The maturation of many secreted proteins in Saccharomyces cerevisiae involves the cleavage of a precursor polypeptide after pairs of basic residues by a specific protease (Kex2p) at a late stage in the secretory pathway. The aim of this project was to determine the order of processing events and the importance of ordered processing in the maturation of a multi-subunit protein toxin (the Kluyveromyces lactis killer toxin) in the S.cerevisiae secretory pathway. The involvement of Kex2p in the maturation of the K.lactis killer toxin was confirmed and the cleavage event was shown to most likely occur at a late stage in the secretory pathway. In order to determine whether the timing of Kex2 processing was critical for efficient maturation and secretion of secretory proteins an attempt was made to localise the Kex2 activity to the endoplasmic reticulum (ER) and thus cause premature Kex2 processing of secretory proteins. In order to localise Kex2 activity to the ER, various constructs were made based on an active portion of the Kex2p that can be secreted. Various sequences were fused to the C-terminus of this truncated Kex2p in an attempt to localise Kex2 activity to the ER. A truncated Kex2 polypeptide with the 60 C-terminal amino acids of yeast protein disulphide isomerase (PDI) added to create a Kex2-PDI hybrid protein was more efficiently retained within the cell than any of the other constructs, including a truncated Kex2p with the yeast HDEL retention signal added to the C-terminus. The implications of these results are discussed in the thesis. Analysis of the effect of the different Kex2 constructs on the processing and assembly of the K.lactis killer toxin was hampered by the complexities of the toxin system and technical difficulties. Therefore in order to determine the activities of the various Kex2 constructs in vivo the production of the peptide mating pheromone α factor was measured. Surprisingly the efficiently retained Kex2-PDI hybrid protein exhibited the lowest in vivo activity in this experiment. Assuming that the Kex2-PDI hybrid protein is localised to the ER as predicted, the results suggest that Kex2p localised in the ER cannot facilitate the efficient maturation of secretory proteins into an active secreted form. Possible reasons for this are discussed in the thesis.