The enzyme 3-deoxy-D-manno-octulosonate 8-phosphate synthase (KDO8PS) catalyses the stereospecific aldol-like condensation between phosphoenolpyruvate (PEP) and the five-carbon sugar D-arabinose 5-phosphate (A5P). This is the first biosynthetic step in the formation of 3-deoxy-D-manno-octulosonate (KDO), an essential lipopolysaccharide component of all Gram-negative bacteria. KDO8PS is evolutionarily related to the shikimate pathway enzyme 3-deoxy-D-arabino-heptulosonate 7-phosphate synthase (DAH7PS), which catalyses a similar condensation reaction between PEP and the four-carbon sugar D-erythrose 4-phosphate (E4P), in the first step of the shikimate pathway to aromatic compounds in plants and microorganisms. As well as being a one-carbon shorter substrate, E4P has the opposite C2-OH configuration to A5P. While there are both metal-dependent and metal-independent forms of KDO8PS, in contrast, all DAH7PS are metal-dependent enzymes. Little is understood about the key sequence features that distinguish KDO8PS and DAH7PS. These features, particularly those that contribute to A5P or E4P binding, are thought to be responsible for the differences in substrate specificity between the two enzymes. This thesis describes the functional and structural studies of KDO8PS mutants to examine the roles of these residues, using the metal-dependent KDO8PS from Acidithiobacillus ferrooxidans and the metal-independent KDO8PS from Neisseria meningitidis. In Chapter 2 an extensive KDO8PS and DAH7PS sequence analysis is presented. The results, which identify sequence conservation in both enzymes, are discussed in the context of the (β/α)8 TIM-barrel structure. Some of the differences in conservation between the two enzymes were highlighted as being obvious in having a role or contributing to the different substrate selection preferences of the two enzymes, such as an extended β7α7 loop in KDO8PS, and motif differences on the β2α2 and β4α4 loops. A similar analysis was also used to compare metal-dependent and metal-independent KDO8PSs, and it was found the two forms differ in the conservation of only three residues. Chapter 3 describes the characterisation of A. ferrooxidans KDO8PS (AfeKDO8PS) and investigates aspects of metal dependency in KDO8PS. The enzyme was found to be metal dependent, and like all other KDO8PS enzymes, to possess a tetrameric quaternary structure, and display tight substrate specificity. The β8α8 loop was found to have a critical role in binding and positioning the substrates, and AfeKDO8PS could not be engineered to be a metal-independent enzyme. The role of the KDO8PS-conserved KANRS motif, present on the β2α2 loop and one of the main contributors to the A5P binding site, is probed in Chapter 4. Individual residues of the motif were mutated to investigate function, and the motif was converted to the equivalent motif found in DAH7PS (KPRS). It was found that the Lys plays a critical role in enzymatic catalysis, and is likely intimately involved in the enzyme mechanism. The Asn residue of the motif in KDO8PS was found to be an important contributor to KDO8PS stereospecificity. The work described in Chapter 5 investigates the role of the β7α7 loop in KDO8PS. This long active-site loop, which exists in a shorter version in DAH7PS, was found not to be essential for catalysis in KDO8PS, but was necessary for efficient catalysis. The two conserved residues on the loop provide interactions to A5P, but the presence of the extended loop as a whole was found to be most important for catalytic efficiency. In Chapter 6 a conserved residue on the re face of PEP is investigated. In KDO8PS the residue is conserved as Asp, and in DAH7PS the same residue is conserved as a Glu. Mutational analysis found that in KDO8PS the Asp residue appears to be important for enzyme activity but unimportant for PEP binding. Mutating this Asp in KDO8PS to Glu was accommodated by KDO8PS, but it was found its introduction could potentially be optimised by coupling the change with mutation to other conserved differences. In KDO8PS, one of the interfaces between adjacent subunits in the tetrameric structure is partially composed of a conserved sequence motif, PAFLxR. In Chapter 7, the roles of the residues in this motif are explored. The Arg of the motif was found to be important for A5P binding. The equivalent (and also conserved) motif in DAH7PS is GARNxQ, and mutation of residues in the KDO8PS motif to the equivalent residues in DAH7PS was tolerated by KDO8PS, but negatively impacted upon the enzyme kinetic parameters. The sequence features investigated in the other chapters were combined with those to the subunit interface to create a DAH7PS-like protein. This extensively engineered protein lost all KDO8PS activity, but nor did it gain DAH7PS activity. Lastly, in Chapter 8 the results from all chapters are reviewed and ideas are discussed for advancing the research presented in this thesis.

Antibiotic resistance is a worldwide threat to human health yet fewer new antibiotics are being approved. New antimicrobial drugs are urgently required. 3 Deoxy-D-manno-2-octulosonate-8-phosphate synthase (KDO8PS) is a target for antimicrobial drug design. KDO8PS catalyzes the condensation of D-arabinose-5 phosphate (A5P) with phosphoenolpyruvate (PEP) to produce KDO8P. KDO8PS catalyzes the first committed step in the lipopolysaccharides (LPS) biosynthesis pathway in Gram-negative bacteria and is critical for bacterial pathogenicity/virulence. We have characterized KDO8PS from Campylobacter jejuni (cjKDO8PS), a new metal-dependent KDO8P synthase (KDO8PS). cjKDO8PS is a tetramer in solution and optimally active at pH 7.5 and 60 °C. We have kinetically established that cjKDO8PS follows a rapid equilibrium sequential ordered ter ter kinetic mechanism, where Mn2+ binds first, followed by PEP, then A5P. Pi dissociates first, before KDO8P, then Mn2+. cjKDO8PS was inhibited by KDO8P oxime, a novel slow tight-binding inhibitor. KDO8P oxime is a competitive inhibitor with respect to PEP and A5P, but uncompetitive with respect to Mn2+, with Ki = 10 ± 1 μM and an ultimate Ki* = 0.28 ± 0.10 μM. KDO8P oxime has a residence time (tR) of 5 days on the enzyme, a parameter that is highly correlated to in vivo efficacy. Crystallization conditions for the cjKDO8PS‧Mn2+‧KDO8P oxime complex have been found and can be optimized to obtain a crystal structure that shows how KDO8P oxime interacts with the active sites. / Thesis / Doctor of Science (PhD) / The relentless increase in global antibiotic resistance is, regrettably, not matched with an increase in new effective antibiotics. New antimicrobial drug discovery strategies are desperately needed. Enzymes are key targets for drug design because they catalyze the majority of biological processes. In this project we sought to study and inhibit the activity of KDO8P synthase (KDO8PS) from Campylobacter jejuni, a common cause of food poisoning. KDO8P synthase is a critical enzyme involved in the lipopolysaccharide (LPS) biosynthesis in Gram-negative bacteria. The LPS acts as a permeability barrier and is crucial for bacterial pathogenicity/virulence. We found that C. jejuni KDO8PS is potently inhibited by KDO8P oxime, a novel inhibitor of KDO8PS. This inhibitor presents a unique opportunity to study these enzymes and a platform from which antibiotics against Gram-negative bacteria can be developed.

The enzyme 3-deoxy-D-manno-octulosonate 8-phosphate (KDO8P) synthase catalyses the aldol condensation of the five-carbon sugar phosphate, arabinose 5-phosphate (A5P), and phosphoenol pyruvate (PEP) to give the eight-carbon phosphorylated sugar, KDO8P. It is the second committed step in the synthesis of KDO, a necessary component of the cell wall of Gram-negative bacteria. This thesis describes the design, synthesis and evaluation of a number of inhibitors of KDO8P synthase that utilise the functionality of one or both substrates. The KDO8P synthase family can be divided based on the requirement of a divalent metal ion. Chapter 2 describes the growth, purification and characterisation of an example from both the metal-independent KDO8P synthases (Neisseria meningitidis, Nme) and metal-dependent KDO8P synthases (Acidithiobacillus ferrooxidans, Afe) in order to utilise these enzymes for the inhibition studies described in this thesis. In Chapter 3, a number of small molecule PEP analogues were selected as mimics of KDO8P synthase reaction intermediates and tested as inhibitors of KDO8P synthase from N. meningitidis and A. ferrooxidans. Glyphosate, (E)-vinyl phosphonate and the fluorinated analogue of (E)-vinyl phosphonate were selected as mimics of the high-energy oxocarbenium intermediate through which the KDO8P synthase reaction is thought to occur. The two enantiomers of phospholactate were selected in order to investigate the chirality of the tetrahedral intermediate and determine the importance of this chirality for inhibition of KDO8P synthase. All five inhibitors were found to be moderate to poor inhibitors of both the KDO8P synthase from N. meningitidis and A. ferrooxidans. Chapter 4 describes the design and synthesis of inhibitors that incorporated structural features of the second substrate, A5P, in order to improve inhibition from that observed for the PEP analogues investigated in Chapter 3. A bisphosphate inhibitor was designed that incorporated a terminal phosphate moiety, representative of the phosphate of A5P. A large increase in inhibition was found, compared to the phospholactates from which it was derived. A structure-activity-relationship study was undertaken on this compound by design of compounds that lacked one of the two phosphate moieties of the bisphosphate inhibitor, in order to determine their relative importance. The inhibition results indicate that the primary terminal phosphate, thought to bind in the A5P phosphate binding site, is more important for inhibition of KDO8P synthase than the secondary phosphate. In Chapter 5 these investigations into the inhibition of KDO8P synthase are discussed in detail, and interpreted using the aid of computational studies. In addition several approaches are described for the completion and advancement of the studies presented here in this thesis.

3-Deoxy-D-arabino-heptulosonate 7-phosphate (DAH7P) synthase and 3-deoxy-Dmanno- octulosonate 8-phosphate synthase (KDO8P) synthase are two enzymes that catalyse very similar reactions. DAH7P synthase is the first enzyme of the shikimate pathway and catalyses the condensation reaction between the four-carbon sugar erythrose 4-phosphate (E4P) 1 and the three-carbon sugar phosphoenolpyruvate (PEP) 2 to give the seven-carbon sugar DAH7P 3. KDO8P synthase catalyses a similar condensation reaction between the five-carbon sugar arabinose 5-phosphate (A5P) 8 and PEP 2 to give the eight-carbon sugar KDO8P 9. Early mechanistic studies have shown the reaction mechanisms of these two enzymes to be very similar and structural and phylogenic analysis has suggested that the two enzymes share a common ancestor. However, there are differences between the two enzymes that have not been explained by the current literature. Whereas all DAH7P synthases require a divalent metal ion for activity, there exists both metallo and non-metallo KDO8P synthases. As well as this, there is the difference in substrate specificity. The natural substrate of KDO8P synthase, A5P, is one carbon longer and has the opposite C2 stereochemistry to E4P, the natural substrate of DAH7P synthase. This study investigates the role of the C2 and C3 hydroxyl groups of E4P and A5P in the enzyme catalysed reactions. The E4P analogues 2-deoxyE4P 38 and 3-deoxyE4P 39 have been synthesised from [beta]-hydroxy-[gamma]-butyrolactone and malic acid respectively. The two analogues were tested as substrates for DAH7P synthase from a variety of organisms, including N. meningitidis, the purification and characterisation of which was carried out during the course of these studies. It was found that both analogues were substrates for DAH7P synthase. 2-DeoxyE4P was found to be the best alternative substrate for DAH7P synthase to date. The analogous study was carried out on KDO8P synthase from N. meningitidis with 2- deoxyR5P 34 and 3-deoxyA5P 40. It was found that removal of the C2 and C3 hydroxyl groups of A5P was much more catastrophic for the KDO8P synthase catalysed reaction. Commercially available 2-deoxyR5P was found to be a very poor substrate, whereas 3-deoxyA5P, which was prepared according to a literature procedure was not a substrate. The difference in substrate specificities of DAH7P synthase and KDO8P synthase is consistent with the hypothesis that despite their similarities, these two related enzymes have different mechanisms. The key step for DAH7P synthase appears to be coordination of the E4P carbonyl to the divalent metal. The metal appears to play a less important role in the KDO8P synthase reaction and the key step is the correct orientation of A5P in the active site.