Genetic manipulation of pyoverdine non-ribosomal peptide synthetases to identify genetic constraints to effective domain recombination
Non-ribosomal peptide synthetases (NRPSs) synthesise small highly diverse peptides with a wide range of activities, such as antibiotics, anticancer drugs, and immunosuppressants. NRPS synthesis often resembles an assembly line, in which each module acts in a linear order to add one monomer to the growing peptide chain. In the basic mechanism of synthesis, an adenylation (A) domain within each module activates a specific monomer. Once activated, the monomer is attached to an immediately downstream thiolation (T) domain via a prosthetic phosphopantheine group, which acts as a flexible arm to pass the substrate between catalytic domains. A condensation (C) domain, upstream to the A-T domains, catalyses peptide bond formation between an acceptor substrate attached to the T domain and a donor substrate attached to the T domain of the upstream module. The peptide remains attached to the T domain of the acceptor substrate, and then acts as the donor substrate for the next C domain. When peptide synthesis reaches the final module, the peptide is released by a thioesterase (TE) domain. The linear mode of synthesis and discrete functional domains within each module gives the potential to generate new products by substituting domains or entire modules with ones that activate alternative substrates. Attempts to create new products using domain and module substitution often result in a loss of activity. The work in this thesis focuses on identifying barriers to effective domain substitution. The NRPS enzyme pvdD, which adds the final residue to the eleven residue non-ribosomal peptide pyoverdine, was developed as a model for domain substitution. The primary benefit for using this model is that pyoverdine creates easily detectible fluorescent products. The first set of experiments focused on testing the limitations of A domain and C-A domain substitutions to alter pyoverdine. Nine A domain and nine C-A domain substitution pvdD variants were constructed and used to complement a P. aeruginosa PAO1 pvdD deletion strain. The A domain substitutions that specified the wild type substrate were highly functional, whereas A domains that specified other substrates resulted in low levels of wild type pyoverdine production. This suggests the acceptor site substrate specificity of the C domain limited the success of A domain substitutions, rather than disruption of the C/A domain junction. In contrast, although C-A domain substitutions in pvdD in some cases synthesised novel pyoverdines, the majority lost function for unknown reasons. The high success rate A domain substitutions (when not limited by the acceptor site specificity of the C domain) suggested the addition of new C domains was a likely cause for loss of function. The second set of experiments investigated whether disrupting the protein interface between C domains and their upstream T domains may cause a loss in function of C-A domain substitutions. However, domain substitutions of T domains were found to have a high rate of success. Therefore, the results thus far confirmed that disrupting interactions of the C domain with A domains or T domains does not have a large affect on enzyme activity. An alternative explanation for the loss in function with C-A domain substitutions is that C domains translocated to a new enzyme are unable to process the new incoming donor peptide chain because of substrate specificity or steric constraints. To develop methods to circumvent limitations caused by the C domain, the final part of this thesis examined acceptor substrate specificity of C domains. Acceptor site substrate specificity was chosen over donor site specificity as it acts on only an amino acid rather than peptide chain. The substrate specificity was narrowed down to a small subsection of the C domain. This was an initial study of C domain substrate specificity, which may guide future development of relaxed specificity C domains.