Optimization of heterologous protein secretion in yeast

Saccharomyces cerevisiae, has proven to be a viable production heterologous host for the production of protein therapeutics; however, despite substantial effort, many proteins remain difficult to produce. For the studies reported here, a neurotrophin known as brain derived neurotropic factor (BDNF) was chosen as the target protein. This important growth factor signals the survival of nerve cells and has considerable therapeutic potential for Alzheimer's, Parkinson's, and stroke patients. In this thesis, several parallel research paths were initiated to improve production of BDNF in yeast. These included engineering the BDNF protein itself, employing an engineered pro-region to assist BDNF folding, and engineering the unfolded protein response pathway of yeast.;First, using a directed evolution approach in partnership with Dr. Michael Burns, BDNF was mutated and, using yeast display, screened for improved expression and folding. BDNF mutants were identified that had up to 5-fold improvements in expression and folding.;While these results were exciting, there would be potential for unwanted immunogenic effects due to the introduction of BDNF mutations. Taking advantage of the fact that BDNF naturally contains a pro-region which substantially improves BDNF production in yeast, this pro-region was selectively mutated and evolved as before.;Evolved pro-regions capable of further improving expression and folding another 2 to 3-fold were obtained. When an evolved pro-region was combined with an evolved BDNF mutant, there was over a 200-fold synergistic improvement in folding over wild-type BDNF. Although effective, both of these methods are neurotrophin isoform specific. Thus, we investigated a new approach targeted to engineering the yeast host itself through directed evolution of the hac1p transcription factor. This transcription factor modulates every part of the secretory pathway to help restore proteostasis. Clones capable of improving wild-type BDNF expression and folding were obtained. These improvements were found to be the result of tuning the UPR response level, rather than mutations to the HAC1 gene.;Low UPR activation resulted in improved folding up to 5-fold, while high UPR activation improved expression 5-fold at the expense of folding. These findings suggest that the secretion for many proteins of interest could be improved by simply tuning UPR activation.