Nitrochemicals are currently widely used as solvents, drugs, biocides, fuels and explosives and are consequently widely distributed in the environment. The reductive nitrite elimination from explosive compounds is catalyzed by two FMN-dependent, xenobiotic reductases (XenA or XenB). These genes for these regiospecific enzymes were cloned from Pseudomonas putida and P. fluorescens I-C respectively and isolated from the soil of a contaminated World War II munitions manufacturing plant. These enzymes enable the microbes to fulfill their nitrogen requirements from nitroglycerin by catalyzing the regiospecific, NADPH dependent, reductive denitration of nitroglycerin with differing selectivities. The two enzymes also transform a number of additional nitrocompounds in vitro, e.g. TNT and metronidazole, a leading drug in the treatment of Helicobacter pylori, a causative agent of human ulcers. Single crystals were obtained for XenA and XenB and complete X-ray diffraction datasets have been collected and analyzed to better understand these characteristics. The 1.6 Å resolution structure of XenA reveals a dimer of β/α)₈-TIM barrels, but the 2.3 Å resolution structure for XenB is a monomer. The (β/α)₈-TIM barrel protein fold is the most common fold in the PDB. However, the XenA structure exhibits a unique, C-terminal domain-swapped topology. Thus a portion of each active site is comprised of residues from the neighboring monomer. To probe the reaction cycle, crystal structures of ligand complexes and the reduced enzyme have been refined. For example, our structure of the XenA-metronidazole complex shows that ligands bind parallel to the FMN si-face. Our 1.5 Å resolution structure for reduced XenA reveals an FMN isoalloxazine ring with an angle of ~165° along the N5-N10 axis. We have also generated models of the reduced enzyme-nitroglycerin complexes by molecular dynamics. The results with both XenA and XenB reveal differences in enzyme-ligand hydrogen bonding. These differences correlate remarkably well with the regiospecific differences observed for nitrite elimination from nitroglycerin and reduction of TNT by the two enzymes.

The technique of X-ray diffraction has been used to investigate compounds which show charge-transfer properties in the solid state. It is shown how the results of structure determination can, under favourable conditions, give information about charge-transfer processes. Two sets of compounds have been studied, both of which contain metal-chloride bonds. The first is a series of mixed-valence cobalt hexammine chlorocuprates in which the average oxidation state of the copper can take any value between 1 and 2. The structure of the single-valence (copper (I)) compound, [Co(NH₃)₆]₄Cu₅Cl₁₇ was solved and data was also collected for a compound with Cu(I)/total Cu ~ 0.25. This was shown to have a structure to the single-valence copper (I) and copper (II) compounds and probable arrangements for the Cu(I) and Cu(II) ions are presented. It is shown that linear Cu-Cl-Cu bridges are almost certainly present and that these can account for the charge-transfer spectrum. A new chlorocuprate anion, [Cu₅Cl</sub>16</sub>]^11- was discovered, which has 23(T) symmetry.

This thesis studies the prolyl-hydroxylase family of 2-oxoglutarate dependent oxygenases from structural, functional and evolutionary perspectives. The role of prolyl-hydroxylation was first identified in collagen, wherein hydroxyproline was found to stabilise the collagen triple helix. In the 1960s, the presence of hydroxyproline in collagen was found to be a result of enzyme catalysed protein modification. An enzyme, now known as collagen prolyl-4-hydroxylase (CP4H), was found to be completely dependent on Fe(II), 2-oxoglutarate (2OG) and molecular oxygen for catalysis, and was the inaugural member of enzyme family known as the Fe(II) and 2OG-dependent oxygenases (2OG oxygenases), the members of which have since expanded dramatically – more than 60 2OG oxygenases are predicted to exist in humans alone. It was not until the 21st century that hydroxyproline was found to play roles in human biology beyond its well-characterised role in collagen stabilisation. In animals, cells adapt to low oxygen conditions (hypoxia) via the upregulation of hundreds of target genes as governed by the hypoxia-inducible transcription factor (HIF). The mammalian hypoxic sensing system was discovered to be regulated by a conserved family of hypoxia-inducible factor prolyl-hydroxylases (PHDs or EGLNs), which catalyse the prolyl-4-hydroxylation of a conserved proline residue in HIF-α under normoxic conditions, so targeting HIF-α for proteasomal degradation via the von Hippel-Lindau (pVHL) E3 ubiquitin ligase pathway. As a result, the PHDs are current therapeutic targets for the treatment of anemia and ischemia-related diseases. Thus, hydroxyproline also plays a critical role in mammalian oxygen sensing. However, the discovery also raised the question of the evolutionary origin of these enzymes and what roles, if any, they may play in other organisms. This thesis begins by describing the identification and biochemical characterisation of the first homologue of the human PHDs in prokaryotes, specifically, in Pseudomonas species, which contains pathogens such as P. aeruginosa. Pseudomonas PHD (PPHD) was discovered to catalyse the prolyl-hydroxylation of a conserved region of elongation factor Tu (EF-Tu), a translational GTPase universally conserved in prokaryotes and known for its critical role in bacterial translation. A crystal structure of PPHD, the first of a prokaryotic prolyl-hydroxylase, was then determined, revealing a striking structural homology of PPHD to the human PHDs. The further determination of crystal structures of Pseudomonas EF-Tu and a PPHD:EF-Tu protein-protein complex, the first of any 2OG oxygenase in complex with its full-length protein substrate, provides important insights into the substrate recognition mechanisms of both the CP4Hs and the PHDs and reveals an evolutionarily conserved pathway of substrate recognition that extends to prokaryotes and will be useful in the design of selective inhibitors of the PHDs. Differences were investigated between the PHDs and a recently discovered subfamily of eukaryotic prolyl-3-hydroxylases, which catalyse the hydroxylation of a conserved proline residue in the small ribosomal subunit S23 (RPS23) and have been implicated in translation accuracy and the stress response. Crystal structures of the RPS23 hydroxylases human OGFOD1 and yeast Tpa1 in complex with 2OG-mimetic inhibitors provide insight into their evolutionary origins. Analyses of the structures will be useful for targeting either OGFOD1 or the PHDs for human therapy. The thesis then describes work on human CP4H, a 240 kDa α2β2 heterotetramer. A novel expression and purification protocol is described for the CP4H complex in addition to the first known reports of its crystallisation and diffraction. Further, the foundations of a high-throughput inhibition assay of the human CP4Hs is presented and will be of immediate interest for assaying inhibitors of the human PHDs in clinical trials, some of which are also predicted to inhibit the CP4Hs. In closing, the thesis attempts to synthesise the results presented in order to provide further insight into the question of the ancestral origins of the prolyl-hydroxylases, a family of enzymes whose range of functions and biological roles likely will continue to expand.

published_or_final_version / Chemistry / Doctoral / Doctor of Philosophy

No description available.

The structure of cytochrome c₅₅₅ from the green sulfur bacterium Chlorobium thiosulfatophilum was determined by using a single isomorphous derivative, K₂HgI₄, in combination with its anomalous signal. The initial 2.25 angstrom map was modified by the technique of Fourier inversion. The smoothing function for the electron density map addressed three different features in the map, the solvent density, the protein density and the volume surrounding the heavy atom binding sites known to contain spurious peaks. This structure determination was undertaken for three reasons. First, Chlorobium thiosulfatophilum is a very primitive sulfur metabolizing bacterium and so its cytochrome c₅₅₅ structure is important for its evolutionary implications. Second, the oxidation-reduction potential of cytochrome c₅₅₅ is significantly different from the oxidation-reduction potential of other cytochromes whose structures have been determined. Comparisons with the other structures would provide information concerning the factors that are important in regulating oxidation-reduction potentials. Finally, the three dimensional structure may aid in explaining the pattern of reactivity cytochrome c₅₅₅ displays with mitochondrial cytochrome c oxidase and reductase, which is reversed when compared to other bacterial c-type cytochromes. The resulting structure contains three alpha helices. These features are consistent with other c-type cytochrome molecules previously determined. Two regions of the map appear to be disordered and are difficult to interpret. Possible causes of this observation are discussed and related to the significance of the structure.

Nucleic acids, as one of the most important macromolecules in living systems, play critical roles in storing, transferring, regulating genetic information, directing proteins synthesis, and catalysis. Understanding the structure of nucleic acid can bring us valuable information for mechanistic study and for drug discovery as well. Among all experimental methods, X-ray crystallography is the most powerful tool in structural biology study to reveal the 3D structure of macromolecules, which has provided over 80% of the highly detailed structural information to date. However, this great technology comes with two disturbing features, crystallization and phasing. The covalent selenium modification of nucleic acids has been proven to be a powerful tool to address both issues in nucleic acid crystallography. First part of this dissertation focuses on the development of novel selenium-modified nucleic acids (SeNA) for crystallization and phasing of B-form DNA containing structures. The novel 2’-SeMeANA modification is the first and currently the only selenium modification, which is fully compatible with X-ray crystallographic study of B-form DNA. Since selenium derivatization at 2’-arabino position dose not affect the B-type 2’-endo sugar conformation, this strategy is suitable for incorporating selenium into DNA for structural studies of B-DNA, DNA-protein complexes, and DNA-drug complexes. Specific base pairing is essential to many biological processes, including replication, transcription, and translation. It is crucial to NA (nucleic acid) sequence-based diagnostic and therapeutic applications as well. By utilizing the unique steric and electronic property of selenium, we designed, synthesized the novel 2-Se-U RNA modification, and demonstrated its highly specific base-pairing property by both biophysical and crystallographic methods. Our studies of 2-Se-U-containing RNAs suggest that this single-atom replacement can largely improve base pairing fidelity against U/G wobble pair, without significant impact on U/A pair.

The structure of rabbit muscle glycogen phosphorylase b, an important regulatory enzyme in glycogen metabolism, has been studied by X-ray crystallographic techniques. This work was carried out as part of a group project, and the crystal structure of the enzyme had already been solved to 3 Å resolution, using the technique of Multiple Isomorphous Replacement. A search for additional heavy atom isomorphous derivatives was carried out, and photographic data to 3 Å resolution were collected for a further ethylmercurythiosalicylate derivative, using a screenless oscillation camera. The collection and reduction of this data, and the refinement of the heavy atom positions is described. The inclusion of this data allowed a new electron density map (with figure of merit = O.63) to be calculated, which enabled previously ambiguous areas in the electron density to be interpreted. Data to 2 Å resolution have been collected on an oscillation camera, using a synchrotron radiation source and cylindrical film cassettes. An intensity gain of up to 13O times, compared to a GX6 rotating anode source, was obtained with the synchrotron radiation source. A reduction in radiation damage, was also observed. The collection and reduction of the 2 Å data is described. The final overall merging R-factor was 15%. Some systematic errors remain in the data, and possible sources of these errors are discussed, and improvements to the data processing procedure suggested. The 2 Å data were empirically scaled to the 3 Å data and have been used in the first stages of the refinement of the phosphorylase b structure. The contribution of the crystallographic results towards an understanding of phosphorylase b as an allosteric protein is discussed.

The Aldo-Keto Reductases (AKR) are a group of oxidoreductase enzymes structurally and mechanistically distinct from the Alcohol Dehydrogenases (ADH). The AKRs are of importance for their ability to produce industrially useful compounds including chiral secondary alcohols. The ADH family have traditionally been exploited for chiral alcohol production; the AKR family have currently been underexploited for chiral alcohol production and present the opportunity to search for novel oxidoreductases with properties and substrate specificities distinct from the ADH enzymes. The AKR studied here, from the hyperthermophilic bacteria Thermotoga maritima has been characterised with respect to its biochemical and structural properties, and its potential as a biocatalyst evaluated. This enzyme is the second example of a thermophilic AKR to have its three dimensional structure solved, the other also being from Thermot. maritima. The AKR studied exhibits high stability with respect to temperature and moderate amounts of organic solvents. A large preference for the reduction reaction compared to the oxidation reaction was found, which has previously been observed in other AKRs. The X-ray crystal structure was solved to 2.6Å resolution in the apo form. The final structure has three loop sections which were not located due to disorder within the crystal, which are expected to become ordered upon cofactor and substrate binding. A section of one of these missing loops was found to bind at the active site of the enzyme, with a glutamate occupying the site of substrate carbonyl binding. The formation of a dimer, increased helix-dipole stabilisation and long distance ion pair interactions all act to increase thermostability of the AKR with respect to its mesophilic homologues. The X-ray crystal structure of Escherichia coli bacterioferritin has also been solved to 1.9Å resolution, which was co-purified along with the recombinant AKR enzyme. This structure shows the symmetrical binding of a heme molecule on the local two-fold axis between subunits and the binding of two metal atoms to each subunit at the ferroxidase centre. These metal atoms have been identified as zinc by the anaylsis of the structure and X-ray data and confirmed by microPIXE experiments. For the first time the heme has been shown to be linked to the internal and external environments via a cluster of waters positioned above the heme molecule. This information has provided a greater insight into the function and mechanism of bacterioferritin.

The Notch receptor family plays a key role in development and disease. In cancer, Notch can act either as an oncogene or as a tumour suppressor, and possibly as a cancer stem-cell factor. Whereas most research has focused on downstream signalling events, little is known about the cell surface organisation of Notch and its ligands. The extracellular part of Notch consists mainly of 36 epidermal growth factor-like domains (EGF-domains), many of which bind calcium. Studies have shown that tandem repeats of calcium-binding EGF domains form a rigid linear arrangement; however, the lack of calcium binding in EGF6, EGF10 and EGF22 led to the hypothesis that these might be sites of flexibility. This thesis addresses the effect of these domains on the organisation of the extracellular region of Notch and provides further insight into the calcium-binding properties of Notch. NMR residual dipolar coupling (RDC) measurements of these regions are presented, together with the X-ray crystallographic data obtained in collaboration. The crystal structure of the human Notch-1 construct EGF4-7 shows a tilt angle of 90° at the EGF5-6 interface which is much larger than the tilt angles of 10-20° observed for the EGF11-13 crystal structure. RDC measurements demonstrated an angle of ~70° in solution. The crystal structures of EGF21-23 and EGF20-23 showed a rod-shaped interface for the EGF21-22 domain, in which a cis-proline forms the packing interaction to a tyrosine at the β-turn in the major β-sheet of EGF22. These two interfaces are novel and demonstrate the possibility of interface formation without Ca²⁺. Crystallisation was unsuccessful for the EGF8-11 construct. However, RDC measurements indicate interdomain motion between EGF9 and EGF10 demonstrating a flexible interface. These data establish new information on the structural organisation and calcium-binding properties of the extracellular region of Notch and identify flexible and rigid interfaces within multiple tandem repeats of EGF domains. This information will be invaluable in constructing models of Notch-ligand complexes for testing in future functional experiments.