The formation of large, complex structures from small building blocks through self-assembly is widely seen in proteins and provides a tool for the creation of functional nanoscale devices. However, the factors controlling protein self-assembly are complex and often poorly understood. Peroxiredoxins are a large family of proteins, many of which are able to form a variety of large structures from a small, basic unit. This assembly has been shown to be strongly influenced by the redox state of the enzyme, which functions as a switch, controlling self-assembly. This thesis uses a protein from this family, human peroxiredoxin 3 (hPrx3) as a model system to investigate whether the self-assembly properties of hPrx3 can be influenced by rational protein engineering. Three forms of hPrx3 were purified and examined. These were the wild type and two variants: a mutant (S78A) and a His-tagged form. Size exclusion chromatography showed that each form showed a different ratio of dimers and larger species. Both variants showed preference for larger species, especially in the His-tagged form. This was shown to be partially dependent on metal binding in the His-tagged form. Larger species formed from multiple rings were also identified. SAXS measurement indicated that in the wild type enzyme, higher order species were dodecameric rings. For the His-tagged variant, SAXS measurement showed that the species observed had a different structure than that of the wild type. Electron microscopy showed that higher order structures seen in both wild type hPrx3 and His-tagged hPrx3 were ring shaped, with dimensions consistent with dodecamers. A competitive assay showed that the wild type, with kcat/km values near 2 x 10⁷, consistent with published results. Both variant forms showed evidence of slightly higher activity than the wild type, indicating a link between activity and assembly. A peroxiredoxin from the thermophilic bacteria Thermus aquaticus, TaqPrx was also examined, in an attempt to investigate a peroxiredoxin capable of self-assembly at high temperatures, which would be very useful for a nanoscale device. TaqPrx was cloned, purified and examined, however, no evidence of self-assembly was observed. Protein modelling and dynamic light scattering measurement indicated that the protein purified was monomeric and had a structure. Sparse matrix crystal screening identified conditions that allowed crystal formation, although strongly diffracting crystals were not produced. A novel assay for peroxiredoxin activity was developed, and suggested that TaqPrx shows peroxiredoxin activity. This thesis shows that peroxiredoxins are a useful model system for the investigation of how protein self-assembly is controlled, and how it can be influenced by protein engineering.