The sialic acids represent a structurally and functionally diverse family of nine-carbon keto sugars. These compounds, derived structurally from neuraminic acid, are central to many of life’s processes at the molecular level. N-Acetylneuraminic acid (NeuNAc) is the most naturally abundant of the sialic acids and acts as the terminal residue of mammalian cell surface glycoconjugates. The negative charge and exposed location of NeuNAc makes this compound intrinsically linked to cellular signalling, recognition and adhesion events. While NeuNAc expression is typically restricted to eukaryotic phyla, some pathogenic bacteria have evolved to express NeuNAc on their own cell surfaces, allowing them to mimic the surface physiology of their mammalian host’s cells and evade detection by the immune system. In bacteria, NeuNAc biosynthesis proceeds via the aldol-like condensation reaction of phosphoenolpyruvate (PEP) with the sugar-substrate N-acetylmannosamine (ManNAc). Alternatively, in mammalian systems ManNAc is first phosphorylated to ManNAc 6-phosphate (ManNAc 6-P) before undergoing condensation with PEP to give NeuNAc 9-phosphate (NeuNAc 9-P). These PEP condensation reactions are catalysed by an evolutionarily related family of homo-dimeric (βα)₈ barrel enzymes (NeuNAc synthase in bacteria and NeuNAc 9-P synthase in mammals), referred to collectively as the sialic acid synthases. In addition to NeuNAc, several pathogenic bacteria synthesise a number of unique ‘bacterial sialic acids’, including legionaminic and pseudaminic acid, which are known to be essential for the motility and pathogenicity of some species. These alternate sialic acids are again biosynthesised via condensation reactions of PEP with variable sugar substrates, catalysed by additional members of the sialic acid synthase family. In chapter two of this thesis, I report a structural and functional characterisation of human NeuNAc 9-P synthase. Modelling and mutagenesis were used to delineate possible sugar-substrate binding modes, with a number of potentially important phosphate binding residues identified. Biophysical analysis reveals the human enzyme adopts a domain-swapped homo-dimeric conformation in solution, as previously observed for the analogous enzyme from Neisseria meningitidis. Chapter three details an investigation into the variable sugar-substrate specificity of mammalian and bacterial sialic acid synthases, which are apparently selective for their respective sugar-substrates depending entirely upon the presence or absence of the C-6 phosphate group. Bioinformatic analysis of bacterial and mammalian sequences revealed the β₂α₂ loop of the catalytic barrel as a putative sugar-substrate selectivity element. Substitution of the bacterial loop with the mammalian loop sequence however was alone insufficient to confer novel activity with the mammalian sugar-substrate. In chapter four, the β₂α₂ loop of the bacterial sialic acid synthases; legionaminic acid synthase (LegS) and pseudaminic acid synthase (PseS), was again identified as a structural element potentially involved in the sugar-substrate selectivity of these enzymes. Generation of chimeric LegS and PseS variants incorporating the β₂α₂ loop from bacterial NeuNAc synthase was unsuccessful at conferring novel NeuNAc synthase activity due to unforeseen disruption of PEP binding and quaternary structure. The work presented in this thesis provides a starting point from which to pursue a comprehensive understanding of the molecular basis of sugar-substrate specificity of the sialic acid synthases. An appreciation of how these enzymes achieve precise sugar-substrate specificity may provide the basis for their exploitation as therapeutic or biosynthetic targets.