Heterologous protein production in bacteria is often a batch process, where the cells are lysed and the protein of interest is purified from the cellular milieu. A frequent approach is to accumulate the protein of interest in the cytoplasm of the cell, requiring extensive purification to separate the protein of interest from other cellular constituents. Secretion of heterologous protein produced with gram-negative bacteria holds many advantages that have not yet been realized due to low yields, and success has been protein-specific. The extracellular space is largely void of proteins, resulting in simplified protein purification and enabling continuous processing for production of a protein of interest. The type III secretion system is an ideal target for engineering generalizable protein secretion at high titer because it is not essential and is proven to secrete heterologous proteins. This allows direct engineering of the secretion system, in contrast to previous efforts that used essential secretion systems.

In this dissertation, I describe approaches taken to characterize and improve the process of protein production using the type III secretion system. In Chapter 2, I describe methods for quantification of secreted protein titer. In Chapters 3 and 4, I describe two complementary approaches to increase product titer. In Chapter 3, I describe a genetic approach to engineer control of the expression of the ~40 genes that comprise the Salmonella pathogenicity island 1 (SPI-1) type III secretion system. The positive transcriptional regulator HilA serves as a node in the regulatory network and is required for expression of the SPI-1 genes. Controlling the expression of hilA allows for control of the many downstream genes required for secretion. This modification increases secreted protein titer by over ten-fold and the effect is generalized for all proteins tested. Importantly, the timing and level of SPI-1 expression is synthetically controlled and are no longer restricted to growth conditions that endogenously induce expression of these genes. In Chapter 4, I describe a protein engineering strategy on the genome to mutate the gene prgI, which codes for a major structural component of the SPI-1 type III secretion system. The structure, termed the secretion apparatus, is thought to be dynamically regulated. I identify amino acid substitutions that result in greater secreted protein titer. The effect of the prgI mutation on secreted protein titer was general for two different model proteins.

In Chapter 5, I characterize product quality by probing the folded state of several different test proteins. Proteins are unfolded during secretion. Secreted proteins are then ejected into the extracellular space in an unfolded state, where refolding takes place in a dilute, aqueous environment. I used protein function as a proxy for protein folding, and demonstrate function in the extracellular space, indicating that secreted proteins indeed refold after secretion. Genetic and chemical methods are used to probe the folded state of the model enzymes beta-lactamase and alkaline phosphatase and a single-chain variable fragment of an antibody to confirm that these proteins are spontaneously adopting a functional conformation. Further, the folding efficiency is a function of the chemical composition of the media, suggesting that a process using secretion to produce proteins must consider media composition to control protein folding.