND METHODS

4.2.1 Preparation of oligonucleotide sequences. RNA sequences were synthesized by in

vitro transcription(51) and purified via denaturing polyacrylamide gel electrophoresis

(PAGE). The RNA was harvested via the crush and soak method by tumbling the gel slice at

4 °C in a solution of 300 mM NaCl, 10 mM Tris-HCl (pH 7.5) and 0.1 mM EDTA. Samples were concentrated with 2-butanol and salt was removed by ethanol precipitation and washed

twice with 70% ethanol. The RNA pellet was dissolved in 10 mM Tris-HCl (pH 7.5) and 0.1

mM EDTA. Concentrations of RNAs were determined based on their absorbance value at

260 nm and extinction coefficients calculated using nearest neighbor parameters (52).

4.2.2 Radiolabeling of RNA Oligonucleotides. Calf-intestinal alkaline phosphatase treated

RNA was 5'-end radiolabeled by treatment with T4 polynucleotide kinase (Promega), [γ-32P]

ATP (Pelkin Elmer) and incubated for 45 min at 37 °C. The reaction was stopped by the

addition of an equal volume of stop buffer (7 M urea, 10 mM Tris-HCl, pH 7.5 and 0.1 mM

EDTA). The radiolabeled full length RNA was isolated by denaturing PAGE. The RNA

was extracted from the gel via the crush and soak method as described above.

4.2.3 Circular Dichroism (CD) studies. All measurements were recorded at room

temperature. RNA was folded by heating the samples in 150 mM KCl, 10 mM Tris-HCl (pH

7.5) and 0.1 mM EDTA at 95 °C for 5 min, followed by slow cooling to room temperature

over a 90 min period. The circular dichroism (CD) spectra were recorded using a Jasco J-

810 spectrophotometer with a 0.1 cm cell at a scan speed of 50 nm/min with a response time

of 1 s. The spectra were averaged over five scans. For each sample, a buffer baseline was

collected in the same cuvette and subtracted from the average scan.

4.2.4 Footprinting by RNase T1. The 5'-end radiolabeled RNA was folded by heating the samples in 10 mM Tris-HCl (pH 7.5), 0.1 mM EDTA and in the presence of 150 mM KCl or

150 mM LiCl at 70 °C for 5 min and slow cooled to 37 °C over a 30 min period. Once reactions attained the appropriate temperature, the RNA was digested with 0.01 U of RNase

T1 (Ambion) for 5 min at 37 °C. The reactions were terminated by using an equal volume of stop buffer (7 M urea, 10 mM Tris-HCl, pH 7.5 and 0.1 mM EDTA). Treated RNA was electrophoresed on a denaturing gel, dried on WhatmanTM paper and exposed to a

phosphorimager screen. The gel images were visualized by scanning the screen on a

Typhoon Phosphorimager 8600 (Molecular Dynamics).

4.2.5 Dimethyl sulfate (DMS) Footprinting. 5′-end radiolabeled IRES A was prepared in 10 mM Tris-HCl pH 7.5, 150 mM KCl, 1 mM MgCl2. To fold the RNA, the reaction mix was

heated for 5 mins at 70 °C in a heating block. The block was then allowed to cool to room

temperature on the bench top over a period of 1 hour. Separate samples were prepared in an

identical way except 150 mM LiCl was added. Dimethyl sulfate was then added to a final

concentration of 2% and the reaction mix was incubated at room temperature for 10 mins at

which point the reaction was stopped with the addition of stop buffer (2 M β-

mercaptoethanol, 300 mM sodium-acetate, 250 µg/mL sheared salmon sperm DNA). Three volumes of 100% ethanol were added immediately and the RNA was precipitated on a dry ice/acetone bath followed by washing the pellet with 70% ethanol. The modified RNA was then reduced with sodium borohydride and then treated with aniline to induce cleavage at the corresponding phosphodiester bond, according to a previously published procedure (135).

4.2.6 Plasmid Construct for IRES-A transcription. We subcloned the relevant portions from

the hVEGF cDNA (hVEGF cDNA is a kind gift from Dr. Judith Abraham, Scios Inc., CA).

The portions encompassing the IRES A sequence were amplified by PCR. PCR Primer

sequences: IRES-A: 5'-GGTACCGCTAGCTCGGGCCGGGAG (sense) and 5'-

TCTAGAGGTTTCGGAGGCCCGACC (anti-sense). The restriction sites used are Kpn 1 and Xba 1 and they were cloned into a pBluescriptII(+) vector (Stratagene) containing the T7

RNA polymerase promoter.

4.2.7 Plasmid construction for the dual luciferase assay. The following four primers were

used for sub-cloning of the VEGF IRES A fragment:

F1 (Forward): 5’-AATCTACTCGAG TCGCGGAGGCTTGGGGCA, R1 (Reverse): 5’-

GGCGTCTTCCAT GGTTTCGGAGGCCCGACC, F2 (Forward): 5’-GCCTCCGAAACC

ATGGAAGACGCCAAAAAC, R2 (Reverse): 5’-CTTATCATGTCTGCTCGAAGCG.

VEGF IRES coding sequences are shown in italics, Luciferase coding sequences are shown

in bold and Xho I restriction site is underlined. VEGF IRES-A PCR fragment (0.3 kb) was

amplified with Pfu Turbo DNA polymerase (Stratagene) using pVEGFS-1 containing VEGF

IRES A as a template and primers F1 and R1. Simultaneously, a PCR fragment (1.65 kb)

encompassing the Firefly luciferase was amplified from pGL3-basic vector (Promega) using

primers F2 and R2. The DNA fragment was digested with Xho I/Xba I restriction enzymes

and cloned into Xho I/Xba I sites of pIRES vector (Clontech). The resulting plasmid was

named (pvIRESLF). The plasmid was digested with Xho I site and then Nhe I/Xba I fragment

containing the Renilla luciferase cDNA from pRL-CMV vector (Promega) were introduced

into Xho I site of pvIRESLF by blunt-end ligation. The resulting plasmid was named

(pLRvIRESLF).

4.2.8 Mutations on IRES-A sequence. Mutations were performed using a QuikChange site directed mutagenesis kit (Stratagene). The incorporation of mutations within the sequences was verified by sequencing performed at the Ohio State University’s Plant-Microbe

Genomics Facility.

4.2.9 Cell Culture. HeLa cells were grown in 96-well plates in Dulbecco’s modified Eagles medium (DMEM) supplemented with 10% fetal bovine serum and the antibiotics streptomycin and penicillin at 37 °C in 5% CO2 in a humidified incubator.

4.2.10 Luciferase and quantitative RT-PCR assays. HeLa cells were transfected with vectors using Lipofectamine™ 2000 (Invitrogen) according to the manufacturer’s protocol.

Twenty-four hours after transfection, Renilla (RL) and firefly (FL) luciferase activities were measured using a Dual-Glo™ Luciferase assay system (Promega) as per the manufacturer’s supplied protocol on a Synergy 2 microplate reader (BioTek Instruments). Total RNA was extracted from transfected HeLa cells using a NucleoSpin® RNA II kit (Clontech). Prior to reverse transcription, DNA was removed from each RNA sample upon treatment with RQ1

RNase-Free DNase (Promega). Renilla and firefly mRNAs were reverse transcribed using

AMV-RT (New England Biolabs) and cDNA was subjected to quantitative real-time PCR using a SYBR® Green PCR Master Mix kit (Applied Biosystems) on an ABI PRISM 7000

Sequence Detection System in the presence of the appropriate set of primers. Forward

primers RL 5'-(GTAACGCTGCCTCCAGCTAC)-3', FL 5'-

(TTCGCTAAGAGCACCCTGAT)-3' and reverse primers RL 5'-

(GTAGGCAGCGAACTCCTCAG)-3' and FL 5'-(GCTGCAGCAGGATAGACTCC)-3'.

4.3 RESULTS

4.3.1 Footprinting reveals a highly protected G-rich segment within hVEGF IRES-A

Figure 4.2 RNase T1 footprinting in the presence of 150 mM K+, 150 mM Li+ and 1 mM MgCl2.

To determine secondary structural features within the hVEGF IRES-A, RNase T1 footprinting was performed on the entire 293 nt sequence in the presence of the divalent and monovalent ions (Figure 4.2, refer to supplemental Figure 4.1 for the entire primary sequence of IRES-A). RNase T1 catalyzes cleavage at guanosines in single-stranded RNA, but not

Figure 4.3 DMS footprinting in the presence of 150 mM K+ or 150 mM Li+. guanosines that are involved in secondary or tertiary structures, including G-quadruplexes

(25). RNase T1 footprinting on the IRES-A sequence shows that in presence of 150 mM K+

a G-rich region between G774 and G790 (nucleotides are numbered based on entire 5'-UTR

of human VEGF mRNA) is protected from cleavage (Figure 4.2), suggesting at a minimum

the presence of some form of secondary structure within this segment. Because of the highly

G-rich nature of the sequence and the tandem G repeats there is a possibility of the formation of a quadruplex structure within this segment. Therefore, we wanted to explore whether the protection persists in the presence of Li+, as the monovalent cation lithium is known to lack

Figure 4.4 Schematic of a subset of G-quadruplex structures that shows the different G-stretches (i) and the proposed utilization of adjacent (ii) and non- adjacent G-tracts (iii).

the ability to promote or stabilize G-quadruplex structures (3) and thus was used as a control

for monovalent metal ion-dependent quadruplex folding. The observation that the protection

of the nucleotides only happens in the presence of K+ ions and not in Li+, combined with the

fact that primary sequence of this protected region harbors more than four tandem G-repeats,

suggests that this segment forms an intramolecular G-quadruplex structure. Furthermore, the

footprinting pattern in the presence of Mg2+ but lacking monovalent metal ions did not show

any protection within the region, suggesting that the protected segment is a K+ induced independently folded domain, most likely adopting a G-quadruplex structure and overall

divalent ion dependent folding of this large RNA molecule is insufficient to cause folding of

the G774-G790 segment (Figure 4.2).

To further test the existence of such structures, the entire IRES-A sequence was

subjected to DMS footprinting. DMS modifies the N7 of guanosine and thus can directly

probe G-quadruplex structure formation (3). Analysis of the footprinting pattern in Figure

4.2, shows that the regions that were protected from RNase T1 cleavage were also protected

from DMS modification in presence of K+ ions (Figure 4.3). However, the protection was

lost when K+ ions were replaced with an equimolar amount of Li+ ions. The K+ ion specific

DMS modification pattern strongly suggests the presence of a G-quadruplex structure within

the G774-G790 section of IRES-A. The protected region consists of a five G-stretch that is flanked by two stretches of 2 Gs on both sides. Upon analysis of this primary sequence it is evident that greater than twenty two-tiered intramolecular G-quadruplexes are theoretically possible by utilizing different combinations of G-tracts. Figure 4.4 shows a schematic representation of a subset of possible structures, however, G-quadruplexes resulting from other combinations of G-stretches are possible. Thus combining the observations made in the RNase T1 enzymatic footprinting and the DMS modification studies it can be concluded that the G774-G790 section of the VEGF IRES-A can adopt one or more quadruplex

structures in a K+ dependent manner, which is not contingent upon divalent metal ion

Figure 4.5 Circular dichroism (CD) spectra of an oligoribonucleotide encompassing the protected region in hVEGF IRES A. All reactions were performed in the presence of either 150 mM KCl, LiCl or NaCl in T10E0.1.

mediated folding of the RNA.

In order to establish that the protected segment can also adopt a G-quadruplex

conformation in isolation (G774-G790), we subjected it to CD analysis. CD experiments

conducted in the presence of 150 mM K+ showed a peak at ~260 nm and a trough around 240

nm (Figure 4.5), which are characteristics of a parallel quadruplex structure (3, 25).

However, the CD spectral features were absent when analyzed in the presence of 150 mM

Li+ or Na+ (Figure 4.5). Thus, the G-rich segment (G774-G790) not only adopts a G-

quadruplex conformation in the context of the entire 293 nt IRES, but is also capable of

adopting a G-quadruplex conformation in isolation.

4.3.2 The G-rich protected region is essential for translation initiation activity of hVEGF

IRES A.

So far the data have shown that a highly G-rich portion within IRES-A RNA is strongly protected from DMS modification and RNase T1 mediated cleavage, suggesting the presence of a G-quadruplex structure, and the segment when studied in isolation adopts a parallel G-quadruplex conformation as determined by circular dichroism. An important and obvious question is whether the protected sequence is critical to translation initiation activity of IRES-A. To directly address this question, a dual luciferase reporter construct was prepared in which the entire IRES-A sequence was placed just upstream of the firefly luciferase, while the renilla luciferase was under the control of a CMV promoter

(hVEGFBicis, Figure 4.6). This plasmid construct allowed quantification of the translation initiation activity of the IRES-A by individually yet simultaneously measuring the dual luciferase activities. To dissect the role of the putative G-quadruplex forming sequence, the

protected segment within the IRES-A was mutated with a series of G to U substitutions

(Figure 4.6). Four G to U mutations were used to decisively eliminate any potential intramolecular G-quadruplex formation (G774,777,781,783U, quadruple mutant, 4MVF).

The activities of the wild-type IRES-A and the 4MVF along with several other mutants (see below) are shown in Figure 4.7. When the activity of the quadruple mutant was measured, the IRES-A function was found to be completely abolished (Figure 4.7). Because the

quadruple mutant lacks the ability to adopt a quadruplex, the observation made above

suggested a potential role of a G-quadruplex structure in IRES-A mediated translation initiation.

Since the putative G-quadruplex can form from any combination of four G-tracts,

mutations were designed to investigate the importance of any specific G-quadruplex structure that may form. In a double mutant (G777,781U), G777 and G781 were replaced with U’s,

Figure 4.6 Schematic of various dual luciferase bi-cistronic constructs. making the second and third G-stretches unavailable for participation in G-quadruplex formation, which in turn can disrupt the formation of the G-quadruplex encompassed by the segment 774-785. However, the dual mutations (G777,781U) did not affect the activity of the IRES as it maintained its functionality on par with the wild type IRES (Figure 4.7; mutants G774,783U, G783,787U, and G774,789U will be discussed below).

Figure 4.7 Histogram showing % activity of the mutant constructs normalized to the wild type construct.

Presumably, the mutant remained fully functional because four stretches of tandem G’s in

close enough proximity were still available, allowing formation of a G-quadruplex

conformation for maintaining the IRES function (see 4.4 below). This, combined with the

observation that the quadruple mutant lacking functional activity as well the ability to form a

G-quadruplex, lead us to propose that the full loss of IRES activity in this mutant was most

likely due to the non-formation of any G-quadruplex structure in that region. Thus, from these results it may be concluded that the quadruplex forming sequence is necessary for

IRES-A mediated cap-independent initiation of translation. To confirm that this loss in activity was indeed occurring at the translational level, RT-PCR was performed on both the wild-type and 4MVF sequences. The results of which indicate that both constructs produced similar mRNA levels and thus the loss in activity of the quadruple mutant (4MVF) can be

attributed to the loss of function at the translation level (Figure 4.8).

Figure 4.8 Histogram representing the ratio of Renilla to firefly luciferase activities in HeLa cells (gray bars). The ratio of Renilla to firefly luciferase of the Wt. plasmid was set to one, and all other values were normalized accordingly. The ratio of CT values for Renilla to firefly determined by qRT-PCR (black bars). All experiments were conducted in triplicate at a minimum.

The above observations can be explained by invoking the notion of the presence of a

‘switchable’ G-quadruplex. For example, assuming that typically contiguous G-stretches are used for G-quadruplex formation, a wild type G-quadruplex would use any set of four G- stretches that are in tandem. However, in case of the double mutant such residues are unavailable because the second and the third G-stretches from the 5'-end of the sequence are mutated. Thus, to adopt a quadruplex conformation and retain its translation initiation activity on par with the wild-type, the structure most likely has to undergo a ‘switch’ forcing the first G-stretch to recruit distal G-stretches. While in the case of the wild type sequence an

ensemble of quadruplexes with similar energy are possible, in the case of the double mutant,

many such structures cannot form and thus a switch resulting in a subset of structures that are

functionally as active as the wild type is likely to be formed.

4.3.3 Footprinting indicates that the double mutant (G777,781U) retains the ability to adopt

G-quadruplex conformation

Figure 4.9 Scanned images of gels showing RNase T1 footprinting of the 293 nt human VEGF IRES-A and its various mutants.

To ensure that the loss in activity associated with the quadruple mutant of IRES-A is truly due to the lack of formation of the quadruplex structure and not a result of disruption of the global folding or structures outside the G-rich segment, protection patterns in the presence of RNase T1 and DMS were monitored. We hypothesized that the lack of quadruplex formation in the inactive mutant (4MVF) would make all of the guanosines in the

774-790 region of IRES-A RNA (four mutations identical to the ones in 4MVF; named as 4

Mut.) susceptible to RNase T1 mediated cleavage and DMS modification, whereas the double mutant sequence (two mutations identical to the ones in G777,781U; named as 2

Mut.) will show footprinting patterns similar to the wild type sequence except at the two mutated guanosines (G777 and G781). These sequences (2 Mut. and 4 Mut.) were subjected to RNase T1 mediated cleavage and DMS modification in the presence of K+ or Li+ ions. In the case of the double mutant (2 Mut.), the protection pattern (774-790) corresponds to the presence of a quadruplex, and is similar to that of the wild type, however, there was no discernible protection in the case of the quadruple mutant (4 Mut. Figure 4.9 and 4.10).

These results directly corroborate the proposal that the double mutant (G777,781U) is active because of the formation of a G-quadruplex via utilization of four available G-stretches, whereas the quadruple mutant (4MVF) lacked the minimum number of G-stretches necessary to support intramolecular G-quadruplex formation.

The quadruplex formation in the case of 2 Mut. perhaps occurs via a different

set of guanosines compared to the wild-type quadruplex, which was sufficient to maintain IRES-A activity in the mutant G777,781U. Taken together, the protection patterns in the presence of K+ and Li+ provide strong evidence for quadruplex

formation in the wild-type and double mutant (2 Mut.) of the IRES-A sequence.

Another observation from the footprinting data was that the sites of

Figure 4.10 Scanned images of gels showing DMS footprinting of the 293 nt human VEGF IRES-A and its various mutants.

cleavage/protection outside of the quadruplex forming region in the wild-type (Wt.),

the double mutant (2 Mut.) and the quadruple mutant (4 Mut.) sequences essentially

remained unaltered, suggesting that the global fold of the RNAs were unperturbed

irrespective of the mutations. Thus, formation or lack of formation of the G-

quadruplex is apparently unrelated to the folding of the remainder of the molecule.

4.3.4 Utilization of different combinations of G-stretches can lead to differential regulation of IRES-A function

In order to precisely constrain the G-stretches being utilized, mutating specific Gs so as to allow determination of stretches that are preferentially used by the IRES was further investigated. For this purpose, three more double mutants: (G774,783U), (G783,787U), and

(G774,789U) of the wild type dual luciferase construct (hVEGFBicis) were prepared and their level of activities were measured. The mutant G774,783U was ~82% active as the wild-type (Figure 4.7). This mutant was chosen to determine whether or not the mutations that differed between the 4MVF and G777,781U mutants were responsible for the complete loss of activity of the IRES-A quadruple mutant. These mutations had little effect on the

IRES function and were not solely responsible for the complete loss of function of the quadruple mutant, but in conjunction with two other mutations were enough to fully abrogate the IRES activity. Thus, it can be argued that the lack of quadruplex formation is the primary cause for the loss of function of the quadruple (4MVF) mutant. Two other mutations were tested to inquire about the ‘switchable’ nature of the structure. Compared to the wild-type, the G783,787U mutant was ~44% active, while the G774,789U mutant was only ~18% active. These results demonstrate that the mutations that were made on the RNA still retained a viable IRES, however, with various degrees of reduction in activity. These observations raise an intriguing possibility that by utilizing different sub sets of G-stretches, the IRES can fine-tune its activity.

4.3.5 An engineered G-quadruplex can rescue activity of a functionally deficient mutant

Figure 4.11 Histogram showing % activity of the dual luciferase rescue mutant construct (Rescue Quad) normalized to its parent construct (G774, 789U).

Since a functional G-quadruplex is required for activity of IRES-A, we rationalized that an inactive or barely active sequence, if engineered, can result in a rescue of the function. We mutated the G774,789U mutant (17% active) to design a new mutant sequence

(A788G, Rescue Quad) that contains an adequate number of G-stretches to allow more flexibility in adoption of a G-quadruplex structure. The reason A788 was chosen to be mutated was because most of the other available A’s if mutated would have resulted in

unusually long G-stretches significantly deviating from the makeup of the wild-type sequence. The dual luciferase activities were measured and plotted in a histogram showing

% activity of the dual luciferase rescue mutant construct (Rescue Quad) normalized to its parent construct (G774, 789U). This new mutant showed about 55% enhancement of IRES-

A activity compared to the G774,789U mutant, indicating a partial rescue of the function

(Figure 4.11). RNase T1 footprinting of the rescue mutant showed a protection pattern in

Figure 4.12 Scanned image of a gel showing RNase T1 footprinting of the mutant Rescue Quad version of the transcribable 293 nt human VEGF IRES-A in the presence of K+ and Li+. NE is the no enzyme lane that contained K+. presence of K+ (Figure 4.12) that is commensurate with the formation of a G-quadruplex structure. However, negligible if any protection was observed in presence of an equimolar

amount of Li+. The data suggest that the engineered sequence is able to adopt a

conformation that is more conducive to cap-independent initiation of translation by hVEGF

IRES-A. This is the first example to our knowledge where an engineered sequence supporting and adopting a G-quadruplex structure was able to rescue translation initiation activity.

4.4 DISCUSSION

4.4.1 Sequence redundancy guarantees the formation of a ‘switchable’ RNA G-quadruplex

critical for IRES-A mediated translation initiation

G-quadruplex forming sequences located in 5'-UTR of mRNA have been almost

exclusively found to be translation inhibitors (17, 23, 25, 136). This report provides a set of

evidence to the contrary. The presented data indicate that G-quadruplex formation is most

likely a prerequisite for successful IRES-A mediated cap-independent translation initiation.

The formation of such a structure may be facilitated by redundancy in the closely located G- stretches embedded within the IRES-A sequence. By employing systematic footprinting analyses, the 293 nt IRES-A was first narrowed down to a ~17 nt (774-790) sequence that adopted a G-quadruplex structure (Figures 4.2and 4.3). However, it may encode for many putative G-quadruplexes, only a subset of which are depicted in Figure 4.4. Analysis of the entire protected segment and at least one of the minimal G-quadruplex forming sequences in isolation revealed that they also adopt a G-quadruplex conformation (Figure 4.5), suggesting that a quadruplex structure can be formed both in the context of the full IRES-A sequence and when analyzed in isolation. When one of the minimal G-quadruplex forming segments

was mutated to disrupt the quadruplex formation in the context of the entire IRES-A

sequence (G777,781U, Figure 4.6), surprisingly the mutation did not result in any change in

activity compared to the wild type (Figure 4.7). However, a quadruple mutant (4MVF),

which lacks a sufficient number of G-stretches to adopt an intramolecular G-quadruplex

conformation, was completely inactive. To explain these results we propose that there is an

obligatory requirement that a G-quadruplex structure must be formed for maintaining the

IRES function. This emphasizes the importance of the G-quadruplex enabling sequence on the function of IRES-A. Previously Bonal et. al. reported that a G-quadruplex structure in conjunction with two stem-loop structures located within the FGF-2 IRES was needed for translation initiation (22). However, the authors concluded that the main regulator of the

IRES function was a 17-nt non-G-quadruplex region located within the domain. The putative

G-quadruplex alone was insufficient to support FGF-2 IRES function (22). Also, according to the authors, the quadruplex structure presumably was detrimental to initiation of translation at one of the internal codons, which is in consonance with the more conventional effect of the RNA G-quadruplex on translation. Thus, the G-quadruplex in hVEGF IRES-A is not only a unique example of a quadruplex exclusively required for translation initiation, but may also represent a more general mechanism of the use of a ‘switchable’ G-quadruplex structure to ensure function by maintaining sufficient redundancy.

There may be two major reasons for the redundancy in the G-stretches within the protected sequence: i) ensure availability of more than an adequate number of G-stretches to guarantee the formation of a G-quadruplex and ii) form G-quadruplexes with differential ability of initiating translation; in effect regulating the translation efficiency. In case of the wild type sequence, the redundancy in the G-stretches allows utilization of adjacent or non-

adjacent G-stretches for G-quadruplex formation (Figure 4.4). However, the quadruplex

Figure 4.13 Histograms representing quantitation of the gel shown in Figure 4.12. The guanosines corresponding to the protected region (774-790) are shown individually. Two guanosines (796 and 798) that are outside the protected region were quantitated and used as controls. formed by the double mutant (G777,781U) must utilize non-adjacent G-stretches (Q4, Figure

4.4). The formation of the quadruplex was supported by the RNase T1 footprinting and

DMS modification data that showed that the cleavage pattern in the double mutant (2 Mut.) is commensurate with a G-quadruplex structure that used the available but non-contiguous

stretches of Gs. These observations fit well if it is assumed that the formed G-quadruplex structure is inherently redundant in nature, because to retain a quadruplex structure in the context of the doubly mutated IRES-A, the original quadruplex had to ‘switch’ to successfully accommodate an alternate G-quadruplex structure (Figure 4.4). It is a possibility that the G-quadruplexes formed by the mutant are a subset of the ensemble of structures with comparable energy formed by the wild type sequence and thus are able to

maintain the level of activity on par with the wild type. This alternate quadruplex structure

uses some of the original nucleotides and recruits new ones that impart the ‘switchable’ nature to the structure. The proposal assumes that not all of the G-stretches used by the

double mutant (G777,781U) are used by the wild type structure, because a G-quadruplex

with shorter loops generally are preferred than a longer loop (in case of G777,781U). To

support the ‘switchable’ quadruplex structures, different combinations of G-stretches and

variable loops sizes need to be accommodated. Although shorter loops are preferred, G-

quadruplex structures with up to a 7 nt loop has been reported.(137, 138) In the double mutant, G777,781U, presumably a 6 nt loop containing a quadruplex is formed, which is supported by the footprinting data (Figure 4.9, 4-10, and 4.13).

Based upon the primary sequence, the G-quadruplex is likely to adopt a two-tiered structure and may not be as stable compared to some of the reported RNA quadruplexes (17,

25). However, the very stable RNA G-quadruplexes reported previously are detrimental to translation, including a two-tiered non-natural quadruplex. The putative two-tiered quadruplex described here is seemingly less stable than some of the reported naturally occurring RNA quadruplexes, and it is also unique because of the fact that it is a functionally essential motif for translation initiation activity. The lower stability may be functionally significant, as it has been demonstrated recently that stronger IRESs have weaker secondary structures. The IRES structure perhaps needs to be malleable enough to adjust to conformational change as it accommodates binding to putative factors necessary for translation initiation. The redundancy in the G-rich stretches in the IRES-A sequence provides a certain degree of flexibility as well as robustness to the system because within this segment the minimum requirement of a quadruplex formation can be met in multiple ways

although it may result in different degrees of functional consequence, as will be discussed

below.

4.4.2 G-quadruplex Structures formed by Strategic Choice of G-stretches May Fine-tune

IRES-A Mediated Translation Initiation

Based upon the differential functionality of a designed set of mutants, a G-quadruplex

structure that is essential for translation initiation is proposed. However, some of the mutants

used in this study did not retain wild type level of activity. For example, the double mutants

G774,783U, G783,787U and G774,789U still have the possibility to form quadruplex

structures based upon analysis of their primary sequence (Figure 4.4). All three mutants

utilize a different combination of G-tracts to form a quadruplex and all show translation initiation activity. The activity, however, is reduced by various degrees in comparison to the

wild type IRES activity. We propose that due to the ‘switchable’ nature of the quadruplex

structure, different mutants may utilize different G-stretches to presumably adopt a G-

quadruplex conformation, but with varying degrees of impact on the translation initiation

function of the IRES. Based upon our data we suggest that by utilizing different subsets of

G-stretches, the IRES could potentially alter its conformation within nucleotides 774-790 so

as to form specific G-quadruplexes, which can potentially be a mechanism for a

conformational switch for regulating translation initiation. Thus, at least one of the possible

quadruplexes can fully support translation initiation, while a sub-set of them support

translation to various degrees. Utilization of the different G-stretches to adopt G-quadruplex conformation, which results in modulation of the translation initiation activity to a varying level, may be a direct or indirect result of interaction with yet to be identified factors. The

mutants in that sense serve as surrogates to structures that may result as the outcome of

protein factor binding. Riboswitches that essentially are stem-loop structures, which change

conformation upon specific metabolite or protein binding to attenuate transcription or initiate

translation have been described (121, 125). To our knowledge, however, there are no reports of conformational switches involving G-quadruplex structures resulting from a naturally occurring sequence.

4.5 CONCLUSION

Despite the growing number of RNA G-quadruplexes being discovered, most of the reports show that these quadruplex structures are inhibitory to translation.(17, 23, 25, 136)

Herein, we provide evidence that a quadruplex forming sequence is essential for IRES-A

mediated translation of the hVEGF mRNA. When mutated, the IRES-A utilizes neighboring

tracts of G-stretches to maintain its activity in eukaryotic cells. The ‘switchable’ character

makes the RNA G-quadruplex described in this report unique. Additionally, the segment may

also act as a G-quadruplex dependent ‘switch’ that can regulate IRES-A function, as different

mutants showed a variety of level of activities. Disruption of the quadruplex formation

interferes with IRES-A function, and because of its importance in translation initiation under

hypoxic stress, a condition encountered in tumor angiogenesis, the quadruplex motif can be a

tumor specific therapeutic target.

CHAPTER 5

RNA DOMAIN SWAPPING SHOW CONTEXT DEPENDENT EFFECT OF G-

QUADRUPLEXES ON TRANSLATION

5.1 INTRODUCTION

The 5'-untranslated (UTR) of the MT3-MMP mRNA harbors a G-quadruplex forming sequence (M3Q) which has been shown to have repressed translational activity of a reporter

Figure 5.1 Schematic of RNA G-quadruplex swapping. mRNA in eukaryotic cells by more than half (25). Another angiogenesis regulator in

humans, the vascular endothelial growth factor (hVEGF) has been associated with

progressive tumor development (126) and has also been a gene of interest in our laboratory.

It has been discovered that the presence of a G-quadruplex within IRES-A of the 5'-UTR of hVEGF is essential for the cap-independent initiation of translation (26). Furthermore, sequence redundancy in this G-rich segment supports a ‘switchable’ RNA G-quadruplex structure that can utilize different sets of specific combinations of G-tracts which allows for a conformational switch implying a tunable regulatory role of the quadruplex structure in translation initiation. It has been established that these two G-quadruplexes play opposite roles in translational regulation. Interestingly, both quadruplex forming sequences are composed of all purine nucleotides and are both 17 nucleotides in length. The similarity of these two sequences would make it an interesting target for RNA engineering (Figure 5.1).

RNA engineering has been gaining interest as an emerging field, particularly for its potential medical applications, however, such examples are few and far between when compared to protein domain swapping. Domain swapping as a mode of RNA engineering has proven to be a useful tool in characterizing the functionality of a gene. In one study, swapping genes between two glycopeptide-resistant Streptomyces species was used to investigate what the bacteria use to determine inducer specificity (140). With the exception of the RNA pseudoknot (141), the nucleic acid domains swapped are structurally less complicated than those previously used in 3D swapping between proteins. The complexity of the G- quadruplex structure and the access to two mRNA systems in which these structures play opposing roles provides a fascinating platform to conduct domain swapping experiments. To our knowledge, this will be the first example where swapping of the modular quadruplex domains will be used for RNA engineering.

5.2 MATERIALS AND METHODS

5.2.1 Plasmid Construction of wt-UTR and mut-UTR MT3-MMP. The entire 282-nucleotide

sequence of the 5'-UTR of MT3-MMP (wt-UTR) was ligated into the Nhe I site of the

psiCHECK-2 vector as previously described (section 2.2) (25).

5.2.2 Plasmid Construction of VEGF IRES constructs. A dual-luciferase reporter construct was prepared in which the entire IRES-A sequence (293 nucleotides) was placed just upstream of the firefly luciferase, while the renilla luciferase was under the control of a CMV promoter. Four G to U mutations were made and previously proven to eliminate any potential intramolecular G-quadruplex formation (G774,777, 781,783U, 4MVF) (26).

Versions of the above mentioned plasmids containing a T7 RNA polymerase promoter upstream of the IRES-A sequence were also previously constructed (section 4.2) (26).

5.2.3 Preparation of Primer Sequences. All primers were synthesized by Integrated DNA

Technologies. The primers were purified via 17% polyacrylamide gel electrophoresis

(PAGE). Full-length products were visualized by UV shadowing and excised from the gel.

The DNA was harvested via the crush and soak method by tumbling the gel slice overnight at

4 °C in a solution of 300 mM NaCl, 10 mM Tris-HCl, and 0.1mM EDTA (pH 7.5). Salt was

removed by ethanol precipitation of the DNA and washed with 70% ethanol. The DNA

pellet was dissolved in DEPC-treated water. DNA concentrations were determined on the

basis of their absorbance value at 260 nm and extinction coefficients.

5.2.4 Deletions and Insertions of G-rich Nucleic Acid Sequences. Deletions of nucleotides

213-229 of the 5'-UTR of MT3-MMP (wt-UTR) were performed using a QuikChange site-

directed mutagenesis kit (Stratagene). The deletion and insertions of nucleic acids were

carried out by using polymerase chain reaction (PCR). The PCR reactions were prepared

according to the manufacture’s protocol using 10 ng dsDNA template, 125 ng forward and

reverse primers, 1.5 µL Quiksolution Reagent (Agilent, CAT#200516-51), and 1µL Pfu Ultra

Fusion HS DNA Polymerase (Agilent, CAT#600670-51). The reaction was PCR amplified using a thermocycler (BIO-RAD) and cycled in the following successive segments: Segment

1 was one cycle at 95 ºC for 1 minute; Segment 2 was one cycle at 95 ºC for 50 seconds, one cycle at 60 ºC for 50 seconds, and one cycle at 68 ºC for 6.5 minutes (Segment 2 was repeated 17 more times); Segment 3 was one cycle at 68 ºC for 7 minutes. The list of primers and dsDNA templates used in each reaction for plasmids containing dual luciferase reporter genes are given in the following table.

Reaction Forward Primer Reverse Primer Deletion of VEGF (G4) 5'- GGA GGA GCC GCA GCC 5'- TCT TCC TTC TCT TCT from VEGF IRES-A AAG AAG AGA AGG AAG A -3' TGG CTG CGG CTC CTC C -3' Insertion of M3Q (G4) 5'- GGA GGA GCC GCA GCC 5'- TCT TCC TTC TCT TCT into VEFG IRES-A GGG AGG GAG GGA GAG GGA TCC CTC TCC CTC CCT CCC (deletion site) AGA AGA GAA GGA AGA -3' GGC TGC GGC TCC TCC -3' Insertion of NRAS 5'- GGA GGA GCC GCA GCC 5'-TCT TCC TTC TCT TCC (G4) into VEGF IRES- GGG AGG GGC GGG TCT GGG CAG ACC CGC CCC TCC A (deletion site) AAG AAG AGA AGG AAG A- 3' CGG CTG CGG CTC CTC C- 3' Deletion of M3Q (G4) 5'- GGA GGA CTT TTT TTT 5'- GTT TTC TCC CTC TCT from MT3-MMP GAA AGG AAA CGA AGA GAG TCG TTT CCT TTC AAA AAA GGA GAA AAC -3' AAG TCC TCC -3' Insertion of VEGF (G4) 5'- GGA GGA CTT TTT TTT 5'- GTT TTC TCC CTC TCT into MT3-MMP GAA AGG AAA CGA GGA GGA CCT CCT CCC CCT CCT CCT (deletion site) GGG GGA GGA GGA GAG AGG CGT TTC CTT TCA AAA GAG AAA AC -3' AAA AGT CCT CC -3'

Table 5.1 Primer sequences used for synthesizing various constructs used in RNA G-

quadruplex domain swapping experiments.

5.2.5 Transformation of Competent Cells with PCR Products. The product of each PCR

reaction underwent parental DNA digestion according to the manufacturer’s protocol using

Dpn I enzyme (Agilent, CAT#500402-51) and incubated for 1 hour at 37 ºC. Transformation

of Dpn I–treated PCR products PCR was performed according to manufacturers protocol

using XL-10 Gold Ultracompetent E. coli cells (Agilent).

5.2.6 Isolation of Plasmid DNA. The isolation of the plasmid was performed using the

PureYield Plasmid Miniprep System (Promega). Concentrations of the isolated DNA

plasmid samples were determined using the absorbance of each sample at 260 nm. The

deletion or insertion of the nucleotide sequences was verified by sequencing performed at the

Ohio State University’s Plant-Microbe Genomics Facility.

5.2.7 Transfection of HeLa Cells. HeLa cells were plated at 27,500 cells per well in 96-well

plates and were grown for 24 hours using Dulbecco’s modified Eagles medium (DMEM)

containing 1% streptomycin/penicillin, and 10% fetal bovine serum. The cells were

incubated at 37 ºC in 5% CO2. Cells were transfected with 1 µg per well plasmid DNA

5.2.8 Dual Luciferase Reporter Assay. Twenty-four hours after transfection, renilla and firefly activities were measured using a Dual-Glo Luciferase Assay System (Promega) according to the manufacturer’s protocol on a Synergy 2 microplate reader (BioTek

Instruments).

5.2.8 RNase T1 footprinting. In vitro transcription and 5'-labeling of RNA oligonucleotides

of mutated IRES-A and wild type MT3-MMP sequences was performed as previously described (25, 26) (section2.2 and 4.2). The 5'-end-labeled RNA was folded by heating the samples in the presence of 100 mM KCl or LiCl in 10 mM Tris-HCl and 0.1 mM EDTA (pH

7.5) at 95 °C for 5 min and then slowly cooling to 37 °C over a period of 60 min. Once reaction mixtures reached the appropriate temperature, the RNA was digested with 0.5 units of RNase T1 (Ambion) for 5 min at 37 °C. The reaction was terminated by using an equal volume of stop buffer as described previously (section 2.2 and 4.2) (25, 26).

5.3. RESULTS AND DISCUSSION

5.3.1 G-quadruplex domain exchange indicates a context dependent function.

Previously, it has been shown that the M3Q sequence when present in the context of the entire 5'-UTR of the MT3-MMP mRNA acts as a translation repressor (Figure 2.8) (25).

Figure 5.2 Histogram showing the translation activity in presence of endogenous VEGF G- quadruplex, the 4MVF sequence, the M3Q G-quadruplex forming sequence, and the NRAS quadruplex forming sequence when present individually within the 5'-UTR of hVEGF IRES-A. The histogram represents the activity percentage, calculated using the ratio of firefly to Renilla activities in HeLa cells. The ratios were normalized to the wild type hVEGF IRES-A (blue bar), which was set to 100%.

To determine how this motif will act in the context of the IRES A of hVEGF mRNA

deletion/insertion experiments were performed on the bicistronic plasmid which harbors the

entire 293 nt hVEGF IRES A sequence before a firefly luciferase reporter gene

(hVEGFBicis) (Figure 4.6) (26). The 17 nt RNA G-quadruplex forming sequence (QVEG)

which was reported to be essential for IRES mediated translation was first deleted then

replaced with the M3Q sequence, a known translation repressor. Additionally, 4MVF, a

plasmid containing four mutations of the G-rich sequence of QVEG, which lacks the ability

to adopt a G-quadruplex structure was also used. All three constructs were transfected into

HeLa cells and both renilla and firefly luciferase activity levels were measured. As is evident

in Figure 5.2, when normalized to the wild-type (hVEGFBicis) the activity of the quadruple

mutant 4MVF was completely abolished as reported previously (Figure 4.7) (26). However, when the M3Q quadruplex forming sequence is planted in place of QVEG there is a 36 ±

11% increase in activity over the wild-type construct, although M3Q sequence is known to be a translation repressor when present in its native context. Clearly, M3Q when present in the context of hVEGF IRES-A has changed its role from a repressor to a ‘helper’ of translation. The possible explanation for this increase in activity could be the hyperstability of the M3Q quadruplex influencing the activity of translation in the context of the IRES-A hVEGF construct. It has been reported that increasing the number of tetrads in a quadruplex structure increases its stability (3). Since the quadruplex located in the hVEGF IRES-A mRNA is essential for translation, increasing the stability of this essential quadruplex could have enhanced its positive effect on translation. The data suggest that the M3Q sequence, which is a known translation repressor (25) reversed its role and instead acted as a ‘helper’ of translation initiation activity of the IRES-A.

To determine whether this is a special case or other RNA G-quadruplexes that act as

translation repressors can also reverse their role according to the context in which they are

placed, we replaced the G-rich segment (G774-790) of IRES-A with the G-rich sequence

from the 5'-UTR of the NRAS mRNA (5'-GGGAGGGGCGGGUCUGGG-3'), which is a

known translation repressor (17). The average activity of firefly to renilla from the plasmid

containing the NRAS sequence inserted into the hVEGF IRES-A 5'-UTR was found to be 81

± 6%. Thus the NRAS sequence was able to support ~80% activity compared to the wild

type G-rich sequence in IRES-A essential for translation. When present in its endogenous

location within the 5'-UTR of NRAS, it was shown to result in approximately 75% repression

in translation activity when normalized to a mutant that does not form a quadruplex (17).

Thus, two different G-rich sequences which are known translation repressors revered their

functional role and acted as a surrogate to the wild type sequence and supported translation

initiation activity. The data emphasize that the RNA G-quadruplex mediated modulation of

translation activity may be dependent upon the context in which they are present.

5.3.2 Enzymatic footprinting shows that the M3Q sequence present in the engineered

hVEGF IRES-A adopts a G-quadruplex structure

It has been previously reported from our laboratory that the G-rich segment of the

wild-type sequence within the hVEGF IRES-A adopts G-quadruplex structure as was established by DMS and RNase T1 footprinting (Figures 4.2 and 4.3) (26). To determine if the inserted M3Q and NRAS sequence also form a RNA quadruplex structure in the context of the hVEGF IRES-A mRNA, RNase T1 footprinting experiments where performed. Figure

5.3 shows the results of all four constructs in the presence of either 150 mM LiCl or KCl. In the case of the wild type, the protection pattern (774-790) corresponds to the presence of a quadruplex, however, expectedly there was no discernible protection in the case of the

quadruple mutant.

Figure 5.3 Scanned image of a gel showing results of RNase T1 footprinting of the various IRES-A RNAs. From left: wild type (wt.) quadruple mutant (4Mut.), M3Q, and the NRAS quadruplex forming sequences inserted into the IRES-A in the presence of Li+ and K+ ions respectively.

The M3Q insertion shows significantly more protection from the RNase T1 mediated

cleavage in the presence of K+ than in the presence of equimolar Li+, suggesting that most

likely a quadruplex structure is being formed. It should be noted, that G786 (in the context

of the inserted M3Q) is not protected from RNase T1 cleavage in either the presence of K+ or

Li+ indicating that the residue does not participate in quadruplex formation. This result

parallels what was observed when the M3Q sequence in isolation was subjected to RNase T1

cleavage (Figure 2.6) (25). Thus, it can be said that the M3Q sequence is folding into a

similar structure in the context of the IRES-A as it does in isolation. Taken together with the

functional data, it can be said that the M3Q sequence can adopt a G-quadruplex structure in

the context the hVEGF IRES-A and supports the translation initiation of the IRES.

Additionally, when the inserted NRAS sequence was placed in the hVEGF IRES-A and

subjected to RNase T1 footprinting in the presence of Li+ and K+, there is strong protection

seen in the G residues that were proposed to be involved in G-quadruplex formation by the

Balsubramanian group (17). Thus, the NRAS sequence, a known translation repressor turns

into a facilitator of translation by adopting a quadruplex structure when it is placed in a

foreign environment.

4.3.3 Replacing M3Q sequence with QVEG into the MT3-MMP 5'-UTR shows repression of translation

It has been previously established that QVEG is essential for translation in the context

of the IRES-A hVEGF mRNA. To determine its role on translation in the context of the

entire 5'-UTR of the MT3-MMP mRNA, the M3Q sequence was deleted from its parent

construct and replaced with the QVEG RNA quadruplex forming sequence. Additionally the

wild-type and mutated construct which contained point mutations that would abolish all

potential to form a RNA quadruplex structure were also transfected into HeLa cells and both

firefly and renilla luciferase activities were measured. The data were interpreted by dividing

the renilla signal by the firefly signal and all data were normalized by setting the mutant

construct equal to 100% active (Figure 5.4).

Figure 5.4 Histograms showing translation activity of endogenous M3Q G-quadruplex, the mutated M3Q sequence and the QVEG G-quadruplex forming sequence present individually within the 5'-UTR of MT3-MMP. The histograms represent the activity percentage, calculated using the ratio of Renilla activity to firefly activity in HeLa cells. The ratios were normalized to the wildtype (blue bar), which was set to 100%.

The average activity percentage for the plasmid containing the endogenous M3Q G-

quadruplex within the 5′-UTR MT3-MMP was found to be 57±3%. This datum compares

well with what has been previously reported (Figure 2.8) (25). The average activity

percentage for the plasmid containing the VEGF G-quadruplex inserted into the 5′-UTR of

MT3-MMP was found to be 50 ± 3%. The functional data implies imply that QVEG, normally essential for translation initiation, is now acting as a repressor of translation when located in the context of the entire 5′-UTR of the MT3-MMP. Because the activity of the engineered 5′-UTR MT3-MMP due to QVEG insertion is very similar to that of the wild- type, and he QVEG is known to from G-quadruplex, it is very possible that the QVEG sequence is forming a G-quadruplex within its new environment and thus repressing translation. The data suggest that function of the QVEG is also context-dependent.

5.4 CONCLUSIONS

The data presented suggest that the role G-quadruplexes play in translation initiation is context-dependent, that is, the RNA sequence environment in which the G-quadruplex is located will determine whether the G-quadruplex will be essential and a facilitator for translation or whether it will repress translation. Thus RNA domain swapping experiments can offer insights into functionality of a complex RNA structure.

CONCLUDING REMARKS

The determination of functional relevance of two G-quadruplex forming sequences from different mRNAs was carried out. In one case, it was shown that the quadruplex forming sequence is inhibitory toward translation in eukaryotic cell. In another, the quadruplex forming sequence was necessary for translation initiation and mutations that made it impossible to form such structure completely abolished all activity. It was once thought that the role of RNA G-quadruplex structures located in the 5′-UTR of mRNAs was completely a repressive one. However, we have shown that RNA G-quadruplexes can have a dual role as being both repressive and necessary for translation. This is a context dependent phenomenon. When the M3Q sequence was inserted in place or the QVEG sequence in the context of the hVEGF IRES-A mRNA it acted as an enhancer toward translation.

Furthermore, when the QVEG sequence was inserted in place of the M3Q sequence in the context of the entire 5′-UTR of MT3-MMP it acted as a repressor toward translation. We have also shown that RNA-quadruplex structures can be targeted with small molecules.

When the TmPyP4 was used to target the M3Q sequence both alone and in the context of the entire 5′-UTR we reported a destabilization of the structure and subsequent relief of the translational repression caused by this structure.

REFERENCES

1. Watson, J. D., and Crick, F. H. C. (1953) A Structure for Deoxyribose Nucleic Acid, Nature 171, 737-738. 2. Gellert, M., Lipsett, M. N., and Davies, D. R. (1962) Helix formation by guanylic acid, Proc. Natl. Acad. Sci. USA 48, 2013–2018. 3. Neidle, S., and Balasubramanian, S. (2006) Quadruplex Nucleic Acids, RSC Biomolecular Sciences. 4. Huppert, J. L., and Balasubramanian, S. (2005) Prevalence of quadruplexes in the human genome, Nucleic Acids Res. 33, 2908-2916. 5. Eddy, J., and Maizels, N. (2008) Conserved elements with potential to form polymorphic G- quadruplex structures in the first intron of human genes, Nucleic Acids Res. 36, 1321-1333. 6. Mani, P., Yadav, V. K., Das, S. K., and Chowdhury, S. (2009) Genome-Wide Analyses of Recombination Prone Regions Predict Role of DNA Structural Motif in Recombination, PLoS ONE 4, e4399. 7. Todd, A. K., Johnston, M., and Neidle, S. (2005) Highly prevalent putative quadruplex sequence motifs in human DNA, Nucleic Acids Res. 33, 2901-2907. 8. Stephen, N. (2009) The structures of quadruplex nucleic acids and their drug complexes, Curr. Opin. Struct. Biol. 19, 239-250. 9. Siddiqui-Jain, A., Grand, C. L., Bearss, D. J., and Hurley, L. H. (2002) Direct evidence for a G- quadruplex in a promoter region and its targeting with a small molecule to repress c-MYC transcription, Proc. Natl. Acad. Sci. USA 99, 11593-11598. 10. Simonsson, T., Pecinka, P., and Kubista, M. (1998) DNA tetraplex formation in the control region of c-myc, Nucleic Acids Res. 26, 1167–1172. 11. Dexheimer, T. S., Sun, D., and Hurley, L. H. (2006) Deconvoluting the structural and drug- recognition complexity of the G-Quadruplex-forming region upstream of the bcl-2 P1 promoter, J. Am. Chem. Soc. 128, 5404-5415. 12. Cogoi, S., and Xodo, L. E. (2006) G-quadruplex formation within the promoter of the KRAS proto-oncogene and its effect on transcription, Nucleic Acids Res. 34, 2536-2549. 13. Sun, D., Guo, K., Rusche, J. J., and Hurley, L. H. (2005) Facilitation of a structural transition in the polypurine/polypyrimidine tract within the proximal promoter region of the human VEGF gene by the presence of potassium and G-quadruplex-interactive agents, Nucleic Acids Res. 33, 6070–6080. 14. Cheong, C., and Moore, P. B. (1992) Solution structure of an unusually stable RNA tetraplex containing G- and U-quartet structures, Biochemistry 31, 8406-8414. 15. Kim, J., Cheong, C., and Moore, P. B. (1991) Tetramerization of an RNA oligonucleotide containing a GGGG sequence, Nature 351, 331-332. 16. Sacca, B., Lacroix, L. and Mergney, J.-L. (2005) The effect of chemical modifications on the thermal stability of different G quadruplex forming oligonucleotides., Nucleic Acids Res. 33, 1182-1192. 17. Kumari, S., Bugaut, A., Huppert, J. L., and Balasubramanian, S. (2007) An RNA G-quadruplex in the 5'-UTR of the NRAS proto-oncogene modulates translation, Nat. Chem. Biol. 3, 218- 221.

18. Liu, H., Matsugami, A., Katahira, M., and Uesugi, S. (2002) A dimeric RNA quadruplex architecture comprised of two G:G(:A):G:G(:A) hexads, G:G:G:G tetrads and UUUU loops, J. Mol. Biol. 322, 955-970. 19. Kostadinov, R., Malhotra, N., Viotti, M., Shine, R., D’Antonio, L., and Bagga, P. (2006) GRSDB: a database of quadruplex forming G-rich sequences in alternatively processed mammalian pre-mRNA sequences, Nucleic Acids Res. 34, D119–D124. 20. Sundquist, W. I., and Heaphy, S. (1993) Evidence for interstrand quadruplex formation in the dimerization of human immunodeficiency virus 1 genomic RNA, Proc. Natl. Acad. Sci. USA 90, 3393-3397. 21. Schaeffer, C., Bardoni, B., Mandel, J. L., Ehresmann, B., Ehresmann, C., and Moine, H. (2001) The fragile X mental retardation protein binds specifically to its mRNA via a purine quartet motif, EMBO J. 20, 4803-4813. 22. Bonnal, S., Schaeffer, C., Creancier, L., Clamens, S., Moine, H., Prats, A.-C., and Vagner, S. (2003) A single internal ribosome entry site containing a G quartet RNA structure drives fibroblast growth factor 2 gene expression at four alternative translation initiation codons, J. Biol. Chem. 278, 39330-39336. 23. Arora, A., Dutkiewicz, M., Scaria, V., Hariharan, M., Maiti, S., and Kurreck, J. (2008) Inhibition of translation in living eukaryotic cells by an RNA G-quadruplex motif, RNA 14, 1290-1296. 24. Christiansen, J., Kofod, M., and Nielsen, F. C. (1994) A guanosine quadruplex and two stable hairpins flank a major cleavage site in insulin-like growth factor II mRNA, Nucleic Acids Res. 22, 5709-5716. 25. Morris, M. J., and Basu, S. (2009) An unusually stable G-quadruplex within the 5'-UTR of the MT3 Matrix Metalloproteinase mRNA represses translation in eukaryotic cells, Biochemistry 48, 5313-5319. 26. Morris, M. J., Negishi, Y., Pazsint, C., Schonhoft, J. D., and Basu, S. (2010) An RNA G- Quadruplex Is Essential for Cap-Independent Translation Initiation in Human VEGF IRES, J. Am. Chem. Soc. 132, 17831-17839. 27. Gomez, D., Guedin, A., Mergny, J.-L., Salles, B., Riou, J.-F. o., Teulade-Fichou, M.-P., and Calsou, P. (2010) A G-quadruplex structure within the 5'-UTR of TRF2 mRNA represses translation in human cells, Nucleic Acids Res. 28. Wieland, M., and Hartig, J. S. (2007) RNA quadruplex-based modulation of gene expression, Chem. Biol. 14, 757-763. 29. Gray, N. K., and Hentze, M. W. (1994) Regulation of protein synthesis by mRNA structure, Molecular Biology Reports 19, 195-200. 30. Mignone, F., Gissi, C., Liuni, S., and Pesole, G. (2002) Untranslated regions of mRNAs, Genome Biology 3, reviews0004.0001 - reviews0004.0010. 31. Kumari, S., Bugaut, A., and Balasubramanian, S. (2008) Position and Stability Are Determining Factors for Translation Repression by an RNA G-Quadruplex-Forming Sequence within the 5′ UTR of the NRAS Proto-oncogene†, Biochemistry 47, 12664-12669. 32. Derecka, K., Balkwill, G. D., Garner, T. P., Hodgman, C., Flint, A. P. F., and Searle, M. S. (2010) Occurrence of a Quadruplex Motif in a Unique Insert within Exon C of the Bovine Estrogen Receptor alpha Gene (ESR1), Biochemistry 49, 7625-7633. 33. Shahid, R., Bugaut, A., and Balasubramanian, S. (2010) The BCL-2 5 ' Untranslated Region Contains an RNA G-Quadruplex-Forming Motif That Modulates Protein Expression, Biochemistry 49, 8300-8306.

34. Thomas, J. R., and Hergenrother, P. J. (2008) Targeting RNA with Small Molecules, Chemical Reviews 108, 1171-1224. 35. Arora, A., and Maiti, S. (2009) Differential Biophysical Behavior of Human Telomeric RNA and DNA Quadruplex, J. Phys. Chem. B 113, 10515-10520. 36. Bugaut, A., Rodriguez, R., Kumari, S., Hsu, S.-T. D., and Balasubramanian, S. (2010) Small molecule-mediated inhibition of translation by targeting a native RNAG-quadruplex Org. Biomol. Chem., 2771-2776. 37. Collie, G., Reszka, A. P., Haider, S. M., Gabelica, V., Parkinson, G. N., and Neidle, S. (2009) Selectivity in small molecule binding to human telomeric RNA and DNA quadruplexes, Chem. Commun., 7482-7484. 38. Halder, K., Largy, E., Benzler, M., Teulade-Fichou, M.-P., and Hartig, J. S. (2011) Efficient Suppression of Gene Expression by Targeting 5′-UTR-Based RNA Quadruplexes with Bisquinolinium Compounds, ChemBioChem 12, 1663-1668. 39. Thomas, J. R., and Hergenrother, P. J. (2008) Targeting RNA with Small Molecules, Chem. Rev. 108, 1171-1224. 40. Monchaud, D., and Teulade-Fichou, M.-P. (2008) A hitchhiker's guide to G-quadruplex ligands, Org. Biomol. Chem. 6, 627-636. 41. Werstuck, G., and Green, M. R. (1998) Controlling Gene Expression in Living Cells Through Small Molecule-RNA Interactions, Science 282, 296-298. 42. Egeblad, M., and Werb, Z. (2002) New functions for the matrix metalloproteinases in cancer progression, Nat. Rev. Cancer 2, 161-174. 43. McQuibban, G. A., Gong, J. H., Tam, E. M., McCulloch, C. A., Clark-Lewis, I. & Overall, C. M. (2000) Inflammation dampened by gelatinase A cleavage of monocyte chemoattractant protein-3., Science 289, 1202-1206. 44. Hotary, K., Li, X.Y., Allen, E., Stevens, S.L., & Weiss, S. J. (2006) A cancer cell metalloprotease triad regulates the basement membrane transmigration program, Gene Dev. 20, 2673-2686. 45. Lowy, A. M., Clements, W. M., Bishop, J., Kong, L., Bonney, T., Sisco, K., Aronow, B., Fenoglio-Preiser, C., and Groden, J. (2006) β-catenin/Wnt signaling regulates expression of the membrane type 3 matrix metalloproteinase in gastric cancer, Cancer Res. 66, 4734- 4741. 46. Jung, M., Romer, A., Keyszer, G., Lein, M., Kristiansen, G., Schnorr, D., Loening, S. A., and Jung, K. (2003) mRNA expression of the five membrane-type matrix metalloproteinases MT1-MT5 in human prostatic cell lines and their down-regulation in human malignant prostatic tissue, Prostate 55, 89-98. 47. Daja, M. M., Niu, X., Zhao, Z., Brown, J. M., and Russell, P. J. (2003) Characterization of expression of matrix metalloproteinases and tissue inhibitors of metalloproteinases in prostate cancer cell lines, Prostate Cancer P. D. 6, 15-26. 48. Kitagawa, Y. (1999) Expression of messenger RNAs for membrane-type 1, 2, and 3 matrix metalloproteinases in human renal cell carcinomas, J. Urology 162, 905. 49. Stetler-Stevenson, W. G. (1999) Matrix metalloproteinases in angiogenesis: a moving target for therapeutic intervention., J. Clin. Invest. 103, 1237-1241. 50. Plaisier, M., Kapiteijn, K., Koolwijk, P., Fijten, C., Hanemaaijer, R., Grimbergen, J. M., Mulder- Stapel, A., Quax, P. H. A., Helmerhorst, F. M., and van Hinsbergh, V. W. M. (2004) Involvement of membrane-type matrix metalloproteinases (MT-MMPs) in capillary tube formation by human endometrial microvascular endothelial cells: role of MT3-MMP, J. Clin. Endocrinol. Metab. 89, 5828-5836.

100

51. Milligan, J. F., Groebe, D. R., Witherell, G. W., and Uhlenbeck, O. C. (1987) Oligoribonucleotide synthesis using T7 RNA polymerase and synthetic DNA templates, Nucleic Acids Res. 15, 8783-8798. 52. Puglisi, J. D., Tinoco, I., James, E. D., and Abelson, J. N. (1989) Absorbance melting curves of RNA, In Methods in Enzymol., pp 304-325, Academic Press. 53. Marky, L. A., and Breslauer, K. J. (1987) Calculating thermodynamic data for transitions of any molecularity from equilibrium melting curves, Biopolymers 26, 1601. 54. Mergny, J.-L., and Lacroix, L. (2003) Analysis of thermal melting curves, Oligonucleotides 13, 515. 55. Nagatoishi, S., Tanaka, Y., and Tsumoto, K. (2007) Circular dichroism spectra demonstrate formation of the thrombin-binding DNA aptamer G-quadruplex under stabilizing-cation- deficient conditions, Biochem. Biophys. Res. Commun. 352, 812-817. 56. Mergny, J.-L., Phan, A.-T., and Lacroix, L. (1998) Following G-quartet formation by UV- spectroscopy, FEBS Letters 435, 74-78. 57. Rachwal, P. A., and Fox, K. R. (2007) Quadruplex melting, Methods 43, 291. 58. Gray, D., Liu, J., Ratliff, R., and Allen, F. (1981) Sequence dependence of the circular dichroism of synthetic double-stranded RNAs, Biopolymers 20, 1337-1382. 59. Davis, J. T. (2004) G-quartets 40 years later: from 5’-GMP to molecular biology and supramolecular chemistry, Angew. Chem. Int. Ed. 43, 668 - 698. 60. Risitano, A., and Fox, K. R. (2004) Influence of loop size on the stability of intramolecular DNA quadruplexes, Nucleic Acids Res. 32, 2598. 61. Rachwal, P. A., Brown, T., and Fox, K. R. (2007) Effect of G-tract length on the topology and stability of intramolecular DNA quadruplexes, Biochemistry 46, 3036-3044. 62. Rachwal, P. A., Brown, T., and Fox, K. R. (2007) Sequence effects of single base loops in intramolecular quadruplex DNA, FEBS Lett. 581, 1657-1660. 63. Darnell, J. C., Jensen, K. B., Jin, P., Brown, V., Warren, S. T., and Darnell, R. B. (2001) Fragile X mental retardation protein targets G quartet mRNAs important for neuronal function, Cell 107, 489-499. 64. Xu, Y., Kaminaga, K., and Komiyama, M. (2008) G-quadruplex formation by human telomeric repeats-containing RNA in Na+ solution, J. Am. Chem. Soc. 130, 11179-11184. 65. Tang, C. F., and Shafer, R. H. (2006) Engineering the quadruplex fold: nucleoside conformation determines both folding topology and molecularity in guanine quadruplexes J. Am. Chem. Soc. 128, 5966-5973. 66. Deng, J., Xiong, Y., and Sundaralingam, M. (2001) X-ray analysis of an RNA tetraplex 2+ (UGGGGU) 4 with divalent Sr ions at subatomic resolution (0.61 Ã), Proc. Natl. Acad. Sci. U. S. A. 98, 13665-13670. 67. Menon, L., and Mihailescu, M.-R. (2007) Interactions of the G quartet forming semaphorin 3F RNA with the RGG box domain of the fragile X protein family, Nucleic Acids Res. 35, 5379- 5392. 68. Marin, V. L., and Armitage, B. A. (2006) Hybridization of complementary and homologous peptide nucleic acid oligomers to a guanine quadruplex-forming RNA, Biochemistry 45, 1745-1754. 69. Zanotti, K. J., Lackey, P. E., Evans, G. L., and Mihailescu, M.-R. (2006) Thermodynamics of the fragile X mental retardation protein RGG box interactions with G quartet forming RNA, Biochemistry 45, 8319-8330.

101

70. Phan, A. T., Kuryavyi, V., Burge, S., Neidle, S., and Patel, D. J. (2007) Structure of an unprecedented G-quadruplex scaffold in the human c-kit promoter J. Am. Chem. Soc. 129, 4386-4392. 71. Gao, N., Li, Y. H., Li, X., Yu, S., Fu, G. L., and Chen, B. (2007) Effect of alpha-synuclein on the promoter activity of tyrosine hydroxylase gene, Neurosci. Bull. 23, 53-57. 72. Phan, A. T., Modi, Y. S., and Patel, D. J. (2004) Propeller-type parallel-stranded G- quadruplexes in the human c-myc promoter, J. Am. Chem. Soc. 126, 8710-8716. 73. Rachwal, P. A., Findlow, I. S., Werner, J. M., Brown, T., and Fox, K. R. (2007) Intramolecular DNA quadruplexes with different arrangements of short and long loops, Nucleic Acids Res. 35, 4214-4222. 74. Matsugami, A., Okuizumi, T., Uesugi, S., and Katahira, M. (2003) Intramolecular higher order packing of parallel quadruplexes comprising a G:G:G:G tetrad and a G(:A):G(:A):G(:A):G heptad of GGA triplet repeat DNA, J. Biol. Chem. 278, 28147-28153. 75. Dai, J., Punchihewa, C., Ambrus, A., Chen, D., Jones, R. A., and Yang, D. (2007) Structure of the intramolecular human telomeric G-quadruplex in potassium solution: a novel adenine triple formation, Nucleic Acids Res. 35, 2440-2450. 76. Kozak, M. (1991) Structural features in eukaryotic mRNAs that modulate the initiation of translation, J. Biol. Chem. 266, 19867-19870. 77. Brenet, F., Dussault, N., Delfino, C., Boudouresque, F., Chinot, O., Martin, P. M., and Ouafik, L. H. (2006) Identification of secondary structure in the 5'-untranslated region of the human adrenomedullin mRNA with implications for the regulation of mRNA translation, Oncogene 25, 6510-6519. 78. Hess, M. A., and Duncan, R. F. (1996) Sequence and structure determinants of Drosophila Hsp70 mRNA translation: 5'-UTR secondary structure specifically inhibits heat shock protein mRNA translation, Nucleic Acids Res. 24, 2441-2449. 79. Manzella, J. M., and Blackshear, P. J. (1990) Regulation of rat ornithine decarboxylase mRNA translation by its 5'- untranslated region, J. Biol. Chem. 265, 11817-11822. 80. Zuker, M., Jaeger, J. A., and Turner, D. H. (1991) A comparison of optimal and suboptimal RNA secondary structures predicted by free energy minimization with structures determined by phylogenetic comparison, Nucleic Acids Res. 19, 2707-2714. 81. de Armond, R., Wood, S., Sun, D., Hurley, L. H., and Ebbinghaus, S. W. (2005) Evidence for the presence of a Guanine quadruplex forming region within a polypurine tract of the Hypoxia Inducible Factor 1 alpha promoter, Biochemistry 44, 16341–16350. 82. Guo, K., Pourpak, A., Beetz-Rogers, K., Gokhale, V., Sun, D., and Hurley, L. H. (2007) Formation of Pseudosymmetrical G-Quadruplex and i-Motif Structures in the Proximal Promoter Region of the RET Oncogene, J. Am. Chem. Soc. 129, 10220-10228. 83. Qin, Y., Rezler, E. M., Gokhale, V., Sun, D., and Hurley, L. H. (2007) Characterization of the G- quadruplexes in the duplex nuclease hypersensitive element of the PDGF-A promoter and modulation of PDGF-A promoter activity by TMPyP4, Nucleic Acids Res. 35, 7698-7713. 84. Sun, D., Liu, W.-J., Guo, K., Rusche, J. J., Ebbinghaus, S., Gokhale, V., and Hurley, L. H. (2008) The proximal promoter region of the human vascular endothelial growth factor gene has a G-quadruplex structure that can be targeted by G-quadruplex interactive agents, Mol. Cancer Ther. 7, 880-889. 85. Ren, J., and Chaires, J. B. (1999) Sequence and structural selectivity of nucleic acid binding ligands, Biochemistry 38, 16067-16075.

102

86. Luedtke, N. W. (2009) Targeting G-quadruplex DNA with small molecules, CHIMIA 63, 134- 139. 87. Weisman-Shomer, P., Cohen, E., Hershco, I., Khateb, S., Wolfovitz-Barchad, O., Hurley, L. H., and Fry, M. (2003) The cationic porphyrin TMPyP4 destabilizes the tetraplex form of the fragile X syndrome expanded sequence d(CGG)n, Nucleic Acids Res. 31, 3963-3970. 88. Pasternack, R. F., Huber, P. R., Boyd, P., Engasser, G., Francesconi, L., Gibbs, E., Fasella, P., Cerio Venturo, G., and Hinds, L. d. (1972) Aggregation of meso-substituted water-soluble porphyrins, J. Am. Chem. Soc. 94, 4511-4517. 89. Keating, L. R., and Szalai, V. A. (2004) Parallel-Stranded Guanine Quadruplex Interactions with a Copper Cationic Porphyrinâ€ Biochemistry 43, 15891-15900. 90. Piotto, M., Saudek, V., and Sklenář, V. (1992) Gradient-tailored excitation for single- quantum NMR spectroscopy of aqueous solutions, J. Biomol. NMR 2, 661-665. 91. Dash, J., Shirude, P. S., Hsu, S.-T. D., and Balasubramanian, S. (2008) Diarylethynyl Amides That Recognize the Parallel Conformation of Genomic Promoter DNA G-Quadruplexes, J. Am. Chem. Soc. 130, 15950-15956. 92. Waller, Z. A. E., Sewitz, S. A., Hsu, S.-T. D., and Balasubramanian, S. (2009) A Small Molecule That Disrupts G-Quadruplex DNA Structure and Enhances Gene Expression, J. Am. Chem. Soc. 131, 12628-12633. 93. Flynn, S., George, S., White, L., Devonish, W., and Takle, G. (1999) Water-soluble, meso- substituted cationic porphyrins--a family of compounds for cellular delivery of oligonucleotides., Biotechniques 26, 736-742. 94. Ofer, N., Weisman-Shomer, P., Shklover, J., and Fry, M. (2009) The quadruplex r(CGG)n destabilizing cationic porphyrin TMPyP4 cooperates with hnRNPs to increase the translation efficiency of fragile X premutation mRNA, Nucleic Acids Res., gkp130. 95. Millet, O., Loria, J. P., Kroenke, C. D., Pons, M., and Palmer, A. G. (2000) The Static Magnetic Field Dependence of Chemical Exchange Linebroadening Defines the NMR Chemical Shift Time Scale, J. Am. Chem. Soc. 122, 2867-2877. 96. Fielding, L. (2003) NMR methods for the determination of protein-ligand dissociation constants, Curr Top Med Chem. 3, 39-53. 97. Johnson, C. E., and Bovey, F. A. (1958) Calculation of Nuclear Magnetic Resonance Spectra of Aromatic Hydrocarbons, 29, 1012-1014. 98. Perkins, S. J., and Wüthrich, K. (1979) Ring current effects in the conformation dependent NMR chemical shifts of aliphatic protons in the basic pancreatic trypsin inhibitor, Biochim. Biophys. Acta. 576, 409-423. 99. Case, D. A. (1995) Calibration of ring-current effects in proteins and nucleic acids, J. Biomol. NMR 6, 341-346. 100. Wheelhouse, R. T., Sun, D., Han, H., Han, F. X., and Hurley, L. H. (1998) Cationic Porphyrins as Telomerase Inhibitors: the Interaction of Tetra-(N-methyl-4-pyridyl)porphine with Quadruplex DNA, J. Am. Chem. Soc. 120, 3261-3262. 101. Yamashita, T., Uno, T., and Ishikawa, Y. (2005) Stabilization of guanine quadruplex DNA by the binding of porphyrins with cationic side arms, Bioorg. Med. Chem. 13, 2423-2430. 102. Iverson, B. L., Shreder, K., Král, V., Smith, D. A., Smith, J., and Sessler, J. L. (1994) Interactions between expanded porphyrins and nucleic acids, Pure Appl. Chem. 66, 845-850. 103. Ghazaryan, A. A., Dalyan, Y. B., Haroutiunian, S. G., Tikhomirova, A., Taulier, N., Wells, J. W., and Chalikian, T. V. (2006) Thermodynamics of Interactions of Water-Soluble Porphyrins with RNA Duplexes, J. Am. Chem.Soc. 128, 1914-1921.

103

104. Parkinson, G. N., Ghosh, R., and Neidle, S. (2007) Structural Basis for Binding of Porphyrin to Human Telomeresâ€ ,â€¡, Biochemistry 46, 2390-2397. 105. Shkirman, S., Solov’ev, K., Kachura, T., Arabei, S., and Skakovskii, E. (1999) Interpretation of the soret band of porphyrins based on the polarization spectrum of N- methyltetraphenylporphin fluorescence, J. Appl. Spectros. 66, 68-75. 106. Fernando, H., Reszka, A. P., Huppert, J., Ladame, S., Rankin, S., Venkitaraman, A. R., Neidle, S., and Balasubramanian, S. (2006) A Conserved Quadruplex Motif Located in a Transcription Activation Site of the Human c-kit Oncogene, Biochemistry 45, 7854-7860. 107. Hellen, C. U. T., and Sarnow, P. (2001) Internal ribosome entry sites in eukaryotic mRNA molecules, Gene. Dev. 15, 1593-1612. 108. Gilbert, W. V., Zhou, K., Butler, T. K., and Doudna, J. A. (2007) Cap-Independent Translation Is Required for Starvation-Induced Differentiation in Yeast, Science 317, 1224-1227. 109. Macejak, D. G., and Sarnow, P. (1991) Internal initiation of translation mediated by the 5' leader of a cellular mRNA, Nature 353, 90-94. 110. Jang, S. K., Krausslich, H. G., Nicklin, M. J., Duke, G. M., Palmenberg, A. C., and Wimmer, E. (1988) A segment of the 5' nontranslated region of encephalomyocarditis virus RNA directs internal entry of ribosomes during in vitro translation, J. Virol. 62, 2636-2643. 111. Pelletier, J., and Sonenberg, N. (1988) Internal initiation of translation of eukaryotic mRNA directed by a sequence derived from poliovirus RNA, Nature 334, 320-325. 112. Pestova, T. V., Kolupaeva, V. G., Lomakin, I. B., Pilipenko, E. V., Shatsky, I. N., Agol, V. I., and Hellen, C. U. T. (2001) Molecular mechanisms of translation initiation in eukaryotes, Proc. Natl. Acad. Sci. USA 98, 7029-7036. 113. Pfingsten, J. S., and Kieft, J. S. (2008) RNA structure-based ribosome recruitment: Lessons from the Dicistroviridae intergenic region IRESes, RNA 14, 1255-1263. 114. Kieft, J. S. (2008) Viral IRES RNA structures and ribosome interactions, Trends. Biochem. Sci. 33, 274-283. 115. Akiri, G., Nahari, D., Finkelstein, Y., Le, S. Y., Elroy-Stein, O., and Levi, B. Z. (1998) Regulation of vascular endothelial growth factor (VEGF) expression is mediated by internal initiation of translation and alternative initiation of transcription, Oncogene 17, 227-236. 116. Huez, I., Creancier, L., Audigier, S., Gensac, M. C., Prats, A. C., and Prats, H. (1998) Two independent internal ribosome entry sites are involved in translation initiation of vascular endothelial growth factor mRNA, Mol. Cell. Biol. 18, 6178-6190. 117. Vagner, S., Waysbort, A., Marenda, M., Gensac, M.-C., Amalric, F., and Prats, A.-C. (1995) Alternative Translation Initiation of the Moloney Murine Leukemia Virus mRNA Controlled by Internal Ribosome Entry Involving the p57/PTB Splicing Factor, J. Biol. Chem. 270, 20376- 20383. 118. Nanbru, C., Lafon, I., Audigier, S., Gensac, M.-C., Vagner, S., Huez, G., and Prats, A.-C. (1997) Alternative Translation of the Proto-oncogene c-myc by an Internal Ribosome Entry Site, J. Biol. Chem. 272, 32061-32066. 119. Stoneley, M., Paulin, F. E., Le Quesne, J. P., Chappell, S. A., and Willis, A. E. (1998) C-Myc 5' untranslated region contains an internal ribosome entry segment, Oncogene 16, 423-428. 120. Bernstein, J., Sella, O., Le, S.-Y., and Elroy-Stein, O. (1997) PDGF2/c-sis mRNA Leader Contains a Differentiation-linked Internal Ribosomal Entry Site (D-IRES), J. Biol. Chem. 272, 9356-9362. 121. Winkler, W., Nahvi, A., and Breaker, R. R. (2002) Thiamine derivatives bind messenger RNAs directly to regulate bacterial gene expression, Nature 419, 952-956.

104

122. Serganov, A. (2009) The long and the short of riboswitches, Curr. Opin. Struct. Biol. 19, 251- 259. 123. Garst, A. D., and Batey, R. T. (2009) A switch in time: Detailing the life of a riboswitch, BBA- Gene. Regul. Mech. 1789, 584-591. 124. Holbrook, S. R. (2005) RNA structure: the long and the short of it, Curr. Opin. Struct. Biol. 15, 302-308. 125. Ray, P. S., Jia, J., Yao, P., Majumder, M., Hatzoglou, M., and Fox, P. L. (2009) A stress- responsive RNA switch regulates VEGFA expression, Nature 457, 915-919. 126. Harmey, J. (2004) VEGF and Cancer, Springer. 127. Miller, D. L., Dibbens, J. A., Damert, A., Risau, W., Vadas, M. A., and Goodall, G. J. (1998) The vascular endothelial growth factor mRNA contains an internal ribosome entry site, Febs Lett. 434, 417-420. 128. Shweiki, D., Itin, A., Soffer, D., and Keshet, E. (1992) Vascular endothelial growth factor induced by hypoxia may mediate hypoxia-initiated angiogenesis, Nature 359, 843-845. 129. Banai, S., Shweiki, D., Pinson, A., Chandra, M., Lazarovici, G., and Keshet, E. (1994) Upregulation of vascular endothelial growth factor expression induced by myocardial ischaemia: implications for coronary angiogenesis, Cardiovasc. Res. 28, 1176-1179. 130. Huez, I., Bornes, S., Bresson, D., Creancier, L., and Prats, H. (2001) New vascular endothelial growth factor isoform generated by internal ribosome entry site-driven CUG translation initiation, Mol. Endocrinol. 15, 2197-2210. 131. Bornes, S. p., Boulard, M., Hieblot, C., Zanibellato, C., Iacovoni, J. S., Prats, H., and Touriol, C. (2004) Control of the Vascular Endothelial Growth Factor Internal Ribosome Entry Site (IRES) Activity and Translation Initiation by Alternatively Spliced Coding Sequences, J. Biol. Chem. 279, 18717-18726. 132. Stein, I., Itin, A., Einat, P., Skaliter, R., Grossman, Z., and Keshet, E. (1998) Translation of Vascular Endothelial Growth Factor mRNA by Internal Ribosome Entry: Implications for Translation under Hypoxia, Molecular and Cellular Biology 18, 3112-3119. 133. Cobbold, L. C., Spriggs, K. A., Haines, S. J., Dobbyn, H. C., Hayes, C., de Moor, C. H., Lilley, K. S., Bushell, M., and Willis, A. E. (2008) Identification of internal ribosome entry segment (IRES)-trans-acting factors for the Myc family of IRESs, Mol. Cell. Biol. 28, 40-49. 134. Kieft, J. S., Zhou, K., Jubin, R., and Doudna, J. A. (2001) Mechanism of ribosome recruitment by hepatitis C IRES RNA, RNA 7, 194-206. 135. Peattie, D. A. (1979) Direct chemical method for sequencing RNA, Proc. Natl. Acad. Sci. USA 76, 1760-1764. 136. Balkwill, G. D., Derecka, K., Garner, T. P., Hodgman, C., Flint, A. P. F., and Searle, M. S. (2009) Repression of Translation of Human Estrogen Receptor Î± by G-Quadruplex Formation, Biochemistry 48, 11487-11495. 137. Dai, J., Dexheimer, T. S., Chen, D., Carver, M., Ambrus, A., Jones, R. A., and Yang, D. (2006) An Intramolecular G-Quadruplex Structure with Mixed Parallel/Antiparallel G-Strands Formed in the Human BCL-2 Promoter Region in Solution, Journal of the American Chemical Society 128, 1096-1098. 138. Phan, A. T., Modi, Y. S., and Patel, D. J. (2004) Propeller-Type Parallel-Stranded G- Quadruplexes in the Human c-myc Promoter, Journal of the American Chemical Society 126, 8710-8716. 139. Xia, X., and Holcik, M. (2009) Strong Eukaryotic IRESs Have Weak Secondary Structure, PLoS ONE 4, e4136.

105

140. Hong, H.-J. (2010) Studying Gene Induction of Glycopeptide Resistance Using Gene Swapping

Antibiotic Resistance Protocols, (Gillespie, S. H., and McHugh, T. D., Eds.), pp 45-62, Humana Press. 141. Chen, J.-L., and Greider, C. W. (2005) Functional analysis of the pseudoknot structure in human telomerase RNA, Proc Natl. Acad. Sci. USA 102, 8080-8085.