Thesis (Ph. D.)--Cornell University, January, 2007. / Vita. Includes bibliographical references (leaves 163-181).

The bacterio-opsin (bop) gene encodes the protein component of the purple membrane (Pum) of the halophilic Archaeum, Halobacterium halobium. We have studied the role of oxygen, DNA supercoiling, and putative regulatory genes (brp and bat) in the induction of bop gene expression. Studies on the wild-type strain NRC-1 showed that the bop gene was induced more than 20-fold at the transcriptional level and the induction was blocked by aeration of cultures. For a Pum overproducer strain S9, the bop gene was transcribed constitutively at a level 10-fold greater than for NRC-1. Similar results were obtained for the divergently transcribed brp gene located upstream of bop. In the absence of a DNA gyrase inhibitor, novobiocin, superhelical density in H. halobium was found to be 50% more negative than in E. coli. Addition of novobiocin at concentrations subinhibitory for growth resulted in a reduction of bop and brp transcription in both strains, and also partially relaxed plasmid DNA supercoiling to the E. coli level. Slightly higher negative supercoiling of plasmids was observed in NRC-1 under microaerobic conditions, whereas, S9 showed slightly higher supercoiling under aerobic conditions. These results support a model where DNA supercoiling mediates the activation of bop gene transcription in response to oxygen limitation. The bop promoter was subsequently cloned in a halobacterial plasmid and introduced into NRC-1, S9, and S9 mutants, which have insertions in brp or a putative oxygen sensor gene, bat, located downstream of brp. The bop promoter on the plasmid showed similar activity as in the chromosomal locus with high transcription level in S9 and no activity in the brp and bat mutants. These results indicate that bop promoter function is not context-dependent and transcriptional activation of bop occurs when the regulatory genes are located in trans. The possible mechanisms for gene regulation mediated by changes in DNA supercoiling include facilitating the interactions between regulators and RNA polymerase, modulating the rate of open complex formation, changing the rate of abortive initiation, or stabilizing unusual DNA structures, which may be favored by 5M salt and high DNA superhelical density found in H. halobium.

Proteins and enzymes are large and complex biological molecules, characterized by unique three-dimensional structure are highly flexible and dynamic nature. Thorough understanding of protein and enzyme function requires studying of their conformational flexibility, because important physiological processes, such as ligand binding and catalysis rely on an enzyme’s dynamic nature and their ability to adopt a variety of conformational states. Computational methods are widely applied in studying enzymes and proteins structure and function providing a detailed atomistic-level of resolution data about the dynamics and catalytic processes, mechanisms in biomolecules, therefore even more nowadays a term ‘computational enzymology’ has emerged. Experimental methods often have difficulty in predicting dynamic motions of proteins. Computational simulations techniques, such as Molecular Dynamics simulations, have proven successful in simulating the conformational flexibility of proteins in studying structure-function relationships. Additionally, the binding events between two molecules, e.g. an enzyme and its substrate, can be computationally predicted with molecular docking methods. Enzymes are proteins that catalyse almost all biochemical reactions and metabolic processes in all organisms. In order to study the conformational flexibility of proteins we apply molecular dynamics simulations, and in order to simulate their reaction mechanisms we apply quantum mechanical simulations. Quantum mechanical simulations can also be used to predict the electronic structure of organic compounds, by calculating their electronic structures we perform orbital analyses and predict their optical properties. The results gained from our computational simulations can give new insights into explanation of experimental findings and data and can inspire and guide further experiments.

Studies of the distribution of macroamphiphiles in Gram-positive bacteria are interesting in relation to understanding their functions and are of chemotaxonomic value at the supergeneric level. Previous studies have revealed the distribution of lipoteichoic acid (LTA) in the low G+C phylum Firmicutes and typically parallel that of the teichoic acid as a secondary cell wall polymer (SCWP). The present study had focused on the distribution of macroamphiphiles in the high G+C phylum Actinobacteria, where most previous studies have revealed liopglycans as the macroamphiphiles in various different lineages of the phylum. The present study has, for the first time, investigated macromaphiphiles in a thermophilic Actinobacterium, Thermobifida fusca. The study detected the presence of LTA, with detailed structural analysis. This confirms the compatibility of these macromolecules with membrane adaptation at higher temperatures. A second thermophilic Actinobacteria, Rubrobacter xylanophilus (which had moved recently to the most distant lineage of the phylum) was found to lack of typical macroamphiphiles. The present study also detected a novel LAM-like molecule in Kineococcus radiotolerans suggesting a close relationship between the distribution of LAM or LAM-like molecules and that of the SCWP, arabinogalactan (AG). The study also identified the presence both LTA and lipoglycan in two representatives of the genus Streptomyces (Streptomyces coelicolor and Streptomyces sp. DSM40537). This allows a re-evaluation of the hypothesis (Fischer, 1994) that a single type of major macroamphiphile is normally present in a single organism. Extending these findings, comparative genomic analyses suggest that the LTA biosynthesis pathway in Actinobacteria might be different from that in the Firmicutes. The alanine substitution pathway for LTA and teichoic acids (TA) was also found to be absent from the phylum Actinobacteria. Moreover, the comparative genomic analyses for Kineoccus radiotolerans were consistent with the practical results for this organism, illustrating the potential of predicting macroamphiphile composition utilizing genome databases. The study had also confirmed that the distribution of macroamphiphiles and SCWPs has importance chemotaxonomic value especially at the supra-generic level; it can be hypothesised from the study that: LTA containing Actinobacteria contain TA as a SCWP, whilst lipoglycan containing Actinobacteria generally contains other SCWPs, such as AG rather than TA.

Glycosaminoglycans are major components of the extracellular matrix, mostly in the form of proteoglycans. They are involved in a diversity of biological processes, ranging from cell signaling to blood coagulation. Hyaluronan and chondroitin sulphate comprise a biologically important subset of glycosaminoglycans. Hyaluronate lyases are glycosaminoglycan degrading enzymes that act as eliminases. They degrade hyaluronan the main polysaccharide component of the host connective tissues, thereby destroying the normal connective tissue structure and exposing the tissue cells to a variety of bacterial toxins. Two members of family 8 polysaccharide lyases were cloned, expressed in E. colt and purified to homogeneity using immobilised metal affinity and gel filtration chromatography methods. The first lyase was SC I C2.15 (82.9 kDa) from the soil bacterium Streptomyces coelicolor A3(2). Characterisation of the N-terminal hexahistidine tagged protein revealed that the enzyme displayed hyaluronate lyase activity, and it exhibited activity toward hyaluronan, chondroitin-4 and 6- sulphate with the highest activity toward HA. The enzyme displayed an optimum activity at pH 5.2 against hyaluronan and 4.8 against chondroitin-4 and 6-sulphate and an optimum temperature at 57 °C. The kinetic parameters of the enzyme against sodium hyaluronan, potassium hyaluronan, chondroitin-4-sulphate and chondroitin-6-sulphate were determined in the absence and presence of calcium. Additionally, HPAEC analysis showed that the enzyme has an exolytic mode of action. The enzyme was crystallised and its three-dimensional structure solved at a resolution of 2.7A. It was shown to be composed of two domains, an a-helical N-terminal catalytic domain and a f3- sheeted C-terminal domain. Through site directed mutagenesis, it was found that amino acid residues Tyr253, His244 and Asn194 are crucial for the enzyme, as substitution of these amino acids with alanine resulted in the loss of activity of the enzyme toward all substrates, demonstrating that these residues are pivotal in the catalysis by the Streptomyces coelicolor hyaluronate lyase. Y253A and N194A were co-crystallised with hyaluronan disaccharide, hyaluronan tetrasaccharide, chondroitin-4-sulphate disaccharide and L-ascorbic acid 6-hexadecanoate. The complexes were bound in the cleft located on the N-terminal domain of the enzyme in front of the interface between the N-and C- terminal domains revealing that the active site (substrate-binding site) is found in the cleft. The a and 13- domains of the Streptomyces coelicolor hyaluronate lyase were each cloned separately. Both subunits were expressed in insoluble form, and therefore both were refolded under denatured conditions. Characterisation of the N-terminal a-domain reveals the presence of small amounts of activity, representing approximately 10% of activity compared with the wild type enzyme, indicating that this domain is the functional domain of Streptomyces coelicolor hyaluronate lyase. Another ORF hylA of Streptococcus pyogenes SF370 that encodes hyaluronate lyase was cloned using a ligase independent cloning system and hyper-expressed in E. coll. Biochemical characterization of HylA showed that the enzyme is active only toward hyaluronan. It worked most efficiently at pH and temperature of 6 and 47 °C respectively, with a Km of 0.206 ± 0.01 mg nil-1 and a kat of 17.5 f 0.56 s-1. Moreover, the mode of action of the enzyme is endolytic as judged by analysis of digestion products by HPAEC. The enzyme was crystallised and sufficient quality diffraction data were derived enabling the determination of crystal parameters. Furthermore, four mutant proteins were generated via the mutagenesis strategy: Y327A, Y327F, N257A and H307A. These were expressed and purified in the same way as the native enzyme. The mutants Y327A, Y327F and N257A were completely inactive toward HA, while the mutant H307A showed very little activity against HA which was approximately 5%.

MicroRNAs (miRNAs) are small RNA sequences 22-25 nucleotides in length which play a role in post-transcriptional gene regulation by binding to target mRNA sequences preventing translation. Changes in miRNA expression can contribute to disease pathogenesis, including gliomagenesis. The release of glioblastoma specific exosomes containing miRNA into the circulation of patients provides a source of biomarkers which could be utilised in a relatively non-invasive diagnostic test. The primary aim of this thesis was to identify secreted biomarkers in biofluids of glioblastoma patients for the diagnosis and prognosis of glioblastoma mulitforme (GBM). This is the first report of miRNA expression based on the age and sex of GBM patients. GBM and non-cancerous control patients were grouped into age categories (20-39, 40-59 and 60+ years old) and gender. Initial analysis was performed using miScript Brain Cancer Array (n=3 per category) and a total of 28 dysregulated miRNAs were identified in GBM patient serum as candidate biomarkers for further study. Using a new patient cohort (n=3 per category), further analysis of the 28 miRNAs by qPCR identified five miRNAs with altered expression in GBM serum: miR-34a-5p, miR-92a-3p, miR-20a-5p, miR-30c-5p, and miR-150-5p. Further validation following power analysis identified four of the five miRNA biomarkers: miR-34a-5p, miR-92a-3p, miR-20a-5p and miR-30c-5p to be significantly dysregulated in the serum of GBM patients. Grouping of patients by age and gender identified miR-34a-5p as significantly increased in aged 60+ patients (p < 0.05); miR-92a-3p expression was significantly higher in male GBM patients compared to female GBM patients (p < 0.05) and miR-20a-5p was significantly higher in a sub group of GBM patients (p < 0.01). Moreover, increased expression of miR-20a-5p in the serum of GBM patients was associated with a better median survival compared to those with no change in miR-20a-5p expression. Investigation into the origin of the serum miRNA biomarkers using qPCR, in situ hybridisation and GBM tissue data from The Cancer Genome Atlas (TCGA) identified potential differences in origin of the four miRNA biomarkers; miR-20a-5p in the serum of GBM patients was likely to originate from the GBM. MiR-34a-5p showed increased expression in the GBM tissue analysed by TCGA. In contrast, analysis of matched patient serum and tissue lysate samples using qPCR demonstrated a higher expression of miR-34a-5p in the serum of GBM patients compared to tissue expression, possibly due to increased miRNA secretion by neighbouring non-cancerous cells, or from leukocytes as part of an immune response. Further work utilising larger patient samples could confirm the origin of miR-34a-5p. Overall four miRNAs were identified in this thesis with altered expression in GBM patients. Further studies could evaluate their use as diagnostic and prognostic serum biomarkers for glioblastoma which could provide a relatively non-invasive alternative to current diagnostic methods requiring surgery.

Possession of specific enzymes can be diagnostic and more reliable than detection of the end products of a metabolic pathway for the enumeration of coliform organisms. A number of chromogenic 13-galactosidase substrates were evaluated with a range of coliform organisms with regard to the intensity of the colour formed upon hydrolysis of the substrate and potential inhibitory effect on coliform growth. A dual substrate membrane filtration medium was developed to detect coliforms and E. coli on the same membrane filter, constituting a highly cost-effective method without the need for extensive confirmatory tests. The medium was evaluated with a number of coliform organisms and simulated chlorine-stressed contaminated water samples. The intense fluorescence of coumarin-based molecules has enabled them to be incorporated into enzyme-based tests for the quantitative assay of indicator bacteria. Several novel derivatives were evaluated in this study and were found to be more promising than the commercially available cournarin giving a combination of greater fluorescence over a broad pH range and reduced growth inhibition with representative coliform strains On conversion to 13-galactoside derivatives, the substrates were evaluated and incorporated into a miniaturised broth assay format, based on most probable number (MPN). The chromogenic and fluorogenic substrates described here have been evaluated and have shown their potential as powerful tools in diagnostic microbiology, utilising specific enzymatic activities of conforms, either in addition to or in place of standard methodologies.

To investigate the interactions between a family 6 carbohydrate-binding module (CBM) from Clostridium thermocellum Xynl 1A (CtCBM6) and its target ligands and to identify the location of the ligand binding site(s) through a mutagenesis strategy, the protein was expressed in Escherichia coli and purified to homogeneity. CtCBM6 was shown previously to interact with xylan (Fernandes et al., 1999) and, informed by the crystal structure, it was found that CtCBM6 was unusual, as it contained two potential ligand-binding clefts (Clefts A and B). Qualitative ligand specificity studies through affinity gel electrophoresis (AGE) demonstrated that CtCBM6 bound preferentially to xylans, interacts weakly with 13-glucan and some soluble substituted forms of cellulose. Quantitative analysis of ligand binding by isothermal titration calorimetry showed that CtCBM6 bound xylooligosaccharides from xylobiose to xylohexaose, with affinity increasing with chain length. The affinity of CtCBM6 for soluble xylan of varying degrees of substitution was judged to be similar. NMR spectroscopy (Dr M. Czjzek at CNRS, Marseille) indicated that xylohexaose interacts with the two solvent exposed aromatic amino acids (Tyr-34 and Trp-92) and a polar amino acid (namely Asn-120) in cleft A of CtCBM6. Site-directed mutagenesis revealed that hydrophobic stacking interactions and hydrogen bonds potentiate the binding of CtCBM6 to xylan. Surface aromatic residues Tyr-34 and Trp-92 of CtCBM6 are pivotal in the interaction between this module and its ligand, as substitution of these amino acids with alanine and methionine resulted in an 8-fold and 50-fold respective decrease in affinity of CtCBM6 for oat spelt xylan, as judged by quantitative AGE. Hydrogen-bonding interactions also made pivotal contributions to the overall binding in CtCBM6. Asn-120 was critical to ligand binding, as the mutant N120A showed —130-fold loss of binding affinity. This suggests that this residue directly participates in ligand binding via hydrogen bonds. Collectively, mutagenesis and NMR studies showed that cleft A can accommodate xylooligosaccharides and xylan, while cleft B was unable to interact with target ligands. Three hyaluronidases (Hy1P1, HylP2 and HylP3) of glycoside hydrolase family 69 (GH69) were cloned from the genome sequenced organism Streptococcus pyogenes SF370, expressed in E. coli and purified to homogeneity. Characterisation of the N-terminally tagged HylP1 (38.4 kDa), Hy1P2 (42.0 kDa) and Hy1133 (41.8 kDa) revealed activity against sodium hyaluronate with a KM for Hy1P1, HylP2 and Hy1P3 of 0.90, 2.07 and 4.35 ml mg 1, and a kcat of 1390.90, 742.01 and 1253.04 s-1, respectively. HylP1, HylP2 and Hy1P3 displayed an optimum pH of 6.5, 6.0 and 5.5, respectively, and an optimum activity at 37 °C. Moreover, PAGE analysis showed each enzyme was endo-cleaving. All three enzymes have been crystallised and sufficient quality diffraction data obtained for Hy1P1 and HylP3 (data collection and processing was performed by Dr Edward Taylor). The 3D structure of HylP1 has been solved at a resolution of 1.8 A and is composed of three monomeric strands that are intertwined to form a trimer.

This study investigated the effects of various factors, namely exposure to brass/copper, sunlight, high temperature, chlorine, low pH, or starvation, on the enumeration of faecal indicator bacteria in water, performed using non-selective & selective media under conventional aerobic conditions & under conditions designed to neutralise reactive oxygen species e.g. by the addition of 0.05% w/v sodium pyruvate as a scavenger of peroxides &/or anaerobic incubation, to encourage fermentative metabolism. These methods were compared with standard US & UK methods for the recovery of injured bacteria. The information gained was used for the development of a novel broth-based medium for the enumeration of sub-lethally injured Escherichia coli under field conditions. The novel broth-based field medium ('Coliblack') was tested against a currently used method (H2S test) & evaluated under field conditions with unskilled personnel in rural locations in India. The results showed that growth conditions designed to neutralise reactive oxygen species could enhance the colony count of faecal indicator bacteria, whether enumerated using a non-selective medium, a selective medium, a resuscitative medium, or a chromogenic growth medium, following exposure to stressors such as storage of water in a brass container, sunlight, high temperature & chlorine. Such conditions may induce sub-lethal injury in E. coli & E. faecalis, inactivating the bacterial cells under conventional aerobic enumeration conditions as a result of their oxygen-sensitivity. However, exposure to some other stressors namely low pH & starvation was not strongly oxygen-sensitive, as no substantial differences between counts were observed on various growth media or enumeration conditions. The results from enumeration experiments were used to develop a field-based broth medium with a peroxide-neutralising resuscitative agent (sodium pyruvate), a selective agent showing minimum inhibition (Tergitol 7) & a chromogenic diagnostic agent based on defined substrate technology (8 hydroxyquinoline-glucuronide). The Coliblack medium indicates the presence of Escherichia coli in the drinking water by a colour change of the medium to black. Furthermore, the preliminary trial carried out for the evaluation of the novel broth-based field medium gave positive feedback from users. In this respect the results obtained in the present study represents an advancement in understanding of how to maximise the enumeration of faecal indicator bacteria in water & how to apply this information to improve detection under field conditions.

Glycosyltransferases (GTs) are essential for the biosynthesis and diversification of many therapeutically important natural products. Of these, UDP-sugar: sterol glucosyltransferases (UGTs) (2.4.1.173) catalyse the synthesis of therapeutically important steryl glycosides (SGs). Guided by the sequence similarity with a previously characterised N-terminally truncated UGT from Saccharomyces cerevisiae (UGT51), this study reports the cloning of the gene fragment encoding the C-terminal catalytic domains from related yeasts and the expression and characterisation of their encoded products produced. N-terminally histidine tagged proteins were purified for in vitro assays against a panel of sterol and steroidal acceptors. Liquid chromatography-mass spectrometry (LC-MS) and kinetic analysis led to the successful characterisation of two novel UGTs from Pichia angusta and Kluyveromyces lactis. In addition, testosterone was shown to be utilized by all UGTs, including the previously characterised S. cerevisiae UGT51. Random mutagenesis of UGTs and homology modelling of the S. cerevisiae UGT revealed structural similarities with family 1 bacterial glycopeptide GTs. Given the structural and mechanistic similarities among GT family 1 UGTs, this approach may provide a template for genetic manipulation of novel UGTs from other members of the GT superfamily with a better understanding of catalytic domains and for broadening their scope in drug development. It may also aid the development of a generic process in the synthesis of SGs.