G-quadruplex secondary structures are four-stranded globular nucleic acid structures that form in specific DNA and RNA G-rich sequences with biological significance, such as those found in human telomeres, oncogene promoter regions, replication initiation sites, and 5’- and 3’-untranslated (UTR) regions, which have been identified as novel drug targets. The non-canonical G-quadruplex secondary structures readily form under physiologically relevant ionic conditions, and exhibit great diversity in their topologies and loop conformations depending on the DNA or RNA sequences at hand. The structural diversity of these unique secondary structures is essential to their specific recognition by different regulatory proteins or small molecule compounds. A significant amount of research has been done in this field that provides compelling evidence for the existence, biological significance, and potential druggability of G-quadruplexes. In this dissertation, I explore G-quadruplex formation in the promoters of BCL2, PDGFR-β and c-Myc oncogenes and their interactions with small molecule compounds or proteins. Firstly, I investigated a newly-identified G-quadruplex (P1G4) forming immediately upstream of the human BCL2 gene, which has been found to be overexpressed in several human tumors. In this research, I have found that P1G4 acts as a transcription repressor, and that its inhibitory effect can be enriched by the G-quadruplex-interactive compound, TMPyP4. Both P1G4 and the previously reported Pu39 G-quadruplexes form independently in adjacent regions within the BCL2 P1 promoter, but P1G4 appears to play a more dominant role in repressing transcriptional activity. NMR and CD studies have shown that the P1G4 G-quadruplex appears to comprise a novel dynamic equilibrium of two parallel structures, one regular, with two 1-nt loops and a 12-nt middle loop, and another broken-stranded, with three 1-nt loops and an 11-nt middle loop; both structures adopt a novel hairpin (stem-loop duplex) conformation in the long central loop. This dynamic equilibrium of two closely-related G-quadruplex structures with a unique hairpin loop conformation may provide a specific target for small molecules to modulate BCL2 gene transcription. I also explored the 3’ end G-quadruplex that forms within the core promoter of PDGFR-β, which has also been observed to be present at abnormal levels in a variety of clinical pathologies, including malignancies. The 3′-end G-quadruplex formed in the PDGFR-β promoter NHE appears to be selectively stabilized by an ellipticine analog, GSA1129, which can shift the dynamic equilibrium in the full-length sequence to favor the 3′-end G-quadruplex, and can repress PDGFR-β activity in cancer cell lines. NMR studies in combination with biophysical experiments have shown that in the wild-type extended 3ʼ-end NHE sequences, two novel intramolecular G-quadruplexes can be formed in a potassium solution, one with a 3’-flanking distant guanine inserted into the 3’-external tetrad (3’-insertion G-quadruplex), and another with a 5’-flanking distant guanine inserted into the 5’-external tetrad (5’-insertion G-quadruplex). Further investigation of the elongated PDGFR-β 3′-end sequence containing both the 5’- and 3’- flanking guanine sequences showed the formation of a combination of the two G-quadruplexes existing in equilibrium. Importantly, it was observed that GSA1129 can bind to and increase the stability of each of the end-insertion G-quadruplexes, raising their Tₘ by 25 degrees. This study highlights the dynamic nature of the 3′-end NHE sequence and the importance of identifying the proper sequence for the formation of biologically relevant G-quadruplex structures. Significantly, the dynamic nature of the 3′-end G-quadruplex suggests that it may be an attractive target for drug regulation. I then analyzed two proteins, Nucleolin and NM23-H2, which interact with the c-Myc G-quadruplex structure that forms in the proximal promoter region of the c-Myc gene; this is one of the most commonly deregulated genes in the human neoplasm. Nucleolin is known to be a transcriptional repressor for c-Myc, binding to and stabilizing the c-Myc G-quadruplex, whereas NM23-H2 is known to be a transcriptional activator that unwinds and destabilizes the c-Myc G-quadruplex. An investigation of the molecular mechanisms of the interaction between the c-Myc G-quadruplex and nucleolin showed that the minimal binding domains required for a tight binding of the protein to the c-Myc G-quadruplex are the four RNA binding domains (RBDs) of nucleolin, referred to as Nuc1234, and that the RGG domain is unnecessary for c-Myc G-quadruplex binding. The stable G-quadruplex formed within Pu27 using G-tract runs I, II, IV and V was determined to be the best substrate (Myc1245T) for nucleolin binding, showing the highest affinity. 3D NMR experiments performed on the free protein Nuc1234 and its complex with the Myc1245T G-quadruplex have shown that upon complex formation, only the disordered linker regions of the protein display significant chemical shift changes, whereas most other residues show chemical shift values similar to those of the free protein. The c-Myc G-quadruplex has three loops that flip outward in a solvent containing K⁺, according to its structure. The hypothesis for this association is that nucleolin wraps around the G-quadruplex and interacts specifically with the flipped-outward loop regions of the c-Myc G-quadruplex via its own inter-RBD linker regions, with little structural change in the RBDs themselves. A definitive determination of the 3D molecular structure of nucleolin and its complex with Myc1245T is currently in development. Biophysical and structural studies were then conducted to investigate the interactions of the protein NM23-H2/NDP kinase B with the c-Myc G-quadruplex. NM23-H2 binds to single-stranded guanine- and cytosine-rich sequences, but not to double-stranded DNA in the NHE III₁ region; the binding therefore appears structure-specific, rather than sequence-specific. Moreover, increasing concentrations of the strong G-quadruplex-interactive compound TMPyP4, a porphyrin-based drug, inhibits the binding of NM23-H2 to the NHE III₁ region; this suggests that the stabilization of the G-quadruplex hinders the recognition and remodeling function of the NM23-H2. By conducting Forster Resonance Energy Transfer (FRET) assays in combination with Circular Dichroism (CD) studies, I demonstrated that NM23-H2 can actively resolve the c-Myc G-quadruplex. Taken together, these results show that the use of small molecules to prevent NM23-H2 from binding to and resolving the NHE III₁ region G-quadruplex may have the potential to inhibit c-Myc transcription for cancer therapeutic purposes. This underlines the importance of understanding the mechanism of function operating between NM23-H2 and the c-Myc G-quadruplex. Understanding molecular mechanism between NM23-H2 and c-Myc is under investigation.

Defective DNA repair capacity has been shown to be a common feature of cancer, and loss of function mutations in 'stability' genes that normally maintain the integrity of the genome may prove a key rate-limiting step in carcinogenesis. Since even genetically unstable cells require some repair functionality to maintain viability, these cancers likely exhibit an over-reliance on other DNA repair pathways for survival. Therapeutically targeting backup repair processes in such tumours represents a novel means by which to achieve selective tumour toxicity. Full exploitation of these synthetic lethal interactions will require an in-depth knowledge of the genetic basis of DNA repair in combination with an armoury of small molecule inhibitors of cellular targets. To this end, we have designed, optimised and run two high-throughput cell-based screens to identify genes and small molecules that can modulate mismatch repair (MMR) activity. Key to these screening strategies are the resistance of cells with dysfunctional MMR to a range of cytotoxic drugs, including the alkylating agent N-methyl-N'-nitro-N-nitrosoguanidine (MNNG). By exploiting this MMR-dependent toxicity we have assayed for siRNA and small molecules that permit the survival of MNNG-treated MMR-proficient cells to levels comparable to MMR-deficient cells, and which therefore represent putative MMR modulating agents. A screen of 571 siRNA for gene depletions that reduce MNNG sensitivity by at least two population standard deviations identified 10 genes of potential interest, and included the four canonical MMR genes, MSH2 (2.87 ± 0.28 (Z ± SE)), MSH6 (4.87 ± 0.06), MLH1 (3.42 ± 0.43) and PMS2 (3.36 ± 0.44). TDG represented an unexpected hit that decreased MNNG sensitivity by 2.55 ± 0.04 population standard deviations. However, clonogenic survival experiments found TDG depletion to be contextual synthetic lethal within an MMR-null background when treated with MNNG, reducing HCT116 clonogenicity by 37% (p < 0.001). Moreover, TDG knockdown increased the number of 53 binding protein 1 (53bp1) foci in MMR-proficient cells by 40% and MMR–deficient cells by 27% following MNNG exposure (p < 0.001). Combined with a failure to replicate the primary screen result, the role for TDG in the response to MNNG could be explained solely through its established role as a member of the base excision repair pathway. A second screen of the NCI Diversity I and II small molecule libraries (n=1786) was conducted to identify putative MMR inhibitors. Subsequent analysis revealed NSC197049 to increase cellular viability of MNNG treated cells by 3.60±0.32 population standard deviations and was successfully validated as a hit. Co-treatment of NSC197049 with MNNG conferred dose-dependent chemoprotection independently of MMR status and cell line, an effect that was lost if NSC197049 was pre- or post-treated. The protection was associated with a reduction in MNNG-dependent 53bp1 foci of 60% in MMR proficient cells and 15% in MMR deficient cells (p < 0.001), together with a marked reduction of > 80% in subG1 content at 48 hours post-MNNG that was independent of MMR status. Interestingly, the characteristic G2/M arrest of MNNG-treated MMR-proficient cells remained intact (~40% arrested). Taken together, these observations are not consistent with NSC197049 acting as an inhibitor of MMR. Dithiolthiones have been described as chemoprotective agents that induce antioxidant defences, whilst we have found NSC197049 phenocopies known antioxidants ascorbic acid and glutathione in protecting against MNNG-induced toxicity. NSC197049 may therefore act by bolstering cellular antioxidant defences. The precise mechanism may be novel, since the proto-typical dithiolthione, Oltipraz, failed to be protective in this study. In summary, we have confirmed that MMR is the primary determinant of MNNG sensitivity, and found that TDG is unlikely to be involved in MMR. We have also identified a novel chemoprotective small molecule that is unlikely to represent an MMR inhibitor, but that might be useful in cancer chemoprevention.

Gene regulation is vital to the success of all living organisms. Understanding this complex process is crucial to our knowledge of how cells function and how in some cases they can lead to debilitating or even fatal disease. In this thesis I focus on a set of DNA-binding proteins known as transcription factors (TFs), proteins fundamental to the process of gene regulation at the level of transcription. I develop assays and techniques for the detection and quantitation of TFs in vitro and in vivo as well as a method for TF encapsulation and release. The advantages of the TF detection assays in this thesis are made possible through the use of single-molecule (sm) fluorescence. This methodology enables detection of individually labeled molecules allowing discrimination of sample heterogeneities inaccessible with ensemble techniques. Here I present two different TF assays based on two sm observables: relative probe stoichiometry and Förster resonance energy transfer (FRET). The first assay design, based on stoichiometry, detects TFs using TF-dependent coincidence of two distinctly labelled DNA ‘half-sites’. I demonstrate sensitive detection (~ pM) in solution and on surfaces, multiplexed detection of multiple TFs, and detection in cell lysates. A kinetic model of the system is also developed, verified experimentally and used to quantify TF concentrations without the need for a calibration curve. The second assay design, based on FRET, is a novel approach to TF detection using TFmediated DNA bending. TFs are detected by bending the sensor and monitored with FRET at the single-molecule or ensemble level. I demonstrate TF detection in purifed form and expressed in cell lysates. As this sensor was designed for use in vivo, methods to hinder nuclease degradation are explored. For TF detection in vivo, I describe a successful strategy to internalise fluorescently labeled molecules into live E.coli. Viability and internalisation efficiency are characterised and ensemble measurements with FRET standards are demonstrated. Importantly, sm FRET measurements in vivo are achieved opening many exciting possibilities. The FRET based TF sensor is then internalised as a step towards real-time in vivo monitoring of TF concentrations. Finally a system based on DNA nanotechnology is presented for the non-covalent encapsulation and release of TFs. Such a system could be delivered into a cell to alter levels of gene expression using external stimuli as inputs. We believe these tools will generate valuable information in the study of prokaryotic gene expression as well as providing a potential commercial avenue towards diagnostics.

The homologous recombination (HR) pathway of DNA repair is essential for the faithful repair of double-stranded DNA breaks (DSBs) in all organisms and as such helps maintain genomic stability. Furthermore, HR is instrumental in the cellular response to exogenous DNA damaging agents such as those used in the clinic for chemo- and radiotherapy. HR in humans is a complex, incompletely understood process involving numerous stages and diverse biochemical activities. Advancing our knowledge of the HR pathway in humans aids the understanding of how chemo- and radiotherapies act and may be used to develop novel therapeutic strategies. Recent studies have identified inhibition of HR as one of the mechanisms via which a number of recently developed chemotherapeutics have their effect. Accordingly, the clinical potential of HR inhibitors is under investigation. My work has centred around the identification of both novel HR proteins and novel, small molecule HR inhibitors. To further these aims, I have successfully employed high-throughput RNAi and small molecule screening strategies. RNAi screens are commonly used to identify genes involved in a given cellular process via genetic loss of function, whilst small molecule, cell based screens are a powerful tool in the drug discovery process.

Single-molecule detection has provided insights into how molecules behave. Without the averaging effect of ensemble measurements, the stochastic behaviour of single molecules can be observed and intermediate steps in multistep transformations can be clearly detected. The single-molecule reactants range from small molecules (e.g. propene) to proteins of several tens of kDa (e.g. myosin). One single-molecule detection technique is single-channel electrical recording. This approach is based on the measurement of the transmembrane ionic current flowing through a nanoscale transmembrane pore under an applied potential. In this thesis, the protein α-hemolysin was employed as a nanoreactor. α-Hemolysin is a toxin secreted by Staphylococcus aureus. Its transmembrane pore (~100 Å in length and ≥14 Å in diameter) allows ions, water and small molecules to pass through its lumen. Under an applied potential, chemical changes in reactants attached to the internal wall of the pore modulate the flow of ions, leading to changes in the transmembrane ionic current. Analysis of this current provides information about the reaction kinetics and mechanisms. Chapter 1 – Single-Molecule Chemistry and α-Hemolysin is an introductory chapter that is divided into two parts. Section 1.1 provides an overview of the different techniques for the detection of chemical reactions at the single-molecule level. Section 1.2 gives a brief review of the protein pore α-hemolysin, including its structure, properties and various applications. Chapter 2 – S-Nitrosothiol Chemistry applies cysteine-containing α-hemolysins to study the biologically relevant chemistry of S-nitrosothiols (RSNO). RSNO are important molecules involved in cell signalling, which control physiological processes such as vasodilation and bronchodilation. Three reactions, namely transnitrosation (the transfer of the ‘NO’ group from RSNO to a thiol), S-thiolation (the formation of a disulfide from RSNO and thiol) and S-sulfonation (the generation of an S-sulfonate (RSSO₃⁻) from RSNO and sulfite ion), were investigated at the single-molecule level. The pH-dependency of the two competing reactions (transnitrosation and S-thiolation), the lifetime of the proposed transnitrosation intermediate, and nature of the chemical reaction between RSNO and sulfite (a bronchoconstrictor) were determined. Chapter 3 – Silver(I)-thiolate and cadmium(II)-thiolate complexes describes the kinetics of the formation and breakdown of these two metal-thiolate complexes. Ag⁺ and Cd²⁺ are commonly used in probing the membrane topology and gating properties of ion channels using the scanning cysteine accessibility method (SCAM). The binding of two Ag⁺ ions per thiol group and the stepwise build-up and dissociation of Cd²⁺-glutathione complexes were unambiguously characterized. Chapter 4 – Copper(II)-Catalyzed Diels-Alder Reactions reports the attempt to carry out copper(II)-catalyzed Diels-Alder reactions inside an engineered α-hemolysin. An iminodiacetate ligand was covalently attached within the lumen of the α-hemolysin pore. This ligand chelates Cu²⁺ ion, which can bind bidentate dienophiles and activate them towards Diels-Alder reaction with dienes. However, due to the ‘slow’ reaction rate of the Diels-Alder reaction (rate constant ~10⁻¹ M⁻¹s⁻) relative to the time-scale of the single-molecule experiment, we failed to observed chemical conversion at the single-molecule level. Nevertheless, the engineered metal-binding α-hemolysin may be useful for sensing molecules bearing metal-coordinating groups.

mRNA localization ensures correct spatial and temporal control of protein synthesis in the cell. Using a single molecule in vitro approach, we provide insight into the mechanisms by which localizing mRNAs are carried by molecular motors on cytoskeletal tracks to their destination. Budding yeast serves as a model system for studying the mechanisms of mRNA transport because localizing mRNAs are moved on actin tracks in the cell by a single class V myosin motor, Myo4p. Molecular motors that specialize in cargo transport are generally double-headed so that they can "walk" for many microns without dissociating, a feature known as processivity. Thus, is was surprising when Myo4p purified from yeast was shown by in vitro assays to be non-processive. The reason for its inability to move processively is that the Myo4p heavy chain does not dimerize with itself, but instead binds tightly to the adapter protein She3p to form a single-headed motor complex. The mRNA-binding adapter protein She2p links Myo4p to mRNA cargo by binding She3p. To understand the molecular mechanisms of mRNA transport in budding yeast, we fully reconstituted a messenger ribonucleoprotein (mRNP) complex from purified proteins and a localizing mRNA (ASH1) found in budding yeast. Using single molecule in vitro assays, we find that She2p recruits two Myo4p-She3p complexes, forming a processive double-headed motor complex that is stabilized by mRNA at physiological ionic strength. Thus, only in the presence of mRNA is Myo4p capable of continuous mRNA transport, an elegant mechanism that ensures that only cargo bound motors are motile. We next wished to understand if the principles of mRNA transport in budding yeast are conserved in higher eukaryotes. In Drosophila, mRNA is transported on microtubule tracks by cytoplasmic dynein, and the adapters that link the motor to localizing transcripts are well-defined. The adapter protein bicaudal D (BicD) coordinates dynein motor activity with mRNA cargo binding. The N-terminus of BicD binds dynein, and the C-terminus interacts with the mRNA-binding protein Egalitarian. Unlike mammalian dynein alone, it was recently shown that an N-terminal fragment of BicD (BicD2CC1), in combination with a large 1.2MDa multi-subunit accessory complex called dynactin, forms a complex (DDBCC1) that is activated for long processive runs. But unlike the constitutively activated BicD2CC1 fragment, the full-length BicD molecule fails to recruit dynein-dynactin because it is auto-inhibited by interactions between the N-terminal dynein binding domain and the C-terminal cargo binding domain. To understand how dynein is activated by native cargo and full-length adapters, we fully reconstituted a mRNP complex in vitro from tissue-purified dynein and dynactin, expressed full-length adapters BicD and Egalitarian, and a synthesized localizing mRNA found in Drosophila. We find that only mRNA-bound Egalitarian is capable of relieving BicD auto-inhibition for the recruitment of dynein-dynactin, and activation of mRNA transport in vitro. Thus, the presence of an mRNA cargo for activation of motor complexes is a conserved mechanism in both budding yeast and higher eukaryotes to ensure that motor activity is tightly coupled to cargo selection.

Doctor of Philosophy / Department of Chemistry / Daniel A. Higgins / This dissertation describes the application of single molecule spectroscopy and tracking to investigations of the nanoscale properties of thin-film chemical gradients and the transport dynamics of molecules dispersed within and upon these gradients. Chemical gradients are surface bound materials that incorporate gradually changing chemical and/or physical properties. A continuous and gradual change in the properties of gradients are expected and often required for their intended applications, which range from directed growth of cell colonies to combinatorial materials science. In reality, such conditions are almost never met due to spontaneous demixing and dewetting processes that can lead to properties variations on microscopic length scales. A better understanding on the properties of chemical gradients on microscopic length scales will aid in the production of better engineered materials. Single molecule spectroscopy (SMS) allows for gradient properties to be probed on nanometer-to-micrometer length scales. In this dissertation, quantitative measurements of gradient polarity (i.e., dielectric properties) are made along a sol-gel derived thin film that incorporates a macroscopic polarity gradient. These measurements report on the microscopic heterogeneity of the gradient film, and point to the occurrence of phase separation of the polar and nonpolar components along the gradient. Single molecule tracking (SMT) provides an important means to examine the dynamics of molecular mass transport in thin films and on surfaces. In this dissertation, SMT is employed to study mass transport in thin water films condensed over monolayer wettability gradients under ambient environments. The results show that the rate and the mechanism of molecular transport depend on the surface wettability, and on the ambient relative humidity. Finally, wettability gradients have been broadly used to drive the transport of liquid droplets. In this dissertation, these applications are extended to achieve spontaneous stretching of DNA by the propulsion of liquid droplets along the gradient. Single molecule fluorescence imaging of DNA stretched along these gradients demonstrates that hydrophobic surfaces play an important role in DNA stretching. The study also shows the surface tension force acting at the gradient-droplet contact line (interface) to be responsible for DNA elongation and alignment. Overall, single molecule methods have been shown to be highly useful for better understanding the properties of chemical gradients as described in this dissertation.

There is a strong need to develop a mechanical method to apply external torque to the bacterial flagellar motor. Such a method will allow us to probe the behaviour of the motor at a range of different speeds under different external conditions. In this thesis, I explored various methods to deliver torque at the single-molecule level, in particular the use of angular optical trapping and magnetic tweezers. I have identified rutile particles as suitable handles for use in angular optical trapping due to their high birefringence. Further progress was not achieved using angular optical trapping due to the lack of a suitable method to attach birefringent particles to the bacterial flagellar motor. On the other hand, I was able to make further progress using magnetic tweezers. A highly-reproducible and high-yielding magnetic bead assay was developed along with electromagnets capable of generating fast-rotating magnetic fields at magnitudes on the order of tens of mT. Using the system of delivering magnetic torque developed, I was able to stall and rotate the motor forward at speeds up to 220 Hz and in the reverse direction. Stalling experiments carried out on the motor revealed the stator mechanosensing depends on torque and not rotation. Signatures of stators dropping out at low load experiments further confirm the load dependence of stators.

Doctor of Philosophy / Department of Chemical Engineering / Keith L. Hohn / The autothermal reforming and partial oxidation of hexadecane via Pt/Ni bimetallic nanoparticles on various ceria-based supports were investigated. Nanoparticles with Pt/Ni molar ratios ranging from 0/100 to 10/90 were loaded on ceria-based supports including cerium oxide, gadolinium-doped cerium oxide and cerium-doped zirconium oxide. The effect of the Pt/Ni molar ratio and the promotional effect of the support were studied by comparing the hydrogen yield. TPR and XPS analysis showed that there was a strong interaction between Ni and the CeO₂-ZrO₂ support, which led to enhancement of catalyst performance when the Pt/Ni ratio was low. The strong interaction between Ni and CeO₂-ZrO₂ support was induced by the formation of a solid solution between NiO and ZrO₂. In the case of bimetallic catalysts loaded on Gd₂O₃-CeO₂, no significant improvement in the catalytic activity of autothermal reforming was achieved until the Pt/Ni ratio reached 10/90. With C-snarf-1 as a pH-sensitive fluorescent probe, the local acidity on the surface of gradient aluminosilicate thin films and in the pore structure of mesoporous silicate films was explored. The single molecule emission ratio (I₅₈₀/I₆₄₀) of C-snarf-1 on the gradient aluminosilicate films showed similar results as previously reported for aluminosilicate mesoporous films. As the Al/Si ratio increases, the emission ratio declines, indicative of increased material acidity. In the case the mesoporous silicate films, much broader distributions of emission ratios were observed and are suggestive of significant heterogeneity in the pore structure of these films. The average emission ratio increased with a rise in pH until pH 6 or 7. A further rise in pH leads to a decline in emission ratio. Molecules with high mobility showed a narrow distribution and slightly lower average emission ratio when compared to data from all detected molecules. This observation implies a reduced heterogeneity for mesopores in which the molecules rapidly diffuse. The narrow distribution and lower average value of emission ratio at low pH, combined with the decrease in emission ratio induced by an increase in ionic strength may further indicate that the interaction between dye molecules and the pore surface impacts the emission ratio of the dye molecules.

<p>Proteins are specialized molecules that catalyze most of the reactions that can sustain life, and they become functional by folding into a specific 3D structure. Despite their importance, the question, "how do proteins fold?" - first pondered in in the 1930's - is still listed as one of the top unanswered scientific questions as of 2005, according to the journal Science. Answering this question would provide a foundation for understanding protein function and would enable improved drug targeting, efficient biofuel production, and stronger biomaterials. Much of what we currently know about protein folding comes from studies on small, single-domain proteins, which may be quite different from the folding of large, multidomain proteins that predominate the proteomes of all organisms.</p><p>In this thesis I will discuss my work to fill this gap in understanding by studying the unfolding and refolding of large, multidomain proteins using the powerful combination of single-molecule force-spectroscopy experiments and molecular dynamic simulations.</p><p>The three model proteins studied - Luciferase, Protein S, and Streptavidin - lend insight into the inter-domain dependence for unfolding and the subdomain stabilization of binding ligands, and ultimately provide new insight into atomistic details of the intermediate states along the folding pathway.</p> / Dissertation